Advanced electron microscopy techniques for mechanistic studies of the growth and transformation of nanocrystals

Transcription

1 The University of Manchester Advanced electron microscopy techniques for mechanistic studies of the growth and transformation of nanocrystals Edward A. Lewis

2 Contents Publication list List of Abbreviations... 8 List of Figures Chapter 1 Figures Chapter 2 Figures Appendix 1 Figures Appendix 2 Figures Appendix 3 Figures Appendix 4 Figures Appendix 5 Figures Abstract Declaration Copyright statement Acknowledgements Chapter 1. Theory TEM imaging STEM imaging High angle annular dark field STEM EDX and EELS EDX vs. EELS Electron beam-induced processes Introduction Beam damage Contamination In situ electron microscopy In situ heating holders Liquids and gases in the electron microscope... 41

3 1.6 Beam-induced processes in liquid-cell experiments Beam-induced nanocrystal growth Radiolysis of water Closed-cell and flow experiments Bubble formation Charging Synthesis of hollow nanostructures Hard templating Soft templating Sacrificial templating Galvanic replacement reactions The nanoscale Kirkendall effect Quantum dot properties Hybrid solar cells Chapter 2. Literature review TEM of nanomaterials HAADF STEM of nanomaterials Spectrum imaging of nanomaterials In situ (scanning) transmission electron microscopy of nanomaterials Introduction In situ transmission electron microscopy In situ heating experiments Liquid-cell experiments In situ studies of hollow nanocrystal formation Hybrid photovoltaics Use of lead chalcogenides in hybrid photovoltaics In situ fabrication of hybrid films Molecular precursors for metal sulphides Challenges for in situ synthesis Summary Conclusions and future work References Appendix 1. Real-time imaging and local elemental analysis of nanostructures in liquids. 137 Introduction

4 Manuscript: Real-time imaging and local elemental analysis of nanostructures in liquids Supporting information: Real-time imaging and local elemental analysis of nanostructures in liquids Appendix 2. Real-time imaging and elemental mapping of AgAu nanoparticle transformations Introduction Manuscript: Real-time imaging and elemental mapping of AgAu nanoparticle transformations Supporting information: Real-time imaging and elemental mapping of AgAu nanoparticle transformations Appendix 3. Further discussion of shape evolution of AgAu nanocrystals during oxidation, analysis of HAADF and EDX line scans Appendix 4. In Situ Synthesis of PbS Nanocrystals in Polymer Thin Films from Lead(II) Xanthate and Dithiocarbamate Complexes: Evidence for Size and Morphology Control Introduction Manuscript: In Situ Synthesis of PbS Nanocrystals in Polymer Thin Films from Lead(II) Xanthate and Dithiocarbamate Complexes: Evidence for Size and Morphology Control Supporting Information: In Situ Synthesis of PbS Nanocrystals in Polymer Thin Films from Lead(II) Xanthate and Dithiocarbamate Complexes: Evidence for Size and Morphology Control Appendix 5. Direct growth of ultra-narrow PbS nanowires in electrospun polymer nanofibres: a low temperature ligand-free synthesis Introduction Manuscript: Direct growth of ultra-narrow PbS nanowires in electrospun polymer nanofibres: a low temperature ligand-free synthesis Introduction Experimental Results and discussion Conclusion References Supporting information: Direct growth of ultra-narrow PbS nanowires in electrospun polymer nanofibres: a low temperature ligand-free synthesis Word Count 40,940 4

5 Publication list This alternative format thesis is built around four papers (3 published, 1 unpublished), where I was central to the experimental work, interpretation of results, and wrote the manuscript. In addition to these papers I have been involved in a variety of other projects; in most of these projects my contribution has been the characterisation of nanomaterials with transmission electron microscopy. A full list of the papers arising from my work during my PhD are listed below Lewis, E. A.; Downie, H.; Collins, R. F.; Prestat, E; Lloyd, J. R.; Haigh, S. J. Imaging the hydrated microbe-metal interface using nanoscale spectrum imaging. Part. Part. Syst. Charact. 2016, DOI: /ppsc Al-Dulaimi, N.; Lewis, E. A.; Lewis, D. J.; Howell, S. K.; Haigh, S. J.; O'Brien, P., Sequential bottom-up and top-down processing for the synthesis of transition metal dichalcogenide nanosheets: the case of rhenium disulfide (ReS 2 ). Chem. Commun. 2016, DOI: /C6CC03316D. Savjani, N.; Lewis, E. A.; Bissett, M. A.; Brent, J. R.; Dryfe, R. A. W.; Haigh, S. J.; O Brien, P., Synthesis of Lateral Size-Controlled Monolayer 1H-MoS2@Oleylamine as Supercapacitor Electrodes. Chem. Mater. 2016, 28, (2), Slater, T. J.; Lewis, E. A.; Haigh, S. J., Recent progress in scanning transmission electron microscope imaging and analysis: application to nanoparticles and 2D nanomaterials. Nanoscience Volume , 3, 168. Green, M.; Haigh, S. J.; Lewis, E. A.; Sandiford, L.; Burkitt-Gray, M.; Fleck, R.; Vizcay-Barrena, G.; Jensen, L.; Mirzai, H.; Curry, R. J.; Dailey, L. A., The Biosynthesis of Infrared-Emitting Quantum Dots in Allium Fistulosum. Sci. Rep. 2016, 6, Lewis, E. A.; McNaughter, P. D.; Yin, Z.; Chen, Y.; Brent, J. R.; Saah, S. A.; Raftery, J.; Awudza, J. A. M.; Malik, M. A.; O Brien, P.; Haigh, S. J., In Situ Synthesis of PbS Nanocrystals in Polymer Thin Films from Lead(II) Xanthate and Dithiocarbamate Complexes: Evidence for Size and Morphology Control. Chem. Mater. 2015, 27, (6),

6 Brent, J. R.; Lewis, D. J.; Lorenz, T.; Lewis, E. A.; Savjani, N.; Haigh, S. J.; Seifert, G.; Derby, B.; O Brien, P., Tin(II) Sulfide (SnS) Nanosheets by Liquid-Phase Exfoliation of Herzenbergite: IV VI Main Group Two-Dimensional Atomic Crystals. J. Am. Chem. Soc. 2015, 137, (39), Lewis, D. J.; Tedstone, A. A.; Zhong, X. L.; Lewis, E. A.; Rooney, A.; Savjani, N.; Brent, J. R.; Haigh, S. J.; Burke, M. G.; Muryn, C. A.; Raftery, J. M.; Warrens, C.; West, K.; Gaemers, S.; O Brien, P., Thin Films of Molybdenum Disulfide Doped with Chromium by Aerosol-Assisted Chemical Vapor Deposition (AACVD). Chem. Mater. 2015, 27, (4), Da Silva, A. G. M.; Lewis, E. A.; Rodrigues, T. S.; Slater, T. J. A.; Alves, R. S.; Haigh, S. J.; Camargo, P. H. C., Surface Segregated AgAu Tadpole-Shaped Nanoparticles Synthesized Via a Single Step Combined Galvanic and Citrate Reduction Reaction. Chem. Eur. J. 2015, 21, (35), Cant, D. J. H.; Syres, K. L.; Lunt, P. J. B.; Radtke, H.; Treacy, J.; Thomas, P. J.; Lewis, E. A.; Haigh, S. J.; O Brien, P.; Schulte, K.; Bondino, F.; Magnano, E.; Flavell, W. R., Surface Properties of Nanocrystalline PbS Films Deposited at the Water Oil Interface: A Study of Atmospheric Aging. Langmuir 2015, 31, (4), Lewis, E. A.; Haigh, S.; O'Brien, P., The synthesis of metallic and semiconducting nanoparticles from reactive melts of precursors. J. Mater. Chem. A 2014, 2, (3), Lewis, E. A.; Slater, T. J. A.; Prestat, E.; Macedo, A.; O'Brien, P.; Camargo, P. H. C.; Haigh, S. J., Realtime imaging and elemental mapping of AgAu nanoparticle transformations. Nanoscale 2014, 6, (22), Lewis, E. A.; Haigh, S. J.; Slater, T. J. A.; He, Z.; Kulzick, M. A.; Burke, M. G.; Zaluzec, N. J., Real-time imaging and local elemental analysis of nanostructures in liquids. Chem. Commun. 2014, 50, Savjani, N.; Lewis, E. A.; Pattrick, R. A. D.; Haigh, S. J.; O'Brien, P., MoS 2 nanosheet production by the direct exfoliation of molybdenite minerals from several type-localities. RSC Adv. 2014, 4, (67), Withers, F.; Yang, H.; Britnell, L.; Rooney, A. P.; Lewis, E. A.; Felten, A.; Woods, C. R.; Sanchez Romaguera, V.; Georgiou, T.; Eckmann, A.; Kim, Y. J.; Yeates, S. G.; Haigh, S. J.; Geim, A. K.; Novoselov, K. S.; Casiraghi, C., Heterostructures Produced from Nanosheet-Based Inks. Nano Lett 2014, 14, (7),

7 Cardoso, M. B. T.; Lewis, E. A.; Castro, P. S.; Dantas, L. M. F.; De Oliveira, C. C. S.; Bertotti, M.; Haigh, S. J.; Camargo, P. H. C., A Facile Strategy to Support Palladium Nanoparticles on Carbon Nanotubes, Employing Polyvinylpyrrolidone as a Surface Modifier. Eur. J. Inorg. Chem. 2014, (9), Yang, H.; Withers, F.; Gebremedhn, E.; Lewis E. A.; Britnell, L.; Felten, A.; Palermo, V.; Haigh, S. J.; Beljonne, D.; Casiraghi, C., Dielectric nanosheets made by liquid-phase exfoliation in water and their use in graphene-based electronics. 2D Mater. 2014, 1, (1), Brent, J. R.; Savjani, N.; Lewis, E. A.; Haigh, S.; Lewis, D.; O'Brien, P., Production of Few-Layer Phosphorene by Liquid Exfoliation of Black Phosphorus. Chem. Commun. 2014, 50, (87), Chen, Y. Q.; Slater, T. J. A.; Lewis, E. A.; Francis, E. M.; Burke, M. G.; Preuss, M.; Haigh, S. J., Measurement of size-dependent composition variations for gamma prime (γ ) precipitates in an advanced nickel-based superalloy. Ultramicroscopy 2014, 144, (0), 1-8. Page, R. C.; Espinobarro-Velazquez, D.; Leontiadou, M. A.; Smith, C.; Lewis, E. A.; Haigh, S. J.; Li, C.; Radtke, H.; Pengpad, A.; Bondino, F.; Magnano, E.; Pis, I.; Flavell, W. R.; O'Brien, P.; Binks, D. J., Near- Unity Quantum Yields from Chloride Treated CdTe Colloidal Quantum Dots. Small 2014, 11, (13), McElroy, N.; Page, R. C.; Espinbarro-Valazquez, D.; Lewis, E. A.; Haigh, S.; O Brien, P.; Binks, D. J., Comparison of solar cells sensitised by CdTe/CdSe and CdSe/CdTe core/shell colloidal quantum dots with and without a CdS outer layer. Thin Solid Films 2013, 560,

8 List of Abbreviations 1D One-dimensional 2D Two-dimensional 3D Three-dimensional AACVD Aerosol assisted chemical vapour deposition ADF Annular dark field BF Bright field BSE Back scattered electron CB Conduction band CBS Concentric back scattered CNT Carbon nanotube DSC Differential scanning calorimetry EDX Energy dispersive X-ray EELS Electron energy loss spectroscopy ELNES Energy loss near-edge structure ETEM Environmental transmission electron microscope FPS Frames per second GLC Graphene liquid-cell HAADF High angle annular dark field HOMO Highest occupied molecular orbital HRTEM High resolution transmission electron microscope IR Infra-red LUMO Lowest unoccupied molecular orbital MEMS Micro-electro-mechanical systems NC Nanocrystal NMR Nuclear magnetic resonance NW Nanowire 8

9 ORR Oxygen reduction reaction PCE Power conversion efficiency PECS Precision Etching and Coating System PVP Polyvinylpyrrolidone QD Quantum dot SEM Scanning electron microscope STEM Scanning transmission electron microscope SWCNT Single walled carbon nanotube TEM Transmission electron microscope TGA Thermogravimetric analysis UV Ultra-violet VB Valence band XRD X-ray diffraction 9

10 List of Figures Chapter 1 Figures Figure 1. (a) Simplified Ray-diagram of a conventional TEM. Unscattered electrons are shown in blue while electrons scattered through an angle θ are shown in orange. The either the back focal plane or the image plane can be projected onto the camera to record diffraction patterns or images respectively. Apertures can be inserted into both planes. Selected area apertures in the image plane can be used to form diffraction patterns from a specific area of an image. Objective apertures in the back focal plane can be used to form (b) bright field images (selecting only the undiffracted beam) and (c) dark field images (selecting a diffracted beam, which has ideally been tilted onto the optic axis). Figure 2. Bright field (BF) and annular dark field (ADF) image formation in a STEM. A focused electron probe is rastered across the specimen and the intensity of electrons falling upon a detector is recorded at each position to build an image pixel-by-pixel. The choice of detector geometry dictates the contrast mechanism. Annular detector inner collection semi angles (α ADF-inner ) greater than 50 mrad are typically referred to as high angle annular dark field (HAADF), with collection of Rutherford scattered electrons leading to incoherent Z-contrast images, at smaller inner angles ADF imaging is coherent due to Bragg scattered beams falling on the detector. Figure 3. (a) A core ionization event, in EELS the energy of the transmitted electron is measured, while in EDX the energy of the X-ray emitted is measured; both energies allow the element to be identified. (b) The nomenclature of characteristic X-rays. Inspired by figure found in Brydson et al. Aberration-Corrected Analytical Transmission Electron Microscopy, Wiley 2011, Figure 4. (a) The Auger effect: energy released in filling a core-vacancy is used to eject another electron. (b) K- shell fluorescence and Auger yields plotted against atomic number, for light elements Auger emission dominates while for heavier elements X-ray fluorescence dominates, adapted with permission from Krause Journal of Physical and Chemical Reference Data 1979, 8 (2), Figure 5. Plot of energy transferred from a beam electron to a nucleus as a function of the angle (θ) through which the electron is scattered (equation 3), maximum energy transfer occurs when the electron is scattered at an angle of 180 degrees. Figure 6. Plot showing the maximum energy transferred to a selection of nuclei, Fe (Z=26), Al (Z=13), O (Z=8), C (Z=6), and Li (Z=3), by electrons accelerated to a range of voltages from kev (equation 4). This plot demonstrates why knock-on damage is most problematic for light atoms and high accelerating voltages. Figure 7. Schematic of a MEMS based heating chip (Protochips Aduro), viewed in cross section (a), plan-view (b), and with heating membrane enlarged (c). The sample is supported on electron transparent membranes spanning the holes in the low conductivity ceramic heating membrane. The holder can place a potential across the metal contacts, resulting in resistive heating of the ceramic heating membrane. Diagram inspired by figure from Asoro et al. ACS Nano 2013, 7, (9),

11 Figure 8. Cross sectional diagram of liquid flow-cell. 1) screw, 2) top plate, 3) holder body, 4) O-ring (outer), 5) flow channel (in), 6) top chip (SiN x window), 7) spacer (gold), 8) bottom chip (SiN x window), 9) O-ring (inner), 10) flow channel (out). A thin layer of liquid (typically ranging from 50 nm to over 1 µm depending on the choice of spacer) is trapped between the windows of the top and bottom chips, imaging electrons pass through both windows and this liquid layer. The flow channels are connected to tubing which can be connected to a syringe pump and used to flow new solutions into the cell during imaging. Figure 9. Schneider et al. s model of electron beam radiolysis of water suggests that radiolysis produces reach steady state concentrations within ~10-3 s of commencing imaging; these steady state concentrations depend strongly on the dose rate of electron irradiation. (a) Concentration of radiolysis products as a function of time in pure deaerated water irradiated homogeneously at a dose rate of 7.5 x 10 7 Gy/s. (b) The steady state concentrations of radiolysis products as a function of dose rate in pure deaerated water subject to homogeneous irradiation. Figure adapted with permission from Schneider et al. J. Phys. Chem. C 2014, 118 (38), Copyright 2014 American Chemical Society. Figure 10. Schematic illustrating the hard templating approach to the synthesis of hollow nanostructures. A template of the desired size and shape is synthesised (a), the surface of the template is modified to allow deposition of shell material on the template (b), and a shell of the target material (red) is grown (c), the template material (blue) is then selectively etched, leaving a hollow shell of the target material (d). Figure 11. Schematic illustrating key mechanistic features of the galvanic replacement reaction between Ag nanocubes and an aqueous solution of HAuCl 4. The silver cube (a) is initially oxidised at specific point on its surface (b), the oxidation of Ag is accompanied by the deposition of Au, a homogeneous shell of Au-Ag alloy forms and the Ag interior of the particle continues to be removed as this shell thickens (b-d). When the silver core is depleted (d) dealloying of the alloyed shell occurs, accompanied by the formation of pores in the box s walls (e-f), ultimately the dealloying can lead to fragmentation of the structure (g). Diagram inspired by figure from Sun et al. Journal of the American Chemical Society 2004, 126 (12), Figure 12. (a) The bulk nanoscale Kirkendall effect, for example when A is zinc, B is copper, and AB is brass. J A, J B and J vac represent the fluxes of A, B, and vacancies respectively. (b) The nanoscale Kirkendall effect: a solution or gas phase species reacts with a metal (M) template to form a compound material (MX n ). Outward diffusion of M cations through the MX n shell is faster than inward diffusion of X anions, resulting in void formation in the particle s core. Figure inspired by diagrams found in Buriak J. M. Science 2004, 304 (5671), Figure 13. (a) Sulfidation of cadmium shows asymmetric intermediate morphologies where the metal-void interface is minimised. While sulfidation of cobalt at >120 C (b) shows a symmetrical intermediate with a central metal core connected to the shell by filaments. At room temperature (c) cobalt sulfidation is accompanied by the formation and coalescence of multiple voids. Figure inspired by Yin et al. Advanced Functional Materials 2006, 16 (11), and Cabot et al. ACS Nano 2008, 2 (7), Figure 14. Cross sectional schematics of hybrid solar cells with conjugated polymer shown in blue and QDs in red. (a) The key processes involved in photocurrent generation, with a photon generating an electron hole pair which is separated at the organic-inorganic interface, with the hole transported through the polymer and the electron extracted through a network of QDs. (b-d) Possible hybrid film morphologies with (b) spherical, (c) rod shaped and (d) hyper-branched QDs, the hyper-branched particles are believed to offer the best morphology for 11

12 charge separation and extraction. Diagrams inspired by those found in Gao et al. Energy & Environmental Science 2013, 6 (7), Figure 15. Energy level diagram of hybrid solar cell with conjugated polymer shown in blue and inorganic semiconductor in red. The diagram shows the case of a photon being absorbed by the organic component, causing excitation of an electron from the polymer s highest occupied molecular orbital (HOMO) to its lowest unoccupied molecular orbital (LUMO). In an alternative scenario (not shown) the inorganic component can absorb a photon, exciting an electron from the valence band (VB) to the conduction band (CB). In both cases the energy level alignment causes charge separation at the organic-inorganic interface, with electrons extracted by the inorganic phase and holes by the polymer phase. Inspired by diagram found in Xu & Qiao Energy & Environmental Science 2011, 4 (8), Chapter 2 Figures Figure 1. Examples of TEM imaging of NCs. (a-c) Bright field, dark field, and HRTEM images of polytypic CZTSSe nanocrystals clearly distinguish between wurtzite and zinc blende domains within the same crystal, adapted with permission from Fan et al. Scientific Reports 2012, 2, 952. (d-e) HRTEM images of a gold nanowire, revealing the direction of elongation and showing extensive twinning and surface faceting, reproduced with permission from Wang et al. Nature Communications 2013, 4, Figure 2. Examples of HAADF STEM imaging of nanocrystals where Z-contrast allows regions with different compositions to be distinguished. (a) During observations of the coalescence of Au (Z=79) and Pd (Z=46) NCs, regions of Au and Pd are clearly distinguishable, as Au appears brighter than Pd, adapted with permission from Mariscal et al. Nanoscale 2011, 3 (12), (b) Au-Pd core-shell NC, reproduced with permission from Kiely C. Nature Materials 2010, 9 (4), (c) FePt-CdS core-shell NC, Reproduced with permission from Trinh et al. RSC Advances 2011, 1 (1), (d,e) Tungstated zirconia catalyst, W (Z=74) apprears bright against a ZrO 2 support (Zr Z=40), black circles indicate single W atoms, black squares identify surface polytungstate species, white circles show WO x clusters. Adapted with permission from Zhou et al. Nature Chemistry 2009, 1 (9), Figure 3. Atomic resolution image of doped graphene, the Z contrast in HAADF STEM images allows the chemical identity of individual atoms to be determined. The bright atom in the centre of the highlighted region is a nitrogen atom introduced by ion implantation. Adapted with permission from Bangert et al. Nano Letters 2013, 13, Copyright 2013 American Chemical Society. Figure 4. Examples of EELS spectrum imaging revealing elemental distribution in bimetallic nanocrystals. (a) Spectrum image of Cu 3 Pt nanocrystals showing almost entirely complete alloying with Cu surface segregation of one or two atomic layers on some facets. (b) Shows PdCo NCs after annealing in a H 2 atmosphere at 500 C, resulting in the formation a Pd shell. Figure adapted with permission from Wang et al. Chemistry of Materials 2012, 24 (12), and Wang et al. Nano Letters 2012, 12 (10), Copyright 2012 American Chemical Society. Figure 5. Example of EDX spectrum imaging of bimetallic AgAu nanocrystal. Cliff Lorimer quantification reveals the particle s composition is ~13% Ag 87% Au. EDX spectrum imaging (a, c, d) and line scans (b) show a ~1 nm 12

13 thick shell covering the entirety surface of the structure. Adapted from Da Silva et al. Chemistry A European Journal 2015, 21 (35), Figure 6. EELS spectrum images of PtCo fuel cell catalyst (a) before and (b) after voltage cycles. Spectrum imaging of a large population of nanocrystals allows the changes in morphology and elemental distribution responsible for loss of catalytic performance to be revealed. Reprinted with permission from Xin et al. Nano Letters 2012, 12 (1), Copyright 2012 American Chemical Society. Figure 7. Bright field TEM images of Ag nanoparticles during in situ heating. Smaller particles sublime at lower temperatures. Reprinted with permission from Asoro et al. ACS Nano 2013, 7 (9), Copyright 2013 American Chemical Society. Figure 8. HRTEM imaging during in situ heating experiments. (a) nanocrystals formed by heating PbSe-CdSe core-shell nanoparticles in situ at 200 C. (b) nanocrystals formed by heating Fe x O-CoFe 2 O 4 core-shell nanocubes in situ at 335 C. In both examples, HRTEM images allow domains of different materials to be distinguished and their crystallographic orientation determined. Figure adapted with permission from Yalcin et al. Nanotechnology 2014, 25, (5), and Grodzinska et al. Journal of Materials Chemistry 2011, 21, (31), Figure 9. Beam-induced growth of Pt 3 Fe NW observed by liquid-cell TEM, an oriented attachement mechanism involving the coalescence of NC building blocks is observed, initially kinked wires are observed to straighten. Adapted from Liao et al. Science 2012, 336 (6084), Reprinted with permission from AAAS. Figure 10. Solution phase Coalescence of gold nanocrystals revealed by liquid-cell TEM. Lattice resolution imaging reveals two mechanistic pathways for coalescence: (A) Coalescing particles have a lattice mismatch less than the critical angle and a defect free bond is formed or (B) angle of attachement is greater than this critical angle and defects form at the interface between the fused nanocrystals. Reprinted with permission from Aabdin et al. Nano Letters 2014, 14 (11), Copyright 2014 American Chemical Society. Figure 11. TEM images of a Pt nanocrystal growing in a GLC. Growth by monomer attachment is observed at atomic resolution; high index facets are observed to disappear due to their faster growth. Figure adapted with permission from Jeong et al. Chemistry of Materials 2015, 27 (9), Copyright 2015 American Chemical Society. Figure 12. The in situ approach to hybrid film fabrication. A molecular precursor and polymer are dissolved in a common solvent, the resulting solution is spin coated onto a substrate, forming a thin film, the polymerprecursor film is then heated, causing the thermal decomposition of the precursor and the growth of nanocrystals. Figure 13. Comparison of the key steps in the fabrication of polymer-qd films for hybrid solar cells using the in situ and ex situ approaches. This in situ approach is simpler as it removes the need for ligand-exchange procedures and eliminates problems such as QD aggregation and low QD loadings, however, precise morphological control of QDs if far easier ex situ. Figure 14. Metal xanthate decomposition is believed to occur via a Chugaev-like mechanism, the relatively lowtemperature decomposition of these precursors is attractive as it produces volatile by-products (H 2 S, COS, and an alkene). 13

14 Figure 15. A selenolate precursor for the in situ synthesis of cadmium selenide nanocrystals. Combining such precursors with metal xanthates could offer a route to ternary MS x Se 1-x materials. Table 1. Summary of literature reports of metal xanthate precursors used to synthesise metal sulphide nanocrystals in situ in a polymer matrix; the heating temperature, polymer and inorganic phases, and the resulting size and morphology of nanocrystals reported are provided alongside the precursor structure. Appendix 1 Figures Figure 1. (a) Schematic of the modified liquid e-cell TEM specimen holder (a detailed description of the holder design can be found in Ref. 11). (b) Contrast variation in HAADF STEM image due to the presence of liquid, the bright area in the lower right hand corner corresponds to a liquid-filled region, the darker area is a liquiddepleted region. (c) Shows the O K α spectrum image. X-ray energy dispersive spectra from two regions of this spectrum image are shown in (d), note the large variation in O K /Si K peak ratios (4.78 vs 0.19 for liquid-filled and liquid-depleted areas respectively). Figure 2. A bimetallic nanostructure in water was imaged before (b) and after (c) EDX data acquisition, nanoparticle deposition has clearly occurred on and around the Ag-NW during data acquisition. Elemental maps (a) (d) and (e) extracted from the EDX spectrum image, for the region of interest indicated by the dotted line in (b), shows an Ag-NW coated in Cu nanoparticles. An X-ray line profile (f) extracted from the spectrum image shows elevated Cu concentration on and around the Ag-NW. Figure 3. (a-c) show a sequence of HAADF STEM images taken at times t=0 s, t= 31.4 s, and t=62.9 s. This image sequence demonstrates beam-induced Cu nanoparticle growth in an area where pre-synthesised nanostructures (Ag-NWs and Pd-CNTs) are present (full image sequence available as Video S6). (d) Shows the areal growth rate, calculated by image thresholding (Fig. S3). Figure 4. (a) HAADF STEM image and EDX hyperspectral images demonstrating simultaneous mapping of multiple elements at the nanoscale in liquid. The full range of nanostructures can be clearly differentiated in the EDX spectrum images (a). The HAADF image shown in (a) was taken before imaging, when little Cu deposition had occurred. The Cu deposition, which occurs during EDX acquisition, is inhomogeneous with elevated Cu concentration along the outer walls of the Pd-CNT. The Cu coated outer walls and Fe core of the Pd-CNT can be seen in the line scan (c) taken from the region indicated in (b). Supporting Figure S1. a) Shows a region where all three components of the soup are present, b) shows a region of Palladium-decorated carbon nanotubes (Pd-CNT), c) shows a cluster of Gold nanoparticles (Au-NP), and d) shows a Silver nanowire (Ag-NW). In each case, the HAADF STEM image (left column), the spectrum image (middle column), and the 0-2 kev region of the EDX spectrum are shown. Note the presence of iron in (c), this is the Fe catalyst left over from the CNT synthesis. The O k /Si k ratio for a dry sample is found, based on a total of 8 spectra, to be ± Supporting Figure S2. A selection of spectrum images acquired from liquid-filled region of the e-cell. a) and d) are analysed in more detail in figures 2 and 4. In each case, the HAADF STEM image (left column), the spectrum image (middle column), and the 0-2 kev region of the EDX spectrum are shown. Comparing these EDX spectra to those in Figure S1, a dramatic change in the O k /Si k ratio is seen. The larger oxygen signal in these wet samples 14

15 arises from the presence of water. While O k /Si k is almost constant for the dry references there is large variation form map-to-map, suggesting the amount of liquid present varies across the e-cell. This is consistent with the tidal waves observed during imaging (Supporting Video 7). In all spectrum images from liquid-filled regions of the e-cell, significant beam-induced Cu nanoparticle growth is observed. Supporting Figure S3. HAADF STEM images were processed using ImageJ software to determine the rate of beam induced copper growth (Fig. 3 and Supporting Video 6). A threshold was applied to every 5 th image to calculate the area of copper particles in the region of interest. Selected frames and the corresponding thresholded images are shown for times: t=8.4 s (a), t=24.10 s (b), t=39.83 s (c), t=55.54 s (d), t=68.64 s (e), and t=84.36 s (f). The resulting data (Fig. 3) shows a higher initial growth rate which decreases to a roughly constant rate of 550 nm 2 s -1 after t=15 s. Supporting Figure S4. EELS spectra from a liquid-filled (a) and liquid-depleted (b) region of the e-cell were acquired and their t/λ values calculated. Inelastic mean free path (λ) values based of Z eff calculations were found for e-cells containing different thicknesses of water (c) and used to estimate the thickness of the liquid layers in the two regions. Estimated liquid thicknesses of 180 nm and 0 nm were calculated for the liquid-filled and the liquid-depleted regions respectively. Supporting Figure S5. A strong X-ray signal is observed for a wide range of positions. a) Si K-edge signal as a function of holder tilt angle for a central point in the window demonstrating that greater shadowing occurs at low tilt angles. b) Si K edge signal as a function of position along the window long axis (perpendicular to the axis of the in-situ specimen holder) demonstrating that a wide range of positions allow viable spectral data collection The data points in red for (b) were taken from regions where due to breaks in one of the windows only a single 50 nm SiN x window was present instead of the two windows that were present for the other (black) data points. To correct for this the Si K counts from single window regions have been doubled. Supporting Video S6. This video shows the growth of copper nanoparticles in the presence of other presynthesised nanostructures (Au-NWs and Pd-CNTs). This process is beam-induced; with electrons from the imaging probe reducing copper ions from the aqueous solution. Quantitative analysis of this video allowed the areal growth rate to be estimated (Fig. 3 and Supporting Fig. 3). Supporting Video S7. The e-cell is incompletely filled with the aqueous nanoparticle soup solution. The e-cell contains air pockets, at atmospheric pressure, resulting in liquid-filled and liquid-depleted regions (Fig. 1 and Supporting Fig. 4). The liquid-depleted regions are transient in nature, with liquid moving around the cell. Due to the presence of the air pockets, liquid can be manipulated to flow into different regions of the e-cell. This video (Supporting Video 7) shows the movement of a liquid wavefront during imaging. The flow of liquid results in the displacement of nanostructures, which are transported as the transient wave front moves across the field of view, relocating the entrapped atmospheric pocket. Such nanoscale tidal waves have been reported in previous e-cell studies. 15

16 Appendix 2 Figures Figure 1. (a-f) HAADF STEM images showing the transformation of a single nanoparticle, from an Au surface segregated Ag-Au structure to a hollow Au-core Ag 2 O-shell structure. (g-i) EDX spectrum images acquired at different stages of the process reveal the compositional changes associated with the morphological changes visible in the HAADF images. The elemental maps in (g) correspond to the period between (a) and (b), (h) corresponds to the period between (c) and (d), and (i) corresponds to the period between (e) and (f). Videos of image sequences acquired between (b) and (c) and between (d) and (e) can be found in the supplementary information (supplementary videos S1 and S2 respectively). The process is accompanied by a dramatic increase in the oxygen signal (with the O/Ag at. % ratio, found by EDX quantification, increasing 153% from 0.86 to 2.18), implying the transformation is driven by oxidation. (j) Shows quantitative analysis of void growth and coalescence gained by automated image analysis of HAADF STEM image sequences acquired between (d) and (e) (supplementary information Fig. S2 and video S2). Figure 2. Illustration of observed structural transformations: (a) Au surface segregated starting structure, (b) beam induced oxidation results in Ag 2 O formation, with oxide nucleating at multiple sites on the particle s surface. Ag diffuses faster than O through AgAu resulting in outward diffusion of Ag and an increased concentration of vacancies in the core which drives both inward diffusion of Au and the nucleation of voids. (c) Ag 2 O islands grow so as to cover the whole surface, continued injection of vacancies into the core result in void growth and further inward diffusion of Au. (d) The final structure has a thick Ag 2 O shell, the voids in the core have coalesced, and Au is now concentrated in the centre of the particle. Figure 3. HAADF STEM images (a-c) show the evolution of an Ag-Au particle under election beam irradiation. (d) and (e) show high resolution images of this particle, with (d) taken just after (a) and (e) taken between (b) and (c), Fourier transforms for these images are shown inset and the areas highlighted by red boxes are shown enlarged below in (f) and (g) together with identified lattice spacings. The structure before transformation ((a), (d) and (f)) displays only lattice spacings corresponding to Ag and Au ({111} spacing of both Ag and Au ~ 0.24 nm) while the structure after beam induced shell growth ((b), (e) and (g)) clearly shows lattice spacing corresponding to Ag 2 O ({111} ~0.27 nm and {110} ~0.33 nm). Corresponding EDX spectrum images can be found in supplementary information (Fig. S4). Figure 4. Homogeneously alloyed AgAu particles, formed by in situ heating of Au surface segregated particles, undergo the same Kirkendall transformations as surface segregated particles at room temperature (Video S3) but are unreactive at elevated temperature (400 C). (a-e) HAADF STEM images (top row) and simultaneously acquired EDX spectrum images (second row, Ag=red and Au=green). (a) At 100 C all particles initially show Au surface segregation, (b) upon heating to 400 C homogeneous alloying occurs but no oxidation reaction occurs, on quenching to room temperature a beam induced 16

17 process of Ag 2 O shell growth and Au core-hollowing is observed (c-e). Heating to higher temperatures can lead to preferential loss of silver, localised melting and dramatic changes in morphology (Fig. S7). Figure 5. (a-c) HAADF images showing an Ag-Au particle undergoing the beam induced oxidation reaction described previously,, forming a hollow Au-Ag 2 O structure. Upon heating to 400 C the hollow structure is lost (d), the shell shrinks and EDX data (f) shows a reduced oxygen signal suggesting some Ag 2 O is reduced to silver metal. The post heating, non-hollow, Au-core Ag-shell particle ((d) and (h)) undergoes a second beam induced oxidation reaction, in this reaction a yolk-shell (Au-yolk, Ag 2 O- shell) particle is formed ((e) and (i)). Figure 6. AgAu bimetallic nanoparticles can undergo a range of controlled in situ transformations by sequential application of beam induced oxidation and heating. Au surface segregated AgAu nanoparticles (a) are transformed into alloyed AgAu nanoparticle (b) by heating. Both alloyed (b) and Au surface segregated (a) AgAu nanoparticles can be transformed by electron beam induced oxidation into hollow Au-core Ag 2 O-shell structures (c), which can then be reduced to an Au-core Agshell structure (d) by heating. Au-core Ag-shell nanoparticles can be oxidised to from Au-yolk Ag 2 O-shell structures (e). Supporting Figure S1. Shell growth and void formation are observed for a range of initial morphologies. Sequences of HAADF STEM images show the beam induced structural evolution of 5 particles. In all cases the particles were exposed to a constant dose rate of ~2550 eå 2 s -1. All particles belong to the same sample of Ag- Au nanoparticles synthesised by the galvanic replacement reaction and containing ~6at.% Au 94at.% Ag. While some particles (e.g. a and b) are solid and approximately spherical, other particles showing some initial hollowing (c) or faceting (d) in their initial structure. Despite different initial morphologies a number of universal mechanistic features are observed during the beam induced reaction. Initially shell growth is non uniform, with islands of Ag 2 O emerging on the particle s surface which eventually coalesce to form a complete shell. Shell growth is accompanied by the formation and growth of voids in the core. In some cases a single void grows in the particles core, in other cases several voids form initially before coalescing to form a single void. Supporting Figure S2. Quantitative analysis of void growth and coalescence. The nominal areas of the voids within a particle were measured with automated image thresholding. A self-consistent thresholding limit was applied to a series of 23 HAADF STEM images (Video S2) acquired as beam induced void growth and coalescence occurred. The particle studied is that shown in Figure 1. The total areal growth rate is found to remain approximately constant at ~0.74 nm 2 s -1. However, the trajectories of individual voids are less predictable, for example void 4 doesn t increase in size at any point, shrinking before eventually coalescing with Void 1. The growth of Void 1 shows distinct jumps in size corresponding to the points where it consumes neighbouring voids. Supporting Figure S3. EDX spectrum imaging reveals growth of an Ag and O rich shell and inward diffusion of Au. EDX spectrum images of the same particles were recorded at different stages of the beam induced oxidation. (a) shows the particle before significant changes have occurred and (b) and (c) show later points of the beam induced reaction. Unlike the particle shown in fig. 1 this particle has some initial hollowing, this is a consequence of the galvanic replacement reaction used to synthesise the Ag-Au starting material. The spectrum images reveal that as the reaction proceeds, the area containing Ag grows. The Ag maps correlate strongly with 17

18 the O maps suggesting the outward diffusion of Ag is driven by oxidation. Simultaneously the area containing Au decreases, it is believed that inward diffusion of Au occurs due to the increased concentration of vacancies in the core which accompanies the outward diffusion of Ag. Throughout this process the Ag: Au ratio remains constant (94at. % Ag, 6at. % Au) however, the amount of oxygen present (O/Ag at. % ratio) increases ~137% between (a) and (c), further supporting the oxidation hypothesis. The growth of an Ag 2 O shell and the inward diffusion of Au are further demonstrated by the line scans shown in the right hand column. Supporting Figure S4. The lattice resolution images in figure 2 confirm that shell growth is due to Ag 2 O formation. EDX spectrum images acquired for the particle shown in figure 2 at various stages of the beam induced reaction show the changes in elemental distribution that accompany the formation of Ag 2 O. The changing distributions of Ag, Au and O are similar to those seen in Figures 1 and S3, with Ag diffusing outwards while Au diffuses inwards and a strong correlation between the distributions of Ag and O. Supporting Figure S5. Oxidation is substrate dependent. Upon heating to 400 C particles on amorphous carbon immediately alloy. However, after quenching, no beam induced restructuring occurred when exposed to the same dose rates that caused the transformations shown in Figures 1 and 4. The absence of void formation and hollowing cannot be attributed to a shortage of oxygen; quantification of spectrum images reveals that O/Ag at. % ratio increased 622% from an initial value of 0.67 (a) to a final value of 4.82 (c) (probably a result of the electron beam drawing PVP to the scan area) while the Ag: Au ratio remains constant at 97at. % Ag 3at. % Au. In fact, the final O/Ag ratio is actually considerably higher than that seen in comparable examples where reactions have occurred on Si 3 N 4 grids (e.g. in Fig. 1 the final ratio is 2.18), but despite more oxygen accumulating in the scanned area during imaging there is no evidence of shell growth and hollowing. Mapping the oxygen counts reveals that, in contrast to samples on Si 3 N 4 (where O and Ag counts are clearly correlated), the support surrounding the particles contains more O than the area occupied by the AgAu particle. It is therefore postulated that the amorphous carbon support reacts preferentially with oxygen containing species, preventing Ag 2 O formation. Supporting Figure S6. Elemental maps corresponding to figure 4. (a) shows the initial Au surface segregated particles, (b) shows the homogeneously alloyed AgAu particles formed by heating to 400 C, (c-e) show the different stages of the beam induced shell growth and hollowing that occurs after quenching to room temperature. The maps of O and Ag counts for (e) suggest that the increase in oxygen counts (Fig 4f) that accompanies the beam induced reaction is due to the growth of an Ag 2 O shell. Supporting Figure S7. Heating Au surface segregated particles to 400 C results in alloying and some morphological changes (rounding of faceted edges and some loss of hollowing), however, the size and shape of the particle is largely retained and the ratio of Ag:Au remains identical to that measured at room temperature. In contrast, we observe a dramatic change in particle size, shape, and composition upon heating to 650 C. The particle shown in this figure shrinks dramatically upon heating, EDX quantification reveals that this shrinkage corresponds to a dramatic loss of Ag. The particle is initially 78at.% Ag 22at.% Au (coming from a different sample to the lower Au content particles used elsewhere in this paper) but the heated particle is 3at.% Ag 97at.% Au. This is consistent with previous reports which show that at high temperatures the lower melting point component of a bimetallic nanoparticle system will leech out, improving the particles high temperature stability. 7 The melting points of bulk Ag and Au are C and 1064 C respectively but nanoparticles show significant melting point depression. Pure Ag particles of a similar size to those studied here have been observed, by in situ TEM heating experiments, to sublime at C. 8 18

19 Supporting Video S1. Video of early stages of growth. Video comprised of a sequence of HAADF STEM images capturing the particle shown in Figure 1 during the early stages of the beam induced reaction. This video covers the period between the images shown Figures 1a and 1c. The particle is subjected to a constant dose rate of ~4990 eå 2 s -1 during the entire video. The Ag 2 O shell appears to emerge from at least 3 distinct sites upon the particle s surface. The video clip, available in the electronic supplementary material, plays at 4x real speed. Supporting Video S2. Video of void coalescence. Video comprised of a sequence of HAADF STEM images showing the growth and coalescence of voids in the particle shown in Figure 1. This video covers the period between the images shown Figures 1d and 1e. The particle is subjected to a constant dose rate of ~4990 eå 2 s -1 during the entire video. By this point in the growth trajectory, the multiple Ag 2 O islands seen in Video S1 have combined to form a complete shell and multiple voids have formed in the core. As the beam induced reaction proceeds the voids continue to grow and eventually coalesce. A quantitative analysis of this image sequence can be found in supplementary figure S2. The video clip, available in the electronic supplementary material, plays at 4x real speed. Supporting Video S3. Video comprised of a sequence of HAADF STEM images showing shell growth and void formation in a cluster of alloyed AgAu nanoparticles. This video covers the period between the images shown in Figures 4d and 4e. The particles were exposed to a constant dose rate of ~2550 eå 2 s -1.Particles alloyed by heating undergo a beam induced reaction after they are returned to room temperature. This reaction shows similar characteristics to the reaction observed in Au surface segregated particles. However, the proximity of particles to other particles undergoing the same beam induced reaction influences the final morphology, with the growing shells of neighbouring particles contacting and merging. The video clip, available in the electronic supplementary material, plays at ~9x real speed. Appendix 3 Figures Figure 1. Thresholded outlines of the same AgAu nanocrystal (a) before and (b) after beam induced oxidation, projected areas of (a) and (b) are calculated to be 802 and 966 nm 2 respectively. Thresholding and area measurements were performed using ImageJ software. Figure 2. Plan view diagram of a torus illustrating the dimensions r and R (used in equation 2). Figure 3. Calculated volumes for the initial spherical particle provided with its extrapolated volume upon oxidation of Ag to Ag 2 O with oxidations ranging from 10% to 100% of the available Ag considered. These volumes are compared to those calculated for the final structure using a toroidal model and a model of a spherical void within a spherical particle. Figure 4. HAADF intensity line profiles taken from the same particle at increasing reaction times (a-d) The final structure (d) shows a grey value in the void region equal to those measured on the surrounding support membrane, supporting a torus-like structure. Figure 5. (a-c) EDX spectrum images of the same particle at increasing extents of oxidation, each image maps the distribution of Ag and Au counts. (d and e) show the line scans indicated in (a-c) with (d) showing Au counts and (e) showing Ag counts, in both cases the line scan data has been smoothed using 25 point adjacent 19

20 averaging. Maxima and minima of the lines shown in (d and e) were determined using Origin software and used to calculate the separation between the onsets for both Ag and Au counts, as well as the separation between the peaks in Au intensity, these values are plotted in (f). Appendix 4 Figures Figure 1. (a) A schematic of the in situ approach employed to fabricate the PbS-polymer films and the precursor families tested: (b) lead(ii) dithiocarbamate complexes, (c) a 1,10-phenanthroline adducts of lead(ii) dithiocarbamate complexes and (d) lead(ii) xanthate complexes shown with butyl groups. Figure 2. A plot of precursor atom economy against solubility in chloroform with tested xanthate (circles), dithiocarbamates (squares) and 1,10-phen adducted dithiocarbamate (diamonds) complexes shown. Figure 3. A thermal ellipsoid plot of lead(ii) dibutyldithiocarbamate 1,10-phenanthroline adduct, Pb(S 2 CNBu 2 ) 2 (1,10-Phen), at 50 % probability. CCDC reference number Figure 4. SEM BSE images of films made at 275 C for 30 minutes: (a) Pb(S 2 CNBu 2 ) 2, (b) Pb(1,10-phen)(S 2 CNBu 2 ) 2, (c,d) Pb(S 2 COBu) 2, (e,f) Pb(S 2 COHex) 2, and (g,h) Pb(S 2 COOct) 2. Xanthate samples (c-h) shown with low (left column) and high magnification images (right column) to illustrate both the film homogeneity and detail of the PbS nanostructures. Figure 5. TEM size analysis of individual nanocrystals extracted from the PbS-polymer composite films. The size analysis of over 50 particles (left column) and a typical TEM image (right column) are shown for: (a) Pb(S 2 COBu) 2 (film shown in Figure 4(c,d)), (b) Pb(S 2 COHex) 2 (film shown in Figure 4(e,f)), and (c) Pb(S 2 COOct) 2 (film shown in Figure 4(g,h)). Figure 6. HRTEM images of PbS nanocrystals formed from (a) Pb(S 2 COBu) 2, (b) Pb(S 2 COOct) 2 and (c) Pb(S 2 COHex) 2. The HRTEM images are accompanied by the Fourier transform (upper right of main image) and an area enlarged area indicated by the dashed box (lower right of main image) to clearly show the lattice fringes and atomically intimate contact between nanocrystals. Figure 7. XRD patterns for films from Pb(S 2 COBu) 2. All films were heated for 30 minutes at the temperatures shown. Peaks corresponding to the unreacted precursor are not present after heating above 150 C. The reference peak positions for PbS galena are shown with dashed lines. Figure 8. Backscattered SEM images of films made with the Pb(S 2 COBu) 2 precursor. All films were heated for 30 minutes at (a) 100 C, (b) 200 C, and (c) 275 C. Figure 9. Backscattered SEM images of films made at low heating temperatures. Films made by heating precursor-polymer films at 150 C for 30 minutes with (a) Pb(S 2 COBu) 2, (b) Pb(S 2 COHex) 2, and (c) Pb(S 2 COOct) 2. XRD shows that in all films the precursor has been fully converted to PbS (Figure 7 & SI Figure 6). SI Figure 1: TGA profiles for (top) polystyrene compared to a typical xanthate, dithiocarbamate complex and adducted dithiocarbamate complex, (middle) the lead(ii) xanthate complexes tested and (bottom) the lead(ii) dithiocarbamate complexes tested and their adducts. SI Figure 2: Pb(S 2 CNBu 2 ) 2 films annealed in air and nitrogen at 275 C for 1 hour. The films annealed under dry nitrogen show pure PbS, while those annealed in air show a mixture of lead sulphide (PbS) and lead sulphate 20

21 (PbSO 4 ) phases, illustrating the need for an oxygen free atmosphere. All other results shown in this paper are of films annealed in nitrogen. SI Figure 3: Typical appearance of film. The left column shows an uncoated glass slide as a colour reference, the middle column shows slides coated with a precursor-polymer film, and the right column shows annealed slides where a PbS-polymer film has formed. SI Figure 4: XRD shows that for all precursors annealing the films at 275 C for 30 minutes causes complete conversion to PbS. Corresponding SEM images can be found in Figure 3. SI Figure 5: Low magnification bright field TEM images showing cube terminated nanowires. The structures support nanowire growth from nanocube seeds, growing out of the (011) facets of the cubes corners. SI Figure 6: XRD of precursor-polymer films containing the Pb(S 2 COHex) 2 or Pb(S 2 COOct) 2 precursors and annealed for 30 minutes at either 100 C or 150 C. For both precursors 150 C gives complete decomposition while 100 C gives partial decomposition. SI Figure 7: XRD of films containing the Pb(S 2 COBu) 2 precursor annealed at 100 C for various times. Increased annealing times lead to a reduction in the intensity of peaks associated with unreacted precursor. SI Figure 8: Size distribution histograms based on TEM size analysis of over 50 particles extracted from Pb(S 2 COBu) 2 based films annealed at 100 C for 30 minutes and 275 C for 30 minutes. Average sizes are found to be 46 ± 14 nm and 104 ± 18 nm for the 100 C and 275 C samples respectively. SI Table 1: Summary of properties of all precursor complexes investigated in this work. SI Table 2: Crystal structure and structure refinement for compound 8. SI Table 3: Selected distances for compound 8. Appendix 5 Figures Figure 1. Grazing incidence XRD of precursor-polystyrene thin films after heating at 90 C for up to 8 hours, the time required for complete conversion of the precursor to PbS. The reference peak positions for PbS galena are shown. Figure 2. SEM images of polymer fibre mats produced by electrospinning before (a,b) and after (c,d) heating at 90 C for 8 hours. Figure 3. HAADF STEM images of PbS nanowires grown in polymer nanofibres at 90 C. (a) Low magnification image showing a large number of typical nanowires. (b,c) High resolution images show that the wires are highly crystalline PbS. A region of (c) is enlarged in (d) for clarity and (e) shows the Fourier transform of (c), all nanowires imaged were found to be elongated in the [110] direction. Figure 4. Evidence for rectangular cross section of nanowires (a) HAADF contrast shows that the narrower wire is thicker than the broader wires that surround it. Regions of the narrow wire (green out line) and broad wire 21

22 (red outline) are enlarged to clearly show their lattice resolution structure. (b) Fourier transform of (a) showing that PbS crystals with two different orientations are present, the spots circled in red are identified as arising from the broad wire, while those circled in green arise from the narrow wire. (c) shows a simulated electron diffraction patterns for [1-10] and [001] oriented PbS. Figure 5. (a) Nanocubes synthesised in a film with ~5 wt. % PbS loading, average diameter= 10.3 ± 1.9. (b) Nanocubes synthesised in a film with ~30 wt. % PbS loading, average diameter= 17.7 ± 2.8 nm. (c) Nanocubes synthesised in a film with ~70 wt. % average diameter= 25.1 ± 4.9 nm. The diameter of 100 representative nanocubes from each samples were measured. High resolution images can be found in SI figure 10. Figure 6. Cross sectional HAADF STEM images of the same heated polymer nanofibre at several different magnifications, red boxes indicate the regions shown in higher magnification images. PbS nanoparticles are distributed uniformly within the polymer fibre. The PbS nanowires show no clear orientation preference. Further cross sectional images can be found in SI figure 11. SI Figure 1. Thermogravimetric analysis of lead(ii)octylxanthate performed at two different heating rates: 1 C/ min (red) and 10 C/ min (blue). The apparent onset of decomposition is lower at the slower heating rate, suggesting that the precursor can decompose slowly at temperatures lower than previously expected. In both experiments a two-step decomposition is observed. SI Figure 2. Additional SEM images of nanofibres heated for 8 hours at 90 C. Red arrows in the upper left image highlight beading. A Histogram, based on a representative sample of 60 fibres, shows the distribution of fibre diameters, the mean diameter is found to be 1.8 ± 0.5 μm. SI Figure 3. Low magnification bright field TEM image of nanowires extracted and selected area diffraction pattern from the same region. The ring pattern is characteristic of a polycrystalline sample (as a large number of randomly oriented nanowires are present in the selected area) and indexes to PbS, confirming that the nanowires produced in the polymer fibres are crystalline PbS. SI Figure 4. EDX spectrum imaging of a bundle of nanowires. (a) HAADF STEM image of the region and Pb (b) and S (c) elemental maps. The spectrum image shows clear colocalisation of Pb and S. Cliff Lorimer quantification confirms that the elements are present in close to the expected 1:1 stoichiometry. SI Figure 5. Low magnification bright field TEM (a) and HAADF STEM (b-d) images of nanowires extracted from nanofibre mat after heating at 90 C. The PbS nanostructures grown are predominantly narrow, high aspect ratio, PbS nanowires (a-c). However, some large nanocubes are also found (d). Based on analysis of typical regions of the grid we estimate that less than 4% of nanocrystals formed are nanocubes. SI Figure 6. High resolution HAADF STEM images and corresponding Fourier transforms of PbS nanocrystals grown in polymer fibres. (a-d) show nanowires which are all found to be elongated in the [110] direction. (e-f) shows a PbS nanocube with {100} faces. SI Figure 7. Histograms showing size distributions of nanowire width and length, measurements were made from HAADF STEM images using representative populations of 100 nanowires. The average width is 3.4 ± 1.8 nm and the average length is 46.2 ± 20.1 nm. 22

23 SI Figure 8. Data form figure 4 presented with additional Fourier transform analysis. (b) shows the Fourier transform of the entire image in (a), (c) is taken from broad wire on the left of (a), while (d) is taken is taken form the bright narrow wire on the right of (a). SI Figure 9. Another example of the coexistence of narrow and broad wires, as in figure 4 (and SI figure 7) the broader wire (top of the image) is viewed down the [001] zone axis, while the narrow wires (below it) are viewed down the [110] axis. SI Figure 10. High resolution HAADF STEM images of PbS nanocubes grown in polymer thin films with (a) 5 wt. % PbS and (b) 30 wt. % PbS. Fourier transforms and enlarged regions of the atomic resolution image are also shown. In both cases the cubes are single crystalline PbS with {100} faces. SI Figure 11. HAADF STEM images of cross sectioned nanofibres. (a) shows 3 cross sectioned fibres, the fibre in the top right hand corner is shown in higher magnification in (b) and the area indicated by the red box in (b) is shown in (c). 23

24 Abstract Advanced electron microscopy techniques for mechanistic studies of the growth and transformation of nanocrystals A thesis submitted to The University of Manchester for the degree of Doctor of Philosophy in the Faculty of Engineering and Physical Sciences 2016 Edward Lewis School of Materials, The University of Manchester The morphology, composition, and distribution of elements within nanocrystals are critical parameters which dictate the material s properties and performance in a diverse array of emerging applications. The (scanning) transmission electron microscope ((S)TEM) represents a powerful tool for probing the structure and chemistry of materials on the nanoscale. Understanding of the mechanisms by which nanocrystals grow, transform, and degrade is vital if we are to develop rational synthesis routes and hence control the properties of the resulting materials. Electron microscopy represents a key tool in developing such an understanding. In situ techniques, where the material of interest is subjected to stimuli such as heat or a chemically reactive environment in the microscope, allow direct observation of dynamic transformations. Ex situ approaches, where multiple samples are prepared in the lab with the reaction parameters systematically altered, can also give important mechanistic insights. This thesis explores the use of both in situ and ex situ (S)TEM to gain insights into the growth and transformation of nanocrystals. Ex situ TEM is used to assess the structure of PbS nanocrystals in a polymer matrix, revealing new methods of morphological control through reaction temperature, precursor structures (appendix 4), and the processing of the polymer matrix (appendix 5). In situ techniques are used to observe the solution phase growth and shelling of nanocrystals (appendix 1) as well as the transformations of nanocrystals during heating in vacuum (appendices 2 and 3). The subjects of my in situ investigations are systems with heterogeneous distributions of elements. Historically, in situ electron microscope has been largely limited to imaging. However, to understand many dynamic transformations knowledge of changing elemental distributions is vital. For this reason, I have focused on the use of energy dispersive X-ray (EDX) spectroscopy to reveal changes in composition and elemental distributions during in situ experiments (appendices 1-3). This type of in situ elemental mapping is especially challenging for liquid-cell experiments, and my results represent the first report of EDX spectrum imaging for nanomaterials in liquid (appendix 1). 24

25 Declaration No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university of institute of learning. 25

26 Copyright statement I. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the Copyright ) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes. ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made. iii. The ownership of certain Copyright, patents, designs, trade marks and other intellectual property (the Intellectual Property ) and any reproductions of copyright works in the thesis, for example graphs and tables ( Reproductions ), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions. iv. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see in any relevant Thesis restriction declarations deposited in the University Library, The University Library s regulations (see and in The University s policy on Presentation of Theses. 26

27 Acknowledgements I would like to thank Dr. Sarah Haigh who has been an exceptionally enthusiastic, inspirational, and supportive supervisor. I have learnt a huge amount from working with her and couldn t have hoped for a better supervisor. I would also like to thank my co-supervisor Prof. Paul O Brien. I am extremely grateful for the support, guidance and expert knowledge Paul has given me during my PhD. A number of post-doctoral researchers have also helped guide my research. Dr. David Lewis has taught me a lot, sharing both his laboratory know-how and general wisdom. He is generous with his time, ideas, and when buying drinks. Dr. Paul McNaughter has been central to my work on PbS nanocrystals, his diligence and attention to detail in the laboratory is an inspiration. I would also like to thank Dr. Nicky Savjani for useful discussions about chemical synthesis and health and safety, and Dr. Eric Prestat for his expertise in all matters TEM related. I am extremely fortunate to have started my PhD in the Haigh group at the same time as Thomas Slater and Yiqiang (Kelvin) Chen who both proved to be great collaborators, teachers, and friends. I d particularly like to thank Kelvin for his help with SEM and Tom for his EDX expertise. A number of other PhD students have been a great source of help; I d like to thank Aidan Rooney, Lan Nguyen, and Mr. Aleks Tedstone for their comradery and support. I would also like to thank Nick Clark who kindly shared his clean room skills and graphene making expertise with me. Jack Brent has been an excellent collaborator providing me with exciting new 2D materials to study, Dr. Chris Page kindly allowed me to image his quantum dots, and Ben Coverdale electrospun polymer fibers for me. As an electron microscopist you are often reliant on collaborators, I have been extremely fortunate to work with some excellent scientists. I have greatly enjoyed working with Profs. Pedro Camargo, Jon Lloyd, Mark Green, and Richard Patrick, and Dr. Cinzia Casiraghi. In particular I d like to thank Prof. Nestor Zaluzec who was central to the liquid-cell EDX work and an extremely generous collaborator; I d also like to thank Prof. Grace Burke and Dr. Matt Kulzick for their key roles in this project. I have had the good fortune to work with some excellent undergraduate and Masters students. I learnt a great deal through working with Fiazal Hussain, Zheyang He, Zhongjie (Shawn) Yin, and Yadong (Albert) Zheng, I hope that I was also able to teach them something. 27

28 Matt Smith has been a huge help, both his training and his work looking after the microscopes are greatly appreciated. I d like to thank Dr. John Warren for his help with powder XRD, Dr. Jim Raftery for performing single crystal XRD, and Teruo Hashimoto for making ultramicrotome sections. I d like to thank all of those who taught and encouraged me throughout my education. In particular I d like to thank my Part II supervisors, Dr. Chris Blanford and Prof. John Foord, who provided me with an excellent introduction to research. I am especially grateful to Chris as without his guidance I wouldn t have ended up studying in Manchester. A rang of funding sources allowed me to carry out the research presented in thesis, I would like to acknowledge the funding which the Haigh group has received from the EPSRC, UK Grants EP/G035954/1 and tep/j021172/1, and the Defence Threat Reduction Agency, Grant HDTRA I would particularly like to thank the North West Nanoscience Doctoral Training Centre (NOWNano DTC) for training, funding and support, and the EPSRC doctoral prize for allowing me to continue my research post-phd. Finally, I d like to thank my parents, John and Joyce, and my girlfriend, Monica, for their love and support. 28

29 Chapter 1. Theory 1.1 TEM imaging To form a TEM image, a parallel beam of electrons illuminates a thin specimen. In the process of passing through the specimen some of these electrons undergo scattering events. The objective lens focuses beams emerging from a point in the specimen to a point in the image plane, creating a magnified image (chapter 1, figure 1). Beams scattered through the same angle are focused to a point in the back focal plane, forming a diffraction pattern. Either the image plane or back focal plane can be projected onto the phosphor screen or camera and recorded to produce a real space image or diffraction pattern. The electrons used to form the image or diffraction pattern can be selected through the use of apertures. The objective aperture is inserted into the back focal plane; it is used to select which diffracted beams will form an image, resulting in contrast based on whether regions of the specimen satisfy a certain Bragg condition. Figure 1. (a) Simplified Ray-diagram of a conventional TEM. Unscattered electrons are shown in blue while electrons scattered through an angle θ are shown in orange. The either the back focal plane or the image plane can be projected onto the camera to record diffraction patterns or images respectively. Apertures can be inserted into both planes. Selected area apertures in the image plane can be used to form diffraction patterns from a specific area of an image. Objective apertures in the back focal plane can be used to form (b) bright field images (selecting only the undiffracted beam) and (c) dark field images (selecting a diffracted beam, which has ideally been tilted onto the optic axis). 29

30 High resolution TEM (HRTEM) imaging allows individual atomic columns to be resolved. To form a HRTEM image a large objective aperture is used, allowing a large number of diffracted beams to interfere, generating a high resolution image in which the atomic structure of the specimen can be observed. The phase contrast mechanism underlying HRTEM makes interpretation of atomic resolution images complex. The sinusoidal contrast transfer function means that contrast inversions will occur at features with different spatial frequencies or if imaging conditions (such as defocus) are altered. 1-3 Consequently HRTEM images cannot generally be interpreted as simply as a direct representation of atomic positions. Image simulation is often required to draw precise structural conclusions from HRTEM data STEM imaging The STEM produces a focused electron probe which is scanned across the sample; the electron intensity detected at each point is used to build up an image pixel-by-pixel. The geometry of the detector(s) used determines the contrast mechanism of the image. Figure 2. Bright field (BF) and annular dark field (ADF) image formation in a STEM. A focused electron probe is rastered across the specimen and the intensity of electrons falling upon a detector is recorded at each position to build an image pixel-by-pixel. The choice of detector geometry dictates the contrast mechanism. Annular detector inner collection semi angles (α ADF-inner ) greater than 50 mrad are typically referred to as high angle annular dark field (HAADF), with collection of Rutherford scattered electrons leading to incoherent Z-contrast images, at smaller inner angles ADF imaging is coherent due to Bragg scattered beams falling on the detector. 30

31 STEM has the advantage of allowing several signals to be collected simultaneously; multiple detectors can be used, for example simultaneously acquiring bright field and dark field images (chapter 1, figure 2). Chemical information can also be acquired; spectrum images are formed by collecting an electron energy loss spectroscopy (EELS) or EDX spectrum at each pixel High angle annular dark field STEM In high angle annular dark field (HAADF) STEM only transmitted electrons that have been scattered to high angles (typically 2-3 time the probe convergence semi-angle or mrad) are detected. HAADF STEM has two important advantages over other imaging modes: it is an incoherent technique and image contrast is strongly dependent on atomic number (Z-contrast). 7-8 Electrons can be considered to have both particle and wave like properties. Rutherford scattering treats the electron as a charged particle interacting with the point charge of an atomic nucleus, the resulting elastic scattering has a differential cross section, describing the probability of scattering to a small solid angle dω at an angle θ, which is defined by equation 1. 9 dσ(θ) dω = e 4 Z 2 16(4πε 0 E 0 ) 2 (sin θ 2 )4 (1) Z is the atomic number of the nucleus, E 0 is the incident electron beam energy, ε 0 is the vacuum permittivity, and e is the elementary charge. 10 This equation predicts that the electron intensity at a scattering angle θ will be proportional to Z 2 of the nucleus. A more sophisticated model would account for the shielding of the nuclear potential by the atom s electrons. This can be represented by replacing (sin θ 2 )2 with (sin θ 2 )2 + ( θ 0 2 )2 to give equation 2. dσ(θ) dω = e 4 Z 2 16(4πε 0 E 0 ) 2 ((sin θ 2 ) 2 + ( θ 0 2 ) 2 ) 2 (2) Where θ 0 is the screening parameter. The screening parameter is dependent on Z and in practice acts to reduce the power of the Z dependence, particularly for low-angle 31

32 scattering. Consequently HAADF intensity doesn t show a Z 2 dependence, but a Z n relationship, with n values typically ranging from depending on the detector geometry used, the exact value depends upon the element and the detector angles If the above equations were an accurate description of all beam-specimen interaction one would expect a detector or aperture positioned to select electrons scattered to a limited range of angles to produce an image with Z-contrast. However, this is not the case, the interaction of electron s with a specimen required consideration of both the electron s particle and wave-like nature. When considering the electron as a wave, scattering process may be coherent or incoherent. Coherent scattering involves the summing of the amplitudes of scattered waves, resulting in constructive and destructive interference. This is the basis of both phase contrast and diffraction contrast in conventional TEM imaging as well as bright field (BF) and annular dark field (ADF) STEM. To understand the intensity seen in coherent images we must consider how the sample alters the electron wavefunction, 13 making direct interpretation of contrast challenging. If, however, the scattering intensities from individual atoms can be summed, the scattering is said to be incoherent. High-angle annular dark field (HAADF) imaging can be considered as an incoherent imaging mode with intensity arising from the sum of Rutherford scattering from individual atoms. incoherence of HAADF imaging is often explained as arising from the high inner angle of the annular detector (the minimum inner collection semi-angle for HAADF imaging is generally considered to be ~50 mrad) resulting in the exclusion of Bragg scattered beams The major advantage of incoherent imaging over coherent imaging is ease of interpretation. 17 In an atomic resolution HAADF STEM image bright dots directly correspond to atomic column positions, unlike coherent imaging modes where interference effects mean that modelling is required for detailed interpretation. 8 For liquid-cell experiments (see sections and and appendix 1) the Z-contrast available with HAADF STEM is especially advantageous as the goal is often to image higher Z material such as metal nanocrystals against the thick but low Z background of the water layer and SiN x windows. It has been shown that nanometre-diameter particles in micrometre thick liquid layers can be resolved by HAADF STEM but are not detected in bright field TEM images due to chromatic aberrations EDX and EELS Analytical information in the electron microscope arises from inelastic interactions between the electron beam and the specimen. Energy dispersive X-ray (EDX) spectroscopy and electron energy loss spectroscopy (EELS) both identify elements through core-ionisation 32 The

33 events. However, they do so by recording different signals: EELS measures the energy lost by incident electrons during scattering events, while EDX measures the energy of X-rays emitted as ionised atoms relax (chapter 1, figure 3). Figure 3. (a) A core ionization event, in EELS the energy of the transmitted electron is measured, while in EDX the energy of the X-ray emitted is measured; both energies allow the element to be identified. (b) The nomenclature of characteristic X-rays. Inspired by figure found in Brydson et al. Aberration-Corrected Analytical Transmission Electron Microscopy, Wiley 2011, EDX vs. EELS While EDX and EELS both offer the potential for chemical analysis in the (S)TEM, the two techniques have their own advantages and disadvantages. In EDX the complete spectrum is acquired simultaneously, this is useful for complex system containing multiple elements, as the peaks for all elements can be collected in a single spectrum. It is advantageous if unexpected elements are present, as they will not be overlooked by selecting energy windows based on the expected composition. EDX spectroscopy usually requires longer to acquire a significant signal than EELS. 20 While most electrons are forward scattered and collected by the EELS spectrometer, X-ray generation is isotropic and the design of the TEM makes collection of all X-rays a geometric impossibility. 21 In reality X-ray detectors cover only a very small percentage of the sphere through which characteristic X-rays propagate. The four detector ChemiSTEM TM system can achieve a 0.9 srad collection angle, 22 meaning ~7% of X-rays generated fall on the detectors. However, most current EDX systems have considerably smaller collection angles (~0.1 srad). Low X-ray counts also arise due to the fact that not all ionisation events will result in the generation of characteristic X-rays. 23 An ionized atom can either relax either via X-ray emission or via the Auger effect, a process where the energy released in filling a core vacancy is used to eject another electron from the same atom (chapter 1, figure 4a). 23 In light atoms the most likely consequence of a core 33

34 ionization event is Auger electron emission, while for heavier element X-ray fluorescence becomes the dominant effect (chapter 1, figure 4b). For example, in carbon (Z=6) the K- Shell X-ray fluorescence yield is ~0.3%, while in gold (Z=79) it is ~96%. 23 Consequently, EDX spectroscopy is better suited to the detection of heavier elements than light elements, 24 for which poor fluorescence yields and high absorption (by both the specimen and the detector window) present a problem In contrast, EELS is more sensitive to light elements and can be used to study specimens containing B, C, N and even H. 20, However, detection of light elements such as oxygen is possible with EDX, especially if windowless or ultrathin 22, 25 windowed detectors are used. Figure 4. (a) The Auger effect: energy released in filling a core-vacancy is used to eject another electron. (b) K- shell fluorescence and Auger yields plotted against atomic number, for light elements Auger emission dominates while for heavier elements X-ray fluorescence dominates, adapted with permission from Krause Journal of Physical and Chemical Reference Data 1979, 8 (2), While both EDX and EELS can be used to determine the localised composition of a specimen, EELS can give additional information that EDX cannot. Energy loss near edge fine structure (ELNES) in core-loss EELS peaks are characteristic of the local chemical bonding. 29, The bonding environment can be found by comparison with known fingerprints or modelling can be used to determine the origin of ELNES features. 26, 29, 31-32, 34 The oxidation state of an element can be seen in changes in the edge onset energy and ELNES features. 26, 33, 35 The low loss EELS spectrum (energy losses less than 50 ev) contains further useful information about the specimen, corresponding to the excitation of bulk- and surfaceplasmons, and interband-transitions. 9, 26 The energy of the bulk plasmon peak depends upon the density of valence electrons, consequently localised changes which affect the valence electron density can be detected in the low loss spectrum, for example, shifts in the 26, plasmon peak can be used to measure specimen temperature (see section 2.4.3). 34

35 Interband transitions, corresponding to the excitation of an electron from the valence band 26, to the conduction band, allow measurement of a material s band gap. However, EDX spectroscopy has a significant advantage over EELS if is necessary to study thick samples. While beam broadening will reduce spatial resolution, EDX spectra can be acquired regardless of how thick the specimen is (as demonstrated by the technique s widespread use in scanning electron microscopes (SEMs)) making the technique attractive for liquid cell experiments (appendix 1). In contrast thin specimens are crucial for EELS analysis. In thicker samples, electrons undergo plural scattering with a high probability that electrons that have experienced a core-loss scattering event will also undergo one or more low-loss scattering events. 40 The effect of plural scattering on the shape and intensity of core-loss edges means that quantitative EELS analysis becomes challenging or impossible While in many cases thick samples can be avoided by good sample preparation, they are often unavoidable in liquid-cell experiments (see section and 2.4.4) While the two techniques have different strengths and weaknesses, in many cases they are 24, 45 complementary and conveniently both signals can be gathered simultaneously. 1.4 Electron beam-induced processes Introduction The elastic and inelastic interactions of the electron beam with the specimen, which are vital for all imaging and analytical techniques in the electron microscope, can lead to chemical and structural changes. The nature and extent of beam induced changes will vary dramatically with both the imaging conditions and the nature of the specimen. 9 For certain classes of sample (for example polymers) or certain imaging conditions (for example high magnifications or extended acquisition times) beam damage can be a significant problem, resulting in the specimen under investigation no longer being representative of the sample from which it was taken. In such cases an understanding of the mechanism of beam damage can allow rational modification of sample preparation techniques and imaging conditions to mitigate this damage. 9 Beam-induced changes are generally considered to be an unwelcome source of artefacts, however, they can be useful for in situ studies where the electron beam can be used to drive dynamic transformations (as discussed in section 1.6) In situ experiments involving beam-induced transformations have been an important part of my research. I have used the electron beam to drive nanocrystal shell growth in liquid-cell studies (appendix 1) 44 and hollowing and inversion of bimetallic particles during in in situ heating experiments (appendices 2-3), 51 35

36 This section will discuss the principle forms of electron beam damage; looking at the damage mechanism, symptoms, and methods of prevention Beam damage Radiolysis Electron-electron interactions lead to inelastic scattering events which can break chemical 9, 52 bonds, a process known as radiolysis. Inelastic scattering events cause electronic excitations; in semiconducting or insulating materials electrons are excited from the valence band to the conduction band, forming an electron hole pair. The valence band vacancies in such materials are relatively long lived, meaning there is sufficient time for bonds to break. 53 In contrast, radiolysis is not generally a problem in samples with good electrical conductivity as vacancies are rapidly filled due to the ready supply of mobile electrons. 53 Radiolysis is common in polymers which either break down or cross link under the beam (a problem for electron microscopy but vital for electron beam lithography 9, 54 resists). The problem of charging is closely related to radiolysis: when inelastic scattering events occur near the surface of a specimen, secondary electrons escape into the vacuum and the specimen accumulates a positive charge. In an electrically insulating specimen this build-up of positive charge can be detrimental to imaging or even destroy the sample. 53 Low dose acquisition procedures can be employed to reduce radiolysis damage. Techniques to minimise the electron dose include minimising pre-image exposure by focusing using a 53, 55 nearby region, or averaging over multiple structurally identical objects. Cryogenic temperatures also appear to reduce radiolysis damage, with mass loss reduced in cooled specimens. 52, 56 Encapsulation of specimens also prevents both mass loss and amorphisation. 52, 56 Encapsulation with both insulating and conducting materials is equally successful in preventing damage; suggesting that the main role of encapsulation is to provide a diffusion barrier, assisting recombination of broken bonds and preventing the loss 53, 57 of volatile species. Thick films may be more radiolysis resistant due to the same diffusion barrier effect. 53, 58 Encapsulation has recently proved useful for reducing electron beam damage in two-dimensional (2D) materials, where MoS 2 has been protected by graphene encapsulation Radiolysis is expected to be independent of accelerating voltage as although damage cross sections increase at lower voltages so do imaging and analysis signals

37 Heating Inelastic scattering events can also deposit energy in the specimen by exciting phonons. The extent of sample heating will depend dramatically on the thermal conductivity of the sample, as only a relatively small amount of the specimen is irradiated the rest of the sample can act as a heat sink. 52 For samples with high thermal conductivity temperature increases are usually insignificant. 53 However, for insulating samples temperature increases can be large; this is especially relevant in polymers and biological specimens where thermal conductivity is poor and only relatively modest temperature changes are required to induce structural deformation Significant beam-induced temperature increases are believed to have been observed in some metal nanoparticle samples, this is believed to be due to low thermal diffusion rates due to the poor contact between a spherical nanoparticle and the support membrane Displacement Electron nuclear interactions lead to elastic scattering events which can displace atomic nuclei. 52 This is the principle damage mechanism in conducting materials (where radiolysis damage is prevented by the ready supply of mobile electrons). 9, 63 Generally cross sections for displacement are far smaller than those for radiolysis, so for insulating materials radiolysis far outweighs displacement damage. 63 specimens only one damage mechanism is dominant. 63 This conveniently means that in most In electron-nucleus interactions energy transfer is scattering angle dependent, with maximum energy transfer (E) occurring when the scattering angle θ=180 (equation 3). 63 E = E max sin 2 ( θ 2 ) (3) Figure 5. Plot of energy transferred from a beam electron to a nucleus as a function of the angle (θ) through which the electron is scattered (equation 3), maximum energy transfer occurs when the electron is scattered at an angle of 180 degrees. 37

38 The maximum energy transferred (E max ) depends upon the energy of the incident electron and the atomic mass of the nucleus it is interacting with. E max increases with increasing accelerating voltage and decreases with atomic mass. An approximate relationship for the dependence of E max on incident electron energy (E 0 ) and nuclear mass (M) is given in equation 4, where m 0 and c refer to the electron rest mass and the speed of light respectively. 63 E max = 2E 0(E 0 + 2m 0 c 2 ) Mc 2 (4) Figure 6. Plot showing the maximum energy transferred to a selection of nuclei, Fe (Z=26), Al (Z=13), O (Z=8), C (Z=6), and Li (Z=3), by electrons accelerated to a range of voltages from kev (equation 4). This plot demonstrates why knock-on damage is most problematic for light atoms and high accelerating voltages. Displacement damage can either result in a nuclei being moved to an interstitial site, termed knock-on damage, or if the displaced atom is near the surface, ejection of the atom from the sample, a process known as sputtering. 9 The bonding environment and atomic number of an element will determine the energy required to displace it. Due to their 9, 63 unfulfilled valence, surface atoms are more readily displaced than those in the bulk. Knock on displacement of bulk atoms generally requires energy transfers in the ev range, while sputtering may only require a few ev. 63 Consequently knock on damage at common TEM voltages will only be a problem in samples containing light elements, while sputtering will affect a far wider range of samples (chapter 1, figure 6). 63 Interestingly carbon coating has been shown to reduce sputtering, however, protection is only temporary as the carbon is eventually sputtered away

39 While generally viewed as an artefact to avoid, knock-on damage in high voltage TEMs can be useful for studying nuclear materials as it can be treated as analogous to neutron 9, damage. An important characteristic of displacement damage is the damage threshold; a beam energy below which no damage will occur i.e. E max is less than the displacement energy. Once the damage threshold is known, displacement damage can be eliminated by performing experiments at a suitably low accelerating voltage. However, lower voltages also have associated disadvantages, most notably reduced imaging resolution and specimen thickness requirements. 67 The severity of this problem has been considerably reduced by the advent of aberration correction which means that atomic resolution imaging and spectroscopy now possible at voltages as low as 30 kv In recent years one of the most prominent areas of (S)TEM research has been studies of the two-dimensional material graphene. 32, Graphene is made of carbon and is atomically thin, so essentially all atoms are surface atoms. Consequently displacement damage is a significant problem and atomic resolution imaging of monolayer graphene at typical voltages of kv is almost impossible. This has led to the widespread use of aberration corrected machines operated at voltages of 80 kv or lower. Pristine graphene is stable during high resolution imaging at 60 kv, 67, 72 however, edges and defects will still damage at his voltage. 27 Lower accelerating voltages have other advantages besides reducing displacement damage. The increased elastic and inelastic cross sections increase signal for both imaging and spectroscopy, which is advantageous for very thin specimens. 73 Lowering the accelerating voltage also reduces EELS delocalisation effects, making single atom spectroscopy possible Contamination Hydrocarbon molecules on the surface of the specimen can be polymerised by the electron beam, forming a layer of contamination over the imaged area. Surface diffusion of mobile carbon to the illuminated region causes this contamination to accumulate as imaging continues, building a thick layer over the region of interest. 9, 52, 74 However, the presence of surface contamination can severely limit imaging, as the signal from the contamination swamps that from the features being investigated. Contamination generally increases with both probe current and approximately linearly with magnification. 74 However, at very high probe current hydrocarbon build-up is mitigated by sputtering

40 Historically hydrocarbon contamination was intrinsic to much electron microscopy due to poor vacuum systems. 9, Improvements in the vacuum system s pressure and cleanliness and the use of oil-free pumps means that the microscope is now rarely a source of contamination However, contamination is often still a problem if hydrocarbons are present on the specimen and in such cases, procedures to clean the sample, such as plasma 9, 52, 74 cleaning or UV-ozone treatment, are necessary prior to imaging. 1.5 In situ electron microscopy In situ electron microscopy has been a central focus of my research. I have used liquid-cells to study nanocrystal growth, focusing on the development of a new technique which allows EDX spectrum images to be acquired in situ (appendix 1). And I have in situ heating holders to study the changing elemental distributions and morphology of bimetallic nanocrystals at elevated temperatures (appendices 2 and 3) In situ heating holders TEM sample holders which allow the specimen to be heated to elevated temperatures in the microscope, can be divided into two design which I ll refer to as traditional holders and MEMS based holders. Traditional holders, have a heating element built into the holder tip and heat a standard 3 mm disc TEM sample, whereas microelectromechanical systems (MEMS) holders require the sample to be transferred to specially designed chips. The MEMS based holder passes an electrical current through a semiconducting membrane, resulting in resistive heating of a small region of the chip (chapter 1, figure 7) When studying solution dispersible nanomaterials sample preparation is relatively easy with both setups, however, when samples must be thinned from the bulk, attaching a suitable specimen to the heating chip membrane can be challenging. 77 Traditional holders heat a far larger thermal mass and this limits the rate of temperature change and introduces greater sample drift during heating. MEMS based holders are therefore well suited to high 75, resolution imaging, where minimal drift at high temperatures is critical, and achieve rapid temperature changes (over 1000 C in a few milliseconds). 82 can Another disadvantage of the traditional holder design is that the furnace surrounding the sample blocks the line of sight between the specimen and the microscope s EDX detector(s), 83 in contrast EDX analysis during heating experiments is possible when using MEMS based holders due to the flat membrane based heater. 51, 82, While membrane based heaters open up the possibility of in situ EDX analysis, there is still potential for artefacts; infra-red (IR) emission from the hot sample can cause spectral shifts and loss of resolution. 83 IR radiation only becomes a significant problem above ~500 C and is more severe when using windowless detectors. 83

41 Figure 7. Schematic of a MEMS based heating chip (Protochips Aduro), viewed in cross section (a), plan-view (b), and with heating membrane enlarged (c). The sample is supported on electron transparent membranes spanning the holes in the low conductivity ceramic heating membrane. The holder can place a potential across the metal contacts, resulting in resistive heating of the ceramic heating membrane. Diagram inspired by figure from Asoro et al. ACS Nano 2013, 7, (9), Liquids and gases in the electron microscope TEMs typically require a vacuum of 10-6 mbar or better at the specimen: these high vacuum conditions are necessary both to protect the electron source and to minimise unwanted scattering events. However, the majority of important processes in the physical sciences (such as the growth of crystals, the corrosion of materials, or the action of catalysts) do not occur in high vacuum. The ability to perform TEM studies on samples in a liquid or gaseous environment is therefore attractive. There are two solutions to this problem, one involves extensive modification to the microscope column in order to create a dedicated environmental TEM (ETEM), and the other involves using sealed specimen holders with electron transparent windows which can be used in conjunction with unmodified 46, microscopes. ETEMs are a powerful tool for in situ observation of dynamic processes in a gaseous atmosphere, they employ apertures and differential pumping to maintain a thin layer of gas 88, 93 around the specimen while retaining high vacuum in the remainder of the column. 41

42 Using such microscopes it is possible to directly observe transformations of nanocrystals in gas at atomic resolution and this approach has been used to gain important insights into the structure and properties of nanocatalysts. 87, 89 However, the ETEM has a number of drawbacks, the requirement for a complex and expensive dedicated instrument means that ETEMs are relatively rare. Additionally, ETEMs are typically operated at low gas pressures (often <5 mbar) Considering that many industrially relevant catalytic processes are carried out at pressures of over 10 5 mbar a considerable pressure gap exists between in situ observations and real world conditions. 91, 99 Also, although liquids can be injected into an ETEM the low pressure of the system means that this generates a vapour around the 88, 90 specimen instead of immersing the specimen in liquid. Closed-cell specimen holder designs allow imaging of samples in a liquid environment (liquid-cell), 47, 90, 100 or in gas at pressures of up to 1000 mbar (gas-cell) This approach relies on two impermeable but electron transparent windows, the windows protect the microscope vacuum and differential pumping is not required. No modifications need be made to the microscope, the same holder can often be used in multiple microscopes. In 2003 Ross and co-workers demonstrated the first modern liquid-cell design, using MEMS fabrication techniques to produce a silicon nitride windowed liquid-cell to image dynamic processes occurring in liquid in the TEM. 90 While early liquid-cell systems were homemade, 90 in subsequent years several commercial holders have become available allowing widespread access to the technology. A number of variations to the basic liquid-cell design exist: closed-cells trap a fixed volume of liquid between two silicon nitride windows, 46, 101 while flow-cells allow liquid to be flowed through the imaged region via tubing built into the TEM holder, providing the possibility of introducing new reactants during an in situ TEM experiment. 19, Chapter 1, figure 8 shows a cross sectional diagram based on the Protochips TM Poseidon design of commercially available flow cells (as used in appendix 1). More sophisticated liquid-cell designs also allow electrochemistry to be performed in situ using electrodes patterned onto 90, the cell s windows. This technique has been used to observe electrochemical processes relevant to the performance of lithium ion batteries, , , the growth of lead dendrites on a gold electrode, 110 and the corrosion of aluminium

43 Figure 8. Cross sectional diagram of liquid flow-cell. 1) screw, 2) top plate, 3) holder body, 4) O-ring (outer), 5) flow channel (in), 6) top chip (SiN x window), 7) spacer (gold), 8) bottom chip (SiN x window), 9) O-ring (inner), 10) flow channel (out). A thin layer of liquid (typically ranging from 50 nm to over 1 µm depending on the choice of spacer) is trapped between the windows of the top and bottom chips, imaging electrons pass through both windows and this liquid layer. The flow channels are connected to tubing which can be connected to a syringe pump and used to flow new solutions into the cell during imaging. 1.6 Beam-induced processes in liquid-cell experiments Beam-induced nanocrystal growth In liquid-cell experiments beam-induced transformations are not necessarily undesirable, for example, most studies of nanocrystal growth, to date, have deliberately used the electron beam to initiate the nucleation and growth of nanocrystals , 103, A wide range of nanocrystals have been observed to grow in situ by exploiting electron beaminduced deposition. 47 Nanostructures grown by beam-induced processes in liquid-cell experiments include Ag, Au, 116 PbS, 103 Si, 117 ZnO, 118 Pt 3 Fe, 112 and Pt 46. Beam induced deposition of Cu is observed in my work (appendix 1). Beam induced reduction of metals is often attributed to the action of the hydrated electron (equation 5). 47 However, hydrogen radicals can also act as reducing species (equation 6), 114 while the hydroxide radical can oxidise metal atoms, slowing nanocrystal growth (equation 114, 119 7). M n + e M n-1 (5) M n + H + H 2 O M n-1 + H 3 O + (6) M n-1 + OH M n + OH (7) 43

44 The growth of silver nanocrystals from aqueous solutions of Ag(I) salts has been used as a model system for beam-induced nanocrystal growth , 120 Nanocrystal growth is found to require imaging above a threshold dose rate; hence sub-threshold conditions allow imaging without growth. 114, 120 Electron irradiation does not immediately form nanocrystals, particles are only detected after a nucleation induction time has elapsed. This nucleation induction time is dose rate dependent with higher dose rates giving shorter induction times. 114 During imaging the number of particles increases, eventually reaching a maximum where no new particles are formed. 114 After this point the number of particles in the imaged area decreases due to coalescence events and the repulsion of particles from the imaged region. 114 After new particles have stopped being produced the existing particles continue to grow, however, growth rates are usually highest initially and slow as imaging proceeds. 114 These observations are all consistent with a mechanism where an initial supersaturation condition must be met to initiate nucleation and subsequent growth depletes the solution of precursor molecules. 114 By altering the imaging conditions it is possible to influence the growth mechanism. Electron dose rates only slightly above the threshold give faceted nanoparticles while spherical particles are produced by higher rates. 120 At high dose rates growth occurs by a diffusion limited mechanism with an excess of reducing agents (hydrated electrons) meaning that growth is dictated by how fast precursor can diffuse to the growing nanocrystals. In contrast, at lower does rates, growth occurs in a reaction limited regime where there is a shortage of reducing agents, consequently precursors do not immediately reduce upon reacting the surface of a growing nanocrystal and reduction only occurs on low energy facets. 114 Considering only dose rate and precursor concentration is, however, insufficient. Different accelerating voltages, but otherwise identical imaging conditions, give very different growth mechanisms, it is found that particles grown at 300 kv are far more faceted than those grown at 80 kv, suggesting that using a higher voltage is in some ways analogous to using a lower dose rate, this would be consistent with the lower cross sections of inelastic scattering at higher energies. 120 The difference between TEM and STEM imaging modes and even different dwell times in STEM imaging can also dramatically affect the mechanism of nanocrystal growth , 120 The fact that the current literature is comprised of reports of beam-induced processes observed under dramatically different imaging conditions makes useful comparison between studies challenging. Beam driven reactions will not necessarily replicate the mechanistic features of lab scale chemical synthesis. The reliance on beam driven processes is in part a reflection on the infancy of the field; processes commonly used in the chemistry laboratory to initiate 44

45 reactions, such as the rapid mixing of two solutions or heating of a solution are still extremely challenging or impossible to perform in situ. 101 Future developments in liquid-cell holder design are likely to address these shortcomings. The current focus on beam-induced reactions can also be viewed as crucial ground work for the future development of the field; it is vital that the effect of the electron beam is understood so that artefacts can be , 121 identified and minimized through optimal choice of imaging parameters Radiolysis of water The field of radiation chemistry predates the recent surge of interest in imaging liquids in the TEM and consequently a large body of information about the reactive species generated during water radiolysis already exists The chemistry of water radiolysis is assumed to be identical regardless of whether gamma radiation or electron beam irradiation is involved, however, it should be noted that the dose rates frequently reported in radiation chemistry literature are orders of magnitude lower than those routinely encountered during (S)TEM imaging. 124 Work on nanoparticle growth frequently identifies the hydrated electron as the key radiolysis product. 47, 113, However, it is just one of many reactive species generated during the radiolysis of water. The Initial radiolysis of water generates H 3 O +, OH, H, hydrated electrons (e (aq)), and excited water (H 2 O*) within seconds of the initial ionisation event (equations 8-11). 126 H ionizing radiation H 2 O + + e - H 2 O* (8) H 2 O + + H 2 O H 3 O + + OH (9) H 2 O* OH + H (10) e + nh 2 O e (aq) (11) These radiolysis products may diffuse and further react in a cascade of further reactions to form additional species such as OH, H 2, and H 2 O 2 (equations 12-19) e (aq) + OH OH (12) e (aq) + H 3 O + H + H 2 O (13) 45

46 2e (aq) + 2H 2 O H 2 + 2OH (14) e (aq) + H + H 2 O H 2 + OH (15) OH + OH H 2 O 2 (16) H + H H 2 (17) H + OH H 2 O (18) H 3 O + + OH 2H 2 O (19) All of the radiolysis products listed above are potentially reactive, with the exception of molecular hydrogen which can form gas bubbles. 120, 124 Considering how many liquid-cell studies rely on radiolysis products to drive the dynamic processes occurring, 46, 112, 114, 116 It is surprising that in most of these studies no attempt has been made to estimate the local concentration of the key reactive species. To quantitatively interpret beam-induced processes in liquid knowledge of the concentrations of radiolysis products is required. 127 This is a complex problem requiring consideration of multiple interrelated chemical reactions. Further complications lie in the fact that the electron microscopist will not be imaging pure water, liquid-cell experiments will either be hoping to observe a pre-existing structure in a hydrated environment or the growth of solid structures from the solution, which will entail precursor molecules to be present in the solution, further complicating the chemistry. Scavenging of one radiolysis product by an additive will affect the concentration 119, 126, 128 of other radiolysis products. Initial progress towards quantitative understanding of radiation chemistry in liquid-cell experiments was made by Schneider et al. who developed a model for irradiated water by considering the kinetics of dozens of interrelated chemical reactions. 127 An important result that emerges from their model is that under continuous irradiation, regardless of the electron dose rate, the concentrations of all radiolysis products will reach a steady state (chapter 1, figure 9a) at which the concentrations is dose rate dependent (chapter 1, figure 9b). The time required to reach a steady state, during homogeneous irradiation, is on the order of 1 ms: while fast compared to typical frame rates in liquid-cell (S)TEM experiments (~1s) it is considerably longer than typical pixel-dwell-times. 127 Consequently it is unclear whether steady state concentrations are a reasonable assumption during STEM imaging. 46

47 Figure 9. Schneider et al. s model of electron beam radiolysis of water suggests that radiolysis produces reach steady state concentrations within ~10-3 s of commencing imaging; these steady state concentrations depend strongly on the dose rate of electron irradiation. (a) Concentration of radiolysis products as a function of time in pure deaerated water irradiated homogeneously at a dose rate of 7.5 x 10 7 Gy/s. (b) The steady state concentrations of radiolysis products as a function of dose rate in pure deaerated water subject to homogeneous irradiation. Figure adapted with permission from Schneider et al. J. Phys. Chem. C 2014, 118 (38), Copyright 2014 American Chemical Society. In a typical liquid-cell experiment only a small proportion of the total volume of liquid is subjected to electron irradiation. This is particularly true of STEM image where only the 114, 121 scan area is irradiated, Schneider s model also attempts to consider the case of inhomogeneous irradiation, where diffusion of radiolysis products out of the irradiated region needs to be considered. 127 reached but takes longer to become established. 127 It is found that in such a case a steady state is still Under typical STEM irradiation conditions, steady state concentrations are predicted to be reached in a few seconds for most reactive species. 127 Modelling of the concentration of reactive species promises a more quantitative understanding of beam-induced processes, potentially allowing more sensible comparisons between studies. It may also help improve the links between mechanistic insights gained in the microscope and rational synthesis in the lab. For example, in cases of metal nanoparticle growth by reduction of a metal salt, Woehl et al. have argued that electron beam-induced growth of nanocrystal is nearly analogous to conventional chemical reduction of nanocrystals. 114 If this is the case, knowing the concentration of radiolysis products responsible for specific growth mechanism or nanocrystal morphology should be able to guide the choice and concentration of reducing agents employed in ex situ chemical synthesis of the same system. Schneider s model also suggests that different radiolysis products have vastly different diffusion lengths. 127 For example, the hydrated electron is a highly reactive species with a lifetime of approximately 5x10-4 s, consequently it is consumed in chemical reactions 47

48 before it can diffuse far from the irradiated region (diffusion lengths on the order of 10 nms). In contrast the far less reactive radiolysis product H 2 are predicted be found in significant levels many micrometres away from the irradiated region. 127 The short diffusion length of the hydrated electron is convenient for those interested in using the reduction of metal salts to pattern fine features with the electron beam. 124 The initial presence of various solutes can also affect the steady state concentrations of radiolysis products. Unless water is deaerated, aqueous solutions will contain dissolved molecular oxygen. It has been shown that should act as a scavenger of hydrated electron, reducing their concentration at low dose rates. 127 Dissolved oxygen is also expected to decrease the concentrations of OH and H, while increasing the levels of H 2 and H 2 O The presence of secondary alcohols is known to scavenge the oxidising hydroxyl radical (equation 20), yielding products that are reducing (equation 21), 119 while dissolved nitrous oxide is known to scavenge hydrated electrons (equation 22). 128 R-HCOH-R + OH R- COH-R + H 2 0 (20) M + + R- COH-R R-CO-R + M 0 + H + (21) e aq + N 2 O + H 2 O N2 + OH + OH (22) Although Schneider s model represents a valuable first step towards a more empirical approach, there are a number of questions regarding its accuracy. There is currently a complete absence of experimental measurements obtained in situ, consequently there is no experimental data to build a model of the electron beam-induced radiolysis of liquids from or to test the validity of the model against. The development of methods for directly measuring concentrations in situ would represent an important breakthrough in the field; it is conceivable the electrochemical liquid-cells could be used to perform such measurements. The dose rates generated in a typical liquid-cell experiment are approximately 7 orders of magnitudes greater than the radiation sources commonly used to study radiolysis. 127 As the G values (the number of molecules created or destroyed per 100 ev of energy deposited) depend upon the energy and type of radiation used, the accuracy of the values upon which Schneider s model is build are unclear. Furthermore, the model only considers the primary interaction of the electron beam with the liquid layer. It seems possible that interactions with the SiN x windows, and radiolysis events caused by secondary electrons and X-rays could also be important. It is well known from synchrotron 48

49 experiments that X-ray radiation can alter the structure and chemistry of a specimen For example, it has been shown 10 kev X-ray radiation (X-rays in this energy range are frequently generated in the electron microscope and detected in EDX spectra) effect the crystallisation of lithium disilicate glass, increasing the number of nucleation sites formed during heat treatment. 129 A more sophisticated model of beam induced radiolysis should include not just the interaction of the primary beam with liquid layer but the effect of secondary radiation such as X-rays and secondary electrons. Although liquid-cell experiments can and have been performed using a wide range of solvents, to-date attempts at modelling beam-induced radiation chemistry have been limited to water. This partly reflects the importance of water (as a solvent form metal nanoparticle synthesis, 113 battery electrolytes, 108 biological structures, 131 corrosion processes, 111 etc.). Experiments carried out using other solvents typically show similar behaviour to their aqueous counterparts, such as the beam-induced growth of metal nanocrystals from solutions of the metal salt, suggesting at least qualitative similarities with aqueous radiation chemistry. However, future work will need to aim for quantitative understanding of radiation chemistry in non-aqueous liquid-cell experiments Closed-cell and flow experiments The majority of liquid-cell experiments have employed a closed cell, where a volume of reaction solution is sealed in the cell prior to the experiment and cannot be replenished or removed during the experiment. 46, 101, 112, 118, Due to the accumulation of reactive species and depletion of reactants (for example silver ions in the silver nanocrystal growth experiments described previously), the outcome of closed liquid-cell experiments is dependent on the imaging history of the sample. 120 If a second experiment is carried out in a different area of the window, the initial concentration of reactants will be lower than in the first experiment carried out in the same cell, while the initial concentration of radiolysis products is likely to be higher. This phenomenon is dramatically illustrated by Abellan et al. using the model system of Ag(I) reduction, identical STEM imaging experiments carried out on a fresh cell and on the same cell after extended imaging have been found to grow considerably different nanoparticle morphologies. 120 The fresh cell yields a large number of nanoparticles some of which are faceted and others are rounded, the later experiments in the depleted cell yield slower growth rates, far fewer particles, and spherical particles with no faceting. 120 The problem of previous experiments altering solution chemistry can be overcome by using a flow-cell where a stock solution is continually flowed into the holder 103, 114, tip, meaning that reactants can be constantly replenished and by-products removed; assuming enough time is left between consecutive experiments the initial 49

50 121, 137 solution chemistry should be the same each time. Flow-cell the advantage of minimising charging and temperature gradients. 19 experiments also have Bubble formation 121, 124 A commonly observed liquid-cell artefact is the formation of gas bubbles in the liquid. This can be a useful phenomenon, in some cases gas bubbles are deliberately used to create a thinner liquid layer and therefore improve imaging resolution For example, Klein et al. are able to image 1.4 nm diameter gold particles in water by generating a gas bubble with intense TEM irradiation. 19 artefact. 141 However, gas bubbles are often an undesirable Initially some questions existed about the nature of bubbles observed in electron irradiated water, with the possibility that they were either radiolytically generated hydrogen gas or water vapour due to electron beam heating. 19 The latter suggestion can be discounted as calculations suggest that typical electron beam heating in a liquid-cell sample should be less than 2.5K. 124, Bubbles are therefore attributed to the beam-induced formation of molecular hydrogen causing dissolved H 2 levels to exceed the saturation concentration resulting in nucleation and growth of H 2 bubbles. 145 Bubble growth rate increases as electron dose rates increase. 124 Experimental evidence for bubble composition come from EELS studies of frozen sucrose and protein solutions, which show that bubbles formed under electron irradiation contain molecular hydrogen. 146 Whether the bubble are solely a consequence of hydrogen evolution is questionable; liquid-cell experiments using degassed solutions show reduced bubble formation, suggesting that the bubble formation process may also involve dissolved gasses. 120 As with other radiolysis processes, solution chemistry will affect the levels of radiolysis products. 127 It is known that aqueous solutions containing organic molecules are more prone to beam-induced bubble formation than pure water. 143, 146 It is believed that this is a consequence of the reaction of hydroxide radicals with carbon-hydrogen bonds (equation 23). 143 R-H + OH RO + H 2 (23) Interestingly bubbles can be manipulated using the electron beam. 144 While bubbles grow at high magnifications they have been observed to shrink and disappear under lower magnification imaging conditions. 144 Huang et al. were able to observe the growth trajectories of nanobubbles in an aqueous protein solution. 143 They show that bubbles demonstrate Ostwald ripening, with H 2 molecules lost by smaller bubbles and gained by 50

51 larger nearby bubbles, with the distance of smaller bubbles from a larger bubble dictating their growth rate. 143 Grogan et al. observe periodic nucleation, growth and detachment of bubbles from a specific point on the SiN x window of the liquid cell, suggesting defects or the contamination may act as nucleation sites Charging Electron specimen interactions can generate secondary and Auger electrons, some of these electrons can escape from the specimen resulting in the build-up of a net charge. 121 Studies of the SiO 2 -water interface show that electrons become solvated in the water while holes remain trapped in the solid, 147 a similar processes is likely to occur in liquid-cell windows, leading to positively charged SiN x windows and a negatively charged liquid layer. 121 Furthermore, charged radiolysis products can react with nanostructures in solution, altering their surface charge. 148 Consequently electric fields and charged particles are likely to be present in many liquid-cell experiments and can influence the observed motion of nanostructures. 121, 144, 148 As well as attracting charged nanocrystals, strong electrostatic attraction between opposing windows can result in the collapse of the cell. Charging related artefacts can potentially be reduced by depositing thin films of conducting material on the SiN x windows to prevent build-up of charge and by adding an electrolyte to solution to screen charged particles from one another Synthesis of hollow nanostructures Templates are frequently employed to achieve hollow nanostructures of a desired size and shape. These reactions can be divided into three classes: hard templating, soft templating, and sacrificial templating. 149 Hollow nanostructures are of particular relevance to the work presented in appendices 2 and 3, where the AgAu starting material was synthesised by Galvanic replacement and the nanoscale Kirkendall effect is observed in situ Hard templating The hard templating approach consists of four distinct reaction steps: a template of the desired shape and dimensions is first synthesised, the surface of this template is then modified to overcome any incompatibility between the template and shell materials, a target material is then coated onto the template to form a shell, finally the template is selectively removed, leaving the hollow shell intact (Chapter 1 Figure 10). 149 This approach was pioneered by Caruso et al. who coated polystyrene nanospheres with silica and then removed the organic core material by heating to 500 C or dissolving in organic solvents. 150 The hard templating approach has been used to synthesise hollow metal oxide, metal sulphide, and metal nanostructures. 155 While offering excellent control over size and 51

52 shape, it has some disadvantages: the multistep synthesis required with hard templating makes high product yields challenging, 149 and aggressive treatments like HF etching are often required to remove the template. 155 There may also be problems with the structural integrity of the shell after template removal; for example, swelling of a polymer template as it is dissolved in an organic solvent can rupture the shell. 149, 151 It is also challenging to encapsulate guest molecules in hollow structures made by hard templating, an important requirement for drug delivery applications. 149 Figure 10. Schematic illustrating the hard templating approach to the synthesis of hollow nanostructures. A template of the desired size and shape is synthesised (a), the surface of the template is modified to allow deposition of shell material on the template (b), and a shell of the target material (red) is grown (c), the template material (blue) is then selectively etched, leaving a hollow shell of the target material (d) Soft templating Soft templating refers to the use of a liquid or gaseous template (as opposed to the solid templates used in hard templating). 149 Oil-in-water or water-in-oil emulsions can be used as soft templates, where the shell is deposited at the oil-water interface For example, toluene microdroplets containing Ti(OBu) 4 form hollow TiO 2 microspheres in 1-butyl-3- methylimidazolium hexaflurophosphate ([C 4 mim]pf 6 ) by hydrolysis of the titanium precursor at the interface between the two phases, a consequence of water in the [C4mim]PF The fact that the template in an emulsion based synthesis is a liquid means that species which are soluble in the nanodroplet phase can be readily incorporated into the hollow nanostructure, 149 in the above example it was demonstrated that both carboxylic acids and gold nanoparticles could be introduced into the hollow TiO 2 through their addition to the emulsion s toluene phase. 156 Other widely used soft templates include gas bubbles, and micelles Compared to hard templating approaches, soft templating approaches typically allow less control over morphology and monodispersity is hard to achieve Sacrificial templating As for hard templating approaches, a solid nanostructure is used as the template in 149, 161 sacrificial approaches. However, in sacrificial reactions the template is directly involved in and consumed during the shell forming reaction. 149, 161 In common with hard

53 templating, established methods of shape control in the synthesis of solid nanocrystals mean that a wide range of morphologies and excellent monodispersity are achievable in template structures However, sacrificial approaches are considerably simpler and more elegant than hard templating: as they are typically performed in a one-step synthesis 161, procedure there is no requirement for surface modification and template etching. Two of the most powerful and widely used sacrificial routes to hollow nanocrystals, the galvanic replacement reaction and the nanoscale Kirkendall effect, which have both featured in the research presented in appendix Galvanic replacement reactions Galvanic replacement relies on a redox reaction between two species, generally a sacrificial metal template and a dissolved metal salt The redox couple is chosen so that the simultaneous oxidation of the template material and reduction of the metal salt is 166, 168, 172 thermodynamically favourable. Reactions 167, 169, 173 especially common, the involving two noble metals are reaction between Ag nanoparticles and HAuCl 4 was the first reported example of a nanoscale galvanic replacement reaction, 171 and is amongst the 167, 170, most widely studied galvanic couples in nanochemistry. It is of particular relevance to my research as the work presented in appendix 2 uses AgAu nanocrystals synthesised by this galvanic replacement reaction. 51 For these reasons I will use this reaction to illustrate the essential features of galvanic replacement reactions. Mechanism The surface of a silver nanostructure is oxidised e when reacted with a solution containing AuCl 4 -, while metallic gold is simultaneously deposited onto the surface forming a gold shell in the shape of the silver template , 174 Despite this reaction being widely studied, the literature is unclear on the redox chemistry of this process. The standard reduction potentials of AuCl 4 - /Au, AgCl/Ag, and Ag + /Ag are 0.99V, 0.22V, and 0.80V respectively. 167 is unclear whether the oxidation of the silver template forms Ag + (aq) or AgCl (s), with key publications from Younan Xia s group suggesting both the former (equation 24), 170 and the latter (equation 25). 167 It 3Ag (s) + AuCl 4 (aq) 3Ag (s) + AuCl 4 (aq) + Au (s) + 3Ag (aq) Au (s) + 3AgCl (s) + Cl (aq) (24) (25) Given the low solubility of AgCl in water, its formation seems likely, however, AgCl crystals are not seen in TEM images of AgAu nanoparticles made in this manner and neither is the material detected by other characterisation techniques. It has been suggested that the 53

54 reactions may involve the initial formation of AgCl nanocrystals which subsequently dissolve. 167 It should also be noted that the standard reduction potentials reported above are relative to the standard hydrogen electrode at 25 C, however, the reactions being discussed are typically performed at elevated temperatures. 167 Chapter 1, figure 11 shows a schematic of the galvanic reaction between a silver nanocube template and AuCl 4 - illustrating the key mechanistic stages of the galvanic replacement reaction. 170 The initial formation of a thin gold shell on the surface of the template (chapter 1, figure 11b) creates a barrier, preventing AuCl 4 - coming into direct contact with the Ag template, however, the shell is imperfect and pinholes allow diffusion of AuCl 4 - and Ag + in and out of the interior of the particle, enlarging the internal cavity (chapter 1, figure 11bd). 170 Alloying is thermodynamically favourable for all AgAu compositions; consequently the shell formed is expected to be an alloy. 166, 178 The hollowing of the interior proceeds until a smooth, continuous, hole-free, alloyed shell in the shape of the template is formed (chapter 1, figure 11d). However, if sufficient AuCl 4 - is present the reaction can proceed further, with oxidation and dissolution of Ag from the alloyed nanoshell. 170 This process, known as dealloying, creates vacancies in the shell which coalesce to form pores, eventually leading to hollow, porous-walled structures known as nanocages (chapter 1, figure 11e,f). 167, 179 Complete dealloying often results in the shells collapsing; forming fragments of 170, 179 shattered Au nanocages (chapter 1, figure 11g). Figure 11. Schematic illustrating key mechanistic features of the galvanic replacement reaction between Ag nanocubes and an aqueous solution of HAuCl 4. The silver cube (a) is initially oxidised at specific point on its surface (b), the oxidation of Ag (red) is accompanied by the deposition of Au (yellow), a homogeneous shell of Au-Ag alloy (orange) forms and the Ag interior of the particle continues to be removed as this shell thickens (bd). When the silver core is depleted (d) dealloying of the alloyed shell occurs, accompanied by the formation of pores in the box s walls (e-f), ultimately the dealloying can lead to fragmentation of the structure (g). Diagram inspired by figure from Sun et al. Journal of the American Chemical Society 2004, 126 (12), The reaction can be halted at different points along this pathway of hollowing and dealloying by altering the amount of AuCl 4 - used This allows considerable control over both morphology (extent of hollowing and porosity of walls) but also over composition 166, 170 (Ag/Au ratio).

55 1.7.5 The nanoscale Kirkendall effect Mechanism The Kirkendall effect refers to a well-known process in metallurgy that introduces diffusion across the interface between two metals. 180 When the diffusion rates of the two metal ions are unequal there is a net flow of matter across the interface. An excess of vacancies will build up on the side of the interface initially containing the faster diffusing metal and these 165, 180 vacancies can condense to form voids (chapter 1, Figure 12a). While generally an undesirable process in the bulk, 181 an analogous process in nanoscience, exploiting unequal 161, 165, diffusion rates across an interface, provides a useful route to hollow nanostructures The nanoscale Kirkendall effect has been widely exploited to synthesise hollow metal oxide, 163, 182, metal chalcogenide, 182, metal phosphide, 193 and metal nitride nanostructures. 194 These reactions typically involve the reaction between a solid metal (M) template and liquid or gas phase reactants, at elevated temperatures, to form a compound material (MX n ). 165, 182, 195 Initially a surface layer of MX n will form and once formed this prevents direct reaction of the metal core with the reactants such that the reaction must proceed via diffusion through the shell. 165 If outward diffusion of metal cations is faster than inward diffusion of anions, as is commonly the case due to differences in ionic radii, there will be a net flow of matter across the interface and a build-up of vacancies in the core, which will tend to become supersaturated and coalesce to form voids. 165, 196 If allowed to run to completion the reactions generally form hollow MX n nanostructures with shapes 162, 186 resembling the template. In the first reported example of the nanoscale Kirkendall effect, cobalt nanospheres were reacted with elemental sulphur in o-dichlorobenzene at 180 C to give spherical cobalt sulphide nanospheres. 182 While outward transport of cobalt dominates, size analysis of products suggests some inward diffusion of S, as the final central voids are smaller than the initial Co templates. 197 As well as solution phase reactions, solid templates can be reacted with a gas, for example Ni nanocrystals can be heated ( C) in air to form hollow NiO structures, ammonia at 1000 C to form hollow AlN nanospheres. 200 and Al nanocrystals can be heated in 55

56 Figure 12. (a) The bulk nanoscale Kirkendall effect, for example when A is zinc, B is copper, and AB is brass. J A, J B and J vac represent the fluxes of A, B, and vacancies respectively. (b) The nanoscale Kirkendall effect: a solution or gas phase species reacts with a metal (M) template to form a compound material (MX n ). Outward diffusion of M cations through the MX n shell is faster than inward diffusion of X anions, resulting in void formation in the particle s core. Figure inspired by diagrams found in Buriak J. M. Science 2004, 304 (5671), To convert metal nanoparticles to metal oxide nanocrystals it is critical that the selfdiffusion coefficient of the metal cation (D M ) is greater than that of the oxygen anion (D O ) in the metal oxide. 165, For example, D Cu >D O in Cu 2 O, while D Pb <D O in PbO, consequently it is possible to convert Cu nanoparticles to hollow Cu 2 O shells by heating at 100 C in air but heating Pb nanoparticles in air results in the formation of solid PbO nanostructures. 196 In the former case outward diffusion of Cu ions through the Cu 2 O shell is faster than inward diffusion of O anions, in the latter case oxygen anions can diffuse inwards faster than Pb can diffuse out. 196 Diffusion coefficients also need to be considered when attempting to rationalise the different intermediate morphologies observed. 201 By ending the reaction before the conversion of the template material is complete, structures with a void filled metal core and a compound shell can be isolated. 198, 201 The shape of the void and core material found in these partially hollowed structures varies considerably. 197, It is found that during the sulfidation of Cd nanocrystals (~250 nm diameter spheres) intermediate structures with a single, off centre hemispherical void and an opposing hemisphere of Cd form within the CdS shell (chapter 1, figure 13a). 201 In contrast, reactions like the sulfidation of Co proceed via symmetrical intermediates containing a metal core connected to the shell by multiple filaments (chapter 1, figure 13b). 182, 197 These different morphological evolutions can be attributed to the relative rates of metal diffusion within the core and across the shell. 201 In 56

57 the Cd/CdS system Cd self-diffusion in the core is considerably faster than diffusion through the CdS shell, allowing the metal core to adopt a low energy structure where the Cd-void interface is minimised. 201 Similar off-centre voids have been seen in Ni nanoparticle oxidation: it is found that a single off centre void is formed when small Ni nanosphere templates (9 or 26 nm) are used and this can be attributed to faster self-diffusion of Ni compared to Ni diffusion through the NiO shell. 198 However, when large (96 nm) particles are used multiple voids form, it is suggested that diffusion processes in the core are not fast enough to allow voids to combine on this length scale. 198 Reaction temperature can also have an effect on the morphology of intermediates and 197, 203 products. The rate of hollowing shows strong temperature dependence, with sulfidation of cobalt nanocrystals giving hollow nanocrystals within 1 minute at 182 C, within 25 minutes at 120 C, but requiring ~19 hours at room temperature. 197 morphological evolution of the particles depends on the reaction temperature. 197 At higher temperatures a symmetrical intermediate morphology is seen, a gap develops between the cobalt core and the cobalt sulphide shell, with the core and shell being connected by filaments. 197 As growth proceeds the central core shrinks and the filaments disappear, leading to a hollow particle with a single central spherical void (chapter 1, figure 13b). 197 In contrast, at room temperature multiple small voids are formed in each particle and these voids gradually coalesce, although the final voids are often not central and may have irregular, non-spherical, shapes (chapter 1, figure 13c). 197 These temperature dependent differences in morphological evolution are attributed to differences in vacancy mobility at low temperature. Temperature can also affect the stability of the products as hollow nanostructures have been shown to be unstable at high temperatures Hollow NiO and CuO nanoparticles have both been shown to shrink and collapse when heated in vacuum, undergoing a reduction reaction to form solid metal nanocrystals. 205 Diffusion processes are fundamental to the nanoscale Kirkendall effect. However, quantitative discussion of diffusion processes in nanocrystals cannot rely on the use bulk values, diffusion coefficients in nanocrystals can be orders of magnitude higher than bulk values and may depend upon nanoparticle size and crystallinity The Fan et al. have proposed that, after the initial Kirkendall void formation, surface diffusion of core material along the surface of the void is the critical mass transport process in the nanoscale Kirkendall effect

58 Figure 13. (a) Sulfidation of cadmium shows asymmetric intermediate morphologies where the metal-void interface is minimised. While sulfidation of cobalt at >120 C (b) shows a symmetrical intermediate with a central metal core connected to the shell by filaments. At room temperature (c) cobalt sulfidation is accompanied by the formation and coalescence of multiple voids. Figure inspired by Yin et al. Advanced Functional Materials 2006, 16 (11), and Cabot et al. ACS Nano 2008, 2 (7), Yolk shell nanostructures One attractive feature of the nanoscale Kirkendall effect is the ability to create nanostructures comprised of multiple materials. In particular, there is interest in combining metallic and semiconducting crystals in one structure. 182, 198, 207 This can be achieved by an incomplete reaction where not all the core material is converted to the compound product, 198, 201 or by using bimetallic templates where only one of the metals is reactive. 182, 187 If the template has an inert core material and a reactive shell material, Kirkendall voids will coalesce at the core shell interface and the central metal nanoparticle can become completely separated from the compound shell, forming what is known as a yolk-shell 182, 207 structure. 58

59 1.8 Quantum dot properties Nanoparticles of semiconducting materials are referred to as quantum dots (QDs). The work in appendices 4 and 5 focuses on new methods to synthesise PbS QDs. Semiconductors are materials characterised by a relatively small band gap which allows thermal excitation of electrons to their conduction band, leading to properties that are intermediate to those of conductors and insulators. Photons with energies greater than a semiconductor s band gap can be absorbed by the material resulting in the formation of a hole in the valence band and an electron in the conduction band. The Bohr radius of a material is the length an electron and a hole can be separated and still remain bound by their Coulombic interaction. 208 When the dimensions of a nanocrystal are less than the Bohr radius the dimensions of the nanostructure now define the length of excitation, leading to the exciton being spatially confined Confinement of the electronic wave function leads to an increase in the semiconductors band gap, blue-shifting both the absorption onset and emission frequency. 210 This size dependent band gap allows 209, 211 significant opportunities for tuning of the materials optical and electronic properties. In addition to a tuneable band gap, the nature of the energy levels in quantum dots is distinct from that found in bulk semiconductors. While the bulk material s energy levels consist of bands containing a continuum of energy levels, quantum dots show a finite number of discrete atom-like energy levels. 210 Another interesting optical property displayed by QDs is a large Stokes shift, with the luminescence red-shifted compared to the absorption edge. This property derives from the splitting of the quantum dot s lowest excitonic state due to a large electron-hole exchange interaction arising from exciton confinement. The state is split into a higher energy spin allowed and lower energy spin forbidden state. 212 The absorption edge is a consequence of the allowed excitation to the higher energy state, however, excited electrons can cool to the lower energy state from which radiative recombination is allowed. 212 The size of the splitting is size dependent, with a larger shift for smaller QDs. 213 The large Stokes shift is an important property for in vivo molecular imaging applications. The signal from a dye with a small Stokes shift is challenging to detect against the strong background of autofluorescence common in biomedical specimens; in contrast, when a large Stokes shift is involved, signal and background can be separated on the basis of wavelength The absorption of light by quantum dots is also distinct from that of their bulk counterparts, with QDs exhibiting size dependent extinction coefficients. 216 Far above band gap the electronic structure tends to that of the bulk and consequently, when using excitation 59

60 energies considerably higher than the material s band gap, the extinction coefficient is bulklike and quantum confinement has no influence. 217 However, at energies close to the band gap extinction coefficients show a strong dependence on QD dimensions At such energies a reduced density of states and size dependence of the oscillator strength are both relevant factors Studies of PbSe QDs have shown that while the absorption coefficient is independent of nanocrystal size at high photon energies, the absorption coefficient at the band gap shows an inverse quadratic dependence on the particle diameter. 218 Consequently, at energies close to the band gap smaller PbSe QDs are more efficient absorbers. 218 Quantum dots have attracted interest in a range of optoelectronic applications, their narrow emission is attractive for use in lasers, displays, and lighting. 209 For solar cell applications their size tuneable properties allow band gaps to be optimised for the absorption of solar radiation (1.35 ev for a single junction solar cell). 219 Furthermore, it is possible to produce layers of different size QDs, resulting in devices with a band gap gradient, creating a built in field to drive charge extraction. 220 Quantum dots have been widely used instead of conventional dyes (for example polypyridyl complexes of ruthenium and osmium) in dye-sensitized solar cells. 219, QDs have the advantage over conventional dyes of having considerably higher extinction coefficients, allowing the use of thinner mesoporous oxide films. 219 By synthesising core-shell structures where there is a band-gap offset between the core-and-shell material it is possible to manipulate exciton lifetimes, either encouraging electron and hole co-localisation in the core (type-i systems), 223 leading to high photoluminescence quantum yields, or spatial separation of charges (type-ii systems), leading to increased exciton lifetimes, advantageous if carrier extraction is desirable A phenomena of particular interest for photovoltaic applications is multiple exciton generation (MEG), a process where a photon with 226, 228- more than twice the band gap energy can generate two or more electron hole pairs. 229 MEG has potential to dramatically improve photovoltaic efficiency when using narrow band gap materials, as photon energies in excess of the band gap would normally be wasted as heat Hybrid solar cells Concerns about climate change, finite reserves of fossil fuels, and energy security make greater use of renewable energy sources attractive Commercial solar cells are predominantly based on silicon technology; 233 however, alternative technologies represent 231, a large research area. Cost represents a barrier to wider adoption of photovoltaic technology, with the dollar per Watt ($/W) price of different technologies often used for

61 comparisons There is particular interest in technologies that have the potential to , 234 achieve either higher efficiencies or lower cost compared to existing products. Furthermore, numerous new applications could be envisaged for photovoltaics if they were lighter weight and flexible. 234 The optical and electronic properties of quantum dots (QDs) such as their size-tuneable band gaps and high electron mobilities could potentially give impressive photovoltaic performance. 235 Semiconducting polymers are another class of material that show great promise for future photovoltaic applications. Unlike single crystal silicon, polymers can be solution processed at low temperatures to coat large areas, and devices based on such 234, polymers could be light weight, flexible, and cheap to manufacture. Hybrid photovoltaics refer to a class of solar cell where the active layer is a composite thin film, 230, 236, containing intimately mixed domains of semiconducting polymer and quantum dots. 239 The photoactive layer of a hybrid solar cell is essentially a bulk heterojunction with an organic hole transporter and an inorganic electron transporter. 230 Hybrid solar cells can potentially combine the best properties of the polymer and the QDs, allowing broad and tuneable absorption across the solar spectrum. 230, 236 The work in appendices 4 and 5 on the growth of PbS nanocrystal in a polymer matrix is potentially revelant to the synthesis of new hybrid photovoltaic materials. Chapter 1, figure 14a shows the principle of photocurrent generation in a hybrid solar cell. An incident photon generates an exciton which then diffuses to an interface where charge separation occurs. The charges are then transported to opposite electrodes, with electrons travelling through networks of semiconducting nanocrystals, while holes are transported through polymer domains. 236 One of the main advantages of addition of nanocrystals to polymer solar cells is the relatively high electron mobilities of inorganic nanocrystals, compared to the extremely low electron mobilites of most conjugated polymers, 240 this combined with fast electron hole separation at polymer-quantum dot interfaces has potential to considerably improve performance of polymer solar cells. 241 An additional advantage of hybrid structures is the ability to harvest a broader range of photon energies by selecting organic and inorganic semiconductors with bandgaps that exploit different regions of the solar spectrum

62 Figure 14. Cross sectional schematics of hybrid solar cells with conjugated polymer shown in blue and QDs in red. (a) The key processes involved in photocurrent generation, with a photon generating an electron hole pair which is separated at the organic-inorganic interface, with the hole transported through the polymer and the electron extracted through a network of QDs. (b-d) Possible hybrid film morphologies with (b) spherical, (c) rod shaped and (d) hyper-branched QDs, the hyper-branched particles are believed to offer the best morphology for charge separation and extraction. Diagrams inspired by those found in Gao et al. Energy & Environmental Science 2013, 6 (7), The nanoscale three-dimensional morphology of the hybrid layer will affect both charge separation and extraction If the exciton is generated too far from an interface other relaxation processes compete with charge separation. The exciton diffusion length in Poly(3-hexylthiophene-2,5-diyl) (P3HT, one of the most widely studied semiconducting polymers) is ~10 nm. 246 Consequently, if there are polymer domains in a P3HT based device where the distance to the nearest QD domain is considerably greater than 10 nm, the charge separation and consequently device efficiency will be poor. Suitable routes for charge carriers to travel from the interface to the electrodes are needed for efficient extraction; thus isolated islands of either material are undesirable. Chapter 1, figure 14 (bd) shows a range of possible hybrid film morphologies in cross-section. The spherical particles in (b) are not ideal as electron extraction will require percolation through a network of poorly connected nanocrystals. The wires in (c) should be better as fewer hops between adjacent crystals are needed to extract an electron. 236 Hyperbranched structures are close to ideal, with both phases being finely interspersed (short exciton diffusion lengths) and providing unhindered pathways for both electron and hole extraction. 242 It has been shown that hybrid cells made with nanorods outperform analogous devices made using spherical QDs

63 Figure 15. Energy level diagram of hybrid solar cell with conjugated polymer shown in blue and inorganic semiconductor in red. The diagram shows the case of a photon being absorbed by the organic component, causing excitation of an electron from the polymer s highest occupied molecular orbital (HOMO) to its lowest unoccupied molecular orbital (LUMO). In an alternative scenario (not shown) the inorganic component can absorb a photon, exciting an electron from the valence band (VB) to the conduction band (CB). In both cases the energy level alignment causes charge separation at the organic-inorganic interface, with electrons extracted by the inorganic phase and holes by the polymer phase. Inspired by diagram found in Xu & Qiao Energy & Environmental Science 2011, 4 (8), Suitable alignment of energy levels is also necessary for dissociation of excitons into free charge carriers A band gap diagram for a typical hybrid solar cell is shown in chapter 1, figure Suitable energy level alignment of the polymer s highest occupied molecular orbital (HOMO) with the quantum dot s valence band (VB) is necessary for extraction of holes. While a suitable energy level alignment between the polymer lowest unoccupied molecular orbital (LUMO) and the quantum dots conduction band (CB) is necessary for 236, 244, 250 electron extraction. 63

64 Chapter 2. Literature review 2.1 TEM of nanomaterials Electron microscopy is one of the most powerful techniques available for understanding the structure and properties of nanomaterials; it is a central tool in all of the work presented in the thesis (appendices 1-5). Bright field TEM imaging is widely used to assess the size, shape, and assembly of nanocrystals Diffraction contrast can reveal changes in crystal structures or orientation either within a group of particles or within a single particle For example bright field and dark field TEM images show the polytypic nature of Cu 2 ZnSnS x Se 4-x nanocrystals, with the particles wurtzite centres and zinc blende ends clearly distinguishable (chapter 2, figure 1 a-b). 256 interpretation, 3 Despite the challenges of image HRTEM is an important tool for studying nanocrystals, allowing the crystallinity of individual nanoparticles to be analysed For example, HRTEM can reveal the direction of nanowire elongation, 261 twinning in nanocrystals, and the facets prevalent at a catalyst s surface observed in HRTEM images of nanocrystals Features, like core-shell structures can also be Chapter 2, figure 1c shows a HRTEM image of the Cu 2 ZnSnS x Se 4-x nanocrystals shown in 2a-b, clearly showing the wurtzite and zinc blende domains. High resolution images of the gold nanowire shown in chapter 2, figure 1d-e reveals not only that the rod is elongated in the [111] direction but also shows an extremely high density of twins which are believed to be responsible for the materials impressive tensile strength (up to 3.12 GPa)

65 Figure 1. Examples of TEM imaging of NCs. (a-c) Bright field, dark field, and HRTEM images of polytypic CZTSSe nanocrystals clearly distinguish between wurtzite and zinc blende domains within the same crystal, adapted with permission from Fan et al. Scientific Reports 2012, 2, 952. (d-e) HRTEM images of a gold nanowire, revealing the direction of elongation and showing extensive twinning and surface faceting, reproduced with permission from Wang et al. Nature Communications 2013, 4, HAADF STEM of nanomaterials The Z-contrast mechanism makes HAADF STEM well suited to locating individual high Z atoms on low Z supports In their pioneering work in the 1970s, Crewe and coworkers imaged single heavy atoms (uranium, mercury, and silver) on carbon supports. 268 In 94, many catalyst materials small clusters and individual atoms may act as active sites. HAADF STEM is therefore, an ideal tool for imaging such features, for example: individual Pt and Rh atoms have been imaged on an Al 2 O 3 support, 269 single atoms of Pt have been detected on an FeO x support, 271 and single W atoms have been imaged on a ZrO 2 support (chapter 2, figure 2 d-e). 270 Despite examples of images of isolated single atoms dating from the earliest days of STEM, 268 the ability to routinely achieve atomic resolution images of crystalline samples 65

66 required the development of aberration corrected electron optics Aberration corrected STEM is particularly vital for atomic resolution imaging at low accelerating voltages. Low voltages (30-80 kv) are widely used to prevent knock-on damage when studying 2D materials in the STEM. 27, 32, 67, 277 Lower kv imaging conditions also increase the likelihood of beam-sample interactions, due to increased interaction cross-sections, giving better signal to noise ratios in both images and spectra. 73 However, low voltages also increase the necessity of having ultra-thin samples. By focusing more electrons into a smaller area aberration corrected STEMs can produce larger probe currents, allowing shorter acquisition times for EDX and EELS spectra. 4 The atomic number sensitivity of HAADF STEM contrast can be interpreted as a form of elemental mapping, 17 for example, in a bimetallic nanoparticles regions of the heavier metal will appear brighter than regions of the lighter element. 278 If there is a significant different between the atomic number of the two components, phase segregated particles, such as core-shell or Janus particles, can be clearly distinguished from alloyed particles 81, (chapter 2, figure 2 a-c). In 2D crystals, HAADF contrast has been demonstrated to allow direct chemical identification of single atoms, 24, 27, 277 individual B, C, N, and O atoms could be distinguished in a BN crystal, despite their similar Z values. 277 Chapter 2, figure 3 shows a HAADF STEM image of nitrogen doped graphene, N atoms (Z=7) can be distinguished from the surrounding C atoms (Z=6). 284 However, in a complex threedimensional (3D) structure where atomic columns contain multiple atoms of different elements, directly determining the number and identity of atoms in a column from a single HAADF image projection is extremely challenging

67 Figure 2. Examples of HAADF STEM imaging of nanocrystals where Z-contrast allows regions with different compositions to be distinguished. (a) During observations of the coalescence of Au (Z=79) and Pd (Z=46) NCs, regions of Au and Pd are clearly distinguishable, as Au appears brighter than Pd, adapted with permission from Mariscal et al. Nanoscale 2011, 3 (12), (b) Au-Pd core-shell NC, reproduced with permission from Kiely C. Nature Materials 2010, 9 (4), (c) FePt-CdS core-shell NC, Reproduced with permission from Trinh et al. RSC Advances 2011, 1 (1), (d,e) Tungstated zirconia catalyst, W (Z=74) apprears bright against a ZrO 2 support (Zr Z=40), black circles indicate single W atoms, black squares identify surface polytungstate species, white circles show WO x clusters. Adapted with permission from Zhou et al. Nature Chemistry 2009, 1 (9), Figure 3. Atomic resolution image of doped graphene, the Z contrast in HAADF STEM images allows the chemical identity of individual atoms to be determined. The bright atom in the centre of the highlighted region is a nitrogen atom introduced by ion implantation. Adapted with permission from Bangert et al. Nano Letters 2013, 13, Copyright 2013 American Chemical Society. 67

68 2.3 Spectrum imaging of nanomaterials Elemental distributions are of critical importance to the catalytic, optical, thermal, and electronic properties of nanomaterials. 4, 82, While HAADF STEM contrast is sensitive to atomic number, it also depends on sample thickness. Thus, the morphological complexity of most NC samples and the consequent imperfect knowledge of the particle s 3D structure makes accurate nanoscale compositional identification impossible using only HAADF imaging. Consequently, spectroscopic techniques are often necessary to confidently understand the chemical composition and elemental distributions of nanostructures. A STEM equipped with either an EDX detector or an EELS spectrometer is able to simultaneously acquire atomic-resolution images and highly spatially resolved chemical information. 33, 290 In particular, spectrum imaging, where a spectrum is acquired for each pixel of an image, is an extremely powerful tool for understanding chemical composition, elemental distributions, and bonding at the nanoscale. 4, 33, 291 Both EDX and EEL spectrum imaging can be performed at atomic resolution and can even achieve single atom 6, 24, sensitivity. 289, Bimetallic nanocrystals are of huge interest in the field of heterogeneous catalysis. Due to the relationship between catalytic activity and surface composition, understanding elemental segregation is of great importance. The work presented in appendix 2 studies changing elemental distributions in AgAu nanocrystals. Elemental distributions in Cu 3 Pt nanoparticles, a promising electrocatalyst, were studied using EEL spectrum imaging. 290 The initially prepared NCs are found to be alloyed in the core with one to two monolayers of Cu surface segregated on some facets (chapter 2, figure 4a). 290 It was found that after electrochemical cycling a ~1.1 nm thick Pt shell was formed, explaining the measured improvements in the oxygen reduction reaction (ORR) activity of the catalyst. 290 Pd x Co NCs are also promising ORR catalysts and EEL spectrum imaging has shown that annealing a PdCo catalyst under hydrogen gas (6%) at 500 C for 2 hours resulted in the formation of core-shell NCs with thin Pd shells (chapter 2, figure 4b). 296 Site-specific segregation has been observed by EEL spectrum imaging of Pt x Ni 1-x nanooctahedra, with Pt enrichment along edges and Ni segregation on the {111} faces. 297 After electrochemical cycling, selective leeching of Ni from the centres of the faces is observed, forming concave octahedra. 297 In all of these examples of ORR catalysts, EEL spectrum imaging is a key tools in understanding the origins of catalytic performance, such structure-property relationships 290, will prove vital for the development of improved catalysts. 68

69 Figure 4. Examples of EELS spectrum imaging revealing elemental distribution in bimetallic nanocrystals. (a) spectrum image of Cu 3 Pt nanocrystals showing almost entirely complete alloying with Cu surface segregation of one or two atomic layers on some facets. (b) Shows PdCo NCs after annealing in a H 2 atmosphere at 500 C, resulting in the formation a Pd shell. Figure adapted with permission from Wang et al. Chemistry of Materials 2012, 24 (12), and Wang et al. Nano Letters 2012, 12 (10), Copyright 2012 American Chemical Society. Recent improvements in the collection efficiency of EDX detectors has led to an increasing number of studies employing EDX spectrum imaging to study elemental distributions in NCs , 291, 294, 298 EDX spectrum imaging has demonstrated that PdAu nanocrystals with a distinct Au core and Pd shell could be transformed into a homogeneous alloy by calcining at 300 C for 1 hour; with the addition of Au preventing Pd oxidation and sintering during calcination. 298 Changing elemental distributions in fuel cell cathode catalysts with a Pd 9 Au 1 core and a monolayer Pt shell have also been studied using EDX spectrum imaging. 291 It was shown that the Pt surface is retained and even thickened after 20,000 cycles, however, cycling does result in a significant loss of Pd from the core. 291 I have used EDX spectrum imaging extensively in my research (see appendices 1-3). 44, 51, 299 Chapter 2, figure 5 shows an EDX spectrum image acquired from a bimetallic AgAu tadpole NC, the presence of an Ag shell covering the entire surface of both the head and tail is surprising as the particles were synthesised via the galvanic reaction of an Au precursor with Ag seed particles. 299 This unexpected surface segregation is impossible to detect from imaging alone, yet is likely to affect both the particles properties, such as catalytic performance and surface plasmon resonance, and our understanding of their growth mechanism. 299 Information from EELS and EDX can often be combined to give more in-depth understanding. For example, to investigate the effect of acid leaching on PtNi catalysts both 69

70 EDX and EELS were used. 300 EDX spectroscopy was used to determine particle compositions, showing that oxygen free leaching conditions lead to less Ni loss, 300 while EELS line scans were used to study elemental distributions, revealing Pt shell formation during leaching. 300 Similarly both techniques were used to study Mo 1-x W x S 2 sheets, with the identify of individual atoms confirmed by atomic resolution EEL spectrum imaging, and large area EDX analysis used to determine the materials average composition. 301 Figure 5. Example of EDX spectrum imaging of bimetallic AgAu nanocrystal. Cliff Lorimer quantification reveals the particle s composition is ~13% Ag 87% Au. EDX spectrum imaging (a, c, d) and line scans (b) show a ~1 nm thick shell covering the entirety surface of the structure. Adapted from Da Silva et al. Chemistry A European Journal 2015, 21 (35), A constant challenge when performing (S)TEM studies of nanomaterials is to accurately reflect the polydispersity of the sample while achieving detailed, high resolution, analysis of individual particles. Conclusions based on a limited number of particles risk observational bias causing outlying members of a population to be treated as representative. Even if the particles analysed are typical, failing to consider the full extent of polydispersity in the sample may make it impossible to account for bulk properties. The problem of undersampling is especially pronounced when spectrum imaging is used, as acquisition times are orders of magnitude greater than for simple imaging. For example, the degradation of PtCo fuel cell catalysts during voltage cycling leads to the formation of highly polydisperse populations, and understanding elemental redistribution during such degradation requires the analysis of statistically significant numbers of particles. 4 Fast EELS acquisition in an aberration corrected STEM allowed elemental maps of hundreds of PtCo NCs to be 70

71 acquired (chapter 2, figure 6). 4 The initial particles were found to have a Pt rich shell, with shell thickness almost independent of particle diameter. Voltage cycling leads to particle coarsening which is typically accompanied by the formation of a thicker shell, some particles also show multiple Co rich cores after cycling. By analysing hundreds of spectrum images it is shown that the average shell thickness increases with particle radius as does the fraction of particles containing multiple cores, possibly a result of coalescence events. 4 The loss of catalytic performance observed during cycling is believed to be a result of both the reduced surface area due to coarsening and Pt redisposition leading to thick Pt shells and rounded morphologies. Figure 6. EELS spectrum images of PtCo fuel cell catalyst (a) before and (b) after voltage cycles. Pt is shown in red and Co in green. Spectrum imaging of a large population of nanocrystals allows the changes in morphology and elemental distribution responsible for loss of catalytic performance to be revealed. Reprinted with permission from Xin et al. Nano Letters 2012, 12 (1), Copyright 2012 American Chemical Society. Atoms of the same element but in different oxidation states can be distinguished by ELNES, and in some cases this technique can be used to perform nanoscale mapping of oxidation state changes. 33, 302 EELS spectrum images have been used to study the surface chemistry of ceria (CeO 2 ) nanoparticles, a catalyst for the removal of NO x, CO, and unreacted hydrocarbons from exhaust fumes. 33 The oxidation states of Ce ions at the nanocrystal surface is believed to affect catalytic performance, EELS spectrum imaging allows Ce 3+ and Ce 4+ ions can be distinguished, allowing surface reduction in ceria nanocrystals to be mapped. 33 On {111} facets a single surface layer is found to be fully reduced while

72 underlying layers are of mixed valency, while in contrast {100} facets show 5-6 fully reduced layers EELS can reveal detailed information about the local bonding of atoms, for example Ramasse et al. were able to show that single atom Si dopants in graphene lie inplane when tetravalent but lie out of plane when trivalent. 32 Surface plasmons are highly sensitive to nanocrystal size, shape and composition. Using STEM low loss EEL spectrum imaging it is possible to map the spatial distribution of surface plasmon resonances for a single nanocrystal While the best energy resolutions achievable with EELS is poor compared to optical techniques, 305 the technique has much better spatial resolution revealing the effect of an individual NC s size and shape on its plasmonic excitation Although ~9 mev energy resolution has recently been demonstrated with a new monochromated instrument, 100 mev is a more typically achievable value In situ (scanning) transmission electron microscopy of nanomaterials Introduction Understanding the growth and transformations of nanocrystals is necessary to develop rational synthesis strategies for high performance NCs. 310 Equally important are transformations that occur post-synthesis, during the nanocrystals use, where changes in 4, 311 structure and chemistry can lead to losses in performance. Understanding such degradation processes can guide the design of more resilient materials. 287, 312 Furthermore, the structure of functional materials at room temperature in a vacuum may be significantly different from that in their operational environment, consequently our understanding of their performance may be limited by the requirements for high vacuum conditions in the majority of conventional (S)TEM investigations. In situ electron microscopy offers the potential to directly observe NC transformations, 47, 89, providing unique mechanistic insights into dynamic processes at the nanoscale. Although in situ techniques have been in use for decades, recent years have seen a considerable increase in interest and applications. In part, this is due to technical developments such as the use of micro-electro-mechanical systems (MEMS) based chips for liquid-cells and heating experiments. 47, 90-91, 315 Liquid-cells have made in situ experiments more widely available, without the need for a dedicated and expensive environmental TEM system. In situ techniques have been used to observe the growth and transformations of nanocrystals as well as their structure in environments relevant to their envisaged 46, 76, 97, 112, 316 applications. My research has looked at both the developments of in situ

73 techniques and their application to study nanocrystal growth processes and transformations (see appendices 1-3). A broad range of in situ (S)TEM techniques now exist: 87, 89, 91, 317 samples can be studied in a gaseous environment, subjected to electrical biasing, 90, 105, or have mechanical forces applied to them. 94, 104, However, this section will focus primarily on the two techniques that are relevant to my work: in situ heating and liquid-cell studies In situ transmission electron microscopy In situ electron microscopy is a broad term covering any experiments where stimuli are applied to a specimen in the microscope; allow dynamic transformations of interest to be directly observed. In the simplest in situ experiments the electron beam is used both to 48-49, image the sample and to drive the transformation. However, beam induced processes are typically regarded as damage and in many cases endeavours are made to minimise their effect (see section 1.4) and to isolate the effect of external stimuli, using specially designed specimen holders or microscopes , 90, 93 In situ techniques have a number of advantages over sequential ex situ observations. One advantage is a dramatic reduction in the number of samples that need to be prepared. For experiments such as heating of bulk materials or interfaces, where different heat treatment needs to be applied to multiple bulk specimens and then each specimen thinned or cross sectioned to make a TEM sample this represents a significant saving in time. 327 nanomaterials this is less of a concern as TEM sample preparation is often straight forward. For example, the progress of a nanocrystal synthesis reaction can be tracked ex situ with TEM by taking aliquots from the reaction mixture at various reaction times and drop-casting particles from each aliquot onto separate TEM grids. 197, 328 However, in situ experiments have important advantages beyond time and cost savings. In situ experiments provide a continuous history, meaning that a single region or nanostructure is tracked over the course of a transformation. 327 This prevents incorrect interpolations of the transformation trajectory. 327 When studying nanocrystals, population inhomogeneity is a likely source of such errors: the researcher assumes that the nanocrystal they observe evolved from a nanocrystal identical to the one characterised in detail in a sample from a shorter reaction time but it may have actually evolved separately from a crystal which was structurally or compositionally different. In an in situ experiment, such errors are eliminated as the same 46, 101 particle can be observed throughout the growth process or transformation. Ensuring that the electron beam is not influencing the transformations observed remains one of the greatest challenges in in situ electron microscopy. 120, 127, 318 Direct observation of a dynamic process during the application of stimuli is not always desirable as the combined 73 For

74 effect of the electron beam and the stimuli may alter the process observed. In some cases quasi in situ experimental protocols, where the same area of the same specimen is observed at various stages during a transformation but is not actually observed as it transforms, have been used to eliminate electron beam induced reactions. 318, For example, when studying the oxidation of carbon nanotubes (CNTs) in an environmental TEM, EEL spectra showed K-shell ionisation of oxygen when O 2 gas is present. 318 To avoid confusing reactions involving oxygen ions with reactions involving molecular oxygen the authors developed a protocol where the CNTs were heated to the reaction temperature in vacuum, imaged and then the beam blanked, O 2 gas was then introduced into the ETEM for a desired reaction time before the microscope was purged of oxygen and the CNTs were imaged again in vacuum. 318 In this manner the same CNT could be imaged before and after a controlled oxidation reaction, without the electron beam altering the chemistry observed. 318 In some examples of quasi in situ reactions the specimen is removed from the microscope between images For example, to study corrosion in aluminium alloys an area of a specimen is imaged and analysed in the TEM, the specimen is then removed from the microscope and exposed to a corrosive environment (oxygen bubbled through aqueous HCl, ph = 3) outside the microscope, and then returned to the microscope for further analysis. 329 This cycle of corrosion and imaging was repeated multiple times, with the same region of the specimen investigated at each stage. 329 However, experiments such as these have associated experimental challenges (for example location of identical specimen areas) and could be described as not truly in situ because the specimen is not directly observed during the transformation In situ heating experiments Nanocrystal coarsening and coalescence Nanomaterials typically have depressed melting points (relative to the bulk material), and a tendency to undergo shape change or coalesce at relatively low temperatures. 331 Polydispersity with respect to both size and shape can effect thermally induced transformations, therefore, bulk measurements (such as XRD) only give limited insights into the behaviour of nanocrystals during heating. 332 Consequently, in situ heating experiments are of particular interest for studies of nanomaterials. Not only does in situ TEM allow one to observe the effect of a given temperature on individual nanostructures, but it allows 76, 80, 315, 333 direct observation of transformation to provide mechanistic insights. The transformation and degradation of nanomaterials at elevated temperatures is relevant to their performance in a number of potential applications, where heating will be applied either during use or manufacture. 74 The behaviour of metal nanoparticles at elevated

75 temperatures is relevant to the performance of catalysts, where sintering reduces surface area and hence activity In situ studies of 2D arrays of Au nanocrystals cast directly onto Si 3 N 4 heating membranes has revealed that the thermal stability and mechanism of degradation depends on a both particle size and heating rate. 337 Smaller particles (2-8 nm) undergo Ostwald ripening, where some large particles grow at the expense of small ones, while larger (15 nm particles ) undergo densification, where particles move closer together and fuse. 337 For the smaller particles it is observed that slow heating leads to the formation of spherical structures, while fast heating gives interconnected string-like structures. 337 For catalytic applications nanocrystals are often dispersed on a support. The thermal stability of Au nanocrystals embedded in a porous Al 2 O 3 flakes has been studied in situ, revealing the structure to be completely thermally stable up to 340 C. Agglomeration of the embedded nanocrystals starts at approximately 530 C, with a loss of porosity in the Al 2 O 3 commencing around 940 C. 335 In situ heating experiments were also used to study Au nanoparticles in a silica support, HAADF STEM imaging of the sintering process reveals a mechanism dominated by particle migration, not Ostwald ripening. 334 While heating experiments in vacuum can give useful insights into sintering mechanisms, catalytic processes require a reactive gaseous environment and this can dramatically alter the NC sintering behaviour. 338 Coalescence of nanoparticles is also potentially a problem for high density magnetic data storage. 333 PtFe nanoparticles in a KCl matrix were studied at 400 C, 500 C, and 600 C; with STEM imaging demonstrating that particle-particle coalescence was the main mechanism of particle growth, although Ostwald ripening was also observed in some regions. 333 Semiconducting nanocrystals can also undergo heating induced coalescence: 10 nm PbSe nanocrystals have been shown to become mobile at 100 C and undergo oriented attachment to form string like structures, while annealing at C leads to crystal unification. 315 The mechanism is characterised by three steps: attachment of particles, rotation in three dimensions, and interfacial relaxation to give a fused structure. 315 Nanocrystal melting and sublimation In situ electron microscopy has been widely used to study the melting behaviour of nanoscale materials The sublimation of polydisperse silver nanocrystals has been observed in situ; the particles sublime at temperatures between 500 and 700 C, with smaller particles subliming at lower temperatures (chapter 2, figure 7). 76 The size of the nanocrystals also effects the mechanism of their sublimation, with larger particles undergoing a continuous process during which they remain spherical but smaller particles subliming in discreet steps forming intermediate faceted structures. 76 Silver nanowires are 75

76 also observed to disappear when heated in situ. 345 When heated at 530 C the nanowires first lose silver from their central region, while at 650 C silver disappears from one end of the nanowire first. 345 Interestingly, the complete vaporisation of a silver nanowire is observed to leave behind a carbon nanotube, formed from the layer of amorphous carbon found on the nanowire s surface. 345 Figure 7. Bright field TEM images of Ag nanoparticles during in situ heating. Smaller particles sublime at lower temperatures. Reprinted with permission from Asoro et al. ACS Nano 2013, 7 (9), Copyright 2013 American Chemical Society. In situ heating of bimetallic and heterostructured nanomaterials Some bimetallic systems show enhanced thermal stability compared to their monometallic counterparts In situ heating experiments comparing Pt nanocubes with Pt-Rh nanocubes show that the bimetallic system maintains its shape at higher temperatures, the spatial distribution of the two elements was mapped using EDX spectrum imaging at 500 C. 82 It is suggested that Rh moves from the corners and edges to cover the {100} faces via surface diffusion. The coating of low meting point Pt surfaces with higher melting point Rh prevents surface pre-melting and helps shape retention. 82 In situ EDX spectrum imaging has also been used to study heating induced changes to the elemental distribution of CdSe/CdS/ZnS core-shell nanostructures. 86 EDX spectrum imaging at 275 C reveals Zn diffusion from the outermost shell into the core and shrinkage due to evaporation of Cd and S, before sublimation occurs at 310 C. 86

77 Au and Fe 2 O 3 nanoparticles fuse at elevated temperatures to form particles with an alloyed surface layer. 85 At higher temperatures the fused particles undergo phase segregation, forming an Au-shell. 85 By performing HAADF STEM imaging and EDX analysis during in situ heating experiments the influence of particle size and Au content on the phase diagram of this complex nanoscale system has been revealed. 85 However, the authors also note the limitations of in situ EDX analysis: due to the long spectral acquisition real-time EDX analysis is challenging. Consequently the frequency at which EDX data is acquired during the experiment must be limited. 85 It is demonstrated that intermediated EDX spectra can be used to calibrate the composition dependence of the HAADF image contrast, allowing realtime information on the compositional changes of individual particles to be determined from the faster HAADF-STEM imaging. 85 Furthermore, due to the significant difference in the (effective) atomic numbers of Au and Fe 2 O 3, phase segregation in particles can be clearly observed from the HAADF-STEM images. 85 Atomic resolution HRTEM or STEM imaging is also a powerful approach for studying the transformations of multi-component nanocrystals in situ, where different crystal structures or lattice parameters can be clearly distinguished In situ HRTEM imaging has been used to observe the changing structure of PbSe-CdSe core-shell nanoparticles. 78 Heating the nanocrystals between 150 and 200 C results in the formation of PbSe and CdSe hemispheres joined by a common {111} Se plane (chapter 2, figure 8a), 78 and this transformation is shown to occur by surface diffusion of CdSe. 78 Fe x O/CoFe 2 O 4 core-shell nanocubes and nanospheres have also been imaged at atomic resolution during heating experiments. 79 At temperatures above 300 C, the Fe x O core material coalesces on the surface of the particle, forming dumbbell structures with a smaller Fe x O domain attached to a larger CoFe 2 O 4 crystal. This phase segregation is apparent from both HRTEM images (chapter 2, figure 8b) and diffraction contrast in lower magnification Bright field (BF) TEM images. 79 Heating CdSe nanorods decorated with gold nanoparticles at C causes the shrinkage and disappearance of small particles on the sides of the rods and the growth of large Au domains on the rod tips, it is found that intrarod Ostwald ripening is responsible for this change. 80 Due to the considerable difference in atomic number, regions of Au and CdS are easily identified from both HAADF STEM and BF TEM images. 80 Furthermore, in situ HRTEM images give detailed information on crystallographic features of the transformation; the interface between the CdSe nanorod and the gold tip is observed to restructure during heating creating a better contact between the two domains. 80 Heating nanoscale octapods, consisting of a CdSe core and CdS pods, to 300 C causes rounding of pod tips and segregation of Cd droplets at the particle surface due to initial non- 77

78 stoichiometry. 331 Heating the octapods at C results in sublimation, while at temperatures just below the sublimation temperature the zinc blend core grows at the expense of the wurtzite pods. 331 Ex situ electron tomography of the heated octapods revealed that the pods in contact with the carbon substrate are more stable than those pointing into the vacuum. 331 Figure 8. HRTEM imaging during in situ heating experiments. (a) nanocrystals formed by heating PbSe-CdSe core-shell nanoparticles in situ at 200 C. (b) nanocrystals formed by heating Fe x O-CoFe 2 O 4 core-shell nanocubes in situ at 335 C. In both examples, HRTEM images allow domains of different materials to be distinguished and their crystallographic orientation determined. Figure adapted with permission from Yalcin et al. Nanotechnology 2014, 25, (5), and Grodzinska et al. Journal of Materials Chemistry 2011, 21, (31), Artefacts in heating experiments Care must be taken so that in situ observations accurately reflect the ex situ transformation 327, 346 of interest. Historically, thin foil effects were a concern in heating experiments and in order to determine if surface effects were influencing the behaviour observed, thin and thick regions of the same specimen were compared. 327 For nanomaterials this is not a concern as the high surface area is an intrinsic property of the specimen not an artefact introduced by sample preparation. However, the effect of the electron beam is pertinent to all specimens. 49, 76, 323 To confirm the authenticity of in situ observations, areas which were exposed to the electron beam during the heating experiment should be compared to areas which were not irradiated. 327, 346 It may also be useful to compare the in situ treatment with 78, 327, 331 an analogous treatment performed ex situ (for example heating in a furnace). Ideally TEM data is compared with bulk measurements from another technique; for example, activation energies for a thermally induced process can be determined in situ and 327, 346 compared to values obtained from differential scanning calorimetry (DSC). 78

79 Accuracy and uniformity of heating During in situ heating experiments the microscope operator must often trust the holder manufacturer s temperature calibration as direct in situ measurement of the specimen s temperature is challenging. Contact thermometers can cause significant changes in the specimen temperature due to its small thermal mass. 37 Non-contact thermometers which rely on the detection of radiation are not built into TEMs, so modified holders are required. 347 Recently Raman spectroscopy has been used to calibrate temperature during an in situ TEM experiment. 347 The microscope was modified so that a parabolic mirror could be inserted below the specimen allowing light to be focused on the specimen and Raman scattered light to be collected. 347 The frequency of Raman peaks decreases with increased temperature, 348 allowing non-contact temperature measurements. 347 Single walled carbon nanotubes (SWCNTs) make an excellent Raman temperature probe due to their high thermal conductivity, large contact area with the support, and sharp G-band excitation peak By dispersing SWCNTs on a MEMs based heating chip (geometry suggests that a Protochips Aduro system was used) the authors were able to measure the chip s temperature in situ. Their results show a considerable temperature gradient across the chip with temperature differences of over 100 C measured across the 7x7 array of holes. 347 As well as requiring equipment modifications, the spatial resolution of optical temperature measurements is considerably worse than the dimensions of many structures studied by electron microscopy. 37 EELS offers an attractive alternative method of temperature calibration, as no microscope modifications are required and nanometre spatial resolution is achievable. 37 The energy required to excite bulk plasmons (E) depends upon the number density (n) of electrons in the specimen (E n 1/2 ). 37 As the temperature increases, thermal expansion decreases n and therefore produces a shift in the plasmon peak. 37 It has been demonstrated that many metals and semiconductors have sufficiently sharp plasmon peaks for EELS spectra to be used to measure their temperature in situ. 37 The temperature of aluminium nanowires during in situ biasing experiments have thus been measured with nanometre spatial resolution. 37 Summary In situ heating allows direct observation of important nanoscale transformations such as 76, 333 nanocrystal coarsening, melting and sublimation. In situ measurements have the unique advantage of allowing individual nanostructures to be continuously observed through the heating process, this is valuable for understanding the properties of polydisperse populations and revealing the mechanisms of transformations. 76, 315 MEMs based holders combined with developments in (S)TEM and EDX detector designs have 79

80 made atomic resolution imaging and/or EDX spectrum image at elevated temperature common features of in situ heating experiments. 51, 81-82, This is especially valuable when studying multi-component nanocrystals, where processes like elemental redistribution, phase segregation, and interface restructuring can be directly observed , Care must be taken in when performing and interpreting in situ heating experiments, to ensure that the dynamic transformations observed are due to the temperature and not the electron beam. 37 Furthermore, inaccurat calibration and temperature gradients may be a significant source of errors, 347 consequently, direct methods of in situ temperature measurement 37, 347 should be applied where appropriate, and in situ data compared to ex situ results Liquid-cell experiments Observations of nanocrystal growth 218, The properties of nanomaterials exhibit strong size and shape dependence, consequently understanding of nanocrystal growth mechanisms is critical to the development of morphologically controlled syntheses. 310, 353 Liquid-cell (S)TEM studies of nanocrystal growth have given unique insights into the mechanisms of nanocrystal growth, allowing the changing size and shape of individual nanocrystals to be tracked in real-time. 46, 102, 112, 354 Nanocrystal growth has traditionally been explained by monomer attachment with facet 310, 350, 353 energies dictating morphology due to the faster growth of higher energy facets. Beam induced growth of Pt nanocrystals in a liquid-cell was observed at atomic resolution, allowing the development of the nanocrystal s facets to be tracked. 354 It is found that in the early stages of growth the particles are truncated octahedra displaying {100}, {110} and {111} faces. Initially all three families of facets grow at a similar rate and the particle retains the truncated octahedral morphology as it grows. 354 However, when the distance from the particle s centre to its {100} faces reaches ~2.5 nm the {100} facets stop growing. 354 The other facets continue to grow until eventually a cube with (100) faces is formed. 354 It is found that {100} facet growth often involves step formation while flat propagation of {100} and {111} facets is generally observed. 354 Density functional theory (DFT) modelling has found that {100} faces have higher energy the {111} faces, even when ligand binding energies are accounted for. 354 This suggests that nanocube growth is dominated by kinetic factors rather than thermodynamics. 354 It is calculated that the energy barrier for a ligand hopping to an adjacent surface site is significantly lower on the {111} surfaces and therefore it is suggested that the resulting greater ligand mobility may be the cause of the faster facet growth

81 Variation of the electron dose rate can be used to mimic the ex situ process of changing the concentration of reducing agents and this approach has been used to explore the mechanism of gold nanocrystal growth. 116 By altering the dose-rate, beam induced growth of gold nanocrystals can be moved between mechanisms dominated by thermodynamic and kinetic control. 116 At high dose rates, branched morphologies are formed, while at lower dose rates faceted nanocrystals and nanoplatelets with large area low energy {111} facets are formed. 116 The critical growth rate below which thermodynamically favourable faceted nanocrystals are formed is found to be ~3 atomic layers per second, at faster rates adatoms have insufficient time to find energetically favourable sites. 116 Nanoplates are observed to transform from triangles to hexagons when their lateral size exceeds ~35 nm. Post-mortem HRTEM imaging suggests that this transformation is due to the formation of additional twin planes parallel to the {111} faces. 116 A number of liquid-cell studies have observed nanocrystal growth by monomer attachment. 46, However, there is an increasing body of evidence to suggest that many nanocrystals in fact grow by an alternative mechanism where nanocrystal building blocks combine and coalesce. 46, 112, Oriented attachment refers to the process where adjacent particles orientate so that they can join at a shared crystal facet; subsequent , 359 fusion can occur, forming a single nanocrystal. Growth by this non-classical mechanism often leads to anisotropic crystal structures with large aspect ratios such as nanorods and nanowires. 357, Growth of nanostructures by oriented attachment can be inconclusive in ex situ studies, for example pearl necklace shaped wires are seen during the formation CdSe nanowires (presumed to be an intermediate state in the oriented attachment mechanism). 363 Liquid-cell studies have played an important role in 46, 112, 358, providing direct evidence of coalescence based growth mechanisms. Observation of beam induced Pt NC growth has shown that growth by monomer attachement and coalesence can both occur simultaneously in the same reaction solution. 46 Interestingly the particles formed by the two different mechanisms arrive at the same final size and morphology, so would be indistinguishable in ex situ studies. 46 The growth of platinum iron nanocrystals in liquid-cells has provided direct evidence for the growth of 112, 366 anisotropic nanostructures by coalescence of nanocrystal building blocks. Beam induced growth of NCs was observed from an organic solution of Pt and Fe precursors in a SiN x liquid-cell, initially many small NCs nucleated and grew by both monomer attachement and coalescence, once they reached ~5 nm in diameter a second growth mechanism took over. 112 As shown in chapter 2, figure 9, the individual NC building blocks interacted to form necked chains, the chains gradually smoothed and relax to form straight Pt 3 Fe nanowires 81

82 (NWs). 112 The growth trajectories observed suggest that dipolar interactions drive the onediemnsional (1D) assembly of NC building blocks. 112 Decreasing the concentration of oleylamine (a common surfactant in NC syntheses) was found to affect the stability of the NWs. 366 A 30% surfactanct concentration lead to the formation of stable NWs, while a 20% concentration apparently destabilised the NWs, and a 50% concentration results in no coalescence, with individual NCs remaining stable in solution. 366 Figure 9. Beam-induced growth of Pt 3 Fe NW observed by liquid-cell TEM, an oriented attachement mechanism involving the coalescence of NC building blocks is observed, initially kinked wires are observed to straighten. Adapted from Liao et al. Science 2012, 336 (6084), Reprinted with permission from AAAS. During the beam induced growth of gold nanocrystals from an aqueous precursor solution coalsecence events may be tracked with lattice resolution HRTEM, revealing the dependence on the angle between the two coalescing particle s crystal structures. 358 It is observed that there are two mechanisms of coalescence, when the angle of attachement is below a critical angle defect free bonding occurs, while if the angle of attachement is greater than the critical angle a defect plane forms at the interface between the two crystals (chapter 2, figure 10). 358 For ~10 nm diameter nanocrystals the critical angle is found to be ~

83 Figure 10. Solution phase Coalescence of gold nanocrystals revealed by liquid-cell TEM. Lattice resolution imaging reveals two mechanistic pathways for coalescence: (A) Coalescing particles have a lattice mismatch less than the critical angle and a defect free bond is formed or (B) angle of attachement is greater than this critical angle and defects form at the interface between the fused nanocrystals. Reprinted with permission from Aabdin et al. Nano Letters 2014, 14 (11), Copyright 2014 American Chemical Society. While the literature to date is dominated by observations of monometallic nanocrystal 46, , , 120, 139, , growth, a 103, 118, 371 compound nanostructures using in situ (S)TEM. For few reports exist studying the growth of example, the growth of PbS nanocrystals has been observed from a solution of lead acetate and thioacetamide. 103 While the rapid motion of growing NCs in solution prevented high resolution STEM imaging of the majority of particles in the solution, some particles become attached to the windows which allowed lattice fringes (0.21 nm, PbS (220) planes) to be resolved in these more static particles. 103 It was shown that varying the ratio of Pb:S precursors can have a pronounced effect on product morphology, with increased S precursor concentrations leading to flower-like morphologies. 103 Shelling reactions The deposition of a shell of one material onto a previously synthesised particle made from another material is an important route to produce core-shell nanostructures Liquid- 102, 113 cell electron microscopy has allowed the first direct observation of shelling reactions. Beam induced deposition of Pd on Au nanocrystals has been investigated with liquid-cell STEM, 113 showing that the morphology of Au particles affects the shelling behaviour: when spherical 5 nm seeds are used uniform Pd shelling is observed, whereas when 15 nm 83

84 icosahedra or triangular prisms are used non uniform shelling occurs with enhanced growth at corners. 113 On isocahedral particles a thin shell grows on facets, while dendritic chains grow from corners. 113 Triangular particles have larger facets and fewer corners and here growth proceeds only at the three corners again forming dendritic structures. 113 The mechanism of dendrite formation is observed to involve attachement of Pd clusters, as the larger particles are dominated by facets (which supress Pd deposition) these particles provide insufficient sites for Pd deposition, Pd clusters nucleate in solution and sinter with Pd filaments to form dendrites. 113 A holder with flow capabilities has been used to study Au deposition on icosahedral Pt NCs. 102 The flow-cell allows the seed particles to be kept separate from the precursor solution until imaging is underway. 102 Fluid flow also allows precursor concentrations to be kept constant and reduces beam induced effects by removing reactive species such as hydrated electrons from the imaging area Imaging conditions under which the electron beam does not cause Au nanoparticle growth in the precursor solution were identified and Pt nanocrystals were loaded into a liquid-cell and imaged. The aqueous Au precursor solution was then flowed through the cells and deposition of Au on the Pt nanocrystals observed. 102 Surface changes due to metal deposition can be seen in TEM images where it was observed that the surface coating is not uniform, with preferential deposition on the corners of the seed particles causing protrusions to appear on the corners. 102 If the flow of Au solution is stopped these corner protrusions are observed to shrink while the faces grow, indicating the role of surface diffusion in redistributing Au from corners to faces. 102 Post-mortem analysis and ex situ synthesis While liquid-cells represent a powerful tool for studying dynamic processes, understanding is often enhanced by additional post-mortem analysis. 102, 116, 375 Drying the windows after a liquid-cell experiment and observing them in a vacuum allows more detailed atomic 112, 116 resolution and spectroscopic characterisation of the nanostructures formed in situ. Gold NCs grown in a liquid-cell were studied post-mortem, using SEM imaging and HAADF- STEM tomography to confirm their three-dimensional structure. 116 Post-mortem HRTEM imaging also revealed stacking faults in the NCs, which helped explain the growth observed in situ. 116 Post-mortem EDX analysis was performed on Pt 3 Fe NWs grown in a liquid-cell to confirm their composition

85 If we wish to relate in situ observations to bench top reactions it is also valuable to characterise NCs synthesised ex situ using similar reaction condition to those found in the liquid-cell. 102, 375 When studying Au deposition of Pd NCs Wu et al. used EDX spectrum imaging of ex situ synthesised Pd-Au NCs to confirm that elemental distributions were consistent with in situ growth. 102 When studying nanocrystals formed via a galvanic replacement reaction, Sutter et al. compared liquid-cell results with ex situ synthesised nanocrystals made using an idential reaction solution, with such experiments helping confirm that in situ observations are not dominated by beam induced effects. 375 Graphene liquid-cells Graphene s single atom thickness, impressive strength, 376 and the low atomic number of carbon mean that it is highly electron transparent For this reason it has been investigated as a TEM support, offering improved contrast when imaging low Z structures, with the additional possibility of subtracting the periodic contribution of the crystalline support by Fourier space masking Graphene has also been shown to be impermeable to gases and liquids Combined with its extraordinary mechanical strength and 376, 383 stiffness this makes it an attractive material for liquid-cell windows. Graphene liquidcells (GLCs) can be fabricated by trapping pockets of liquid between two graphene sheets. 132, 355, 368, Yuk et al. demonstrated the feasibility of a GLC by encapsulating a Pt precursor solution between two graphene sheets. 132 The electron beam induced growth of Pt nanocrystals was observed with HRTEM at atomic resolution, revealing details of the coalescence and surface faceting processes. 132 A subsequent study of Pt NC growth used low precursor dilutions and thin liquid layers found at the edge of the liquid pocket to achieve slow growth rates, reduce the chance of coalescence events, and limit nanocrystal mobility. 355 Under these conditions it was possible to observe nanocrystal growth by monomer attachment at atomic resolution. 355 The evolution of facets was monitored in real-time; initially low and high index facets coexist, however, as the particle grows the high index facets (such as {433} and {233}) shrink and disappear, due to the faster growth of these high-energy surfaces. 355 Eventually all the remaining facets are {111} or {100}: the lowest energy surfaces are those for which monomer attachment is slowest (chapter 2, figure 11). 355 The high spatial resolution and slow growth rate allows the atom by atom growth of surfaces to be followed, ledge-by-ledge growth is observed with surface diffusion occurring to fill imperfect ledges. 355 nanocrystals connected by DNA. 385 GLCs have also been used to study the dynamics of Au These Au NCs can be observed moving in 3D and from their trajectories and interparticle distances it is possible to infer the configuration of the DNA linker, potentially offering a means to study the conformations of soft materials in 85

86 solution. 385 More recently, the 3D motion of freely rotating Pt nanocrystals has also been observed at atomic resolution in a GLC. 384 By acquiring high frame rate videos (50 fps) of a single particle tumbling in solution it is possible to build up a library of 1000s of images of the same crystal at different orientations, such that single particle reconstruction techniques can then be used to determine the nanocrystal s 3D structure. 384 Ferritin molecules in water have also been studied in GLCs, the acquisition of EELS spectra in this study demonstrates the potential to use GLCs to probe chemical bonding in liquids. 386 Figure 11. TEM images of a Pt nanocrystal growing in a GLC. Growth by monomer attachment is observed at atomic resolution; high index facets are observed to disappear due to their faster growth. Figure adapted with permission from Jeong et al. Chemistry of Materials 2015, 27 (9), Copyright 2015 American Chemical Society. It is understood that in liquid-cells, liquid close to the windows will have different properties from that in the bulk. 387 For example, it is believed that interactions between 144, 358, solution phase NCs and the window can lead to a significant damping of NC motion. 387 Factors such as nanocrystal surface coating and window hydrophobicity may affect the strength of this interaction GLCs are reported to have lower substrate-nanocrystal interactions, making them especially attractive for studies such as those outlined above where 3D motion of particles in solution is critical While the extreme thinness of both the window material and liquid layer are attractive for high resolution imaging, 355 it should be noted that graphene liquid-cells have a number of disadvantages, compared to MEMS SiN x windowed designs: they are hard to produce, the size and thickness of the liquid pocket is not well controlled, and experiments involving liquid-flow, electrochemistry, or heating are all impossible. Furthermore, although graphene cells have given some of the highest resolution imaging in liquid, a number of studies have demonstrated lattice 103, 112, 118, 354, 358, 388 resolution imaging in SiN x windowed cells.

87 Summary and future challenges Liquid cells allow direct observation of dynamic nanoscale processes in liquid. This technology has already proved itself to be a powerful tool for observing the growth of nanocrystals in real-time, giving important insights into coalescence based growth mechanisms, 46, 112, 358 facet evolution, , 389 and shell deposition. To date, almost all studies of nanocrystal growth have observed beam induced processes. 46, , 120, 368 This partly reflects the difficulty of applying other stimuli to liquids in situ. The development of holders where heating can be applied to liquids in situ will allow observations that are more directly transferable to and comparable with the laboratory. Liquid-cell experiments at elevated temperatures have been demonstrated by fabricating a closed-cell that can fit into a standard furnace type heating holder (used to observe the oscillatory growth trajectories of bismuth NCs at ~180 C). 101 However, such approaches are not compatible with flow-cell designs. The future integration of heating elements into the liquid-cell windows, as has been successfully demonstrated in atmospheric pressure gas cells, 91 and is likely to prove a more versatile and controllable solution. It has also been demonstrated that a laser pulse can be used to initiate PbS nanocrystal growth in a liquidcell. 103 The prevelance of beam-induced reactions also illustrates the dramatic effect electron irradiation has on solution phase chemistry. 127 Understanding and controlling radiolysis products represents an important challenge, the development of low-dose imaging techniques or the scavanging of reactive radiolysis products are likely to prove important if we are to move beyond beam-induced reactions. 375, 390 Direct electron detectors represent an important technological development, 391 electron doses at high frame rates. 384 offering the ability to image with lower Another potential pitfall in liquid-cell experiments in misinterpretation of the chemistry observed; due to the complexities of beam induced chemistry predicting the chemical reactions that will occur in the liquid-cell is not trivial. Post-mortem analysis of structures grown in situ is an important means of confirming that the nanocrystals grown have the expected chemistry. 112, 116 Until recently analytical techniques have not been possible in 44, 389, situ, with EELS hampered by sample thickness and EDX impossible due to shadowing. 392 Recent work on the modification of liquid-cell holder designs has made EDX analysis and EDX spectrum imaging in liquid-cells possible for the first time. 44, 392 The ability to determine the chemical compositions and monitor changes in elemental distributions in situ should dramatically improve our ability to study complex multicomponent systems in liquid-cells. The EDX capabilities of a new modified liquid-cell holder are demonstrated in appendix

88 2.4.5 In situ studies of hollow nanocrystal formation Traditionally mechanistic insights into the galvanic replacement reaction have been gained through periodically stopping the reaction or systematically changing the reaction 167, 328 conditions and then charactering the products ex situ. However, recently in situ techniques have been investigated as method for directly monitoring galvanic replacement 375, 390 reactions. Liquid-cell electron microscopy has been used to study the galvanic replacement reaction between Ag templates and a PdCl 2 solution, demonstrating that it is possible to directly observe this solution phase reaction in real-time with nanometre resolution. 375 As discussed in sections 1.6 the electron beam can have a significant effect on 47, 127 the chemistry of a liquid-cell and can cause nanocrystal growth in metal salt solutions. Such beam induced processes present a significant challenge when attempting to observe galvanic replacement as the metal salts used typically have a high reduction potential and are therefore readily reduced by the electron beam. Low dose methods and the use of scavengers, to remove reactive radiolysis products such as radicals, are therefore required. 375 Given the bimetallic nature of the reaction and the importance of elemental distributions in both our understanding of the mechanism and the properties of the products, these reactions would be promising candidates for future liquid-cell EDX studies. 44 However, as EDX spectrum images typically require electron doses at least an order of magnitude higher than images avoiding beam-induced artefacts would be extremely challenging. As with galvanic replacement, the majority of mechanistic insight into the nanoscale Kirkendall effect have come from ex situ analysis, comparing structures isolated after different reaction times or using different reaction conditions. 197, 201, 203 Recently, in situ electron microscopy has been used to probe the mechanistic features of the nanoscale Kirkendall effect. 202, 393 The oxidation of nm diameter Bi nanoparticles in a mixture of oleylamine and dichlorobenzene has been observed using liquid-cell TEM, allowing direct observations of morphological evolution during the oxidation process. 202 At elevated temperatures (180 C) and under electron beam irradiation the particles undergo oxidation, forming a Bi 2 O 3 shell (revealed by post-mortem HRTEM). Asymmetrical void formation, similar to the hemispherical voids observed in Cd sulfidation, 201 occurs, however, it was observed that here the shell grows thicker on the void side of the particle, indicating that Bi diffusion is faster through the shell adjacent to the void, as opposed to the shell adjacent to the remaining Bi core. 202 Estimates of the diffusion constant for Bi in Bi 2 O 3 find values three orders of magnitude larger than those for the bulk. 202 When only a small amount of Bi remains in the core it is observed to form small mobile droplet like structures on the inner 88

89 wall of the shell and it is suggested that these may be liquid due to melting point depression in small nanoparticles. 202 The product morphology is found to be strongly temperature dependent, with room temperature oxidation yielding solid oxide nanoparticles, while at a higher temperature of 192 C hollow particles form but are unstable. 202 The Kirkendall effect has also been observed in situ using an environmental transmission electron microscope (ETEM). 393 The oxidation of Ni nanoparticles, and Ni nanoparticles doped with 5 at. % Cr was investigated. 393 Direct observations were made of changes to the particle s morphology during heating to 375 C under a 25% O 2 75% Ar atmosphere at 1 mbar pressure. 393 In the early stages of oxidation faceting is observed at the metal-oxide interface and oxide nucleation initially occurs at multiple sites on the particle s surface. When vacancies become sufficiently saturated in the core a single void forms and grows, it is observed that oxidation accelerates once this void has formed. 393 In addition to in situ imaging, ex situ EDX spectrum imaging was used to study elemental distributions in Cr doped particles before and after oxidation. 393 It is found that exposure to air at room temperature results in the doped particles initially having a thin Cr enriched oxide layer at their surface, indicating preferential oxidation of Cr. 393 During the subsequent heating in air it is believed that Ni diffuses out through this Cr rich oxide layer while O diffuses in. Postheating EDX shows the Cr rich layer embedded in the NiO shell, suggesting that the diffusion process is not entirely dominated by outward diffusion of Ni (which should lead to the Cr rich layer being located on the inner surface of the shell) Hybrid photovoltaics Use of lead chalcogenides in hybrid photovoltaics The goal of the work presented in appendices 4 and 5 is to develop new routes to growing PbS nanocrystals in polymers, with superior control of nanocrystal size and shape. Ultimately the goal of such work is improved hybrid photovoltaic materials. PbS is a IV-VI semiconductor with the rock salt crystal structure; it has a narrow bulk band gap of 0.41 ev and a large exciton Bohr radius of 18 nm Strong quantum confinement effects emerge in semiconductors whose critical dimensions are less than their Bohr radius: bulk electronic bands become quantized and band gaps widen as size decreases. 396 Therefore, PbS nanocrystals with dimensions less than 18 nm show properties distinct from 89

90 that of the bulk material PbSe also has the rock salt structure and an even larger exciton Bohr radius of 46 nm and narrower bulk band gap of 0.28eV. 397 Due to their similar crystal structure (lattice mismatch of ~3%) and chemistry it is not surprising that ternary alloys with the formula PbS x Se 1-x can also be synthesised The optical and electronic properties of lead chalcogenide QDs make them attractive for photovoltaic applications. 240, 399, Most conjugated polymers and commonly used inorganic semiconductors, such as CdSe, have relatively wide band gaps, as light absorbers they are unable to extract energy from the near infra-red region of the solar spectrum. 240 The wavelengths absorbed by lead chalcogenide nanocrystals can be tuned from the visible to the infrared, allowing broader absorption of solar radiation. 241, 244 Due to their large exciton Bohr radii it is possible to synthesise particles with sizes around a tenth of the Bohr radius and in this regime there is strong electronic coupling between adjacent particles, which improves charge transport in devices. 399 A further attractive property of lead chalcogenide QDs is the potential for multiple exciton generation (MEG), MEG has been reported in both PbS and PbSe nanocrystals, 226 and PbSe based solar cells have demonstrated external photocurrent efficiencies greater than 100% as a consequence of MEG. 404 The ternary structure PbS x Se 1-x is an especially attractive photovoltaic material, combining useful properties of both the sulfide and selenide. For example solar cells made with PbS show larger open-circuit voltages, while PbSe tends to give superior short circuit 399, 402 photocurrents due to its higher mobility. However, solar cells made with ternary nanocrystals outperform those made with either PbS or PbSe. 399 The optimum stoichiometry reported varies from PbS 0.7 Se 0.3 to PbS 0.4 Se 0.6 ; this is possibly due to the different polymers used, with the ideal stoichiometry offering good energy level alignment 244, 399 between QDs and polymer. Lead chalcogenide QDs have been incorporated into a wide variety of photovoltaic device architectures, 402, and there have been a number of promising demonstrations of their use in hybrid solar cells. 244, 408 PbSe QDs (~6 nm in diameter) have been shown to improve the infrared response of P3HT based devices, infrared absorption by the PbSe nanocrystals is found to contribute 33% of the total photocurrent under solar irradiation. 241 Devices using PbS x Se 1-x in P3HT (weight ratio 1:1) show a 100 fold increase in performance over P3HT alone, with the hybrid device showing broader energy harvesting due higher absorbance at longer wavelengths. 90

91 Lead chalcogenide based hybrid devices have often shown poor performance compared to devices containing other nanocrystals, for example CdSe/P3HT devices report power conversion efficiencies of 2% while similar PbS/P3HT devices only achieve efficiencies of 0.01% This dramatic difference is likely a consequence of poor energy level alignment, with a small energy gap between the lead chalcogenide valence band and the polymer HOMO. 408 The recent development of low band gap polymers has dramatically improved performance allowing energetically favourable heterojunctions. 244, 408, 411 PbS nanocrystals in the low bandgap polymer PDPPTPT (90:10 wt.%) show efficient charge transfer enabling the fabrication of devices with ~3% power conversion efficiency (PCE), 248 and PbS nanocrystals in the low band gap polymer PDTPBT device have given a PCE of 3.78%. 408 The current record for PCE in a hybrid solar cell is 5.50%, held by a device using PbS x Se 1-x nanocrystals in PDTPBT. 244 The impressive performance of the latter devices is partly due to the optimised nanoscale morphology of the two phases In situ fabrication of hybrid films Nanocrystal surface chemistry is critical for charge separation and transport. 230 chemical synthesis of colloidal nanocrystals uses long chain (typically containing a backbone of 8 or more carbon atoms) organic ligands to control growth, and provide solubility and stability However, the resulting layer of insulating ligands hinders interfacial charge 230, 248, 414 transfer. The The performance of hybrid solar cells can often be improved by performing ligand exchange on the nanocrystals. 230, 410, 415 Charge transfer is improved when the initial long chain ligands used for synthesis are replaced with short conjugated ligands. 409, 415 For example PbS QDs capped with benzene dithiol show hole transfer to PCPDTBT, while analogous oleylamine capped QDs show none. 411 However, such procedures generally result in poor device morphologies as they reduce nanocrystal solubility in the film casting solvent, leading to low loadings, increased aggregation, and 230, 248, 414 coarser phase segregation. Other approaches to improve interfacial charge transfer include the use of thermally cleavable ligands, 416 and solution phase syntheses employing conjugated polymers as capping agents Despite their promise, the performance of hybrid photovoltaics has been relatively disappointing compared to competing technologies. 230, 248, 419 The highest reported power conversion efficiency (PCE) for a hybrid device is 5.5%, 244 whilst other QD based architectures can exceed 7% efficiency All organic polymer solar cells commonly achieve PCEs of 7% or greater, with the best performing devices exceeding 10% The new generation of perovskite solar cells are now routinely achieving efficiencies greater than 10%, with efficiencies approaching 20% recently reported. 426, Finally, single and 91

92 polycrystalline Si cells can both achieve PCEs over 20% and are becoming increasingly cheaper to produce. 426 It is therefore apparent that, to compete with both alternative solar technologies, the efficiency of hybrid devices must improve. Strategies for fabricating hybrid films with controlled nanoscale morphologies and ligand-free organic-inorganic 245, 414, 431 interfaces will be crucial to improving performance. In situ growth of inorganic nanocrystals in a polymer film removes the need for capping ligands. 414, The in situ approach uses molecular precursors many of which are soluble in the same solvents as the polymer. A solution of polymer and precursor can be spin coated to form a polymer-precursor film. This film is then heated, causing the precursor to decompose and metal sulphide nanostructures to grow in the polymer film (chapter 2, figure 12). 414, 432 Problems associated with QD solubility and aggregation, or concerns about presynthesised nanocrystals effecting micro- or nanofluidics (relevant for techniques like inkjet printing or electrospinning) are removed with this in situ synthesis approach. 435 Furthermore, simplicity and scalability are significantly increased. The commonly used hot injection method of nanocrystal synthesis is extremely challenging to scale to industrially relevant quantities and ligand exchange steps add additional complexity to the ex situ 230, approach. In contrast, the process of coating a solution onto a substrate and heating the resulting film is simple and compatible with existing mass production techniques. 438 Figure 12. The in situ approach to hybrid film fabrication. A molecular precursor and polymer are dissolved in a common solvent, the resulting solution is spin coated onto a substrate, forming a thin film, the polymerprecursor film is then heated, causing the thermal decomposition of the precursor and the growth of nanocrystals. 92

93 Figure 13. Comparison of the key steps in the fabrication of polymer-qd films for hybrid solar cells using the in situ and ex situ approaches. This in situ approach is simpler as it removes the need for ligand-exchange procedures and eliminates problems such as QD aggregation and low QD loadings, however, precise morphological control of QDs if far easier ex situ. The biggest attraction of the in situ approach is the possibility of producing ligand-free NCpolymer interfaces. 414, 431 Comparison of CdS-P3HT devices made using in situ and ex situ synthesised quantum dots has shown that in situ films are better at dissociating excitons into free charges, even after ligand exchange has been performed on ex situ nanocrystals Molecular precursors for metal sulphides Precursor molecules containing all the constituent elements of a desired material are known as single source precursors (SSPs). SSPs such as metal xanthates and dithiocarbamates have been widely used in the fabrication of metal sulphide thin films and nanocrystals Common techniques for converting the precursor into the desired material include hot-injection synthesis for the fabrication of nanocrystals and aerosolassisted chemical vapour deposition (AACVD) for the fabrication of thin films In situ growth of metal sulphide nanocrystals in a polymer matrix has generally relied on two families of SSP: metal xanthates and metal thiolates. The xanthates are especially attractive for hybrid photovoltaics: their decomposition temperatures are significantly lower than alternative precursors such as dithiocarbamates and thiolates Metal xanthates are believed to decompose by a Chugaev-like mechanism (chapter 2, figure 14), forming the metal sulphide, an alkene, H 2 S, and COS. 445, Alkenes shorter than hexene are gaseous at room temperature and some longer chain alkenes have boiling points below 93

94 typical reaction temperatures (1-hexene, 1-heptene, and 1-octene have boiling points of 63, 94, and 121 C respectively). Consequently, the major by-products of xanthate decomposition are generally gaseous, leading to the production of high purity target materials. 448 Figure 14. Metal xanthate decomposition is believed to occur via a Chugaev-like mechanism, the relatively lowtemperature decomposition of these precursors is attractive as it produces volatile by-products (H 2 S, COS, and 445, 447 an alkene). The attractive properties of xanthates have led to their use in AACVD where they have allowed low temperature ( C) deposition of PbS on plastic substrates, 448 and the formation of thin NiS films at 300 C. 447 Solution phase synthesis of nanocrystals using metal xanthate precursors allows even lower reaction temperatures, this is especially true when the solvent used is an alkyl amine, which can lower the decomposition temperature by more than 70 C. 445 Hot injection of metal xanthates has been used to synthesise a wide variety of metal sulphide nanocrystals, including CdS, PbS, ZnS, and Cu 2 ZnSnS , Mild reaction conditions can often be used and considerable size and shape control is possible by varying the concentration, time, temperature, and the solvent used in the reaction. 453, Furthermore, simultaneous decomposition of two different metal xanthates can be used to produce alloyed metal sulphide nanocrystals with controlled composition , 448 Lead xanthates have been used as a precursor for PbS thin films and nanocrystals. They are notable for their especially low decomposition temperature in the presence of primary amines, with lead xanthates decomposing at room temperature in decylamine. 445 While monodentate n-donor ligands such as pyridine have been successfully complexed to 414, 434, Ni and Cd xanthates, they cause lead xanthates to decompose at room temperature However, bidentate N-donor ligands can successfully form adducts with lead xanthates, typically lowering the decomposition temperature and melting point. 446 Choice of precursors for in situ nanocrystal synthesis requires consideration of solubility, decomposition temperature, and thermolysis by-products. 245, 457 As the polymer matrix is subjected to the same heating as the precursor molecules, heating conditions that damage 94

95 the polymer are useless (most work with P3HT uses temperatures of , 450 C or lower). Consequently precursors with low decomposition temperatures are attractive as they allow gentle heating. 450 While thiolates have been used for in situ synthesis, thermogravimetric analysis of Cd(SR) 2 precursors suggest that the decomposition onset is over 220 C, 458 and thiolate decomposition in a polymer matrices typically employs temperatures of C Consequently thiolate precursors will be incompatible with many conjugated polymers and have mainly found use in more robust matrices such as polystyrene In contrast, metal xanthates often decompose at temperatures below 160 C. 414, 434, 450 Given the goal of ligand free interfaces the gaseous by-products of xanthate decomposition are also attractive. 450 As the precursor and polymer must be simultaneously coated from the solution phase, high precursor solubility in organic solvents which are also able to solubilise the conjugated polymer is necessary in order to allow fabrication of films with high precursor (and hence high nanocrystal) loadings. Long alkyl chains typically improve precursor solubility in relevant solvents. 432, 450 The solubility of short chain xanthates can also be improved by adducting the complex; pyridine adducts of cadmium xanthates have been widely used. 245, 414, 434 However, this approach is of questionable merit as it undermines arguments based on by-product volatility. It is also not universally applicable due to the instability of the pyridine adducts for some systems, such as lead xanthates. 446 Chapter 2, table 1 summarises the xanthate precursors used in the in situ approach to date. The first report of in situ synthesis using a metal xanthate precursor was by Leventis et al. 434 In this work the pyridine adduct of cadmium ethyl xanthate was used to produce a CdS/P3HT film with a CdS: polymer mass ratio of 4.7:1, from which a device with a PCE of 0.7% was produced. 434 Later work by the group looked at optimising this reaction, and found that increased reaction temperatures lead to smaller CdS domains and higher charge photogeneration and that the deposition of a CdS interface layer further improves performance, 414 the resulting devices were able to achieve improved PCEs up to 2.17%. 414 Moving away from the short chained ethyl xanthate and using longer chained ligands can remove the need for adducting to achieve good solubility; Rath et al. have used the 2,2- dimethylpentan-3-yl xanthates of Cu and In to form copper indium sulphide (CIS) in a polymer matrix, with the resulting device PCE of 2.8% being the best performing CIS/polymer device to date. 432 The low temperature decomposition of xanthate precursors means that this approach is compatible with polymer substrates, and the same precursors were used to make flexible CIS/polymer devices with PCEs of 1.6%. 450 Antimony ethyl xanthate has been to make Sb 2 S 3 /P3HT films via the in situ approach, the resulting devices 95

96 achieved PCEs of 1.29%. 246 Several studies have added non-stoichiometric amounts of a 450, 461 primary amine to the xanthate-polymer solution, as this has been shown to lower the temperature for in situ precursor decomposition and crystal growth, allowing even gentler processing conditions. 450, 461 The solution processability and low decomposition temperature of xanthates have also been exploited in a number of other approached to produce photovoltaic materials Spin coating the bismuth xanthate Bi(S 2 COEt) 3 (with no polymer present) onto a substrate and annealing at low temperatures, produces a mesoporous Bi 2 S 3 film which can then be infiltrated with a hole transporting polymer (P3HT) to make a hybrid film. 462 Another promising way of incorporating quantum dots into solar cells is to use them to sensitize TiO 2 films, 463 various metal xanthates have been spin coated onto mesoporous TiO 2 films and annealed at 200 C to produce TiO 2 films sensitized with CdS, Sb 2 S 3, or Bi 2 S 3 nanocrystals

97 Table 1. Summary of literature reports of metal xanthate precursors used to synthesise metal sulphide nanocrystals in situ in a polymer matrix; the heating temperature, polymer and inorganic phases, and the resulting size and morphology of nanocrystals reported are provided alongside the precursor structure. Nanoparticle Precursor Heating temp. Polymer phase Inorganic Phase size and shape 150 C P3HT CdS 150 C P3HT CdS ~5.5 nm diameter particles 160 C P3HT CdS 60 mins 160 C P3HT CdS 60 mins Reference 414, ±1 nm diameter particles, regardless of precursor structure. 195 C P3HT or Bi 2S 3 nanorods, 7- for 15 PMMA 22 nm in minutes diameter and nm in length 3-6 nm particles, time and temperate effects size and extent of C PBPFO CdS aggregation C P3HT Sb 2S C PCDTBT CIS nm Spheres 2.4 nm spheres C PSiF-DBT 3-5 nm spheres

98 2.5.4 Challenges for in situ synthesis It is well established that the size and shape of nanocrystals as well as their distribution in the polymer influences the performance of hybrid photovoltaics. 244, 247 Ex situ nanocrystal synthesis allow significant size and shape control through use of ligands, temperature, and precursor concentrations. 453, 466 In contrast, understanding of the factors influencing crystal size and film morphology in the in situ approach remains limited. 245, 414 The role of the SSP is underexplored with most studies concentrating on a single precursor system. A notable exception to this oversight is the work of Maclachlan et al. who compared a series of cadmium xanthates with different chain lengths, and demonstrated that precursor structure can have a significant effect on device performance. 245 This result is due to longer chain xanthates reducing nanocrystal aggregation, leading to smaller domain sizes within the film. 245 It has also been shown that higher temperatures can lead to smaller domains. 414 While control over domain size and aggregation is demonstrated, evidence of nanocrystal size and shape control is not evident in this work. 245 Size control is more evident in studies using metal thiolate precursors; both increasing reaction temperature and decreasing alkyl chain length result in increased nanocrystal size. 460 Work on the in situ synthesis of metal nanoparticles from alkyl thiolates has also revealed increased particle size with increased precursor/polymer ratio. 467 In the work presented in appendix 4 I investigate the decomposition of lead(ii) xanthate precursors in a polymer matrix, and demonstrate, for the first time, that size and shape control through precursor design is possible for in situ decomposition of metal xanthate precursors. 457 To fully realise the promise of the in situ approach further work is needed on developing suitable control over particle size, shape, and 3D network morphology. The latter will require the use of sophisticated characterisation; the relationship between a hybrid layer s photovoltaic performance and it three-dimensional morphology has been demonstrated in a number of electron tomography studies. 247, 468 Cross sectional TEM imaging can also be useful in investigating film structure: CdS-P3HT films fabricated using a cadmium-xanthate precursor imaged in cross section reveal vertical phase separation with networks of CdS nanocrystals primarily found in the bottom half of the film, closer to the substrate, while the upper half of the polymer film appears to contain no nanocrystals. 434 Cross sectional TEM imaging has also been used to investigate in situ synthesised polymer-cis films, 432 and ex situ synthesised CdSe-P3HT films. 239 To date, the focus of in situ synthesis has largely been the growth of metal sulphide 414, 431, 434, 465 materials in the polymer. There is only one example of metal selenide formation via an in situ reaction. 249 A cadmium selenolate precursor (chapter 2, figure 15) 98

99 was heated at C in a PBPFO matrix to form 4-8 nm diameter CdSe nancrystals. 249 In many cases ternary systems containing S and Se offer desirable properties, so the in situ growth of materials like PbS x Se 1-x is an important goal. 244, Although ternary sulphides, containing two cation species, have been synthesised in situ by simultaneous decomposition of two different metal xanthates, many useful systems of this type, for example Cd x Zn 1-x S, have yet to be made in situ. 456 Figure 15. A selenolate precursor for the in situ synthesis of cadmium selenide nanocrystals. 249 Combining such precursors with metal xanthates could offer a route to ternary MS x Se 1-x materials. The properties of the polymer may also influence film morphology. Several studies have remarked that the polymer phase is likely to act much like a capping agent does in conventional syntheses, controlling nanocrystal size and aggregation. 432, 434 It is therefore conceivable that in the future the chemistry of conjugated polymers could be tuned to encourage desirable nanocrystal morphologies Summary Polymer-QD hybrid films have potential for use in low-cost, light-weight, flexible photovoltaics , 450 Lead chalcogenides are amongst the most promising QD materials, with the current record for hybrid photovoltaic efficiency held by a PbS x Se 1-X based device. 244 However, the performance of hybrid devices is hampered by the QD s surface ligands. 230 The in situ approach has recently emerged as a simple scalable technique to fabricate ligand free hybrid films , 414 morphological control still remains a major challenge. Systematic This approach is still in its infancy, and studies of the influence of precursor structure and heating conditions on nanocrystal size and film morphology are needed. Furthermore, many metal chalcogenide systems have yet to be fabricated in situ. My work, presented in appendix 4, demonstrates an important development for in situ synthesis of metal sulphide nanocrystals. Using a range of lead(ii) xanthate precursors I demonstrate the first example of in situ synthesis of PbS nanostructures. 457 Furthermore, I demonstrate, for the first time, that control over the size and shape of nanocrystals is possible in the in situ approach by altering the xanthate precursor s structure. 457 I show that longer chained xanthates give smaller nanocrystal dimensions and that anisotropic growth is possible. 457 The formation of PbS nanowires when using lead(ii) octyl xanthate as a 99

100 precursor is especially promising. 457 By widening the library of nanocrystals that can be grown in situ to include the attractive properties of PbS and by establishing the tools required for morphological control, this work may prove to be an important step towards improved hybrid solar cell performance

101 3. Conclusions and future work In situ STEM techniques allow direct observation of nanocrystal growth trajectories and transformations. 46 In situ experiments can record the continuous history of a single region, allowing nanocrystal growth in liquids or transformations during heating to be observed at atomic resolution. 75, 315, 358 However, until recently, it has been challenging to acquire 389, 392, 469 chemical information during in situ experiments. In this thesis I have performed the first nanoscale EDX spectrum imaging on samples in liquid (appendix 1). 44 SiN x windowed liquid-cell holders have previously prevented EDX analysis due to shadowing, however, modified holder designs allow the detection of characteristic X-rays. 44, 392 EDX is well suited to liquid-cell experiments where the intrinsic sample thickness makes EELS analysis challenging or impossible. 389, 469 My work demonstrates simultaneous mapping of the distributions of multiple elements with nanometre resolution. 44 For chemically complex systems this new technique will prove invaluable for unambiguous interpretation of solution phase transformations. The technique has subsequently been used to study bacterial synthesis of bimetallic nanocrystals. 470 My initial work on liquid-cell EDX has been qualitative, using EDX spectrum imaging to identify the elements present and their spatial distributions. 44 However, in future studies quantitative data could be valuable. As the mass attenuation coefficients for water are known, absorption corrections should be possible. 471 However, such calculations will require knowledge of the nanocrystal s location in the liquid layer and hence the X-ray path length. The field of liquid-cell electron microscopy is currently dominated by beam-induced transformations, future work is likely to attempt to minimise beam-induced processes and 101, 127 instead use more relevant stimuli, such as heating, to drive transformations. Understanding the interaction of the electron beam with the specimen remains a major challenge. Although Schneider et al. have recently proposed a model for electron beam radiolysis in a liquid-cell, their work lacks any direct experimental measurements. 127 The ability to directly measure the concentration of reactive radiolysis products would dramatically advance the field, allowing quantitative interpretations of beam-induced phenomena. A promising avenue for future research would be to use the electrochemical capabilities built into some liquid cells to directly probe solution chemistry. It may be possible to detect electrochemical signatures of key reactive species and by comparison 101

102 with standards of known concentrations quantify the beam-induced concentrations created during an in situ experiment. For in situ EDX analysis improved collection efficiency will be desirable, allowing lower electron doses and improved temporal resolution. This could be achieved by increasing detector collection angles and further reducing holder shadowing, allowing faster and more efficient collection of spectra. To-date, only the holder top-plate has been modified; however, the window chips are also a significant source of shadowing, 44 and further work could investigate modified chips designs as a route to improved X-ray detection. In this thesis I have studied beam-induced and heating-induced transformations of AgAu nanocrystals (appendices 2 and 3). 51 MEMs based heating holders allow EDX spectrum imaging to be performed in situ, 82, allowing changes in elemental distributions to be mapped and quantified. I demonstrated that the electron beam can drive oxidation of Ag, resulting in the formation of hollow nanocrystals via the nanoscale Kirkendall effect. 51 HAADF STEM imaging allows the growth and coalescence of Kirkendall voids to be observed, while EDX spectrum imaging reveals accompanying elemental redistributions, such as the inward diffusion of Au during Ag oxidation. 51 Beam induced processes are of limited relevance, with gas or solution phase reactions representing simpler and more scalable routes to hollow materials. Recent developments in SiN x windowed gas-cell holders allow pressures up to 1000 mbar to be combined with heating and EDX analysis. 472 Such a set-up could be used for future studies of the nanoscale Kirkendall effect, allowing metal nanocrystals to be heated in an oxidising atmosphere to induced transformations which can be simply replicated in a furnace ex situ. 198 Direct growth of metal sulphide nanocrystals in polymer matrices using molecular 245, 414, precursors is a simple and scalable route to quantum dot-polymer hybrid materials , 434 This approach is especially attractive for photovoltaic applications where the long chain capping ligands used in ex situ synthesis are avoided, improving interfacial charge transfer. 245, 431 However, control of nanocrystal size and shape remains relatively poor compared to ex situ synthesis. 431 In this thesis I have demonstrated the first example of in situ growth of PbS nanocrystals in polymer thin films (appendix 4). 457 The demonstration of size and shape control through precursor design represents an important development. Long chain xanthates are shown to be a particularly promising class of precursor, allowing high precursor loadings and low decomposition temperatures, and leading to small nanocrystal dimensions

103 I have also investigated growth of PbS nanocrystals in electrospun polymer fibres (appendix 5). It is possible to fully decompose the lead(ii)octyl xanthate precursor at 90 C. Narrow high aspect ratio PbS nanowires grow in polymer nanofibres, in contrast to the cubes formed in otherwise identical thin films, demonstrating another promising tool for in situ morphological control. The experiments on PbS nanocrystal growth presented in this thesis use polystyrene as a model polymer matrix. 457 While this has allowed us to cheaply explore the affects that a wide range of precursor chemistries and heating conditions have on nanocrystal morphology, future work will use polymers which are more relevant to the envisaged photovoltaic applications. 248 By using a semiconducting polymer with a suitable band gap we could perform transient absorption experiments and fabricate devices, allowing the morphological control we have developed to be related to key photovoltaic performance parameters. 245 A number of previous studied have used ex situ TEM to study hybrid photovoltaic materials, using the morphological information revealed to understand the structure-property relationships. For example, plan-view bright field TEM was used by MacLachlan et al. to study the influence of molecular precursor structure on the morphology of CdS-P3HT films synthesised by the in situ approach. 245 They were able to relate the morphologies observed to optical properties and photovoltaic device performances. 245 Given the importance of the layer s 3D morphology, 2D imaging is limited for improving our understanding of structure-property relationships. 247 Hindson et al. used HAADF electron tomography to reveal the 3D structure of hybrid films made with either spherical particles or nanorods, they were able to show that the spherical particles tend to aggregate, while the nanorods result in more homogeneous nanocrystal loadings and highly connected networks. 247 My existing work on PbS-polymer films has exclusively used planview imaging, tomographic techniques have potential to add greater morphological insights to future studies. To-date, the in situ electron microscopy techniques have not been applied to the study of hybrid photovoltaics. This would seem like a promising avenue for future studies, in particular in situ heating could give important mechanistic insights into the growth of nanocrystals in a polymer matrix. The mechanism by which precursors decompose and nanocrystals grow in a polymer thin film remains poorly understood, a direct observation of the morphological evolution that occurs during heating would therefore, be extremely useful. I have recently attempted to perform in situ heating experiments on the composite system studied in appendix 4, however, the electron beam was found to have a dramatic effect on the process. Regions not exposed to the beam underwent the expected precursor decomposition and nanocrystal growth during heating, 103

104 however, those regions which had previously been exposed to the beam or were imaged during heating showed no nanocrystals growth. This unexpected behaviour is attributed to the beam-induced cross-linking of polystryrene, 473 it seems that the direct observation of nanocrystal growth in a polymer matrix will either require the use of a more beam resilient polymer or gentler imaging conditions. As the properties of PbS x Se 1-x are known to be better than those of PbS for photovoltaic applications, 244, 399 introducing controlled Se doping into our synthesis is another promising direction which we are currently investigating. References 1. Smith, D. J., Chapter 1: Characterization of Nanomaterials Using Transmission Electron Microscopy. In Nanocharacterisation, The Royal Society of Chemistry: 2015; pp Ziegler, A.; Kisielowski, C.; Ritchie, R. O., Imaging of the crystal structure of silicon nitride at 0.8 Ångström resolution. Acta Materialia 2002, 50 (3), Plachinda, P.; Rouvimov, S.; Solanki, R., Structure analysis of CVD graphene films based on HRTEM contrast simulations. physica status solidi (a) 2011, 208 (11), Xin, H. L.; Mundy, J. A.; Liu, Z.; Cabezas, R.; Hovden, R.; Kourkoutis, L. F.; Zhang, J.; Subramanian, N. P.; Makharia, R.; Wagner, F. T.; Muller, D. A., Atomic-Resolution Spectroscopic Imaging of Ensembles of Nanocatalyst Particles Across the Life of a Fuel Cell. Nano Letters 2012, 12 (1), Hunt, J. A.; Williams, D. B., Electron energy-loss spectrum-imaging. Ultramicroscopy 1991, 38 (1), Muller, D. A.; Kourkoutis, L. F.; Murfitt, M.; Song, J. H.; Hwang, H. Y.; Silcox, J.; Dellby, N.; Krivanek, O. L., Atomic-Scale Chemical Imaging of Composition and Bonding by Aberration-Corrected Microscopy. Science 2008, 319 (5866), Nellist, P.; Pennycook, S., The Principles and Interpretations of Annular Dark-Field Z- Contrast Imaging. Advances in imaging and electron physics 2000, 113, Pennycook, S. J., Z-contrast stem for materials science. Ultramicroscopy 1989, 30 (1 2), Williams, D.; Carter, B., Transmission Electron Microscopy: A textbook for materials science. Springer Science: Williams, D.; Carter, B., Chapter 3: Elastic Scattering. In Transmission Electron Microscopy, Springer Science: 2009; pp Treacy, M. M. J., Z Dependence of Electron Scattering by Single Atoms into Annular Dark- Field Detectors. Microscopy and Microanalysis 2011, 17 (6),

105 12. Hartel, P.; Rose, H.; Dinges, C., Conditions and reasons for incoherent imaging in STEM. Ultramicroscopy 1996, 63 (2), Nellist, P. D., Theory and Simulations of STEM Imaging. In Aberration-Corrected Analytical Transmission Electron Microscopy, John Wiley & Sons, Ltd: 2011; pp Pennycook, S.; Jesson, D., High-resolution Z-contrast imaging of crystals. Ultramicroscopy 1991, 37 (1), Jesson, D. E.; Pennycook, S. J., Incoherent imaging of crystals using termally scattered electrons. Proceedings of the Royal Society-Mathematical and Physical Sciences 1995, 449 (1936), Howie, A., Image-contrast and localized signal selection techniques. Journal of Microscopy- Oxford 1979, 117, Pennycook, S.; Boatner, L., Chemically sensitive structure-imaging with a scanning transmission electron microscope. Nature 1988, 336 (6199), de Jonge, N.; Poirier-Demers, N.; Demers, H.; Peckys, D. B.; Drouin, D., Nanometerresolution electron microscopy through micrometers-thick water layers. Ultramicroscopy 2010, 110 (9), Klein, K. L.; Anderson, I. M.; De Jonge, N., Transmission electron microscopy with a liquid flow cell. Journal of Microscopy 2011, 242 (2), Isaacson, M.; Johnson, D., The microanalysis of light elements using transmitted energy loss electrons. Ultramicroscopy 1975, 1 (1), Zaluzec, N. J., Detector Solid Angle Formulas for Use in X-Ray Energy Dispersive Spectrometry. Microscopy and Microanalysis 2009, 15 (2), Schlossmacher, P.; Klenov, D. O.; Freitag, B.; von Harrach, H. S., Enhanced Detection Sensitivity with a New Windowless XEDS System for AEM Based on Silicon Drift Detector Technology. Microscopy Today 2010, 18 (4), Krause, M. O., Atomic radiative and radiationless yields for K and L shells. Journal of Physical and Chemical Reference Data 1979, 8 (2), Lovejoy, T. C.; Ramasse, Q. M.; Falke, M.; Kaeppel, A.; Terborg, R.; Zan, R.; Dellby, N.; Krivanek, O. L., Single atom identification by energy dispersive x-ray spectroscopy. Applied Physics Letters 2012, 100 (15), Thomas, L. E., Light-element analysis with electrons and x-rays in a high-resolution STEM. Ultramicroscopy 1985, 18 (1 4), Brydson, R.; Hondow, N., Electron Energy Loss Spectrometry and Energy Dispersive X-ray Analysis. In Aberration-Corrected Analytical Transmission Electron Microscopy, Brydson, R., Ed. Wiley: 2011; pp Suenaga, K.; Koshino, M., Atom-by-atom spectroscopy at graphene edge. Nature 2010, 468 (7327),

106 28. Ci, L.; Song, L.; Jin, C.; Jariwala, D.; Wu, D.; Li, Y.; Srivastava, A.; Wang, Z. F.; Storr, K.; Balicas, L.; Liu, F.; Ajayan, P. M., Atomic layers of hybridized boron nitride and graphene domains. Nature Materials 2010, 9 (5), Brydson, R.; Sauer, H.; Engel, W.; Zeitler, E., EELS as a fingerprint of the chemical coordination of light elements. Microscopy Microanalysis Microstructures 1991, 2 (2-3), Garvie, L. A., Can electron energy-loss spectroscopy (EELS) be used to quantify hydrogen in minerals from the OK edge? American Mineralogist 2010, 95 (1), Schmid, H. K., Phase Identification in Carbon and BN Systems by EELS. Microscopy Microanalysis Microstructures 1995, 6 (1), Ramasse, Q. M.; Seabourne, C. R.; Kepaptsoglou, D.-M.; Zan, R.; Bangert, U.; Scott, A. J., Probing the Bonding and Electronic Structure of Single Atom Dopants in Graphene with Electron Energy Loss Spectroscopy. Nano Letters 2012, 13 (10), Turner, S.; Lazar, S.; Freitag, B.; Egoavil, R.; Verbeeck, J.; Put, S.; Strauven, Y.; Van Tendeloo, G., High resolution mapping of surface reduction in ceria nanoparticles. Nanoscale 2011, 3 (8), Dileep, K.; Panchakarla, L. S.; Balasubramanian, K.; Waghmare, U. V.; Datta, R., Electron energy loss spectroscopy of ZnO nanocrystals with different oxygen vacancy concentrations. Journal of Applied Physics 2011, 109 (6), Tan, H.; Verbeeck, J.; Abakumov, A.; Van Tendeloo, G., Oxidation state and chemical shift investigation in transition metal oxides by EELS. Ultramicroscopy 2012, 116, Daniels, H.; Brydson, R.; Rand, B.; Brown, A., Investigating carbonization and graphitization using electron energy loss spectroscopy (EELS) in the transmission electron microscope (TEM). Philosophical Magazine 2007, 87 (27), Mecklenburg, M.; Hubbard, W. A.; White, E. R.; Dhall, R.; Cronin, S. B.; Aloni, S.; Regan, B. C., Nanoscale temperature mapping in operating microelectronic devices. Science 2015, 347 (6222), Arenal, R.; Stéphan, O.; Kociak, M.; Taverna, D.; Loiseau, A.; Colliex, C., Electron Energy Loss Spectroscopy Measurement of the Optical Gaps on Individual Boron Nitride Single-Walled and Multiwalled Nanotubes. Physical Review Letters 2005, 95 (12), Gu, L.; Srot, V.; Sigle, W.; Koch, C.; van Aken, P.; Scholz, F.; Thapa, S. B.; Kirchner, C.; Jetter, M.; Rühle, M., Band-gap measurements of direct and indirect semiconductors using monochromated electrons. Physical Review B 2007, 75 (19), Egerton, R. F.; Yang, Y. Y.; Chen, F. Y. Y., EELS of thick specimens. Ultramicroscopy 1991, 38 (3), Bourdillon, A. J.; Hall, D. J.; Morrison, C. J.; Stobbs, W. M., The relative importance of multiple inelastic scattering in the quantification of EELS. Journal of Microscopy 1983, 130 (2), Hosoi, J.; Oikawa, T.; Inoue, M.; Kokubo, Y., Thickness dependence of signal/background ratio of inner-shell electron excitation loss in eels. Ultramicroscopy 1984, 13 (3),

107 43. Zaluzec, N. J.; Burke, M. G.; Haigh, S. J.; Kulzick, M. A., X-ray Energy-Dispersive Spectrometry During In Situ Liquid Cell Studies Using an Analytical Electron Microscope. Microscopy and Microanalysis 2014, 20 (2), Lewis, E. A.; Haigh, S. J.; Slater, T. J. A.; He, Z.; Kulzick, M. A.; Burke, M. G.; Zaluzec, N. J., Real-time imaging and local elemental analysis of nanostructures in liquids. Chemical Communications 2014, 50, Longo, P.; Twesten, R. D., Fast STEM Spectrum Imaging Using Simultaneous EELS and EDS. Microscopy Today 2013, 21 (01), Zheng, H.; Smith, R. K.; Jun, Y.-w.; Kisielowski, C.; Dahmen, U.; Alivisatos, A. P., Observation of Single Colloidal Platinum Nanocrystal Growth Trajectories. Science 2009, 324 (5932), Liao, H. G.; Niu, K.; Zheng, H., Observation of growth of metal nanoparticles. Chemical Communications 2013, 49 (100), El Mel, A. A.; Molina-Luna, L.; Buffiere, M.; Tessier, P. Y.; Du, K.; Choi, C. H.; Kleebe, H. J.; Konstantinidis, S.; Bittencourt, C.; Snyders, R., Electron Beam Nanosculpting of Kirkendall Oxide Nanochannels. ACS Nano 2014, 8 (2), van Huis, M. A.; Figuerola, A.; Fang, C.; Béché, A.; Zandbergen, H. W.; Manna, L., Chemical Transformation of Au-Tipped CdS Nanorods into AuS/Cd Core/Shell Particles by Electron Beam Irradiation. Nano Letters 2011, 11 (11), Zheng, H.; Rivest, J. B.; Miller, T. A.; Sadtler, B.; Lindenberg, A.; Toney, M. F.; Wang, L.-W.; Kisielowski, C.; Alivisatos, A. P., Observation of Transient Structural-Transformation Dynamics in a Cu2S Nanorod. Science 2011, 333 (6039), Lewis, E. A.; Slater, T. J. A.; Prestat, E.; Macedo, A.; O'Brien, P.; Camargo, P. H. C.; Haigh, S. J., Real-time imaging and elemental mapping of AgAu nanoparticle transformations. Nanoscale 2014, 6 (22), Egerton, R. F.; Li, P.; Malac, M., Radiation damage in the TEM and SEM. Micron 2004, 35 (6). 53. Egerton, R. F., Control of radiation damage in the TEM. Ultramicroscopy 2013, 127 (0), Wu, B.; Neureuther, A. R., Energy deposition and transfer in electron-beam lithography. Journal of Vacuum Science & Technology B 2001, 19 (6), Sader, K.; Schaffer, B.; Vaughan, G.; Brydson, R.; Brown, A.; Bleloch, A., Smart acquisition EELS. Ultramicroscopy 2010, 110 (8), Egerton, R. F.; Crozier, P. A.; Rice, P., Electron energy-loss spectroscopy and chemical change. Ultramicroscopy 1987, 23 (3 4), Fryer, J. R., Radiation damage in organic crystalline films. Ultramicroscopy 1984, 14 (3), Fryer, J. R.; Holland, F., High Resolution Electron Microscopy of Molecular Crystals. Radiation Processes at Room Temperature. Proceedings of The Royal Society A 1984, 393, Zan, R.; Ramasse, Q. M.; Jalil, R.; Georgiou, T.; Bangert, U.; Novoselov, K. S., Control of Radiation Damage in MoS2 by Graphene Encapsulation. ACS Nano 2013, 7 (11),

108 60. Algara-Siller, G.; Kurasch, S.; Sedighi, M.; Lehtinen, O.; Kaiser, U., The pristine atomic structure of MoS2 monolayer protected from electron radiation damage by graphene. Applied Physics Letters 2013, 103 (20), Howe, J. M.; Yokota, T.; Murayama, M.; Jesser, W. A., Effects of heat and electron irradiation on the melting behavior of Al Si alloy particles and motion of the Al nanosphere within. Journal of Electron Microscopy 2004, 53 (2), Yokota, T.; Murayama, M.; Howe, J. M., In situ transmission-electron-microscopy investigation of melting in submicron Al-Si alloy particles under electron-beam irradiation. Physical Review Letters 2003, 91 (26), Egerton, R. F.; McLeod, R.; Wang, F.; Malac, M., Basic questions related to electron-induced sputtering in the TEM. Ultramicroscopy 2010, 110 (8), Egerton, R. F.; Wang, F.; Crozier, P. A., Beam-Induced Damage to Thin Specimens in an Intense Electron Probe. Microscopy and Microanalysis 2006, 12 (1), Carpenter, G. J. C.; Watters, J. F., A study of electron irradiation damage in Zirconium using a high voltage electron microscope. Journal of Nuclear Materials 1981, 96 (3), Pells, G. P.; Phillips, D. C., Radiation damage of α-al2o3 in the HVEM: I. Temperature dependence of the displacement threshold. Journal of Nuclear Materials 1979, 80 (2), Krivanek, O. L.; Dellby, N.; Murfitt, M. F.; Chisholm, M. F.; Pennycook, T. J.; Suenaga, K.; Nicolosi, V., Gentle STEM: ADF imaging and EELS at low primary energies. Ultramicroscopy 2010, 110 (8), Suenaga, K.; Iizumi, Y.; Okazaki, T., Single atom spectroscopy with reduced delocalization effect using a 30 kv-stem. European Physical Journal-Applied Physics 2011, 54 (3), Zan, R.; Bangert, U.; Ramasse, Q.; Novoselov, K. S., Evolution of Gold Nanostructures on Graphene. Small 2011, 7 (20), Zan, R.; Ramasse, Q. M.; Bangert, U.; Novoselov, K. S., Graphene Reknits Its Holes. Nano Letters 2012, 12 (8), Lee, J.; Zhou, W.; Pennycook, S. J.; Idrobo, J.-C.; Pantelides, S. T., Direct visualization of reversible dynamics in a Si-6 cluster embedded in a graphene pore. Nature Communications 2013, 4, Huang, P. Y.; Meyer, J. C.; Muller, D. A., From atoms to grains: Transmission electron microscopy of graphene. MRS Bulletin 2012, 37 (12), Suenaga, K.; Kobayashi, H.; Koshino, M., Single atom imaging and spectroscopy in nanostructured carbon materials. MRS Bulletin 2012, 37 (1), Mitchell, D. R. G., Contamination mitigation strategies for scanning transmission electron microscopy. Micron 2015, 73 (0), van Huis, M. A.; Young, N. P.; Pandraud, G.; Creemer, J. F.; Vanmaekelbergh, D.; Kirkland, A. I.; Zandbergen, H. W., Atomic imaging of phase transitions and morphology transformations in nanocrystals. Advanced Materials 2009, 21 (48),

109 76. Asoro, M. A.; Kovar, D.; Ferreira, P. J., In situ Transmission Electron Microscopy Observations of Sublimation in Silver Nanoparticles. ACS Nano 2013, 7 (9), Duchamp, M.; Xu, Q.; Dunin-Borkowski, R. E., Convenient Preparation of High-Quality Specimens for Annealing Experiments in the Transmission Electron Microscope. Microscopy and Microanalysis 2014, 20 (06), Grodzinska, D.; Pietra, F.; van Huis, M. A.; Vanmaekelbergh, D.; de Mello Donega, C., Thermally induced atomic reconstruction of PbSe/CdSe core/shell quantum dots into PbSe/CdSe bihemisphere hetero-nanocrystals. Journal of Materials Chemistry 2011, 21 (31), Yalcin, A. O.; de Nijs, B.; Fan, Z.; Tichelaar, F. D.; Vanmaekelbergh, D.; van Blaaderen, A.; Vlugt, T. J.; van Huis, M. A.; Zandbergen, H. W., Core shell reconfiguration through thermal annealing in FexO/CoFe2O4 ordered 2D nanocrystal arrays. Nanotechnology 2014, 25 (5), Figuerola, A.; Huis, M. v.; Zanella, M.; Genovese, A.; Marras, S.; Falqui, A.; Zandbergen, H. W.; Cingolani, R.; Manna, L., Epitaxial CdSe-Au Nanocrystal Heterostructures by Thermal Annealing. Nano Letters 2010, 10 (8), Yalcin, A. O.; Fan, Z.; Goris, B.; Li, W.-F.; Koster, R. S.; Fang, C.-M.; van Blaaderen, A.; Casavola, M.; Tichelaar, F. D.; Bals, S.; Van Tendeloo, G.; Vlugt, T. J. H.; Vanmaekelbergh, D.; Zandbergen, H. W.; van Huis, M. A., Atomic Resolution Monitoring of Cation Exchange in CdSe-PbSe Heteronanocrystals during Epitaxial Solid Solid Vapor Growth. Nano Letters 2014, 14 (6), Lu, N.; Wang, J.; Xie, S.; Xia, Y.; Kim, M. J., Enhanced shape stability of Pd-Rh core-frame nanocubes at elevated temperature: in situ heating transmission electron microscopy. Chemical Communications 2013, 49 (100), Zaluzec, N. J., X-ray Microanalysis of Nanoscale Materials in Liquids & Gases in the Analytical Scanning/Transmission Electron Microscope. At IAM Nano, Hamburg, Germany, Aduro Application Note AA05.2, EDS in the TEM and SEM using Aduro. (accessed 30/1/2015). 85. Baumgardner, W. J.; Yu, Y.; Hovden, R.; Honrao, S.; Hennig, R. G.; Abruña, H. D.; Muller, D.; Hanrath, T., Nanoparticle Metamorphosis: An in Situ High-Temperature Transmission Electron Microscopy Study of the Structural Evolution of Heterogeneous Au:Fe2O3 Nanoparticles. ACS Nano 2014, 8 (5), Yalcin, A. O.; Goris, B.; van Dijk - Moes, R. J. A.; Fan, Z.; Erdamar, A. K.; Tichelaar, F. D.; Vlugt, T. J. H.; Van Tendeloo, G.; Bals, S.; Vanmaekelbergh, D.; Zandbergen, H. W.; van Huis, M. A., Heatinduced transformation of CdSe/CdS/ZnS core/multishell quantum dots by Zn diffusion into inner layers. Chemical Communications 2014, 51, Gai, P. L.; Boyes, E. D.; Helveg, S.; Hansen, P. L.; Giorgio, S.; Henry, C. R., Atomic-Resolution Environmental Transmission Electron Microscopy for Probing Gas Solid Reactions in Heterogeneous Catalysis. MRS Bulletin 2007, 32 (12), Gai, P. L., Development of Wet Environmental TEM (Wet-ETEM) for In Situ Studies of Liquid- Catalyst Reactions on the Nanoscale. Microscopy and Microanalysis 2002, 8 (01),

110 89. Jinschek, J. R., Advances in the environmental transmission electron microscope (ETEM) for nanoscale in situ studies of gas-solid interactions. Chemical Communications 2014, 50 (21), Williamson, M. J.; Tromp, R. M.; Vereecken, P. M.; Hull, R.; Ross, F. M., Dynamic microscopy of nanoscale cluster growth at the solid-liquid interface. Nature Materials 2003, 2 (8), Creemer, J. F.; Helveg, S.; Hoveling, G. H.; Ullmann, S.; Molenbroek, A. M.; Sarro, P. M.; Zandbergen, H. W., Atomic-scale electron microscopy at ambient pressure. Ultramicroscopy 2008, 108 (9), Allard, L. F.; Overbury, S. H.; Bigelow, W. C.; Katz, M. B.; Nackashi, D. P.; Damiano, J., Novel MEMS-Based Gas-Cell/Heating Specimen Holder Provides Advanced Imaging Capabilities for In Situ Reaction Studies. Microscopy and Microanalysis 2012, 18 (04), Gai, P., Developments in in situ Environmental Cell High-Resolution Electron Microscopy and Applications to Catalysis. Topics in Catalysis 2002, 21 (4), Boyes, E. D.; Ward, M. R.; Lari, L.; Gai, P. L., ESTEM imaging of single atoms under controlled temperature and gas environment conditions in catalyst reaction studies. Annalen der Physik 2013, 525 (6), Gai, P. L.; Boyes, E. D., Advances in atomic resolution in situ environmental transmission electron microscopy and 1Å aberration corrected in situ electron microscopy. Microscopy Research and Technique 2009, 72 (3), Hansen, P. L.; Wagner, J. B.; Helveg, S.; Rostrup-Nielsen, J. R.; Clausen, B. S.; Topsøe, H., Atom-Resolved Imaging of Dynamic Shape Changes in Supported Copper Nanocrystals. Science 2002, 295 (5562), Yoshida, H.; Kuwauchi, Y.; Jinschek, J. R.; Sun, K.; Tanaka, S.; Kohyama, M.; Shimada, S.; Haruta, M.; Takeda, S., Visualizing Gas Molecules Interacting with Supported Nanoparticulate Catalysts at Reaction Conditions. Science 2012, 335 (6066), Helveg, S.; Lopez-Cartes, C.; Sehested, J.; Hansen, P. L.; Clausen, B. S.; Rostrup-Nielsen, J. R.; Abild-Pedersen, F.; Norskov, J. K., Atomic-scale imaging of carbon nanofibre growth. Nature 2004, 427 (6973), Stoltze, P.; Nørskov, J. K., Bridging the "Pressure Gap" between Ultrahigh-Vacuum Surface Physics and High-Pressure Catalysis. Physical Review Letters 1985, 55 (22), de Jonge, N.; Ross, F. M., Electron microscopy of specimens in liquid. Nature Nanotechnology 2011, 6 (11), Xin, H. L.; Zheng, H., In Situ Observation of Oscillatory Growth of Bismuth Nanoparticles. Nano Letters 2012, 12 (3), Wu, J.; Gao, W.; Wen, J.; Miller, D. J.; Lu, P.; Zuo, J.-M.; Yang, H., Growth of Au on Pt Icosahedral Nanoparticles Revealed by Low-Dose In Situ TEM. Nano Letters 2015, 15 (4), Evans, J. E.; Jungjohann, K. L.; Browning, N. D.; Arslan, I., Controlled Growth of Nanoparticles from Solution with In Situ Liquid Transmission Electron Microscopy. Nano Letters 2011, 11 (7),

111 104. Zeng, Z. Y.; Liang, W. I.; Liao, H. G.; Xin, H. L. L.; Chu, Y. H.; Zheng, H. M., Visualization of Electrode-Electrolyte Interfaces in LiPF6/EC/DEC Electrolyte for Lithium Ion Batteries via in Situ TEM. Nano Letters 2014, 14 (4), Gu, M.; Parent, L. R.; Mehdi, B. L.; Unocic, R. R.; McDowell, M. T.; Sacci, R. L.; Xu, W.; Connell, J. G.; Xu, P.; Abellan, P.; Chen, X.; Zhang, Y.; Perea, D. E.; Evans, J. E.; Lauhon, L. J.; Zhang, J.- G.; Liu, J.; Browning, N. D.; Cui, Y.; Arslan, I.; Wang, C.-M., Demonstration of an Electrochemical Liquid Cell for Operando Transmission Electron Microscopy Observation of the Lithiation/Delithiation Behavior of Si Nanowire Battery Anodes. Nano Letters 2013, 13 (12), Unocic, R.; Adamczyk, L.; Dudney, N.; Alsem, D.; Salmon, N.; More, K., In-situ TEM Characterization of Electrochemical Processes in Energy Storage Systems. Microscopy and Microanalysis 2011, 17, Unocic, R. R.; Sacci, R. L.; Brown, G. M.; Veith, G. M.; Dudney, N. J.; More, K. L.; Walden, F. S.; Gardiner, D. S.; Damiano, J.; Nackashi, D. P., Quantitative Electrochemical Measurements Using In Situ ec-s/tem Devices. Microscopy and Microanalysis 2014, 20 (2), Holtz, M. E.; Yu, Y.; Gunceler, D.; Gao, J.; Sundararaman, R.; Schwarz, K. A.; Arias, T. A.; Abruña, H. D.; Muller, D. A., Nanoscale Imaging of Lithium Ion Distribution During In Situ Operation of Battery Electrode and Electrolyte. Nano Letters 2014, 14 (3), Mehdi, B. L.; Qian, J.; Nasybulin, E.; Park, C.; Welch, D. A.; Faller, R.; Mehta, H.; Henderson, W. A.; Xu, W.; Wang, C. M.; Evans, J. E.; Liu, J.; Zhang, J. G.; Mueller, K. T.; Browning, N. D., Observation and Quantification of Nanoscale Processes in Lithium Batteries by Operando Electrochemical (S)TEM. Nano Letters 2015, 15 (3), White, E. R.; Singer, S. B.; Augustyn, V.; Hubbard, W. A.; Mecklenburg, M.; Dunn, B.; Regan, B. C., In Situ Transmission Electron Microscopy of Lead Dendrites and Lead Ions in Aqueous Solution. ACS Nano 2012, 6 (7), Chee, S. W.; Pratt, S. H.; Hattar, K.; Duquette, D.; Ross, F. M.; Hull, R., Studying localized corrosion using liquid cell transmission electron microscopy. Chemical Communications 2015, 51, Liao, H. G.; Cui, L. K.; Whitelam, S.; Zheng, H. M., Real-Time Imaging of Pt3Fe Nanorod Growth in Solution. Science 2012, 336 (6084), Jungjohann, K. L.; Bliznakov, S.; Sutter, P. W.; Stach, E. A.; Sutter, E. A., In Situ Liquid Cell Electron Microscopy of the Solution Growth of Au-Pd Core-Shell Nanostructures. Nano Letters 2013, 13 (6), Woehl, T. J.; Evans, J. E.; Arslan, L.; Ristenpart, W. D.; Browning, N. D., Direct in Situ Determination of the Mechanisms Controlling Nanoparticle Nucleation and Growth. ACS Nano 2012, 6 (10), Noh, K. W.; Liu, Y.; Sun, L.; Dillon, S. J., Challenges associated with in-situ TEM in environmental systems: The case of silver in aqueous solutions. Ultramicroscopy 2012, 116, Alloyeau, D.; Dachraoui, W.; Javed, Y.; Belkahla, H.; Wang, G.; Lecoq, H.; Ammar, S.; Ersen, O.; Wisnet, A.; Gazeau, F.; Ricolleau, C., Unravelling Kinetic and Thermodynamic Effects on the 111

112 Growth of Gold Nanoplates by Liquid Transmission Electron Microscopy. Nano Letters 2015, 15 (4), Liu, Y.; Chen, X.; Noh, K. W.; Dillon, S. J., Electron beam induced deposition of silicon nanostructures from a liquid phase precursor. Nanotechnology 2012, 23 (38), Liu, Y.; Tai, K. P.; Dillon, S. J., Growth Kinetics and Morphological Evolution of ZnO Precipitated from Solution. Chemistry of Materials 2013, 25 (15), Belloni, J., Nucleation, growth and properties of nanoclusters studied by radiation chemistry: Application to catalysis. Catalysis Today 2006, 113 (3 4), Abellan, P.; Woehl, T. J.; Parent, L. R.; Browning, N. D.; Evans, J. E.; Arslan, I., Factors influencing quantitative liquid (scanning) transmission electron microscopy. Chemical Communications 2014, 50 (38), Woehl, T. J.; Jungjohann, K. L.; Evans, J. E.; Arslan, I.; Ristenpart, W. D.; Browning, N. D., Experimental procedures to mitigate electron beam induced artifacts during in situ fluid imaging of nanomaterials. Ultramicroscopy 2013, 127 (0), Joseph, J. M.; Seon Choi, B.; Yakabuskie, P.; Clara Wren, J., A combined experimental and model analysis on the effect of ph and O2(aq) on γ-radiolytically produced H2 and H2O2. Radiation Physics and Chemistry 2008, 77 (9), Siefermann, K. R.; Liu, Y.; Lugovoy, E.; Link, O.; Faubel, M.; Buck, U.; Winter, B.; Abel, B., Binding energies, lifetimes and implications of bulk and interface solvated electrons in water. Nature Chemistry 2010, 2 (4), Grogan, J. M.; Schneider, N. M.; Ross, F. M.; Bau, H. H., Bubble and Pattern Formation in Liquid Induced by an Electron Beam. Nano Letters 2013, 14 (1), den Heijer, M.; Shao, I.; Radisic, A.; Reuter, M. C.; Ross, F. M., Patterned electrochemical deposition of copper using an electron beam. APL Materials 2014, 2 (2), Das, S., Critical Review of Water Radiolysis Processes, Dissociation Products, and Possible Impacts on the Local Environment: A Geochemist s Perspective. Australian Journal of Chemistry 2013, 66 (5), Schneider, N. M.; Norton, M. M.; Mendel, B. J.; Grogan, J. M.; Ross, F. M.; Bau, H. H., Electron Water Interactions and Implications for Liquid Cell Electron Microscopy. The Journal of Physical Chemistry C 2014, 118 (38), Henglein, A.; Meisel, D., Radiolytic Control of the Size of Colloidal Gold Nanoparticles. Langmuir 1998, 14 (26), Martis, V.; Nikitenko, S.; Sen, S.; Sankar, G.; van Beek, W.; Filinchuk, Y.; Snigireva, I.; Bras, W., Effects of X-rays on Crystal Nucleation in Lithium Disilicate. Crystal Growth & Design 2011, 11 (7), Stanley, H. B.; Banerjee, D.; van Breemen, L.; Ciston, J.; Liebscher, C. H.; Martis, V.; Merino, D. H.; Longo, A.; Pattison, P.; Peters, G. W. M.; Portale, G.; Sen, S.; Bras, W., X-ray irradiation induced reduction and nanoclustering of lead in borosilicate glass. CrystEngComm 2014, 16 (39),

113 131. Peckys, D. B.; de Jonge, N., Visualizing Gold Nanoparticle Uptake in Live Cells with Liquid Scanning Transmission Electron Microscopy. Nano Letters 2011, 11 (4), Yuk, J. M.; Park, J.; Ercius, P.; Kim, K.; Hellebusch, D. J.; Crommie, M. F.; Lee, J. Y.; Zettl, A.; Alivisatos, A. P., High-Resolution EM of Colloidal Nanocrystal Growth Using Graphene Liquid Cells. Science 2012, 336 (6077), Park, J.; Zheng, H. M.; Lee, W. C.; Geissler, P. L.; Rabani, E.; Alivisatos, A. P., Direct Observation of Nanoparticle Superlattice Formation by Using Liquid Cell Transmission Electron Microscopy. ACS Nano 2012, 6 (3), Zheng, H.; Mirsaidov, U. M.; Wang, L.-W.; Matsudaira, P., Electron Beam Manipulation of Nanoparticles. Nano Letters 2012, 12 (11), de Jonge, N.; Peckys, D. B.; Kremers, G. J.; Piston, D. W., Electron microscopy of whole cells in liquid with nanometer resolution. Proceedings of the National Academy of Sciences of the United States of America 2009, 106 (7), Ring, E. A.; de Jonge, N., Microfluidic System for Transmission Electron Microscopy. Microscopy and Microanalysis 2010, 16 (5), Peckys, D. B.; Veith, G. M.; Joy, D. C.; de Jonge, N., Nanoscale Imaging of Whole Cells Using a Liquid Enclosure and a Scanning Transmission Electron Microscope. PLoS ONE 2009, 4 (12), Jiang, Y.; Zhu, G.; Lin, F.; Zhang, H.; Jin, C.; Yuan, J.; Yang, D.; Zhang, Z., In situ Study of Oxidative Etching of Palladium Nanocrystals by Liquid Cell Electron Microscopy. Nano Letters 2014, 14 (7), zhu, g.; Jiang, Y.; Lin, F.; Zhang, H.; Jin, C.; Yuan, J.; Yang, D.; Zhang, Z., In situ Study of the Growth of Two-Dimensional Palladium Dendritic Nanostructures Using Liquid-Cell Electron Microscopy. Chemical Communications Kraus, T.; de Jonge, N., Dendritic Gold Nanowire Growth Observed in Liquid with Transmission Electron Microscopy. Langmuir 2013, 29 (26), Woehl, T. J.; Park, C.; Evans, J. E.; Arslan, I.; Ristenpart, W. D.; Browning, N. D., Direct Observation of Aggregative Nanoparticle Growth: Kinetic Modeling of the Size Distribution and Growth Rate. Nano Letters 2013, 14 (1), Zheng, H.; Claridge, S. A.; Minor, A. M.; Alivisatos, A. P.; Dahmen, U., Nanocrystal Diffusion in a Liquid Thin Film Observed by in Situ Transmission Electron Microscopy. Nano Letters 2009, 9 (6), Huang, T.-W.; Liu, S.-Y.; Chuang, Y.-J.; Hsieh, H.-Y.; Tsai, C.-Y.; Wu, W.-J.; Tsai, C.-T.; Mirsaidov, U.; Matsudaira, P.; Chang, C.-S.; Tseng, F.-G.; Chen, F.-R., Dynamics of hydrogen nanobubbles in KLH protein solution studied with in situ wet-tem. Soft Matter 2013, 9 (37), White, E. R.; Mecklenburg, M.; Shevitski, B.; Singer, S. B.; Regan, B. C., Charged Nanoparticle Dynamics in Water Induced by Scanning Transmission Electron Microscopy. Langmuir 2012, 28 (8),

114 145. Grogan, J. M.; Schneider, N. M.; Ross, F. M.; Bau, H. H., Bubble and Pattern Formation in Liquid Induced by an Electron Beam. Nano Letters Aronova, M. A.; Sousa, A. A.; Leapman, R. D., EELS characterization of radiolytic products in frozen samples. Micron 2011, 42 (3), Le Caër, S., Water radiolysis: influence of oxide surfaces on H2 production under ionizing radiation. Water 2011, 3 (1), Liu, Y. Z.; Lin, X. M.; Sun, Y. G.; Rajh, T., In Situ Visualization of Self-Assembly of Charged Gold Nanoparticles. Journal of the American Chemical Society 2013, 135 (10), Lou, X. W.; Archer, L. A.; Yang, Z., Hollow Micro-/Nanostructures: Synthesis and Applications. Advanced Materials 2008, 20 (21), Caruso, F.; Caruso, R. A.; Möhwald, H., Nanoengineering of Inorganic and Hybrid Hollow Spheres by Colloidal Templating. Science 1998, 282 (5391), Li, H.; Ha, C.-S.; Kim, I., Facile Fabrication of Hollow Silica and Titania Microspheres Using Plasma-Treated Polystyrene Spheres as Sacrificial Templates. Langmuir 2008, 24 (19), Wang, P.; Chen, D.; Tang, F.-Q., Preparation of Titania-Coated Polystyrene Particles in Mixed Solvents by Ammonia Catalysis. Langmuir 2006, 22 (10), Hosein, I. D.; Liddell, C. M., Homogeneous, Core Shell, and Hollow-Shell ZnS Colloid-Based Photonic Crystals. Langmuir 2007, 23 (5), Wolosiuk, A.; Armagan, O.; Braun, P. V., Double Direct Templating of Periodically Nanostructured ZnS Hollow Microspheres. Journal of the American Chemical Society 2005, 127 (47), Lu, L.; Sun, G.; Xi, S.; Wang, H.; Zhang, H.; Wang, T.; Zhou, X., A colloidal templating method to hollow bimetallic nanostructures. Langmuir 2003, 19 (7), Nakashima, T.; Kimizuka, N., Interfacial Synthesis of Hollow TiO2 Microspheres in Ionic Liquids. Journal of the American Chemical Society 2003, 125 (21), Hua, Z.; Shuyuan, Z.; Shuan, P.; Gongpu, L.; Jianguo, H., A simple solution route to ZnS nanotubes and hollow nanospheres and their optical properties. Nanotechnology 2004, 15 (8), Hadiko, G.; Han, Y. S.; Fuji, M.; Takahashi, M., Synthesis of hollow calcium carbonate particles by the bubble templating method. Materials Letters 2005, 59 (19 20), Liu, Q.; Liu, H.; Han, M.; Zhu, J.; Liang, Y.; Xu, Z.; Song, Y., Nanometer-Sized Nickel Hollow Spheres. Advanced Materials 2005, 17 (16), Xu, H.; Wang, W., Template synthesis of multishelled Cu2O hollow spheres with a singlecrystalline shell wall. Angewandte Chemie International Edition 2007, 46 (9), Anderson, B. D.; Tracy, J. B., Nanoparticle Conversion Chemistry: Kirkendall Effect, Galvanic Exchange, and Anion Exchange. Nanoscale 2014, 6, Raidongia, K.; Rao, C. N. R., Study of the transformations of elemental nanowires to nanotubes of metal oxides and chalcogenides through the Kirkendall effect. Journal of Physical Chemistry C 2008, 112 (35),

115 163. Lian, J.; Anggara, K.; Lin, M.; Chan, Y., Formation of Hollow Iron Oxide Tetrapods via a Shape-Preserving Nanoscale Kirkendall Effect. Small 2014, 10 (4), Gonzalez, E.; Arbiol, J.; Puntes, V. F., Carving at the Nanoscale: Sequential Galvanic Exchange and Kirkendall Growth at Room Temperature. Science 2011, 334 (6061), Wang, W.; Dahl, M.; Yin, Y., Hollow Nanocrystals through the Nanoscale Kirkendall Effect. Chemistry of Materials 2012, 25 (8), Xia, X.; Wang, Y.; Ruditskiy, A.; Xia, Y., 25th Anniversary Article: Galvanic Replacement: A Simple and Versatile Route to Hollow Nanostructures with Tunable and Well-Controlled Properties. Advanced Materials 2013, 25 (44), Au, L.; Lu, X.; Xia, Y., A comparative study of galvanic replacement reactions involving Ag nanocubes and AuCl2 or AuCl4. Advanced Materials 2008, 20 (13), Lu, X.; Chen, J.; Skrabalak, S.; Xia, Y., Galvanic replacement reaction: a simple and powerful route to hollow and porous metal nanostructures. Proceedings of the Institution of Mechanical Engineers, Part N: Journal of Nanoengineering and Nanosystems 2007, 221 (1), Chen, J.; Wiley, B.; McLellan, J.; Xiong, Y.; Li, Z.-Y.; Xia, Y., Optical Properties of Pd Ag and Pt Ag Nanoboxes Synthesized via Galvanic Replacement Reactions. Nano Letters 2005, 5 (10), Sun, Y.; Xia, Y., Mechanistic Study on the Replacement Reaction between Silver Nanostructures and Chloroauric Acid in Aqueous Medium. Journal of the American Chemical Society 2004, 126 (12), Sun, Y.; Mayers, B. T.; Xia, Y., Template-Engaged Replacement Reaction: A One-Step Approach to the Large-Scale Synthesis of Metal Nanostructures with Hollow Interiors. Nano Letters 2002, 2 (5), Skrabalak, S. E.; Chen, J.; Sun, Y.; Lu, X.; Au, L.; Cobley, C. M.; Xia, Y., Gold Nanocages: Synthesis, Properties, and Applications. Accounts of Chemical Research 2008, 41 (12), Teng, X.; Wang, Q.; Liu, P.; Han, W.; Frenkel, A. I.; Wen; Marinkovic, N.; Hanson, J. C.; Rodriguez, J. A., Formation of Pd/Au Nanostructures from Pd Nanowires via Galvanic Replacement Reaction. Journal of the American Chemical Society 2008, 130 (3), Skrabalak, S. E.; Au, L.; Li, X.; Xia, Y., Facile synthesis of Ag nanocubes and Au nanocages. Nature Protocols 2007, 2 (9), Zhang, Q.; Xie, J.; Lee, J. Y.; Zhang, J.; Boothroyd, C., Synthesis of Ag@AgAu Metal Core/Alloy Shell Bimetallic Nanoparticles with Tunable Shell Compositions by a Galvanic Replacement Reaction. Small 2008, 4 (8), Chen, J.; McLellan, J. M.; Siekkinen, A.; Xiong, Y.; Li, Z.-Y.; Xia, Y., Facile Synthesis of Gold Silver Nanocages with Controllable Pores on the Surface. Journal of the American Chemical Society 2006, 128 (46), Zeng, J.; Zhang, Q.; Chen, J.; Xia, Y., A Comparison Study of the Catalytic Properties of Au- Based Nanocages, Nanoboxes, and Nanoparticles. Nano Letters 2010, 10 (1),

116 178. Mallard, W. C.; Bass, R. F.; Slifkin, L. M.; Gardner, A. B., Self-diffusion in silver-gold soild solutions. Physical Review 1963, 129 (2), Sun, Y.; Xia, Y., Alloying and Dealloying Processes Involved in the Preparation of Metal Nanoshells through a Galvanic Replacement Reaction. Nano Letters 2003, 3 (11), Smigelskas, A. D.; Kirkendall, E. O., Zinc diffusion in alpha-bass. Transactions of the American Institute of Mining and Metallurgical Engineers 1947, 171, Zeng, K.; Stierman, R.; Chiu, T.-C.; Edwards, D.; Ano, K.; Tu, K. N., Kirkendall void formation in eutectic SnPb solder joints on bare Cu and its effect on joint reliability. Journal of Applied Physics 2005, 97 (2), Yin, Y.; Rioux, R. M.; Erdonmez, C. K.; Hughes, S.; Somorjai, G. A.; Alivisatos, A. P., Formation of Hollow Nanocrystals Through the Nanoscale Kirkendall Effect. Science 2004, 304 (5671), El Mel, A.-A.; Nakamura, R.; Bittencourt, C., The Kirkendall effect and nanoscience: hollow nanospheres and nanotubes. Beilstein Journal of Nanotechnology 2015, 6, Fan, H. J.; Gösele, U.; Zacharias, M., Formation of Nanotubes and Hollow Nanoparticles Based on Kirkendall and Diffusion Processes: A Review. Small 2007, 3 (10), Liu, B.; Zeng, H. C., Fabrication of ZnO Dandelions via a Modified Kirkendall Process. Journal of the American Chemical Society 2004, 126 (51), Du, W.; Liu, R.; Jiang, Y.; Lu, Q.; Fan, Y.; Gao, F., Facile synthesis of hollow Co3O4 boxes for high capacity supercapacitor. Journal of Power Sources 2013, 227, Shevchenko, E. V.; Bodnarchuk, M. I.; Kovalenko, M. V.; Talapin, D. V.; Smith, R. K.; Aloni, S.; Heiss, W.; Alivisatos, A. P., Gold/Iron Oxide Core/Hollow-Shell Nanoparticles. Advanced Materials 2008, 20 (22), El Mel, A. A.; Buffière, M.; Tessier, P. Y.; Konstantinidis, S.; Xu, W.; Du, K.; Wathuthanthri, I.; Choi, C. H.; Bittencourt, C.; Snyders, R., Highly Ordered Hollow Oxide Nanostructures: The Kirkendall Effect at the Nanoscale. Small 2013, 9 (17), Xiao, G.; Zeng, Y.; Jiang, Y.; Ning, J.; Zheng, W.; Liu, B.; Chen, X.; Zou, G.; Zou, B., Controlled Synthesis of Hollow Cu2-xTe Nanocrystals Based on the Kirkendall Effect and Their Enhanced CO Gas- Sensing Properties. Small 2013, 9 (5), Tan, H.; Li, S.; Fan, W. Y., Core Shell and Hollow Nanocrystal Formation via Small Molecule Surface Photodissociation; Ag@Ag2Se as an Example. The Journal of Physical Chemistry B 2006, 110 (32), Ng, C. H. B.; Tan, H.; Fan, W. Y., Formation of Ag2Se Nanotubes and Dendrite-like Structures from UV Irradiation of a CSe2/Ag Colloidal Solution. Langmuir 2006, 22 (23), Zhu, J.; Shen, Y.; Xie, A.; Zhu, L., Tunable surface plasmon resonance of Au@Ag2S core-shell nanostructures containing voids. Journal of Materials Chemistry 2009, 19 (46), Chiang, R. K.; Chiang, R. T., Formation of Hollow Ni2P Nanoparticles Based on the Nanoscale Kirkendall Effect. Inorganic Chemistry 2006, 46 (2),

117 194. Ma, Y.; Huo, K.; Wu, Q.; Lu, Y.; Hu, Y.; Hu, Z.; Chen, Y., Self-templated synthesis of polycrystalline hollow aluminium nitride nanospheres. Journal of Materials Chemistry 2006, 16 (27), Nakamura, R.; Lee, J. G.; Tokozakura, D.; Mori, H.; Nakajima, H., Formation of hollow ZnO through low-temperature oxidation of Zn nanoparticles. Materials Letters 2007, 61 (4 5), Nakamura, R.; Tokozakura, D.; Nakajima, H.; Lee, J. G.; Mori, H., Hollow oxide formation by oxidation of Al and Cu nanoparticles. Journal of Applied Physics 2007, 101 (7), Yin, Y.; Erdonmez, C. K.; Cabot, A.; Hughes, S.; Alivisatos, A. P., Colloidal Synthesis of Hollow Cobalt Sulfide Nanocrystals. Advanced Functional Materials 2006, 16 (11), Railsback, J. G.; Johnston-Peck, A. C.; Wang, J.; Tracy, J. B., Size-Dependent Nanoscale Kirkendall Effect During the Oxidation of Nickel Nanoparticles. ACS Nano 2010, 4 (4), Nakamura, R.; Lee, J. G.; Mori, H.; Nakajima, H., Oxidation behaviour of Ni nanoparticles and formation process of hollow NiO. Philosophical Magazine 2008, 88 (2), Zheng, J.; Song, X.; Zhang, Y.; Li, Y.; Li, X.; Pu, Y., Nanosized aluminum nitride hollow spheres formed through a self-templating solid gas interface reaction. Journal of Solid State Chemistry 2007, 180 (1), Cabot, A.; Smith, R. K.; Yin, Y.; Zheng, H.; Reinhard, B. M.; Liu, H.; Alivisatos, A. P., Sulfidation of cadmium at the nanoscale. ACS Nano 2008, 2 (7), Niu, K. Y.; Park, J.; Zheng, H. M.; Aivisatos, A. P., Revealing Bismuth Oxide Hollow Nanoparticle Formation by the Kirkendall Effect. Nano Letters 2013, 13 (11), Yang; Kim, D. S.; Knez, M.; Scholz, R.; Berger, A.; Pippel, E.; Hesse, D.; Gösele, U.; Zacharias, M., Influence of Temperature on Evolution of Coaxial ZnO/Al2O3 One-Dimensional Heterostructures: From Core Shell Nanowires to Spinel Nanotubes and Porous Nanowires. The Journal of Physical Chemistry C 2008, 112 (11), Tu, K. N.; Gösele, U., Hollow nanostructures based on the Kirkendall effect: Design and stability considerations. Applied Physics Letters 2005, 86 (9), Nakamura, R.; Tokozakura, D.; Lee, J. G.; Mori, H.; Nakajima, H., Shrinking of hollow Cu2O and NiO nanoparticles at high temperatures. Acta Materialia 2008, 56 (18), Fan, H. J.; Knez, M.; Scholz, R.; Hesse, D.; Nielsch, K.; Zacharias, M.; Gosele, U., Influence of surface diffusion on the formation of hollow nanostructures induced by the Kirkendall effect: The basic concept. Nano Letters 2007, 7 (4), Gao, J.; Liang, G.; Zhang, B.; Kuang, Y.; Zhang, X.; Xu, B., FePt@CoS2 Yolk Shell Nanocrystals as a Potent Agent to Kill HeLa Cells. Journal of the American Chemical Society 2007, 129 (5), Klimov, V. I., Spectral and Dynamical Properties of Multiexcitons in Semiconductor Nanocrystals. Annual Review of Physical Chemistry 2007, 58 (1), Kim, J. Y.; Voznyy, O.; Zhitomirsky, D.; Sargent, E. H., 25th Anniversary Article: Colloidal Quantum Dot Materials and Devices: A Quarter-Century of Advances. Advanced Materials 2013, 25 (36),

118 210. Ekimov, A. I.; Efros, A. L.; Onushchenko, A. A., Quantum size effect in semiconductor microcrystals. Solid State Communications 1985, 56 (11), Kongkanand, A.; Tvrdy, K.; Takechi, K.; Kuno, M.; Kamat, P. V., Quantum Dot Solar Cells. Tuning Photoresponse through Size and Shape Control of CdSe TiO2 Architecture. Journal of the American Chemical Society 2008, 130 (12), Nirmal, M.; Norris, D. J.; Kuno, M.; Bawendi, M. G.; Efros, A. L.; Rosen, M., Observation of the "Dark Exciton" in CdSe Quantum Dots. Physical Review Letters 1995, 75 (20), Efros, A. L.; Rosen, M.; Kuno, M.; Nirmal, M.; Norris, D. J.; Bawendi, M., Band-edge exciton in quantum dots of semiconductors with a degenerate valence band: Dark and bright exciton states. Physical Review B 1996, 54 (7), Gao, X. H.; Yang, L. L.; Petros, J. A.; Marshal, F. F.; Simons, J. W.; Nie, S. M., In vivo molecular and cellular imaging with quantum dots. Current Opinion in Biotechnology 2005, 16 (1), Gao, X. H.; Cui, Y. Y.; Levenson, R. M.; Chung, L. W. K.; Nie, S. M., In vivo cancer targeting and imaging with semiconductor quantum dots. Nature Biotechnology 2004, 22 (8), Cademartiri, L.; Montanari, E.; Calestani, G.; Migliori, A.; Guagliardi, A.; Ozin, G. A., Size- Dependent Extinction Coefficients of PbS Quantum Dots. Journal of the American Chemical Society 2006, 128 (31), Moreels, I.; Lambert, K.; Smeets, D.; De Muynck, D.; Nollet, T.; Martins, J. C.; Vanhaecke, F.; Vantomme, A.; Delerue, C.; Allan, G.; Hens, Z., Size-Dependent Optical Properties of Colloidal PbS Quantum Dots. ACS Nano 2009, 3 (10), Moreels, I.; Lambert, K.; De Muynck, D.; Vanhaecke, F.; Poelman, D.; Martins, J. C.; Allan, G.; Hens, Z., Composition and Size-Dependent Extinction Coefficient of Colloidal PbSe Quantum Dots. Chemistry of Materials 2007, 19 (25), Grätzel, M., Dye-sensitized solar cells. Journal of Photochemistry and Photobiology C: Photochemistry Reviews 2003, 4 (2), Kramer, I. J.; Levina, L.; Debnath, R.; Zhitomirsky, D.; Sargent, E. H., Solar Cells Using Quantum Funnels. Nano Letters 2011, 11 (9), Santra, P. K.; Kamat, P. V., Mn-Doped Quantum Dot Sensitized Solar Cells: A Strategy to Boost Efficiency over 5%. Journal of the American Chemical Society 2012, 134 (5), Lee, Y.-L.; Lo, Y.-S., Highly Efficient Quantum-Dot-Sensitized Solar Cell Based on Co- Sensitization of CdS/CdSe. Advanced Functional Materials 2009, 19 (4), Peng, X.; Schlamp, M. C.; Kadavanich, A. V.; Alivisatos, A. P., Epitaxial Growth of Highly Luminescent CdSe/CdS Core/Shell Nanocrystals with Photostability and Electronic Accessibility. Journal of the American Chemical Society 1997, 119 (30), Kim, S.; Fisher, B.; Eisler, H.-J.; Bawendi, M., Type-II Quantum Dots: CdTe/CdSe(Core/Shell) and CdSe/ZnTe(Core/Shell) Heterostructures. Journal of the American Chemical Society 2003, 125 (38), Reiss, P.; Protiere, M.; Li, L., Core/Shell Semiconductor Nanocrystals. Small 2009, 5 (2),

119 226. Ellingson, R. J.; Beard, M. C.; Johnson, J. C.; Yu, P. R.; Micic, O. I.; Nozik, A. J.; Shabaev, A.; Efros, A. L., Highly efficient multiple exciton generation in colloidal PbSe and PbS quantum dots. Nano Letters 2005, 5 (5), Sambur, J. B.; Novet, T.; Parkinson, B. A., Multiple Exciton Collection in a Sensitized Photovoltaic System. Science 2010, 330 (6000), Hardman, S. J. O.; Graham, D. M.; Stubbs, S. K.; Spencer, B. F.; Seddon, E. A.; Fung, H.-T.; Gardonio, S.; Sirotti, F.; Silly, M. G.; Akhtar, J.; O'Brien, P.; Binks, D. J.; Flavell, W. R., Electronic and surface properties of PbS nanoparticles exhibiting efficient multiple exciton generation. Physical Chemistry Chemical Physics 2011, 13 (45), Nozik, A. J., Multiple exciton generation in semiconductor quantum dots. Chemical Physics Letters 2008, 457 (1-3), Moule, A. J.; Chang, L.; Thambidurai, C.; Vidu, R.; Stroeve, P., Hybrid solar cells: basic principles and the role of ligands. Journal of Materials Chemistry 2012, 22 (6), Sargent, E. H., Colloidal quantum dot solar cells. Nature Photonics 2012, 6 (3), Conibeer, G., Third-generation photovoltaics. Materials Today 2007, 10 (11), Green, M. A.; Zhao, J.; Wang, A.; Wenham, S. R., Progress and outlook for high-efficiency crystalline silicon solar cells. Solar Energy Materials and Solar Cells 2001, 65 (1 4), Pagliaro, M.; Ciriminna, R.; Palmisano, G., Flexible Solar Cells. ChemSusChem 2008, 1 (11), Freitas, J. N.; Goncalves, A. S.; Nogueira, A. F., A comprehensive review of the application of chalcogenide nanoparticles in polymer solar cells. Nanoscale 2014, 6 (12), Gao, F.; Ren, S.; Wang, J., The renaissance of hybrid solar cells: progresses, challenges, and perspectives. Energy & Environmental Science 2013, 6 (7), Mayer, A. C.; Scully, S. R.; Hardin, B. E.; Rowell, M. W.; McGehee, M. D., Polymer-based solar cells. Materials Today 2007, 10 (11), Günes, S.; Neugebauer, H.; Sariciftci, N. S., Conjugated polymer-based organic solar cells. Chemical reviews 2007, 107 (4), Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P., Hybrid nanorod-polymer solar cells. Science 2002, 295 (5564), Nam, M.; Kim, S.; Kim, T.; Kim, S.-W.; Lee, K.-K., Broadband energy-harvesting hybrid solar cells employing nanocomposites of polythiophene: ternary PbSSe nanocrystals. Applied Physics Letters 2011, 99 (23), Cui, D.; Xu, J.; Zhu, T.; Paradee, G.; Ashok, S.; Gerhold, M., Harvest of near infrared light in PbSe nanocrystal-polymer hybrid photovoltaic cells. Applied Physics Letters 2006, 88 (18), Gur, I.; Fromer, N. A.; Chen, C.-P.; Kanaras, A. G.; Alivisatos, A. P., Hybrid Solar Cells with Prescribed Nanoscale Morphologies Based on Hyperbranched Semiconductor Nanocrystals. Nano Letters 2006, 7 (2),

120 243. Huynh, W. U.; Dittmer, J. J.; Libby, W. C.; Whiting, G. L.; Alivisatos, A. P., Controlling the morphology of nanocrystal-polymer composites for solar cells. Advanced Functional Materials 2003, 13 (1), Liu, Z.; Sun, Y.; Yuan, J.; Wei, H.; Huang, X.; Han, L.; Wang, W.; Wang, H.; Ma, W., High- Efficiency Hybrid Solar Cells Based on Polymer/PbSxSe1-x Nanocrystals Benefiting from Vertical Phase Segregation. Advanced Materials 2013, 25 (40), MacLachlan, A. J.; Rath, T.; Cappel, U. B.; Dowland, S. A.; Amenitsch, H.; Knall, A.-C.; Buchmaier, C.; Trimmel, G.; Nelson, J.; Haque, S. A., Polymer/Nanocrystal Hybrid Solar Cells: Influence of Molecular Precursor Design on Film Nanomorphology, Charge Generation and Device Performance. Advanced Functional Materials 2015, 25 (3), Bansal, N.; O'Mahony, F. T. F.; Lutz, T.; Haque, S. A., Solution Processed Polymer Inorganic Semiconductor Solar Cells Employing Sb2S3 as a Light Harvesting and Electron Transporting Material. Advanced Energy Materials 2013, 3 (8), Hindson, J. C.; Saghi, Z.; Hernandez-Garrido, J.-C.; Midgley, P. A.; Greenham, N. C., Morphological Study of Nanoparticle Polymer Solar Cells Using High-Angle Annular Dark-Field Electron Tomography. Nano Letters 2011, 11 (2), Piliego, C.; Manca, M.; Kroon, R.; Yarema, M.; Szendrei, K.; Andersson, M. R.; Heiss, W.; Loi, M. A., Charge separation dynamics in a narrow band gap polymer-pbs nanocrystal blend for efficient hybrid solar cells. Journal of Materials Chemistry 2012, 22 (46), Bansal, A. K.; Antolini, F.; Sajjad, M. T.; Stroea, L.; Mazzaro, R.; Ramkumar, S. G.; Kass, K. J.; Allard, S.; Scherf, U.; Samuel, I. D. W., Photophysical and structural characterisation of in situ formed quantum dots. Physical Chemistry Chemical Physics 2014, 16 (20), Xu, T.; Qiao, Q., Conjugated polymer-inorganic semiconductor hybrid solar cells. Energy & Environmental Science 2011, 4 (8), Poudyal, N.; Chaubey, G. S.; Rong, C.-b.; Liu, J. P., Shape control of FePt nanocrystals. Journal of Applied Physics 2009, 105 (7), 07A Erik, W.; Michael, A.; Arnaud, M.; Jekabs, G.; Dong, W.; Subhasis, R.; Anwar, A.; German, S.- A.; Lennart, B., Precise control over shape and size of iron oxide nanocrystals suitable for assembly into ordered particle arrays. Science and Technology of Advanced Materials 2014, 15 (5), Wang, Z. L., Structural analysis of self-assembling nanocrystal superlattices. Advanced Materials 1998, 10 (1), Caruntu, D.; Rostamzadeh, T.; Costanzo, T.; Salemizadeh Parizi, S.; Caruntu, G., Solvothermal synthesis and controlled self-assembly of monodisperse titanium-based perovskite colloidal nanocrystals. Nanoscale 2015, 7 (30), Wellner, A.; Nellist, P.; Palmer, R.; Aindow, M.; Wilcoxon, J. P., Orientational and translational ordering of sub-monolayer films of passivated multiply-twinned gold clusters. Journal of Physics D: Applied Physics 2000, 33 (2), Fan, F.-J.; Wu, L.; Gong, M.; Chen, S. Y.; Liu, G. Y.; Yao, H.-B.; Liang, H.-W.; Wang, Y.-X.; Yu, S.-H., Linearly arranged polytypic CZTSSe nanocrystals. Scientific Reports 2012, 2,

121 257. Wang, Z. L., Transmission Electron Microscopy of Shape-Controlled Nanocrystals and Their Assemblies. The Journal of Physical Chemistry B 2000, 104 (6), Garcia-Negrete, C. A.; Rojas, T. C.; Knappett, B. R.; Jefferson, D. A.; Wheatley, A. E. H.; Fernandez, A., Shape-defined nanodimers by tailored heterometallic epitaxy. Nanoscale 2014, 6 (19), Som, A.; Pradeep, T., Heterojunction double dumb-bell Ag2Te-Te-Ag2Te nanowires. Nanoscale 2012, 4 (15), Miszta, K.; Dorfs, D.; Genovese, A.; Kim, M. R.; Manna, L., Cation Exchange Reactions in Colloidal Branched Nanocrystals. ACS Nano 2011, 5 (9), Wang, K.; Rai, S. C.; Marmon, J.; Chen, J.; Yao, K.; Wozny, S.; Cao, B.; Yan, Y.; Zhang, Y.; Zhou, W., Nearly lattice matched all wurtzite CdSe/ZnTe type II core-shell nanowires with epitaxial interfaces for photovoltaics. Nanoscale 2014, 6 (7), Luis Elechiguerra, J.; Reyes-Gasga, J.; Jose Yacaman, M., The role of twinning in shape evolution of anisotropic noble metal nanostructures. Journal of Materials Chemistry 2006, 16 (40), Wang, J.; Sansoz, F.; Huang, J.; Liu, Y.; Sun, S.; Zhang, Z.; Mao, S. X., Near-ideal theoretical strength in gold nanowires containing angstrom scale twins. Nature Communications 2013, 4, Shen, M.; Han, A.; Wang, X.; Ro, Y. G.; Kargar, A.; Lin, Y.; Guo, H.; Du, P.; Jiang, J.; Zhang, J.; Dayeh, S. A.; Xiang, B., Atomic Scale Analysis of the Enhanced Electro- and Photo-Catalytic Activity in High-Index Faceted Porous NiO Nanowires. Scientific Reports 2015, 5, Fujita, T.; Guan, P.; McKenna, K.; Lang, X.; Hirata, A.; Zhang, L.; Tokunaga, T.; Arai, S.; Yamamoto, Y.; Tanaka, N.; Ishikawa, Y.; Asao, N.; Yamamoto, Y.; Erlebacher, J.; Chen, M., Atomic origins of the high catalytic activity of nanoporous gold. Nature Materials 2012, 11 (9), Zhou, T.; Lu, M.; Zhang, Z.; Gong, H.; Chin, W. S.; Liu, B., Synthesis and Characterization of Multifunctional FePt/ZnO Core/Shell Nanoparticles. Advanced Materials 2010, 22 (3), Tang, J.; Huo, Z.; Brittman, S.; Gao, H.; Yang, P., Solution-processed core-shell nanowires for efficient photovoltaic cells. Nature Nanotechnology 2011, 6 (9), Wall, J.; Langmore, J.; Isaacson, M.; Crewe, A. V., Scanning Transmission Electron Microscopy at High Resolution. Proceedings of the National Academy of Sciences 1974, 71 (1), Nellist, P.; Pennycook, S., Direct imaging of the atomic configuration of ultradispersed catalysts. Science 1996, 274 (5286), Zhou, W.; Ross-Medgaarden, E. I.; Knowles, W. V.; Wong, M. S.; Wachs, I. E.; Kiely, C. J., Identification of active Zr WOx clusters on a ZrO2 support for solid acid catalysts. Nature Chemistry 2009, 1 (9), Qiao, B.; Wang, A.; Yang, X.; Allard, L. F.; Jiang, Z.; Cui, Y.; Liu, J.; Li, J.; Zhang, T., Single-atom catalysis of CO oxidation using Pt1/FeOx. Nature Chemistry 2011, 3 (8), Wang, Y.-G.; Mei, D.; Glezakou, V.-A.; Li, J.; Rousseau, R., Dynamic formation of single-atom catalytic active sites on ceria-supported gold nanoparticles. Nature communications 2015, 6,

122 273. Yang, X.-F.; Wang, A.; Qiao, B.; Li, J.; Liu, J.; Zhang, T., Single-Atom Catalysts: A New Frontier in Heterogeneous Catalysis. Accounts of Chemical Research 2013, 46 (8), Krivanek, O. L.; Nellist, P. D.; Dellby, N.; Murfitt, M. F.; Szilagyi, Z., Towards sub-0.5 angstrom electron beams. Ultramicroscopy 2003, 96 (3-4), Dellby, N.; Krivanek, O. L.; Nellist, P. D.; Batson, P. E.; Lupini, A. R., Progress in aberrationcorrected scanning transmission electron microscopy. Journal of Electron Microscopy 2001, 50 (3), Nellist, P. D.; Chisholm, M. F.; Dellby, N.; Krivanek, O. L.; Murfitt, M. F.; Szilagyi, Z. S.; Lupini, A. R.; Borisevich, A.; Sides, W. H.; Pennycook, S. J., Direct sub-angstrom imaging of a crystal lattice. Science 2004, 305 (5691), Krivanek, O. L.; Chisholm, M. F.; Nicolosi, V.; Pennycook, T. J.; Corbin, G. J.; Dellby, N.; Murfitt, M. F.; Own, C. S.; Szilagyi, Z. S.; Oxley, M. P.; Pantelides, S. T.; Pennycook, S. J., Atom-byatom structural and chemical analysis by annular dark-field electron microscopy. Nature 2010, 464 (7288), Kiely, C., Electron microscopy: New views of catalysts. Nature Materials 2010, 9 (4), Mariscal, M. M.; Mayoral, A.; Olmos-Asar, J. A.; Magen, C.; Mejia-Rosales, S.; Perez-Tijerina, E.; Jose-Yacaman, M., Nanoalloying in real time. A high resolution STEM and computer simulation study. Nanoscale 2011, 3 (12), Trinh, T. T.; Mott, D.; Thanh, N. T. K.; Maenosono, S., One-pot synthesis and characterization of well defined core-shell structure of FePt@CdSe nanoparticles. RSC Advances 2011, 1 (1), Jin, M.; Zhang, H.; Wang, J.; Zhong, X.; Lu, N.; Li, Z.; Xie, Z.; Kim, M. J.; Xia, Y., Copper Can Still Be Epitaxially Deposited on Palladium Nanocrystals To Generate Core Shell Nanocubes Despite Their Large Lattice Mismatch. ACS Nano 2012, 6 (3), Genki, S.; Yuki, N.; Toru, Y.; Tomohiro, A., Solution plasma synthesis of bimetallic nanoparticles. Nanotechnology 2014, 25 (13), Cassidy, C.; Singh, V.; Grammatikopoulos, P.; Djurabekova, F.; Nordlund, K.; Sowwan, M., Inoculation of silicon nanoparticles with silver atoms. Scientific Reports 2013, 3, Bangert, U.; Pierce, W.; Kepaptsoglou, D. M.; Ramasse, Q.; Zan, R.; Gass, M. H.; Van den Berg, J. A.; Boothroyd, C. B.; Amani, J.; Hofsäss, H., Ion Implantation of Graphene Toward IC Compatible Technologies. Nano Letters 2013, 13 (10), Van Aert, S.; Batenburg, K. J.; Rossell, M. D.; Erni, R.; Van Tendeloo, G., Three-dimensional atomic imaging of crystalline nanoparticles. Nature 2011, 470 (7334), Ong, M. D.; Jacobs, B. W.; Sugar, J. D.; Grass, M. E.; Liu, Z.; Buffleben, G. M.; Clift, W. M.; Langham, M. E.; Cappillino, P. J.; Robinson, D. B., Effect of Rhodium Distribution on Thermal Stability of Nanoporous Palladium Rhodium Powders. Chemistry of Materials 2012, 24 (6), Cao, A.; Veser, G., Exceptional high-temperature stability through distillation-like selfstabilization in bimetallic nanoparticles. Nature Materials 2010, 9 (1),

123 288. McElroy, N.; Page, R. C.; Espinbarro-Valazquez, D.; Lewis, E.; Haigh, S.; O Brien, P.; Binks, D. J., Comparison of solar cells sensitised by CdTe/CdSe and CdSe/CdTe core/shell colloidal quantum dots with and without a CdS outer layer. Thin Solid Films 2013, 560, Jiang, H.-L.; Akita, T.; Ishida, T.; Haruta, M.; Xu, Q., Synergistic Catalysis of Au@Ag Core Shell Nanoparticles Stabilized on Metal Organic Framework. Journal of the American Chemical Society 2011, 133 (5), Wang, D.; Yu, Y.; Xin, H. L.; Hovden, R.; Ercius, P.; Mundy, J. A.; Chen, H.; Richard, J. H.; Muller, D. A.; DiSalvo, F. J.; Abruña, H. D., Tuning Oxygen Reduction Reaction Activity via Controllable Dealloying: A Model Study of Ordered Cu3Pt/C Intermetallic Nanocatalysts. Nano Letters 2012, 12 (10), Sasaki, K.; Naohara, H.; Choi, Y.; Cai, Y.; Chen, W.-F.; Liu, P.; Adzic, R. R., Highly stable Pt monolayer on PdAu nanoparticle electrocatalysts for the oxygen reduction reaction. Nature Communications 2012, 3, D'Alfonso, A. J.; Freitag, B.; Klenov, D.; Allen, L. J., Atomic-resolution chemical mapping using energy-dispersive x-ray spectroscopy. Physical Review B 2010, 81 (10), Suenaga, K.; Tence, T.; Mory, C.; Colliex, C.; Kato, H.; Okazaki, T.; Shinohara, H.; Hirahara, K.; Bandow, S.; Iijima, S., Element-selective single atom imaging. Science 2000, 290 (5500), Slater, T. J. A.; Macedo, A.; Schroeder, S. L. M.; Burke, M. G.; O Brien, P.; Camargo, P. H. C.; Haigh, S. J., Correlating Catalytic Activity of Ag Au Nanoparticles with 3D Compositional Variations. Nano Letters 2014, 14 (4), Liu, J. H.; Wang, A. Q.; Chi, Y. S.; Lin, H. P.; Mou, C. Y., Synergistic effect in an Au-Ag alloy nanocatalyst: CO oxidation. Journal of Physical Chemistry B 2005, 109 (1), Wang, D.; Xin, H. L.; Wang, H.; Yu, Y.; Rus, E.; Muller, D. A.; DiSalvo, F. J.; Abruña, H. D., Facile Synthesis of Carbon-Supported Pd Co Core Shell Nanoparticles as Oxygen Reduction Electrocatalysts and Their Enhanced Activity and Stability with Monolayer Pt Decoration. Chemistry of Materials 2012, 24 (12), Cui, C.; Gan, L.; Heggen, M.; Rudi, S.; Strasser, P., Compositional segregation in shaped Pt alloy nanoparticles and their structural behaviour during electrocatalysis. Nature Materials 2013, 12 (8), Wilson, A. R.; Sun, K.; Chi, M.; White, R. M.; LeBeau, J. M.; Lamb, H. H.; Wiley, B. J., From Core Shell to Alloys: The Preparation and Characterization of Solution-Synthesized AuPd Nanoparticle Catalysts. The Journal of Physical Chemistry C 2013, 117 (34), Da Silva, A. G. M.; Lewis, E. A.; Rodrigues, T. S.; Slater, T. J. A.; Alves, R. S.; Haigh, S. J.; Camargo, P. H. C., Surface Segregated AgAu Tadpole-Shaped Nanoparticles Synthesized Via a Single Step Combined Galvanic and Citrate Reduction Reaction. Chemistry A European Journal 2015, 21 (35), Gan, L.; Heggen, M.; O Malley, R.; Theobald, B.; Strasser, P., Understanding and Controlling Nanoporosity Formation for Improving the Stability of Bimetallic Fuel Cell Catalysts. Nano Letters 2013, 13 (3),

124 301. Dumcenco, D. O.; Kobayashi, H.; Liu, Z.; Huang, Y.-S.; Suenaga, K., Visualization and quantification of transition metal atomic mixing in Mo1-xWxS2 single layers. Nature Communications 2013, 4, Arenal, R.; de la Peña, F.; Stéphan, O.; Walls, M.; Tencé, M.; Loiseau, A.; Colliex, C., Extending the analysis of EELS spectrum-imaging data, from elemental to bond mapping in complex nanostructures. Ultramicroscopy 2008, 109 (1), Nelayah, J.; Kociak, M.; Stephan, O.; Garcia de Abajo, F. J.; Tence, M.; Henrard, L.; Taverna, D.; Pastoriza-Santos, I.; Liz-Marzan, L. M.; Colliex, C., Mapping surface plasmons on a single metallic nanoparticle. Nature Physics 2007, 3 (5), Kociak, M.; Stephan, O., Mapping plasmons at the nanometer scale in an electron microscope. Chemical Society Reviews 2014, 43 (11), Krivanek, O. L.; Lovejoy, T. C.; Dellby, N.; Carpenter, R. W., Monochromated STEM with a 30 mev-wide, atom-sized electron probe. Microscopy 2013, 00 (0), Myroshnychenko, V.; Nelayah, J.; Adamo, G.; Geuquet, N.; Rodríguez-Fernández, J.; Pastoriza-Santos, I.; MacDonald, K. F.; Henrard, L.; Liz-Marzán, L. M.; Zheludev, N. I.; Kociak, M.; García de Abajo, F. J., Plasmon Spectroscopy and Imaging of Individual Gold Nanodecahedra: A Combined Optical Microscopy, Cathodoluminescence, and Electron Energy-Loss Spectroscopy Study. Nano Letters 2012, 12 (8), Guiton, B. S.; Iberi, V.; Li, S.; Leonard, D. N.; Parish, C. M.; Kotula, P. G.; Varela, M.; Schatz, G. C.; Pennycook, S. J.; Camden, J. P., Correlated Optical Measurements and Plasmon Mapping of Silver Nanorods. Nano Letters 2011, 11 (8), Mirsaleh-Kohan, N.; Iberi, V.; Simmons, P. D.; Bigelow, N. W.; Vaschillo, A.; Rowland, M. M.; Best, M. D.; Pennycook, S. J.; Masiello, D. J.; Guiton, B. S.; Camden, J. P., Single-Molecule Surface- Enhanced Raman Scattering: Can STEM/EELS Image Electromagnetic Hot Spots? The Journal of Physical Chemistry Letters 2012, 3 (16), Krivanek, O. L.; Lovejoy, T. C.; Dellby, N.; Aoki, T.; Carpenter, R. W.; Rez, P.; Soignard, E.; Zhu, J.; Batson, P. E.; Lagos, M. J.; Egerton, R. F.; Crozier, P. A., Vibrational spectroscopy in the electron microscope. Nature 2014, 514 (7521), Tao, A. R.; Habas, S.; Yang, P., Shape Control of Colloidal Metal Nanocrystals. Small 2008, 4 (3), Shao, Y.; Yin, G.; Gao, Y., Understanding and approaches for the durability issues of Pt-based catalysts for PEM fuel cell. Journal of Power Sources 2007, 171 (2), Cao, A.; Lu, R.; Veser, G., Stabilizing metal nanoparticles for heterogeneous catalysis. Physical Chemistry Chemical Physics 2010, 12 (41), Xu, T.; Sun, L., Dynamic In-Situ Experimentation on Nanomaterials at the Atomic Scale. Small 2015, 11 (27), Chen, X.; Li, C.; Cao, H., Recent developments of the in situ wet cell technology for transmission electron microscopies. Nanoscale 2015, 7,

125 315. van Huis, M. A.; Kunneman, L. T.; Overgaag, K.; Xu, Q.; Pandraud, G.; Zandbergen, H. W.; Vanmaekelbergh, D., Low-Temperature Nanocrystal Unification through Rotations and Relaxations Probed by in Situ Transmission Electron Microscopy. Nano Letters 2008, 8 (11), Harutyunyan, A. R.; Chen, G.; Paronyan, T. M.; Pigos, E. M.; Kuznetsov, O. A.; Hewaparakrama, K.; Kim, S. M.; Zakharov, D.; Stach, E. A.; Sumanasekera, G. U., Preferential Growth of Single-Walled Carbon Nanotubes with Metallic Conductivity. Science 2009, 326 (5949), Xin, H. L.; Pach, E. A.; Diaz, R. E.; Stach, E. A.; Salmeron, M.; Zheng, H., Revealing Correlation of Valence State with Nanoporous Structure in Cobalt Catalyst Nanoparticles by In Situ Environmental TEM. ACS Nano 2012, 6 (5), Koh, A. L.; Gidcumb, E.; Zhou, O.; Sinclair, R., Observations of Carbon Nanotube Oxidation in an Aberration-Corrected Environmental Transmission Electron Microscope. ACS Nano 2013, 7 (3), Kiener, D.; Minor, A. M., Source Truncation and Exhaustion: Insights from Quantitative in situ TEM Tensile Testing. Nano Letters 2011, 11 (9), Carlton, C. E.; Ferreira, P. J., In situ TEM nanoindentation of nanoparticles. Micron 2012, 43 (11), Espinosa, H. D.; Bernal, R. A.; Filleter, T., In Situ TEM Electromechanical Testing of Nanowires and Nanotubes. Small 2012, 8 (21), Sun, J.; He, L.; Lo, Y.-C.; Xu, T.; Bi, H.; Sun, L.; Zhang, Z.; Mao, S. X.; Li, J., Liquid-like pseudoelasticity of sub-10-nm crystalline silver particles. Nature Materials 2014, 13 (11), Latham, A. H.; Wilson, M. J.; Schiffer, P.; Williams, M. E., TEM-Induced Structural Evolution in Amorphous Fe Oxide Nanoparticles. Journal of the American Chemical Society 2006, 128 (39), Kim, J.-U.; Cha, S.-H.; Shin, K.; Jho, J. Y.; Lee, J.-C., Synthesis of Gold Nanoparticles from Gold(I) Alkanethiolate Complexes with Supramolecular Structures through Electron Beam Irradiation in TEM. Journal of the American Chemical Society 2005, 127 (28), Cazacu, A.; Larosa, C.; Beaunier, P.; Laurent, G.; Nanni, P.; Mitoseriu, L.; Lisiecki, I., Self- Organization and/or Nanocrystallinity of Co Nanocrystals Effects on the Oxidation Process Using High-Energy Electron Beam. Advanced Functional Materials 2014, 24 (1), Yang, Z.; Walls, M.; Lisiecki, I.; Pileni, M.-P., Unusual Effect of an Electron Beam on the Formation of Core/Shell (Co/CoO) Nanoparticles Differing by Their Crystalline Structures. Chemistry of Materials 2013, 25 (11), Ko, D.-H.; Sinclair, R., In-situ dynamic high-resolution transmission electron microscopy: application to Pt/GaAs interfacial reactions. Ultramicroscopy 1994, 54 (2 4), Zhang, H.; Jin, M.; Wang, J.; Li, W.; Camargo, P. H. C.; Kim, M. J.; Yang, D.; Xie, Z.; Xia, Y., Synthesis of Pd Pt Bimetallic Nanocrystals with a Concave Structure through a Bromide-Induced Galvanic Replacement Reaction. Journal of the American Chemical Society 2011, 133 (15),

126 329. Malladi, S. R. K.; Tichelaar, F. D.; Xu, Q.; Wu, M. Y.; Terryn, H.; Mol, J. M. C.; Hannour, F.; Zandbergen, H. W., Quasi in situ analytical TEM to investigate electrochemically induced microstructural changes in alloys: AA2024-T3 as an example. Corrosion Science 2013, 69, Kooyman, P. J.; Buglass, J. G.; Reinhoudt, H. R.; van Langeveld, A. D.; Hensen, E. J. M.; Zandbergen, H. W.; van Veen, J. A. R., Quasi in Situ Sequential Sulfidation of CoMo/Al2O3 Studied Using High-Resolution Electron Microscopy. The Journal of Physical Chemistry B 2002, 106 (45), Goris, B.; Van Huis, M. A.; Bals, S.; Zandbergen, H. W.; Manna, L.; Van Tendeloo, G., Thermally Induced Structural and Morphological Changes of CdSe/CdS Octapods. Small 2012, 8 (6), Peters, K. F.; Cohen, J. B.; Chung, Y.-W., Melting of Pb nanocrystals. Physical Review B 1998, 57 (21), Chen, H.; Yu, Y.; Xin, H. L.; Newton, K. A.; Holtz, M. E.; Wang, D.; Muller, D. A.; Abruña, H. D.; DiSalvo, F. J., Coalescence in the Thermal Annealing of Nanoparticles: An in Situ STEM Study of the Growth Mechanisms of Ordered Pt Fe Nanoparticles in a KCl Matrix. Chemistry of Materials 2013, 25 (8), Liu, Z.; Che, R.; Elzatahry, A. A.; Zhao, D., Direct Imaging Au Nanoparticle Migration Inside Mesoporous Silica Channels. ACS Nano 2014, 8 (10), Padbury, R. P.; Halbur, J. C.; Krommenhoek, P. J.; Tracy, J. B.; Jur, J. S., Thermal Stability of Gold Nanoparticles Embedded within Metal Oxide Frameworks Fabricated by Hybrid Modifications onto Sacrificial Textile Templates. Langmuir 2015, 31 (3), Simonsen, S. B.; Chorkendorff, I.; Dahl, S.; Skoglundh, M.; Sehested, J.; Helveg, S., Ostwald ripening in a Pt/SiO2 model catalyst studied by in situ TEM. Journal of Catalysis 2011, 281 (1), Ristau, R.; Tiruvalam, R.; Clasen, P.; Gorskowski, E.; Harmer, M.; Kiely, C.; Hussain, I.; Brust, M., Electron microscopy studies of the thermal stability of gold nanoparticle arrays. Gold Bulletin 2009, 42 (2), Sharma, R.; Crozier, P., Environmental Transmission Electron Microscopy in Nanotechnology. In Handbook of Microscopy for Nanotechnology, Yao, N.; Wang, Z., Eds. Springer US: 2005; pp Allen, G. L.; Bayles, R. A.; Gile, W. W.; Jesser, W. A., Small particle melting of pure metals. Thin Solid Films 1986, 144 (2), Chen, C. L.; Lee, J.-G.; Arakawa, K.; Mori, H., In situ observations of crystalline-to-liquid and crystalline-to-gas transitions of substrate-supported Ag nanoparticles. Applied Physics Letters 2010, 96 (25), Chen, C. L.; Lee, J.-G.; Arakawa, K.; Mori, H., Comparative study on size dependence of melting temperatures of pure metal and alloy nanoparticles. Applied Physics Letters 2011, 99 (1),

127 342. Buffat, P.; Borel, J. P., Size effect on melting temperature of gold particles. Physical Review A 1976, 13 (6), Zhang, M.; Olson, E. A.; Twesten, R. D.; Wen, J. G.; Allen, L. H.; Robertson, I. M.; Petrov, I., In situ Transmission Electron Microscopy Studies Enabled by Microelectromechanical System Technology. Journal of Materials Research 2005, 20 (07), Zhang, D. L.; Cantor, B., Melting behaviour of In and Pb particles embedded in an Al matrix. Acta Metallurgica et Materialia 1991, 39 (7), Mayoral, A.; Allard, L. F.; Ferrer, D.; Esparza, R.; Jose-Yacaman, M., On the behavior of Ag nanowires under high temperature: in situ characterization by aberration-corrected STEM. Journal of Materials Chemistry 2011, 21 (3), Sinclair, R., In situ high-resolution transmission electron microscopy of material reactions. MRS Bulletin 2013, 38 (12), Picher, M.; Mazzucco, S.; Blankenship, S.; Sharma, R., Vibrational and optical spectroscopies integrated with environmental transmission electron microscopy. Ultramicroscopy 2015, 150, Shohei, C.; Yoichi, M.; Yuhei, M.; Shigeo, M., Temperature Dependence of Raman Scattering from Single-Walled Carbon Nanotubes: Undefined Radial Breathing Mode Peaks at High Temperatures. Japanese Journal of Applied Physics 2008, 47 (4R), Acharya, S.; Sarma, D. D.; Golan, Y.; Sengupta, S.; Ariga, K., Shape-Dependent Confinement in Ultrasmall Zero-, One-, and Two-Dimensional PbS Nanostructures. Journal of the American Chemical Society 2009, 131 (32), Tian, N.; Zhou, Z.-Y.; Sun, S.-G.; Ding, Y.; Wang, Z. L., Synthesis of Tetrahexahedral Platinum Nanocrystals with High-Index Facets and High Electro-Oxidation Activity. Science 2007, 316 (5825), Mohamed, M. B.; Burda, C.; El-Sayed, M. A., Shape Dependent Ultrafast Relaxation Dynamics of CdSe Nanocrystals: Nanorods vs Nanodots. Nano Letters 2001, 1 (11), Orendorff, C. J.; Sau, T. K.; Murphy, C. J., Shape-Dependent Plasmon-Resonant Gold Nanoparticles. Small 2006, 2 (5), Xia, Y.; Xiong, Y.; Lim, B.; Skrabalak, S. E., Shape Controlled Synthesis of Metal Nanocrystals: Simple Chemistry Meets Complex Physics? Angewandte Chemie International Edition 2009, 48 (1), Liao, H.-G.; Zherebetskyy, D.; Xin, H.; Czarnik, C.; Ercius, P.; Elmlund, H.; Pan, M.; Wang, L.- W.; Zheng, H., Facet development during platinum nanocube growth. Science 2014, 345 (6199), Jeong, M.; Yuk, J. M.; Lee, J. Y., Observation of Surface Atoms during Platinum Nanocrystal Growth by Monomer Attachment. Chemistry of Materials 2015, 27 (9), Penn, R. L.; Banfield, J. F., Imperfect oriented attachment: Dislocation generation in defectfree nanocrystals. Science 1998, 281 (5379),

128 357. Niederberger, M.; Coelfen, H., Oriented attachment and mesocrystals: Non-classical crystallization mechanisms based on nanoparticle assembly. Physical Chemistry Chemical Physics 2006, 8 (28), Aabdin, Z.; Lu, J.; Zhu, X.; Anand, U.; Loh, N. D.; Su, H.; Mirsaidov, U., Bonding Pathways of Gold Nanocrystals in Solution. Nano Letters 2014, 14 (11), Banfield, J. F.; Welch, S. A.; Zhang, H. Z.; Ebert, T. T.; Penn, R. L., Aggregation-based crystal growth and microstructure development in natural iron oxyhydroxide biomineralization products. Science 2000, 289 (5480), Pacholski, C.; Kornowski, A.; Weller, H., Self-assembly of ZnO: From nanodots, to nanorods. Angewandte Chemie-International Edition 2002, 41 (7), Halder, A.; Ravishankar, N., Ultrafine single-crystalline gold nanowire arrays by oriented attachment. Advanced Materials 2007, 19 (14), Cho, K. S.; Talapin, D. V.; Gaschler, W.; Murray, C. B., Designing PbSe nanowires and nanorings through oriented attachment of nanoparticles. Journal of the American Chemical Society 2005, 127 (19), Pradhan, N.; Xu, H. F.; Peng, X. G., Colloidal CdSe quantum wires by oriented attachment. Nano Letters 2006, 6 (4), Yu, J. H.; Joo, J.; Park, H. M.; Baik, S. I.; Kim, Y. W.; Kim, S. C.; Hyeon, T., Synthesis of quantum-sized cubic ZnS nanorods by the oriented attachment mechanism. Journal of the American Chemical Society 2005, 127 (15), Niu, K.-Y.; Liao, H.-G.; Zheng, H., Visualization of the Coalescence of Bismuth Nanoparticles. Microscopy and Microanalysis 2014, 20 (2), Liao, H.-G.; Zheng, H., Liquid Cell Transmission Electron Microscopy Study of Platinum Iron Nanocrystal Growth and Shape Evolution. Journal of the American Chemical Society 2013, 135 (13), Park, J.; Kodambaka, S.; Ross, F. M.; Grogan, J. M.; Bau, H. H., In Situ Liquid Cell Transmission Electron Microscopic Observation of Electron Beam Induced Au Crystal Growth in a Solution. Microscopy and Microanalysis 2012, 18, De Clercq, A.; Dachraoui, W.; Margeat, O.; Pelzer, K.; Henry, C. R.; Giorgio, S., Growth of Pt Pd Nanoparticles Studied In Situ by HRTEM in a Liquid Cell. The Journal of Physical Chemistry Letters 2014, 5 (12), Parent, L. R.; Robinson, D. B.; Cappillino, P. J.; Hartnett, R. J.; Abellan, P.; Evans, J. E.; Browning, N. D.; Arslan, I., In Situ Observation of Directed Nanoparticle Aggregation During the Synthesis of Ordered Nanoporous Metal in Soft Templates. Chemistry of Materials 2014, 26 (3), Parent, L. R.; Robinson, D. B.; Woehl, T. J.; Ristenpart, W. D.; Evans, J. E.; Browning, N. D.; Arslan, I., Direct in Situ Observation of Nanoparticle Synthesis in a Liquid Crystal Surfactant Template. ACS Nano 2012, 6 (4),

129 371. Nielsen, M. H.; Li, D. S.; Zhang, H. Z.; Aloni, S.; Han, T. Y. J.; Frandsen, C.; Seto, J.; Banfield, J. F.; Colfen, H.; De Yoreo, J. J., Investigating Processes of Nanocrystal Formation and Transformation via Liquid Cell TEM. Microscopy and Microanalysis 2014, 20 (2), Ghosh Chaudhuri, R.; Paria, S., Core/shell nanoparticles: classes, properties, synthesis mechanisms, characterization, and applications. Chemical Reviews 2011, 112 (4), Greytak, A. B.; Allen, P. M.; Liu, W.; Zhao, J.; Young, E. R.; Popovic, Z.; Walker, B. J.; Nocera, D. G.; Bawendi, M. G., Alternating layer addition approach to CdSe/CdS core/shell quantum dots with near-unity quantum yield and high on-time fractions. Chemical Science 2012, 3 (6), Page, R. C.; Espinobarro-Velazquez, D.; Leontiadou, M. A.; Smith, C.; Lewis, E. A.; Haigh, S. J.; Li, C.; Radtke, H.; Pengpad, A.; Bondino, F.; Magnano, E.; Pis, I.; Flavell, W. R.; O'Brien, P.; Binks, D. J., Near-Unity Quantum Yields from Chloride Treated CdTe Colloidal Quantum Dots. Small 2014, 11 (13), Sutter, E.; Jungjohann, K.; Bliznakov, S.; Courty, A.; Maisonhaute, E.; Tenney, S.; Sutter, P., In situ liquid-cell electron microscopy of silver palladium galvanic replacement reactions on silver nanoparticles. Nature Communications 2014, 5, Lee, C.; Wei, X.; Kysar, J. W.; Hone, J., Measurement of the Elastic Properties and Intrinsic Strength of Monolayer Graphene. Science 2008, 321 (5887), Wilson, N. R.; Pandey, P. A.; Beanland, R.; Young, R. J.; Kinloch, I. A.; Gong, L.; Liu, Z.; Suenaga, K.; Rourke, J. P.; York, S. J.; Sloan, J., Graphene Oxide: Structural Analysis and Application as a Highly Transparent Support for Electron Microscopy. ACS Nano 2009, 3 (9), Panthani, M. G.; Hessel, C. M.; Reid, D.; Casillas, G.; José-Yacamán, M.; Korgel, B. A., Graphene-Supported High-Resolution TEM and STEM Imaging of Silicon Nanocrystals and their Capping Ligands. The Journal of Physical Chemistry C 2012, 116 (42), Nair, R. R.; Blake, P.; Blake, J. R.; Zan, R.; Anissimova, S.; Bangert, U.; Golovanov, A. P.; Morozov, S. V.; Geim, A. K.; Novoselov, K. S.; Latychevskaia, T., Graphene as a transparent conductive support for studying biological molecules by transmission electron microscopy. Applied Physics Letters 2010, 97 (15), Lee, Z.; Jeon, K.-J.; Dato, A.; Erni, R.; Richardson, T. J.; Frenklach, M.; Radmilovic, V., Direct Imaging of Soft-Hard Interfaces Enabled by Graphene. Nano Letters 2009, 9 (9), Berry, V., Impermeability of graphene and its applications. Carbon 2013, 62 (0), Bunch, J. S.; Verbridge, S. S.; Alden, J. S.; van der Zande, A. M.; Parpia, J. M.; Craighead, H. G.; McEuen, P. L., Impermeable atomic membranes from graphene sheets. Nano Letters 2008, 8 (8), Booth, T. J.; Blake, P.; Nair, R. R.; Jiang, D.; Hill, E. W.; Bangert, U.; Bleloch, A.; Gass, M.; Novoselov, K. S.; Katsnelson, M. I.; Geim, A. K., Macroscopic graphene membranes and their extraordinary stiffness. Nano Letters 2008, 8 (8), Park, J.; Elmlund, H.; Ercius, P.; Yuk, J. M.; Limmer, D. T.; Chen, Q.; Kim, K.; Han, S. H.; Weitz, D. A.; Zettl, A.; Alivisatos, A. P., 3D structure of individual nanocrystals in solution by electron microscopy. Science 2015, 349 (6245),

130 385. Chen, Q.; Smith, J. M.; Park, J.; Kim, K.; Ho, D.; Rasool, H. I.; Zettl, A.; Alivisatos, A. P., 3D Motion of DNA-Au Nanoconjugates in Graphene Liquid Cell Electron Microscopy. Nano Letters 2013, 13 (9), Wang, C.; Qiao, Q.; Shokuhfar, T.; Klie, R. F., High-Resolution Electron Microscopy and Spectroscopy of Ferritin in Biocompatible Graphene Liquid Cells and Graphene Sandwiches. Advanced Materials 2014, 26 (21), Lu, J.; Aabdin, Z.; Loh, N. D.; Bhattacharya, D.; Mirsaidov, U., Nanoparticle Dynamics in a Nanodroplet. Nano Letters 2014, 14 (4), Dukes, M. J.; Jacobs, B. W.; Morgan, D. G.; Hegde, H.; Kelly, D. F., Visualizing nanoparticle mobility in liquid at atomic resolution. Chemical Communications 2013, 49 (29), Jungjohann, K. L.; Evans, J. E.; Aguiar, J. A.; Arslan, I.; Browning, N. D., Atomic-Scale Imaging and Spectroscopy for In Situ Liquid Scanning Transmission Electron Microscopy. Microscopy and Microanalysis 2012, 18 (3), Jungjohann, K. L.; Bliznakov, S.; Adzic, R.; Sutter, E., Monitoring Galvanic Replacement of Ag Nanoparticles by Pd using Low Dose In Situ Liquid S/TEM. Microscopy and Microanalysis 2013, 19, Clough, R. N.; Moldovan, G.; Kirkland, A. I., Direct Detectors for Electron Microscopy. Journal of Physics: Conference Series 2014, 522 (1), Zaluzec, N.; Demortiere, A.; Cook, R.; Koritala, R.; Wen, J.; Miller, D.; Kulzick, M., X-ray and Electron Energy Loss Spectroscopy in Liquids in the Analytical S/TEM. Microscopy and Microanalysis 2014, 20, Wang, C. M.; Genc, A.; Cheng, H.; Pullan, L.; Baer, D. R.; Bruemmer, S. M., In-Situ TEM visualization of vacancy injection and chemical partition during oxidation of Ni-Cr nanoparticles. Scientific Reports 2014, 4, Lu, X.; Zhao, Y.; Wang, C., Fabrication of PbS Nanoparticles in Polymer-Fiber Matrices by Electrospinning. Advanced Materials 2005, 17 (20), Zhou, Y.; Itoh, H.; Uemura, T.; Naka, K.; Chujo, Y., Preparation, Optical Spectroscopy, and Electrochemical Studies of Novel π-conjugated Polymer-Protected Stable PbS Colloidal Nanoparticles in a Nonaqueous Solution. Langmuir 2002, 18 (13), Norris, D. J.; Bawendi, M. G., Measurement and assignment of the size-dependent optical spectrum in CdSe quantum dots. Physical Review B 1996, 53 (24), Du, H.; Chen, C.; Krishnan, R.; Krauss, T. D.; Harbold, J. M.; Wise, F. W.; Thomas, M. G.; Silcox, J., Optical Properties of Colloidal PbSe Nanocrystals. Nano Letters 2002, 2 (11), Akhtar, J.; Afzaal, M.; Banski, M.; Podhorodecki, A.; Syperek, M.; Misiewicz, J.; Bangert, U.; Hardman, S. J. O.; Graham, D. M.; Flavell, W. R.; Binks, D. J.; Gardonio, S.; O Brien, P., Controlled Synthesis of Tuned Bandgap Nanodimensional Alloys of PbSxSe1 x. Journal of the American Chemical Society 2011, 133 (14), Ma, W.; Luther, J. M.; Zheng, H.; Wu, Y.; Alivisatos, A. P., Photovoltaic Devices Employing Ternary PbSxSe1-x Nanocrystals. Nano Letters 2009, 9 (4),

131 400. Zhai, G.; Church, C. P.; Breeze, A. J.; Zhang, D.; Alers, G. B.; Carter, S. A., Quantum dot PbS0.9Se0.1/TiO2 heterojunction solar cells. Nanotechnology 2012, 23 (40), Onicha, A. C.; Petchsang, N.; Kosel, T. H.; Kuno, M., Controlled Synthesis of Compositionally Tunable Ternary PbSexS1 x as Well as Binary PbSe and PbS Nanowires. ACS Nano 2012, 6 (3), Ma, W.; Swisher, S. L.; Ewers, T.; Engel, J.; Ferry, V. E.; Atwater, H. A.; Alivisatos, A. P., Photovoltaic Performance of Ultrasmall PbSe Quantum Dots. Acs Nano 2011, 5 (10), McDonald, S. A.; Konstantatos, G.; Zhang, S. G.; Cyr, P. W.; Klem, E. J. D.; Levina, L.; Sargent, E. H., Solution-processed PbS quantum dot infrared photodetectors and photovoltaics. Nature Materials 2005, 4 (2), Semonin, O. E.; Luther, J. M.; Choi, S.; Chen, H.-Y.; Gao, J.; Nozik, A. J.; Beard, M. C., Peak External Photocurrent Quantum Efficiency Exceeding 100% via MEG in a Quantum Dot Solar Cell. Science 2011, 334 (6062), Leventis, H. C.; O'Mahony, F.; Akhtar, J.; Afzaal, M.; O'Brien, P.; Haque, S. A., Transient Optical Studies of Interfacial Charge Transfer at Nanostructured Metal Oxide/PbS Quantum Dot/Organic Hole Conductor Heterojunctions. Journal of the American Chemical Society 2010, 132 (8), Leschkies, K. S.; Beatty, T. J.; Kang, M. S.; Norris, D. J.; Aydil, E. S., Solar Cells Based on Junctions between Colloidal PbSe Nanocrystals and Thin ZnO Films. ACS Nano 2009, 3 (11), Pattantyus-Abraham, A. G.; Kramer, I. J.; Barkhouse, A. R.; Wang, X.; Konstantatos, G.; Debnath, R.; Levina, L.; Raabe, I.; Nazeeruddin, M. K.; Gra tzel, M.; Sargent, E. H., Depleted- Heterojunction Colloidal Quantum Dot Solar Cells. ACS Nano 2010, 4 (6), Seo, J.; Cho, M. J.; Lee, D.; Cartwright, A. N.; Prasad, P. N., Efficient Heterojunction Photovoltaic Cell Utilizing Nanocomposites of Lead Sulfide Nanocrystals and a Low-Bandgap Polymer. Advanced Materials 2011, 23 (34), Radychev, N.; Lokteva, I.; Witt, F.; Kolny-Olesiak, J.; Borchert, H.; Parisi, J., Physical Origin of the Impact of Different Nanocrystal Surface Modifications on the Performance of CdSe/P3HT Hybrid Solar Cells. The Journal of Physical Chemistry C 2011, 115 (29), Jangwon, S.; Sung Jin, K.; Won Jin, K.; Rohit, S.; Marek, S.; Alexander, N. C.; Paras, N. P., Enhancement of the photovoltaic performance in PbS nanocrystal:p3ht hybrid composite devices by post-treatment-driven ligand exchange. Nanotechnology 2009, 20 (9), Kahmann, S.; Mura, A.; Protesescu, L.; Kovalenko, M.; Brabec, C. J.; Loi, M. A., Optoelectronics of PbS quantum dot and narrow bandgap polymer blends. Journal of Materials Chemistry C 2015, 3, Wuister, S. F.; de Mello Donegá, C.; Meijerink, A., Influence of Thiol Capping on the Exciton Luminescence and Decay Kinetics of CdTe and CdSe Quantum Dots. The Journal of Physical Chemistry B 2004, 108 (45),

132 413. Green, M., The nature of quantum dot capping ligands. Journal of Materials Chemistry 2010, 20 (28), Dowland, S.; Lutz, T.; Ward, A.; King, S. P.; Sudlow, A.; Hill, M. S.; Molloy, K. C.; Haque, S. A., Direct Growth of Metal Sulfide Nanoparticle Networks in Solid-State Polymer Films for Hybrid Inorganic Organic Solar Cells. Advanced Materials 2011, 23 (24), Lokteva, I.; Radychev, N.; Witt, F.; Borchert, H.; Parisi, J.; Kolny-Olesiak, J., Surface Treatment of CdSe Nanoparticles for Application in Hybrid Solar Cells: The Effect of Multiple Ligand Exchange with Pyridine. The Journal of Physical Chemistry C 2010, 114 (29), Seo, J.; Kim, W. J.; Kim, S. J.; Lee, K.-S.; Cartwright, A. N.; Prasad, P. N., Polymer nanocomposite photovoltaics utilizing CdSe nanocrystals capped with a thermally cleavable solubilizing ligand. Applied Physics Letters 2009, 94 (13), Warner, J. H.; Watt, A. A. R.; Tilley, R. D., Controlling PbS nanocrystal aggregation in conducting polymers. Nanotechnology 2005, 16 (10), Warner, J. H.; Watt, A. A. R., Monodisperse PbS nanocrystals synthesized in a conducting polymer. Materials Letters 2006, 60 (19), Noone, K. M.; Anderson, N. C.; Horwitz, N. E.; Munro, A. M.; Kulkarni, A. P.; Ginger, D. S., Absence of Photoinduced Charge Transfer in Blends of PbSe Quantum Dots and Conjugated Polymers. ACS Nano 2009, 3 (6), Ip, A. H.; Thon, S. M.; Hoogland, S.; Voznyy, O.; Zhitomirsky, D.; Debnath, R.; Levina, L.; Rollny, L. R.; Carey, G. H.; Fischer, A.; Kemp, K. W.; Kramer, I. J.; Ning, Z. J.; Labelle, A. J.; Chou, K. W.; Amassian, A.; Sargent, E. H., Hybrid passivated colloidal quantum dot solids. Nature Nanotechnology 2012, 7 (9), Chuang, C.-H. M.; Brown, P. R.; Bulović, V.; Bawendi, M. G., Improved performance and stability in quantum dot solar cells through band alignment engineering. Nature Materials 2014, 13 (8), Zhou, H.; Yang, L.; Stuart, A. C.; Price, S. C.; Liu, S.; You, W., Development of Fluorinated Benzothiadiazole as a Structural Unit for a Polymer Solar Cell of 7 % Efficiency. Angewandte Chemie 2011, 123 (13), You, J.; Dou, L.; Yoshimura, K.; Kato, T.; Ohya, K.; Moriarty, T.; Emery, K.; Chen, C.-C.; Gao, J.; Li, G.; Yang, Y., A polymer tandem solar cell with 10.6% power conversion efficiency. Nature Communications 2013, 4, Liang, Y.; Xu, Z.; Xia, J.; Tsai, S.-T.; Wu, Y.; Li, G.; Ray, C.; Yu, L., For the Bright Future Bulk Heterojunction Polymer Solar Cells with Power Conversion Efficiency of 7.4%. Advanced Materials 2010, 22 (20), Price, S. C.; Stuart, A. C.; Yang, L.; Zhou, H.; You, W., Fluorine Substituted Conjugated Polymer of Medium Band Gap Yields 7% Efficiency in Polymer Fullerene Solar Cells. Journal of the American Chemical Society 2011, 133 (12), NREL Best Research-Cell Efficiencies. (accessed ). 132

133 427. Chen, C.-C.; Chang, W.-H.; Yoshimura, K.; Ohya, K.; You, J.; Gao, J.; Hong, Z.; Yang, Y., An Efficient Triple-Junction Polymer Solar Cell Having a Power Conversion Efficiency Exceeding 11%. Advanced Materials 2014, 26 (32), Eperon, G. E.; Burlakov, V. M.; Docampo, P.; Goriely, A.; Snaith, H. J., Morphological Control for High Performance, Solution-Processed Planar Heterojunction Perovskite Solar Cells. Advanced Functional Materials 2014, 24 (1), Liu, D.; Kelly, T. L., Perovskite solar cells with a planar heterojunction structure prepared using room-temperature solution processing techniques. Nature Photonics 2014, 8 (2), Zhou, H.; Chen, Q.; Li, G.; Luo, S.; Song, T.-b.; Duan, H.-S.; Hong, Z.; You, J.; Liu, Y.; Yang, Y., Interface engineering of highly efficient perovskite solar cells. Science 2014, 345 (6196), Reynolds, L. X.; Lutz, T.; Dowland, S.; MacLachlan, A.; King, S.; Haque, S. A., Charge photogeneration in hybrid solar cells: A comparison between quantum dots and in situ grown CdS. Nanoscale 2012, 4 (5), Rath, T.; Edler, M.; Haas, W.; Fischereder, A.; Moscher, S.; Schenk, A.; Trattnig, R.; Sezen, M.; Mauthner, G.; Pein, A.; Meischler, D.; Bartl, K.; Saf, R.; Bansal, N.; Haque, S. A.; Hofer, F.; List, E. J. W.; Trimmel, G., A Direct Route Towards Polymer/Copper Indium Sulfide Nanocomposite Solar Cells. Advanced Energy Materials 2011, 1 (6), Rath, T.; Trimmel, G., In situ syntheses of semiconducting nanoparticles in conjugated polymer matrices and their application in photovoltaics. Hybrid Materials 2014, 1 (1), Leventis, H. C.; King, S. P.; Sudlow, A.; Hill, M. S.; Molloy, K. C.; Haque, S. A., Nanostructured Hybrid Polymer Inorganic Solar Cell Active Layers Formed by Controllable in Situ Growth of Semiconducting Sulfide Networks. Nano Letters 2010, 10 (4), Di Benedetto, F.; Camposeo, A.; Persano, L.; Laera, A. M.; Piscopiello, E.; Cingolani, R.; Tapfer, L.; Pisignano, D., Light-emitting nanocomposite CdS-polymer electrospun fibres via in situ nanoparticle generation. Nanoscale 2011, 3 (10), Park, J.; An, K. J.; Hwang, Y. S.; Park, J. G.; Noh, H. J.; Kim, J. Y.; Park, J. H.; Hwang, N. M.; Hyeon, T., Ultra-large-scale syntheses of monodisperse nanocrystals. Nature Materials 2004, 3 (12), Nightingale, A. M.; Bannock, J. H.; Krishnadasan, S. H.; O'Mahony, F. T. F.; Haque, S. A.; Sloan, J.; Drury, C.; McIntyre, R.; demello, J. C., Large-scale synthesis of nanocrystals in a multichannel droplet reactor. Journal of Materials Chemistry A 2013, 1 (12), Krebs, F. C.; Tromholt, T.; Jorgensen, M., Upscaling of polymer solar cell fabrication using full roll-to-roll processing. Nanoscale 2010, 2 (6), Akhtar, J.; Malik, M. A.; O'Brien, P.; Helliwell, M., Controlled synthesis of PbS nanoparticles and the deposition of thin films by Aerosol-Assisted Chemical Vapour Deposition (AACVD). Journal of Materials Chemistry 2010, 20 (29), Boadi, N. O.; Malik, M. A.; O'Brien, P.; Awudza, J. A. M., Single source molecular precursor routes to lead chalcogenides. Dalton Transactions 2012, 41 (35),

134 441. Malik, M. A.; Afzaal, M.; O'Brien, P., Precursor Chemistry for Main Group Elements in Semiconducting Materials. Chemical Reviews 2010, 110 (7), Fan, D.; Afzaal, M.; Mallik, M. A.; Nguyen, C. Q.; O'Brien, P.; Thomas, P. J., Using coordination chemistry to develop new routes to semiconductor and other materials. Coordination Chemistry Reviews 2007, 251 (13-14), Malik, M. A.; Revaprasadu, N.; O'Brien, P., Air-Stable Single-Source Precursors for the Synthesis of Chalcogenide Semiconductor Nanoparticles. Chemistry of Materials 2001, 13 (3), Pickett, N. L.; O'Brien, P., Syntheses of semiconductor nanoparticles using single-molecular precursors. The Chemical Record 2001, 1 (6), Pradhan, N.; Katz, B.; Efrima, S., Synthesis of high-quality metal sulfide nanoparticles from alkyl xanthate single precursors in alkylamine solvents. Journal of Physical Chemistry B 2003, 107 (50), Clark, J. M.; Kociok-Kohn, G.; Harnett, N. J.; Hill, M. S.; Hill, R.; Molloy, K. C.; Saponia, H.; Stanton, D.; Sudlow, A., Formation of PbS materials from lead xanthate precursors. Dalton Transactions 2011, 40 (26), Alam, N.; Hill, M. S.; Kociok-Köhn, G.; Zeller, M.; Mazhar, M.; Molloy, K. C., Pyridine Adducts of Nickel(II) Xanthates as Single-Source Precursors for the Aerosol-Assisted Chemical Vapor Deposition of Nickel Sulfide. Chemistry of Materials 2008, 20 (19), Akhtar, J.; Afzaal, M.; Vincent, M. A.; Burton, N. A.; Hillier, I. H.; O'Brien, P., Low temperature CVD growth of PbS films on plastic substrates. Chemical Communications 2011, 47 (7), DePuy, C. H.; King, R. W., Pyrolytic Cis Eliminations. Chemical Reviews 1960, 60 (5), Fradler, C.; Rath, T.; Dunst, S.; Letofsky-Papst, I.; Saf, R.; Kunert, B.; Hofer, F.; Resel, R.; Trimmel, G., Flexible polymer/copper indium sulfide hybrid solar cells and modules based on the metal xanthate route and low temperature annealing. Solar Energy Materials and Solar Cells 2014, 124 (0), Pradhan, N.; Efrima, S., Single-precursor, one-pot versatile synthesis under near ambient conditions of tunable, single and dual band fluorescing metal sulfide nanoparticles. Journal of the American Chemical Society 2003, 125 (8), Nair, P. S.; Radhakrishnan, T.; Revaprasadu, N.; Kolawole, G.; O'Brien, P., Cadmium ethylxanthate: A novel single-source precursor for the preparation of CdS nanoparticles. Journal of Materials Chemistry 2002, 12 (9), Li, Y.; Li, X.; Yang, C.; Li, Y., Controlled synthesis of CdS nanorods and hexagonal nanocrystals. Journal of Materials Chemistry 2003, 13 (10), Liu, Y.; Ge, M.; Yue, Y.; Sun, Y.; Wu, Y.; Chen, X.; Dai, N., Colloidal Cu2ZnSnS4 nanocrystals generated by a facile route using ethylxanthate molecular precursors. Physica Status Solidi Rapid Research Letters 2011, 5 (3),

135 455. Pradhan, N.; Efrima, S., Supercrystals of Uniform Nanorods and Nanowires, and the Nanorod-to-Nanowire Oriented Transition. The Journal of Physical Chemistry B 2004, 108 (32), Li, Y. C.; Ye, M. F.; Yang, C. H.; Li, X. H.; Li, Y. F., Composition- and Shape-Controlled Synthesis and Optical Properties of ZnxCd1 xs Alloyed Nanocrystals. Advanced Functional Materials 2005, 15 (3), Lewis, E. A.; McNaughter, P. D.; Yin, Z.; Chen, Y.; Brent, J. R.; Saah, S. A.; Raftery, J.; Awudza, J. A. M.; Malik, M. A.; O Brien, P.; Haigh, S. J., In Situ Synthesis of PbS Nanocrystals in Polymer Thin Films from Lead(II) Xanthate and Dithiocarbamate Complexes: Evidence for Size and Morphology Control. Chemistry of Materials 2015, 27 (6), Antolini, F.; Di Luccio, T.; Laera, A. M.; Mirenghi, L.; Piscopiello, E.; Re, M.; Tapfer, L., Direct synthesis of II VI compound nanocrystals in polymer matrix. Physica Status Solidi (b) 2007, 244 (8), Antolini, F.; Burresi, E.; Stroea, L.; Morandi, V.; Ortolani, L.; Accorsi, G.; Blosi, M., Time and Temperature Dependence of CdS Nanoparticles Grown in a Polystyrene Matrix. Journal of Nanomaterials 2012, 2012, Di Luccio, T.; Laera, A. M.; Tapfer, L.; Kempter, S.; Kraus, R.; Nickel, B., Controlled Nucleation and Growth of CdS Nanoparticles in a Polymer Matrix. The Journal of Physical Chemistry B 2006, 110 (25), Bansal, N.; Reynolds, L. X.; MacLachlan, A.; Lutz, T.; Ashraf, R. S.; Zhang, W.; Nielsen, C. B.; McCulloch, I.; Rebois, D. G.; Kirchartz, T.; Hill, M. S.; Molloy, K. C.; Nelson, J.; Haque, S. A., Influence of Crystallinity and Energetics on Charge Separation in Polymer Inorganic Nanocomposite Films for Solar Cells. Scientific Reports 2013, 3, MacLachlan, A. J.; O'Mahony, F. T. F.; Sudlow, A. L.; Hill, M. S.; Molloy, K. C.; Nelson, J.; Haque, S. A., Solution-Processed Mesoscopic Bi2S3:Polymer Photoactive Layers. ChemPhysChem 2014, 15 (6), Lee, H.; Leventis, H. C.; Moon, S.-J.; Chen, P.; Ito, S.; Haque, S. A.; Torres, T.; Nüesch, F.; Geiger, T.; Zakeeruddin, S. M.; Grätzel, M.; Nazeeruddin, M. K., PbS and CdS Quantum Dot-Sensitized Solid-State Solar Cells: Old Concepts, New Results. Advanced Functional Materials 2009, 19 (17), Lutz, T.; MacLachlan, A.; Sudlow, A.; Nelson, J.; Hill, M. S.; Molloy, K. C.; Haque, S. A., Thermal decomposition of solution processable metal xanthates on mesoporous titanium dioxide films: a new route to quantum-dot sensitised heterojunctions. Physical Chemistry Chemical Physics 2012, 14 (47), Kaltenhauser, V.; Rath, T.; Haas, W.; Torvisco, A.; Muller, S. K.; Friedel, B.; Kunert, B.; Saf, R.; Hofer, F.; Trimmel, G., Bismuth sulphide-polymer nanocomposites from a highly soluble bismuth xanthate precursor. Journal of Materials Chemistry C 2013, 1 (47), Lee, S. M.; Cho, S. N.; Cheon, J., Anisotropic shape control of colloidal inorganic nanocrystals. Advanced Materials 2003, 15 (5),

136 467. Carotenuto, G.; Martorana, B.; Perlo, P.; Nicolais, L., A universal method for the synthesis of metal and metal sulfide clusters embedded in polymer matrices. Journal of Materials Chemistry 2003, 13 (12), Oosterhout, S. D.; Wienk, M. M.; van Bavel, S. S.; Thiedmann, R.; Koster, L. J. A.; Gilot, J.; Loos, J.; Schmidt, V.; Janssen, R. A. J., The effect of three-dimensional morphology on the efficiency of hybrid polymer solar cells. Nature Materials 2009, 8 (10), Holtz, M. E.; Yu, Y.; Gao, J.; Abruna, H. D.; Muller, D. A., In Situ Electron Energy-Loss Spectroscopy in Liquids. Microscopy and Microanalysis 2013, 19 (4), Lewis, E. A.; Downie, H.; Collins, R. F.; Prestat, E.; Lloyd, J. R.; Haigh, S. J., Imaging the Hydrated Microbe-Metal Interface Using Nanoscale Spectrum Imaging. Particle & Particle Systems Characterization 2016, DOI: /ppsc Hubbell, J. H.; Seltzer, S. M. Tables of X-ray mass attenuation coefficients and mass energyabsorption coefficients 1 kev to 20 MeV for elements Z= 1 to 92 and 48 additional substances of dosimetric interest; National Institute of Standards and Technology, Gaithersburg, MD (United States). Ionizing Radiation Division: Long, E.; O'Brien, S.; Lewis, E. A.; Prestat, E.; Downing, C.; Cucinotta, C. S.; Sanvito, S.; Haigh, S. J.; Nicolosi, V., An in-and ex-situ TEM study into the oxidisation of titanium (IV) sulphide. arxiv preprint arxiv: Shoko, M.; Jun-ichi, F.; Yukinori, O.; Eiichi, N.; Shinji, M., Nanometer-Scale Patterning of Polystyrene Resists in Low-Voltage Electron Beam Lithography. Japanese Journal of Applied Physics 1997, 36 (12S),

137 Appendix 1. Real-time imaging and local elemental analysis of nanostructures in liquids Introduction The number of studies employing electron microscopy to study specimens in liquid has grown dramatically in recent years, largely thanks to the development and commercialisation of SiN x windowed liquid-cells. 1-3 This emerging field is reviewed in section While the technique has allowed the imaging of hydrated biological cells, 4 observations of nanocrystal growth and dynamics, 5 corrosion, 6 and electrochemical processes, 1, 7 it has been largely limited to TEM and STEM imaging. As highlighted in sections 1.3 and 2.3 one of the great strengths of the (scanning) transmission electron microscope is its ability not just to perform high resolution imaging, but to simultaneously probe the material s chemical composition and map elemental distributions with high spatial resolution. Due to the absence of complementary spectroscopic analysis techniques, liquid-cell experiments are often limited to simple monometallic systems, as interpretation of the transformations observed in more chemically complex systems is likely to be impossible from images alone. The manuscript presented in this appendix: Real-time imaging and local elemental analysis of nanostructures in liquids was published in Chemical Communications in The manuscript demonstrates the use of a modified holder design to perform EDX spectrum imaging in a liquid-cell. Using a model system, comprised of variety nanocrystals in aqueous solution, we show that spectrum imaging in liquid can be achieved with nanometre spatial resolution. This allows morphologically similar nanocrystals to be distinguished and the chemical origin of transformations to be investigated. The modified liquid-cell holder used in this work was designed and made by Nestor Zaluzec. a I produced the nanoparticle soup sample, the palladium-decorated carbon nanotubes included in this soup were synthesised by Pedro Camargo. b The electron microscopy data was acquired by Sarah Haigh c, Nestor Zaluzec a, and I. I analysed the data, produced the figures and wrote the manuscript. Zheyang He c assisted with the development of image theresholding methods. Thomas Slater c helped analyse the tilt and position dependence experiments. Sarah Haigh, Nestor Zaluzec, Grace Burke, c Thomas Slater and Matthew Kulzick d all provided useful discussions and helped edit the manuscript. The published manuscript, along with supporting information and accompanying videos are available online and can be found using the digital object identifier DOI: /C4CC02743D. 137

138 Affiliations a Electron Microscopy Centre, Argonne National Laboratory, Argonne, Illinois, USA. b Instituto de Química, Universidade de São Paulo, São Paulo, Brazil. c School of Materials, University of Manchester, M13 9PL, UK. d BP Corporate Research Centre, Naperville, Illinois, USA. References 1. Williamson, M. J.; Tromp, R. M.; Vereecken, P. M.; Hull, R.; Ross, F. M., Dynamic microscopy of nanoscale cluster growth at the solid-liquid interface. Nature Materials 2003, 2 (8), Liao, H. G.; Niu, K.; Zheng, H., Observation of growth of metal nanoparticles. Chemical Communications 2013, 49 (100), de Jonge, N.; Ross, F. M., Electron microscopy of specimens in liquid. Nature Nanotechnology 2011, 6 (11), de Jonge, N.; Peckys, D. B.; Kremers, G. J.; Piston, D. W., Electron microscopy of whole cells in liquid with nanometer resolution. Proceedings of the National Academy of Sciences of the United States of America 2009, 106 (7), Liao, H. G.; Cui, L. K.; Whitelam, S.; Zheng, H. M., Real-Time Imaging of Pt 3 Fe Nanorod Growth in Solution. Science 2012, 336 (6084), Chee, S. W.; Pratt, S. H.; Hattar, K.; Duquette, D.; Ross, F. M.; Hull, R., Studying localized corrosion using liquid cell transmission electron microscopy. Chemical Communications 2015, 51, Zeng, Z. Y.; Liang, W. I.; Liao, H. G.; Xin, H. L. L.; Chu, Y. H.; Zheng, H. M., Visualization of Electrode-Electrolyte Interfaces in LiPF6/EC/DEC Electrolyte for Lithium Ion Batteries via in Situ TEM. Nano Letters 2014, 14 (4), Lewis, E. A.; Haigh, S. J.; Slater, T. J. A.; He, Z.; Kulzick, M. A.; Burke, M. G.; Zaluzec, N. J., Real-time imaging and local elemental analysis of nanostructures in liquids. Chemical Communications 2014, 50,

139 Manuscript: Real-time imaging and local elemental analysis of nanostructures in liquids 139

140 140

141 141

142 142

143 143

144 Supporting information: Real-time imaging and local elemental analysis of nanostructures in liquids 144

145 145

146 146

147 147

148 148

149 149

150 150

151 The terms used in equations 1 and 2 are defined as follows: Z eff is the effective atomic number of the material, f n is the atomic fraction of the element n and Z n the atomic mass of the element n. λ is the inelastic mean free path of an electron in the material, F is the relativistic factor at the accelerating voltage used, β is the EELS collection angle, and E 0 is the beam energy. 151

152 Appendix 2. Real-time imaging and elemental mapping of AgAu nanoparticle transformations. Introduction Bimetallic nanocrystals are of great interest for applications in catalysis and plasmonics. The particle s properties depend on their nanoscale morphology: size, shape, surface area, chemical composition, and, in particular, surface chemistry. 1-3 AgAu nanoparticles are found to outperform monometallic Ag or Au analogues in a number of catalytic reactions, 4-5 for example the addition of silver is known to improve the ability of Au nanoparticles to oxidise CO. 4, 6 However, degradation of performance during use is a major problem for nanocrystal based catalysts. Degradation may be a consequence of particle coarsening, which leads to reduced surface area. However, it may also be a consequence of changes to the nanocrystals chemical composition or in the nanoscale distribution of elements. Core-shell nanoparticle structures often show significantly different performance from alloyed analogues, 1 so processes which lead to transformations between core-shell and alloyed structures (or vice versa) can be either a source of degradation or a useful method to improve or regenerate catalytic performance. The phenomenon of nanocrystal coarsening has been widely studied with in situ electron microscopy. 7-8 However, changes in elemental distributions are less widely studied. 9 The recent adoption of MEMs based heating holders has now made EDX spectrum imaging during in situ heating experiments possible, allowing direct observation of the changes in elemental distributions during heating (recent developments in in situ electron microscopy are discussed in section 2.4) In this appendix I study AgAu nanoparticles synthesised by the galvanic replacement reaction (discussed in section 1.7.4). 12 Both beam-induced and heating induced transformations are investigated and a rich variety of dynamic processes are observed, including beam-induced hollowing via the nanoscale Kirkendall effect (discussed in section 1.7.5). 13 HAADF STEM imaging and EDX spectrum imaging are used to observe the transformations; spectrum imaging proves critical to the interpretation of the data as changes in elemental distribution accompany all the transformations observed but cannot be identified from imaging alone. This work also highlights the often overlooked role of the support membrane in electron microscopy: it is shown that beam-induced oxidation of silver occurs when using a SiN x support, but not when using the more common amorphous carbon substrate. 152

153 The manuscript found in this appendix: Real-time imaging and elemental mapping of AgAu nanoparticle transformations was published in Nanoscale in The published manuscript, along with supporting information and accompanying videos are available online and can be found using the digital object identifier DOI: /C4NR04837G. I designed and performed the experiments, analysed the data and wrote the manuscript. Alexandra Macedo a and Pedro Camargo a supplied the AgAu nanoparticles used in the experiments. Sarah Haigh, b Thomas Slater, b Eric Prestat b and Paul O Brien b,c all provided useful discussions and helped in the preparation and editing of the manuscript. Affiliations a Instituto de Química, Universidade de São Paulo, São Paulo, Brazil. b School of Materials, University of Manchester, M13 9PL, UK. c School of Chemistry, University of Manchester, M13 9PL, UK. References 1. Jiang, H.-L.; Akita, T.; Ishida, T.; Haruta, M.; Xu, Q., Synergistic Catalysis of Au@Ag Core Shell Nanoparticles Stabilized on Metal Organic Framework. Journal of the American Chemical Society 2011, 133 (5), Tsao, Y.-C.; Rej, S.; Chiu, C.-Y.; Huang, M. H., Aqueous Phase Synthesis of Au Ag Core Shell Nanocrystals with Tunable Shapes and Their Optical and Catalytic Properties. Journal of the American Chemical Society 2014, 136 (1), Slater, T. J. A.; Macedo, A.; Schroeder, S. L. M.; Burke, M. G.; O Brien, P.; Camargo, P. H. C.; Haigh, S. J., Correlating Catalytic Activity of Ag Au Nanoparticles with 3D Compositional Variations. Nano Letters 2014, 14 (4), Liu, J. H.; Wang, A. Q.; Chi, Y. S.; Lin, H. P.; Mou, C. Y., Synergistic effect in an Au-Ag alloy nanocatalyst: CO oxidation. Journal of Physical Chemistry B 2005, 109 (1), Zhang, C.-L.; Lv, K.-P.; Huang, H.-T.; Cong, H.-P.; Yu, S.-H., Co-assembly of Au nanorods with Ag nanowires within polymer nanofiber matrix for enhanced SERS property by electrospinning. Nanoscale 2012, 4 (17), Iizuka, Y.; Kawamoto, A.; Akita, K.; Daté, M.; Tsubota, S.; Okumura, M.; Haruta, M., Effect of Impurity and Pretreatment Conditions on the Catalytic Activity of Au Powder for CO Oxidation. Catalysis Letters 2004, 97 (3-4), Simonsen, S. B.; Chorkendorff, I.; Dahl, S.; Skoglundh, M.; Sehested, J.; Helveg, S., Ostwald ripening in a Pt/SiO2 model catalyst studied by in situ TEM. Journal of Catalysis 2011, 281 (1),

154 8. Prestat, E.; Popescu, R.; Blank, H.; Schneider, R.; Gerthsen, D., Coarsening of Pt nanoparticles on amorphous carbon film. Surface Science 2013, 609, Xin, H. L.; Mundy, J. A.; Liu, Z.; Cabezas, R.; Hovden, R.; Kourkoutis, L. F.; Zhang, J.; Subramanian, N. P.; Makharia, R.; Wagner, F. T.; Muller, D. A., Atomic-Resolution Spectroscopic Imaging of Ensembles of Nanocatalyst Particles Across the Life of a Fuel Cell. Nano Letters 2012, 12 (1), Lu, N.; Wang, J.; Xie, S.; Xia, Y.; Kim, M. J., Enhanced shape stability of Pd-Rh core-frame nanocubes at elevated temperature: in situ heating transmission electron microscopy. Chemical Communications 2013, 49 (100), Yalcin, A. O.; Goris, B.; van Dijk - Moes, R. J. A.; Fan, Z.; Erdamar, A. K.; Tichelaar, F. D.; Vlugt, T. J. H.; Van Tendeloo, G.; Bals, S.; Vanmaekelbergh, D.; Zandbergen, H. W.; van Huis, M. A., Heatinduced transformation of CdSe/CdS/ZnS core/multishell quantum dots by Zn diffusion into inner layers. Chemical Communications 2014, 51, Xia, X.; Wang, Y.; Ruditskiy, A.; Xia, Y., 25th Anniversary Article: Galvanic Replacement: A Simple and Versatile Route to Hollow Nanostructures with Tunable and Well-Controlled Properties. Advanced Materials 2013, 25 (44), Wang, W.; Dahl, M.; Yin, Y., Hollow Nanocrystals through the Nanoscale Kirkendall Effect. Chemistry of Materials 2012, 25 (8), Lewis, E. A.; Slater, T. J. A.; Prestat, E.; Macedo, A.; O'Brien, P.; Camargo, P. H. C.; Haigh, S. J., Real-time imaging and elemental mapping of AgAu nanoparticle transformations. Nanoscale 2014, 6 (22),

155 Manuscript: Real-time imaging and elemental mapping of AgAu nanoparticle transformations. 155

156 156

157 157

158 158

159 159

160 160

161 161

162 162

163 Supporting information: Real-time imaging and elemental mapping of AgAu nanoparticle transformations. 163

164 164

165 165

166 166

167 167

168 168

169 169

170 170

171 171

172 Appendix 3. Further discussion of shape evolution of AgAu nanocrystals during oxidation, analysis of HAADF and EDX line scans. Transmission electron microscopy provides two-dimensional projections of threedimensional structures. Although tomographic techniques allow determination of three dimensional structures, either by performing a tilt series, 1-3 or by collecting images of multiple identical particles at different orientations, 4-5 these approaches would be challenging or impossible to apply to the system in question (oxidation of AgAu nanocrystals). The beam sensitivity of the specimen means that structural changes during tilt series acquisition would be likely, while the inhomogeneity of the nanocrystal population rules out a single particle approach. However, careful consideration of the images may be able to shed some light on the likely three dimensional structure of the final oxidised product. The transformation of the particle shown in Appendix 2, Figure 1 is considered here. 6 Thresholding of HAADF STEM images provide a simple means of tracking the area of a particle through a series of images. Figure 1 shows the thresholded outlines of the same particle at the onset and at the end of the beam induced transformation, the projected area of the particle increases during the transformation (from 802 to 966 nm 2 ), however, as the shape of the particle doesn t remain constant during the transformation, areas cannot be simple converted to estimated volumes. Figure 1. Thresholded outlines of the same AgAu nanocrystal (a) before and (b) after beam induced oxidation, projected areas of (a) and (b) are calculated to be 802 and 966 nm 2 respectively. Thresholding and area measurements were performed using ImageJ software. Simple geometric models of the particles are considered to assess the likely 3D morphology of the final structure. The initial particle, based on our previous experience of this system, is 172

173 assumed to be approximately spherical. While two idealised models for the final donut shaped projection are proposed: 1. a sphere containing a spherical void and 2. a torus. The volume of a sphere is given by equation 1. V = 4 3 πr3 (1) Where r is the radius of the sphere. The volume of a sphere containing a spherical void can therefore be calculated by finding the difference between the volume of two spheres (the particle and the void). The volume of a torus is given by equation 2. V = (πr 2 )(2πR) (2) Where the r and R are defined by the diagram shown in figure 2. Figure 2. Plan view diagram of a torus illustrating the dimensions r and R (used in equation 2). The radius of the initial particle is measured to be 16.0 ±0.4 nm while the radius of the final particle is measured as 18.2 ±1.5 nm, with the internal void having a radius of 3.6 ±0.4 nm. It is known from lattice resolution imaging and EDX spectrum imaging that the transformation observed is a consequence of Ag(0) being oxidised to Ag 2 O. However, the extent of oxidation which has occurred at the end of our image sequence is unknown. The oxidation of Ag (density g cm -1 ) to Ag 2 O (density 7.14 g cm -1 ) is accompanied by a 46.9% volume increase. Given the particle in question has a composition of 7 at.% Au 93 at.% Ag (based on EDX quantification) and using the densities and atomic masses of the two elements it is calculated that the initial particle should be 93% Ag by volume (this is also apparent when one considers the nearly identical lattice parameters of the two elements). As Au is not oxidised in the reaction, one would expect complete oxidation of Ag to be accompanied by a 43.6% volume increase. Predicted volumes were calculated ranging from oxidation of 100% of the Ag present to oxidation of only 10% of the Ag present. Figure 3 shows a plot of the volumes calculated for various extents of oxidation alongside the volumes calculated for the final structures based upon the torus and spherical void models. 173

174 Figure 3. Calculated volumes for the initial spherical particle provided with its extrapolated volume upon oxidation of Ag to Ag 2 O with oxidations ranging from 10% to 100% of the available Ag considered. These volumes are compared to those calculated for the final structure using a toroidal model and a model of a spherical void within a spherical particle. This analysis proves inconclusive. The fact that the extent of oxidation is unknown means that extrapolation of the initial sphere s volume to a final oxidized volume gives a wide range of possible final volumes. Furthermore, the fact that both the torus and sphere within a sphere model rely on the two initial measurements (radius of internal void and radius of external particle) leads to propagation of errors when estimating the volume, furthermore the projection of the final particle shows considerable deviation from circularity, meaning relatively large errors are associated with the external radius. Consequently, large error bars are associated with both models of the final structure. Predicted volumes (ranging from 10% to 100% oxidation) all lie within the range of feasible volumes calculated using both the toroidal model and the spherical void model. Therefore, it is impossible to confidently assign a final geometry to the oxidised structure from such geometric considerations alone. 174

175 The intensity of the HAADF signal at different points across the particle is another source of information about the particle s 3D shape. Figure 4 shows a series of line scan profiles taken across the same particle as oxidation progresses. The scan of the final structure (Figure 4d) demonstrated that the central void has the same grey value as the support membrane showing that there is no material above or below the void, supporting a toruslike model of the structure. Images at earlier points in the process show that the void initially had material above and/or below it. The shape of the HAADF profile could potentially be used to get further information about the shape of the particle, however, such an interpretation is dramatically complicated by the inhomogeneous and transient composition of the structure in question; due to the large difference in the atomic numbers of Ag, Au, and Ag 2 O (with extent of oxidation unknown) HAADF intensity will be strongly influence by local composition as well as particle thickness. Furthermore, the final image shows that the particle is non-spherically symmetric. It is likely that the final structure is dependent on some initial asymmetry of defects and composition in the original particle. A detailed investigation into the relationship between initial particles structures and compositions and those of the final particles would be an interesting area for further work but is beyond the scope of this thesis. It is likely that such an analysis would require not only a larger data set to account for population inhomogeneity but also the use of quantitative EELS to assess the extent of oxidation and atomic resolution imaging to understand the roles of defects, twinning, and strain

176 Figure 4. HAADF intensity line profiles taken from the same particle at increasing reaction times (a-d) The final structure (d) shows a grey value in the void region equal to those measured on the surrounding support membrane, supporting a torus-like structure. To complement the HAADF line scans, EDX line scans were extracted to reveal changing elemental distributions during oxidation (Figure 5). Measuring the separation between the onsets of Ag counts reveals a steady increase as oxidation occurs, consistent with the outward diffusion of Ag as Ag 2 O is formed at the surface. In contrast, the separation between the onsets of Au counts remains constant, however, the positions of the peaks in Au intensity are observed to move inwards. This supports the inward diffusion of gold during the oxidation, driven by the vacancy gradient created by the outward diffusion of silver. However, despite the inward diffusion of gold, diminished Au concentrations are still found in their original location, meaning the onset is unchanged but the peaks shifts. 176

177 Figure 5. (a-c) EDX spectrum images of the same particle at increasing extents of oxidation, each image maps the distribution of Ag and Au counts. (d and e) show the line scans indicated in (a-c) with (d) showing Au counts and (e) showing Ag counts, in both cases the line scan data has been smoothed using 25 point adjacent averaging. Maxima and minima of the lines shown in (d and e) were determined using Origin software and used to calculate the separation between the onsets for both Ag and Au counts, as well as the separation between the peaks in Au intensity, these values are plotted in (f). Self-diffusion in silver-gold has previously been studied in bulk specimens at a range of compositions and temperatures. 8 The diffusion coefficients in ~8 at.% Au samples at ~710 C are 5.25x10-11 and 1.78x10-11 cm 2 /s for Ag and Au respectively, unfortunately data for lower temperature is not available. 8 For a given composition the diffusion behaviour of the AgAu system obeys the Arrhenius equation (equation 3). D = D 0 exp ( Q RT ) (3) Where D is the diffusion coefficient, D 0 is the frequency factor, and Q is the activation energy. For the 8 at.% Au system, the activation energy for Ag is found to be kcal/mol, while that for Au is found to be kcal/mole. 8 The slower diffusion of Au compared to Ag may in part be responsible for the void formation observed, as inward diffusion of Au is slower than outward diffusion of Ag, leading to increased vacancy concentrations in the particle s core. However, we should be cautious in applying bulk 177

178 measurements to nanoscale specimens as phenomena such as strain in nanocrystals and surface diffusion facilitated by nanoscale voids can have a dramatic influence on diffusion at the nanoscale. 7, 9 Interestingly the diffusion coefficients of both species show a strong compositional dependence in this system, with addition of silver leading to lower diffusion coefficients, 8 this could represent an interesting avenue for future studies on AgAu nanocrystals. References 1. Midgley, P. A.; Weyland, M., 3D electron microscopy in the physical sciences: the development of Z-contrast and EFTEM tomography. Ultramicroscopy 2003, 96 (3-4), Midgley, P. A.; Ward, E. P. W.; Hungria, A. B.; Thomas, J. M., Nanotomography in the chemical, biological and materials sciences. Chemical Society Reviews 2007, 36 (9), Slater, T. J. A.; Macedo, A.; Schroeder, S. L. M.; Burke, M. G.; O Brien, P.; Camargo, P. H. C.; Haigh, S. J., Correlating Catalytic Activity of Ag Au Nanoparticles with 3D Compositional Variations. Nano Letters 2014, 14 (4), Frank, J., Single-Particle Imaging of Macromolecules by Cryo-Electron Microscopy. Annual Review of Biophysics and Biomolecular Structure 2002, 31 (1), Park, J.; Elmlund, H.; Ercius, P.; Yuk, J. M.; Limmer, D. T.; Chen, Q.; Kim, K.; Han, S. H.; Weitz, D. A.; Zettl, A.; Alivisatos, A. P., 3D structure of individual nanocrystals in solution by electron microscopy. Science 2015, 349 (6245), Lewis, E. A.; Slater, T. J. A.; Prestat, E.; Macedo, A.; O'Brien, P.; Camargo, P. H. C.; Haigh, S. J., Real-time imaging and elemental mapping of AgAu nanoparticle transformations. Nanoscale 2014, 6 (22), Pratt, A.; Lari, L.; Hovorka, O.; Shah, A.; Woffinden, C.; Tear, S. P.; Binns, C.; Kröger, R., Enhanced oxidation of nanoparticles through strain-mediated ionic transport. Nature Materials 2014, 13 (1), Mallard, W. C.; Bass, R. F.; Slifkin, L. M.; Gardner, A. B., Self-diffusion in silver-gold soild solutions. Physical Review 1963, 129 (2), Fan, H. J.; Knez, M.; Scholz, R.; Hesse, D.; Nielsch, K.; Zacharias, M.; Gosele, U., Influence of surface diffusion on the formation of hollow nanostructures induced by the Kirkendall effect: The basic concept. Nano Letters 2007, 7 (4),

179 Appendix 4. In Situ Synthesis of PbS Nanocrystals in Polymer Thin Films from Lead(II) Xanthate and Dithiocarbamate Complexes: Evidence for Size and Morphology Control. Introduction Single source precursors are molecules containing all the constituent elements of a target material. They have been used to produce nanoparticles and thin films via a wide variety of techniques, such as hot-injection synthesis and aerosol-assisted chemical vapour deposition (AACVD). Metal dithiocarbamate and metal xanthate complexes are amongst the most commonly used single source precursors for the synthesis of metal sulphide materials. By dispersing these molecular precursors in a polymer matrix it is possible to grow semiconducting nanocrystals directly within the polymer. 1-3 This elegant approach to the fabrication of polymer-quantum dot composites has attracted interest as a route towards producing improved hybrid solar cell materials, as it eliminates problems associated with insulating ligands, agglomeration, and low-loading of nanocrystals (see section 2.5.2). 4 However, despite these advantages, the growth of nanocrystals in a polymer matrix remains poorly understood and the mechanistic toolkit required for precise morphological control has yet to be developed. 5 As morphology is critical to the performance of hybrid devices, 6-7 developing methods for controlling the morphology of the resulting nanostructures is vital if this approach is to realise its potential. The paper presented in this appendix demonstrates the first report of growth of PbS nanocrystals from single source precursors in a polymer matrix. Not only are these the first reports of PbS growth by this methods but the work gives important insights into new methods for achieving morphological control in the resulting composite materials. I have shown for the first time that the structure of the molecular precursor can have a dramatic influence on the morphology of the nanocrystals produced, with the long chain octyl xanthates identified as an especially promising precursor system. The manuscript presented in this appendix was published in Chemistry of Materials in The published manuscript and supporting information are available online and can be found using the digital object identifier DOI: /cm504765z. I designed the experiments and wrote the manuscript. I acquired and analysed all the (S)TEM results and XRD patterns, SEM images were acquired by both myself and Yiqiang 179

180 Chen a. I synthesised and characterised the single source precursors: lead(ii) diethyldithiocarbamate, lead(ii) dibutyldithiocarbamate, lead(ii) diethyldithiocarbamate 1,10-phenanthroline adduct, and lead(ii) dibutyldithiocarbamate 1,10-phenanthroline adduct. Single crystals of the lead(ii) dibutyldithiocarbamate 1,10-phenanthroline adduct were grown by myself and their crystal structure was determined by James Raftery b. Lead(II) ethylxanthate, lead(ii) n-butylxanthate, lead(ii) n-hexylxanthate, and lead(ii) n- octylxanthate were synthesised by Paul McNaughter b, Jack Brent a, and Selina Saah c. The fabrication and heating of thin films was done by myself, Paul McNaughter, and Zhongjie Yin a. Paul McNaughter, Sarah Haigh a, and Paul O Brien a,b all provided useful discussion and were actively involved in the preparation and editing of the manuscript. Affiliations a School of Materials, University of Manchester, M13 9PL, UK. b School of Chemistry, University of Manchester, M13 9PL, UK. c Kwame Nkrumah University of Science and Technology, Kumasi, Ghana. References 1. Leventis, H. C.; King, S. P.; Sudlow, A.; Hill, M. S.; Molloy, K. C.; Haque, S. A., Nanostructured Hybrid Polymer Inorganic Solar Cell Active Layers Formed by Controllable in Situ Growth of Semiconducting Sulfide Networks. Nano Letters 2010, 10 (4), Dowland, S.; Lutz, T.; Ward, A.; King, S. P.; Sudlow, A.; Hill, M. S.; Molloy, K. C.; Haque, S. A., Direct Growth of Metal Sulfide Nanoparticle Networks in Solid-State Polymer Films for Hybrid Inorganic Organic Solar Cells. Advanced Materials 2011, 23 (24), Rath, T.; Edler, M.; Haas, W.; Fischereder, A.; Moscher, S.; Schenk, A.; Trattnig, R.; Sezen, M.; Mauthner, G.; Pein, A.; Meischler, D.; Bartl, K.; Saf, R.; Bansal, N.; Haque, S. A.; Hofer, F.; List, E. J. W.; Trimmel, G., A Direct Route Towards Polymer/Copper Indium Sulfide Nanocomposite Solar Cells. Advanced Energy Materials 2011, 1 (6), Reynolds, L. X.; Lutz, T.; Dowland, S.; MacLachlan, A.; King, S.; Haque, S. A., Charge photogeneration in hybrid solar cells: A comparison between quantum dots and in situ grown CdS. Nanoscale 2012, 4 (5), MacLachlan, A. J.; Rath, T.; Cappel, U. B.; Dowland, S. A.; Amenitsch, H.; Knall, A.-C.; Buchmaier, C.; Trimmel, G.; Nelson, J.; Haque, S. A., Polymer/Nanocrystal Hybrid Solar Cells: Influence of Molecular Precursor Design on Film Nanomorphology, Charge Generation and Device Performance. Advanced Functional Materials 2015, 25 (3),

181 6. Liu, Z.; Sun, Y.; Yuan, J.; Wei, H.; Huang, X.; Han, L.; Wang, W.; Wang, H.; Ma, W., High- Efficiency Hybrid Solar Cells Based on Polymer/PbSxSe1-x Nanocrystals Benefiting from Vertical Phase Segregation. Advanced Materials 2013, 25 (40), Hindson, J. C.; Saghi, Z.; Hernandez-Garrido, J.-C.; Midgley, P. A.; Greenham, N. C., Morphological Study of Nanoparticle Polymer Solar Cells Using High-Angle Annular Dark-Field Electron Tomography. Nano Letters 2011, 11 (2), Lewis, E. A.; McNaughter, P. D.; Yin, Z.; Chen, Y.; Brent, J. R.; Saah, S. A.; Raftery, J.; Awudza, J. A. M.; Malik, M. A.; O Brien, P.; Haigh, S. J., In Situ Synthesis of PbS Nanocrystals in Polymer Thin Films from Lead(II) Xanthate and Dithiocarbamate Complexes: Evidence for Size and Morphology Control. Chemistry of Materials 2015, 27 (6),

182 Manuscript: In Situ Synthesis of PbS Nanocrystals in Polymer Thin Films from Lead(II) Xanthate and Dithiocarbamate Complexes: Evidence for Size and Morphology Control. 182

183 183

184 184

185 185

186 186

187 187

188 188

189 189

190 190

191 Supporting Information: In Situ Synthesis of PbS Nanocrystals in Polymer Thin Films from Lead(II) Xanthate and Dithiocarbamate 191

192 Complexes: Evidence for Size and Morphology Control. 192

193 193

194 194

195 195

196 196

197 197

198 198

199 199

200 200

201 201

202 202

203 Appendix 5. Direct growth of ultra-narrow PbS nanowires in electrospun polymer nanofibres: a low temperature ligand-free synthesis. Introduction The unpublished manuscript included in this appendix builds upon the work presented in appendix 4. 1 While spin coating was used to make thin films in the previous study, this work looks at the possibility of using electrospinning to produce nanofibres from the polymerprecursor solution. I demonstrate that the lead (II) octylxanthate precursor can be fully decomposed at an unprecedentedly low temperature of 90 C. As this temperature is below the glass transition temperature (T g ) of polystyrene it is possible to conserve the polymer nanofibre morphology whilst converting the precursor to PbS in situ. Interestingly narrow high aspect ratio PbS nanowires form in the polymer fibres, in contrast to the nanocubes which grow in thin film controls. This result is of considerable interest as morphological control is challenging when growing nanocrystals in situ. 1-2 It is postulated that the alignment of the polymer chains which occurs during the electrospinning process may be responsible for the anisotropic crystal growth. 3-6 I designed the experiments and wrote the manuscript. I acquired and analysed all the (S)TEM, SEM and XRD results. The lead(ii) n-octylxanthate was synthesised by Paul McNaughter. a Electrospinning was performed by Benjamin Coverdale b who provided useful guidance on solution formulations and spinning parameters. Teruo Hashimoto b performed ultramicrotome cross sectioning of the nanofibres. Sarah Haigh b and Paul O Brien a,b provided useful discussions and were actively involved in the preparation and editing of the manuscript. Affiliations a) School of Chemistry, University of Manchester, M13 9PL, UK. b) School of Materials, University of Manchester, M13 9PL, UK. References 1. Lewis, E. A.; McNaughter, P. D.; Yin, Z.; Chen, Y.; Brent, J. R.; Saah, S. A.; Raftery, J.; Awudza, J. A. M.; Malik, M. A.; O Brien, P.; Haigh, S. J., In Situ Synthesis of PbS Nanocrystals in Polymer Thin Films from Lead(II) Xanthate and Dithiocarbamate Complexes: Evidence for Size and Morphology Control. Chemistry of Materials 2015, 27 (6),

204 2. MacLachlan, A. J.; Rath, T.; Cappel, U. B.; Dowland, S. A.; Amenitsch, H.; Knall, A.-C.; Buchmaier, C.; Trimmel, G.; Nelson, J.; Haque, S. A., Polymer/Nanocrystal Hybrid Solar Cells: Influence of Molecular Precursor Design on Film Nanomorphology, Charge Generation and Device Performance. Advanced Functional Materials 2015, 25 (3), Bellan, L. M.; Craighead, H. G., Molecular orientation in individual electrospun nanofibers measured via polarized Raman spectroscopy. Polymer 2008, 49 (13 14), Richard-Lacroix, M.; Pellerin, C., Molecular Orientation in Electrospun Fibers: From Mats to Single Fibers. Macromolecules 2013, 46 (24), Richard-Lacroix, M.; Pellerin, C., Orientation and Partial Disentanglement in Individual Electrospun Fibers: Diameter Dependence and Correlation with Mechanical Properties. Macromolecules 2015, 48 (13),

205 Manuscript: Direct growth of ultra-narrow PbS nanowires in electrospun polymer nanofibres: a low temperature ligand-free synthesis. E. A. Lewis, a P. D. McNaughter, b B. Coverdale, a T. Hashimoto, a S. J. Haigh, a P. O Brien. a,b a School of Materials, The University of Manchester, Oxford Road, Manchester, M13 9PL, UK. b School of Chemistry, The University of Manchester, Oxford Road, Manchester, M13 9PL, UK. Abstract: The in situ growth of PbS nanocrystals in electrospun polymer nanofibres is demonstrated, using lead(ii) octyl xanthate as a molecular precursor. Narrow, high aspect ratio, PbS nanowires grow in the fibres, in contrast to the nanocubes which grow in analogous polymer thin films. Size can be controlled through precursor loading, with higher precursor: polymer ratios giving larger nanocrystals. Introduction Hybrid solar cells refer to devices where the active layer contains semiconducting nanocrystals (quantum dots) within a conjugated polymer matrix. 1-4 This technology offers the potential for light-weight flexible photovoltaics with tuneable optical properties, which can be produced via solution processing techniques, allowing low cost large-area fabrication. 1, 5-6 In most cases quantum dots are first synthesised ex situ and then blended with a polymer. 1, 4, 7-9 However, the capping ligands used during synthesis of the quantum dots ex situ has the unwanted effect of hindering interfacial charge transfer and this leads to poor photovoltaic performance. 5, 10 Ligand exchange reactions can partially address this issue but also negatively affect nanocrystal solubility, leading to low loadings and nanocrystal aggregation. 3, 5, 11 In situ methods have recently emerged as an alternative synthesis route; by growing quantum dots directly in a polymer film from molecular precursors it is possible the achieve high nanocrystal loadings, ligand free interfaces, and interconnected nanocrystal networks. 3, 10, Furthermore, this approach is inherently simpler and more readily scaled than ex situ methods. However, compared to the ex situ approach, in situ synthesis routes have generally shown poor control of nanocrystal size 12, 18 and shape, limiting device performance. An extensive range of single source precursors have been developed for chemical vapour deposition, and for solution phase and solventless nanocrystal syntheses The first steps towards in situ morphological control have recently been made thanks to systematic studies of the role of precursor structure. 12, 18 Long chain metal xanthates are emerging as 205

206 promising molecular precursors; they have been shown to lead to finer division of organic and inorganic phases and smaller quantum dot dimensions compared to similar short chain structures. 12, 18 Here we demonstrate the potential role of the polymer matrix in influencing nanocrystal morphology during in situ synthesis, an aspect of growth which, to date, has been largely ignored. 25 In this work we study materials produced via traditional spin coating as well as by electrospinning. The latter is a versatile technique for fabricating polymeric fibres with dimensions ranging from nanometres to microns During electrospinning a voltage is applied between the needle of a syringe filled with a polymer solution and a collector plate The solution is then pumped out of the syringe at a constant rate and the electric field causes the jet of polymer solution to thin and elongate, following a looping trajectory This is accompanied by rapid solvent evaporation such that a mat of polymer fibres builds up on the collector plate The resulting nanofibre mats display high surface areas, biocompatibility, and mechanical flexibility Metallic or semiconducting nanocrystals can be incorporated into polymer fibres, with the resulting composite materials having found application in water treatment, 31 antibacterial materials, 32 sensors, solar cells, 35 and catalysis. 36 Hybrid nanocrystal - electrospun nanofibre materials can be produced by three main routes: decorating nanocrystals onto polymer fibre surfaces in a post electrospinning reaction, 32, 37 adding presynthesised nanocrystals into the electrospinning solution, 34, 38 or adding a precursor to the electrospinning solution and using post processing to convert the precursor to nanocrystals in situ within the fibres Previous attempts at incorporating PbS nanocrystals into polymer fibres have relied on the use of the toxic and flammable gaseous reagent H 2 S. 37, 40 Polymer nanofibres have been decorated with PbS nanocrystals by post electrospinning treatment, first grafting lead dimethylacrylate to the fibre surface and then treating the fibres with H 2 S. 37 Lead sulphide has also been grown in polyvinylpyrrolidone fibres by incorporating lead acetate into the electrospinning solution and then exposing the fibres to H 2 S gas. 40 In this work we employ a lead xanthate single source precursor to grow PbS nanocrystals in situ in electrospun polystyrene fibres. The use of a single source precursor eliminates the need for sulfinating agents such as H 2 S and due to the high solubility of long chain metal xanthates in organic solvents, 18 this chemistry should be compatible with a broad range of polymers. While a cadmium thiolate single source precursor has previously been used to grow CdS nanocrystals in polymer nanofibres, 42 this represents the first investigation of xanthate 206

207 decomposition in polymer fibres. Compared to thiolates, xanthates have the advantage of 13, 42 significantly lower decomposition temperatures and volatile biproducts. Experimental The synthesis and characterisation of the lead(ii)octylxanthate molecular precursor is described in a previous publication. 18 For this work several grams of this complex were synthesised and its purity confirmed by H 1 NMR, melting point, and elemental analysis. In an effort to improve our understanding of the molecule s low temperature decomposition additional thermogravimetric analysis (TGA) experiments were performed by the University of Manchester Microanalytical Laboratory using a Seiko SSC/S200. Stock polymer solutions of 34 mg/ml (for spin coating thin films) and 200 mg/ml (for electrospinning) were made by shaking a vial containing appropriate quantities of 280,000 mw polystyrene (Sigma Aldrich) and chloroform (Sigma Aldrich) at 60 C for several hours. The solutions were then allowed to cool to room temperature and a mass of lead(ii)octylxantate required to achieve the desired PbS loading was added. Glass slides (Menzel Gla zer, Thermo Scientific UK) were cut into 15 x 20 mm rectangles. The glass slides were cleaned by sonication in acetone and dried in air. The glass was then mounted on an Ossila spin coater: 150μL of polymer-precursor solution pipetted onto the substrate, which was spun at 1500 rpm for 30 seconds. A chloroform solution containing 200 mg/ml (20 m/v) polystyrene (280K) and 26mg/ml Pb(II)octylxanthate (~5 wt.% PbS) was prepared for electrospinning. Using a liquid infusion pump and a 10ml syringe, solutions were electrospun through a copper electrode from a blunted 21 gauge needle at a rate of 2.5 ml/h. The applied voltage at the needle was 12 kv and the mandrel potential was -5 kv. The spinning distance was 20 cm. Electrospinning took place for roughly 15 minutes. The resulting cream coloured mat was removed from the mandrel and cut into fragments for heating and analysis. For heat treatment, the polymer thin film or fibre mat was loaded into a quartz tube with one closed end and one end connected to a Schlenk line. The tube was then purged of air and held under a nitrogen atmosphere. A Carbolite MTF tube furnace was heated to 90 C and the quartz tube inserted into the hot furnace so that the specimen was positioned in the centre of the furnace. The specimen was held at 90 C for the desired time before the quartz tube was removed from the furnace and allowed to cool. X-ray diffraction (XRD) was used to monitor the extent of precursor decomposition and confirm the phase of the crystalline products. Grazing incidence X-ray diffraction patterns 207

208 of thin films were acquired using a Bruker D8 Advance diffractometer, using a Cu Kα source with a Göbel mirror optic and Soller slits on the detector side; scans were acquired with a grazing incidence of 3 over a 2θ range of 5 80 with 0.02 steps and 2s per step. Fibre dimensions and mat morphologies were determined by scanning electron microscope (SEM) imaging. A portion of mat was attached to an SEM stub with silver paint and then gold coated to prevent charging using an Edwards S150B sputter coater. Secondary electron imaging was performed using a FEI Quanta 650 FEG-SEM operated at 15 kv. For high resolution transmission electron microscope (TEM) imaging the nanocrystals were extracted from the polymer matrix. This was achieved by dissolving the PbS-polymer hybrid material in chloroform, drop casting the resulting solution onto holey carbon TEM grids, and then cleaning the grids with several subsequent drops of clean chloroform. Cross sectional imaging was used to determine the distribution of nanocrystals in the fibres. Fibre mats were embedded in epoxy resin and cross sectioned samples were prepared by ultramicrotome with the resulting sections floated onto copper TEM support grids. Scanning transmission electron microscope (STEM) and high resolution TEM imaging was performed using an FEI Tecnai F30 microscope operated at 300 kv and a probe side aberration corrected FEI Titan G S/TEM ChemiSTEM TM microscope operated at 200 kv. Lattice resolution high angle annular dark field (HAADF) STEM images were acquired with the Titan using a probe current of ~900 pa, a probe convergence semi angle of 26 mrad, and a HAADF inner semi angle of 48 mrad. Indexing and simulation electron diffraction patterns was performed using Java Electron Microscopy Simulation (JEMS) and web-based electron microscopy application software (Web-EMAPS) Energy dispersive X-ray (EDX) spectrum imaging was performed in the Titan with all EDX detectors turned on and the sample un-tilted. Results and discussion In our previous study we identified the lead(ii) octyl xanthate as an especially attractive precursor, allowing controlled growth of PbS nanocrystals in polymer thin films. 18 Here we show that this precursor can be fully decomposed at far lower temperatures than previously reported. The low processing temperatures mean that full decomposition can be achieved below the glass transition temperature of many polymers opening up exciting possibilities for solution processing and mechanical forming of the polymer, so as to produce patterned films or fibres, prior to nanocrystal synthesis in situ. This is attractive for a number of applications where polymer-nanocrystal hybrid materials require shaping. For example, polymer nanofibre materials containing quantum dots are of interest for light 208

209 emitting fibres, waveguides, and optical sensors. 34,42, Furthermore, patterned polymer films with periodic structures at the optical wavelength scale have been shown to increase interaction with visible light, thereby improving the overall power-conversion efficiency of photovoltaic devices Upon heating polymer thin films containing lead(ii) octylxanthate in a nitrogen atmosphere at 90 C a distinct visual change is observed, from light transparent brown to dark opaque brown, characteristic of the decomposition of the precursor to form PbS nanocrystals. Grazing incidence XRD confirms the formation of pure cubic PbS (galena phase), however, for shorter reaction times the XRD patterns also reveal the presence of unreacted precursor, identified by a collection of distinctive peaks in the 2θ=5-20 range that are characteristic of the precursor molecule. As reaction times are increased the intensity of the precursor peaks diminish with respect to the PbS peaks (figure 1). After 8 hours at 90 C the precursor peaks are no longer detectable, suggesting complete conversion to PbS. Figure 1. Grazing incidence XRD of precursor-polystyrene thin films after heating at 90 C for up to 8 hours, the time required for complete conversion of the precursor to PbS. The reference peak positions for PbS galena are shown. 209

210 The reaction temperature of 90 C reported here is significantly lower than those used in previous reports of in situ nanocrystal growth. Temperatures between 140 and 275 C have previously been used for in situ decomposition of metal xanthates. 3, 13-14, For the envisioned photovoltaic applications of these materials low temperature processing is attractive, due to compatibility with plastic substrates, allowing roll-to-roll fabrication of light-weight flexible solar cells, 13, and the prevention of heat induced degradation of the polymer. 13 The fact that full decomposition of the precursor can be achieved at such low temperatures is initially surprising as previous TGA results suggested that the thermal decomposition of the precursor does not commence until ~130 C. 18 This discrepancy can be explained by repeating the TGA measurement with a slower heating rate: SI figure 1 shows a significant decrease in the apparent decomposition onset when the heating rate is decreased from 10 C / min to 1 C, demonstrating that low temperature decomposition does occur at lower temperatures but is too slow to be detected by TGA when using fast heating rates (10 C / min). These findings are consistent with the relatively long reaction times required at 90 C (8 hours for full decomposition compared to 30 minutes at 150 C). 18 The dramatically lowered processing temperature means that the reaction can be carried out below the glass transition temperature (T g ) of the polymer matrix (polystyrene T g =100 C). To demonstrate the potential of sub-t g processing we produced polymer-precursor nanofibres by electrospinning, using precursor loadings sufficient to achieve ~5 wt. % PbS. Heating these mats under N 2 at 90 C resulted in a colour change from cream to brown, indicating that PbS nanocrystals had been formed in situ. Figure 2 shows SEM images of the nanofibre mats before and after heating at 90 C for 8 hours, the dimensions of the fibres and three-dimensional morphology of the mats appear unchanged by the heating process, demonstrating that it is possible for the polymer matrix to retain its morphology while PbS nanocrystals are formed in situ. The polymer fibres are found to have an average diameter of 1.8± 0.5 µm (SI figure 2). 210

211 Figure 2. SEM images of polymer fibre mats produced by electrospinning before (a,b) and after (c,d) heating at 90 C for 8 hours. HAADF STEM imaging reveals that the morphology of the PbS nanocrystals grown in the polymer nanofibres is dominated by high aspect ratio (>10) nanowires with a mean diameter of 3.4 ± 1.8 nm (Figure 3). Atomic resolution HAADF STEM images (Figure 3b-d), electron diffraction (SI figure 3), and EDX spectrum imaging (SI figure 4) shows that these wires are crystalline PbS. Only a small proportion (<4 %) of the nanocrystals formed have the cubic morphology more commonly observed for PbS (SI figure 5). The wires are found to be elongated in the <110> direction while the cubes are found to be terminated by {100} faces (figure 3 and SI figure 6). The <110> direction of nanowire growth is consistent with 18, previous examples of PbS nanowires grown from long chain lead xanthate precursors. Although a range of nanowire diameters are present in the samples (SI figure 7), the wires are often observed to stack with similarly sized wires to form bundles (Figure 3c). When wires of significantly different widths are found together, it is noted that the narrower wires appear brighter than their wider neighbours in the HAADF STEM images, which as the HAADF signal intensity is proportional to atomic number, suggests that these are thicker in the direction of the electron beam (figure 4). Atomic resolution analysis reveals that both have the same <110> growth direction but the bright narrower wires are oriented parallel to [1-10] while the broader wires are oriented along the [001] zone axis (figure 4 and SI figures 8 and 9). The two crystallographic orientations observed represent a 90 rotation suggesting rectangular cross sections. For PbS {100} surfaces are lower in energy than {110} 211

212 surfaces, rectangular cross sections can therefore be explained in terms of surface energy minimisation. 55 Figure 3. HAADF STEM images of PbS nanowires grown in polymer nanofibres at 90 C. (a) Low magnification image showing a large number of typical nanowires. (b,c) High resolution images show that the wires are highly crystalline PbS. A region of (c) is enlarged in (d) for clarity and (e) shows the Fourier transform of (c), all nanowires imaged were found to be elongated in the [110] direction. 212

213 Figure 4. Evidence for rectangular cross section of nanowires (a) HAADF contrast shows that the narrower wire is thicker than the broader wires that surround it. Regions of the narrow wire (green out line) and broad wire (red outline) are enlarged to clearly show their lattice resolution structure. (b) Fourier transform of (a) showing that PbS crystals with two different orientations are present, the spots circled in red are identified as arising from the broad wire, while those circled in green arise from the narrow wire. (c) shows a simulated electron diffraction patterns for [1-10] and [001] oriented PbS. We have performed control experiments in order to understand whether the formation of narrow nanowires, not seen in our previous study, 18 is a consequence of the different heating conditions or due to the influence of the fibre morphology of the polymer matrix. Polymer-precursor thin films with identical precursor loadings were prepared by a traditional spin coating route and after heating these films with identical conditions to those used for the nanofibre mats (90 C for 8 hours) these produced only PbS nanocubes (figure 5a). This demonstrates that the morphology of nanocrystals grown in situ can be 213

214 controlled by appropriate shaping of the polymer matrix. Figure 5. (a) Nanocubes synthesised in a film with ~5 wt. % PbS loading, average diameter= 10.3 ± 1.9. (b) Nanocubes synthesised in a film with ~30 wt. % PbS loading, average diameter= 17.7 ± 2.8 nm. (c) Nanocubes synthesised in a film with ~70 wt. % average diameter= 25.1 ± 4.9 nm. The diameter of 100 representative nanocubes form each samples were measured. High resolution images can be found in SI figure 10. The nanocrystals produced in the control experiments are themselves interesting for two reasons, firstly the small size of the nanocubes produced (average diameter ~10 nm as shown in figure 5a), and secondly the complete absence of any nanowires. In previous work employing the same lead(ii) octyl xanthate precursor, but different precursor loadings and temperatures, average nanocube dimensions of 38 ± 7 nm were achieved and a mixture of cubes and wires (average wire diameter 18 ± 5 nm) were observed. 18 Several reports of in situ synthesis using thiolate precursors suggest that higher precursor : polymer ratios yield larger nanocrystals However, the time and temperature of heating is also observed to affect particle size, with studies of in situ CdS growth in polymers showing larger particles 25, with higher temperatures or longer annealing times. To identify to origin of the small particle size, films were prepared with higher precursor loadings. It was shown that increasing the precursor : polymer ratio leads to an increase in cube dimensions (figure 5a-c), confirming that xanthate loading can be used to control nanocrystal size in situ. In each case the population of nanocubes produced was remarkably uniform, with no nanowires observed to be present in any sample. This is encouraging as it 214

215 suggests that lower reaction temperatures favour more reproducible and consistent nanocrystal growth. Cross-sectional imaging of the heated polymer nanofibres (Figure 6) is also encouraging as it shows a uniform distribution of the PbS nanocrystals within the polymer matrix. The ability to achieve homogeneous nanocrystal distribution is an 42, 48 important advantage of in situ routes compared to ex situ processing routes. The mechanism by which the polymer matrix could influence the morphology of nanocrystals grown in situ requires some discussion. It could be postulated that nanocrystals grown directly within a 1D polymer host would be anisotropic due to confinement effects, such nanoreactor based approaches have been used previously, for example to grow inorganic nanowires inside carbon nanotubes However, the large diameters of the polymer fibres (1.8 ± 0.5 μm) compared to the PbS nanobelts (mean diameter 3.4 ± 1.8 nm) appears to rule out this mechanism for the nanowires produced in this work. A second mechanism could be related to the orientation of the polymer chains within the nanofibres. There is a growing understanding that electrospinning effects the conformation of polymer chains, with the electric field causing alignment and disentanglement, which is retained in the final fibre due to the rapid solvent evaporation. 28, Molecular orientation in polymer nanofibres has often been overlooked due to the random orientation of nanofibres in the nanofibre mats, which means that the anisotropic properties of individual fibres are not apparent in standard bulk characterisation techniques. 28 However, studies of aligned fibres or individual fibres reveal that molecular orientation and disentanglement are common in electrospun fibres, and are likely to be responsible for their size 28, 64 dependent thermal, mechanical, and electronic properties. A number of studies have demonstrated that molecular orientation and disentanglement occurs during electrospinning of polystyrene nanofibres Both processes increase exponentially with decreasing fibre diameter, with the onset diameter of both phenomena (~2.5 µm) suggesting these should be present in the fibres produced in this work. 64 Orientation effects within the polymer matrix could be responsible for the differences in morphology between nanocrystals grown in spin coated thin films and in electrospun fibres. Specifically, molecular orientation could lead to local confinement of the growing crystals or to differences in diffusivities parallel and perpendicular to the wire which may be responsible for the growth of highly anisotropic nanowire structures.. No obvious preferential orientation of the nanowires parallel to the long axis of the fibre is revealed in the cross sectional images (figure 6). However, variations in nanowire length complicate 215

216 interpretation of these images and the beam sensitivity of the polystyrene matrix prevents a deeper three dimensional analysis of the structure via electron tomographic imaging. Despite uncertainty about the mechanism, the fact that electrospun fibres promote anisotropic growth of nanocrystals in situ is an exciting results with wide ranging potential applications; ranging from hybrid photovoltaics, where nanowires improve charge extraction, 1, to wearable electronics, where the incorporation of piezoelectric nanowires into fibres could be used to turn the wearer s motion into electrical power. 70 Figure 6. Cross sectional HAADF STEM images of the same heated polymer nanofibre at several different magnifications, red boxes indicate the regions shown in higher magnification images. PbS nanoparticles are distributed uniformly within the polymer fibre. The PbS nanowires show no clear orientation preference. Further cross sectional images can be found in SI figure 11. Conclusion In this work we have demonstrated the formation of nanocrystals of a semiconductor within intact electrospun polymer nanofibres. The low processing temperature used opens up the possibility of carrying out such reactions in electrically active polymers. Perhaps even more importantly the morphology of the in situ prepared material is profoundly affected by the matrix. The ability to grow anisotropic morphologies and control nanocrystal size in situ should improve the performance of the resulting hybrid materials. 216

217 References 1. Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P., Hybrid nanorod-polymer solar cells. Science 2002, 295 (5564), Gur, I.; Fromer, N. A.; Chen, C.-P.; Kanaras, A. G.; Alivisatos, A. P., Hybrid Solar Cells with Prescribed Nanoscale Morphologies Based on Hyperbranched Semiconductor Nanocrystals. Nano Letters 2006, 7 (2), Dowland, S.; Lutz, T.; Ward, A.; King, S. P.; Sudlow, A.; Hill, M. S.; Molloy, K. C.; Haque, S. A., Direct Growth of Metal Sulfide Nanoparticle Networks in Solid-State Polymer Films for Hybrid Inorganic Organic Solar Cells. Advanced Materials 2011, 23 (24), Liu, Z.; Sun, Y.; Yuan, J.; Wei, H.; Huang, X.; Han, L.; Wang, W.; Wang, H.; Ma, W., High- Efficiency Hybrid Solar Cells Based on Polymer/PbSxSe1-x Nanocrystals Benefiting from Vertical Phase Segregation. Advanced Materials 2013, 25 (40), Moule, A. J.; Chang, L.; Thambidurai, C.; Vidu, R.; Stroeve, P., Hybrid solar cells: basic principles and the role of ligands. Journal of Materials Chemistry 2012, 22 (6), Gao, F.; Ren, S.; Wang, J., The renaissance of hybrid solar cells: progresses, challenges, and perspectives. Energy & Environmental Science 2013, 6 (7), Kahmann, S.; Mura, A.; Protesescu, L.; Kovalenko, M.; Brabec, C. J.; Loi, M. A., Optoelectronics of PbS quantum dot and narrow bandgap polymer blends. Journal of Materials Chemistry C 2015, 3, Seo, J.; Cho, M. J.; Lee, D.; Cartwright, A. N.; Prasad, P. N., Efficient Heterojunction Photovoltaic Cell Utilizing Nanocomposites of Lead Sulfide Nanocrystals and a Low-Bandgap Polymer. Advanced Materials 2011, 23 (34), Seo, J.; Kim, W. J.; Kim, S. J.; Lee, K.-S.; Cartwright, A. N.; Prasad, P. N., Polymer nanocomposite photovoltaics utilizing CdSe nanocrystals capped with a thermally cleavable solubilizing ligand. Applied Physics Letters 2009, 94 (13), Reynolds, L. X.; Lutz, T.; Dowland, S.; MacLachlan, A.; King, S.; Haque, S. A., Charge photogeneration in hybrid solar cells: A comparison between quantum dots and in situ grown CdS. Nanoscale 2012, 4 (5), Piliego, C.; Manca, M.; Kroon, R.; Yarema, M.; Szendrei, K.; Andersson, M. R.; Heiss, W.; Loi, M. A., Charge separation dynamics in a narrow band gap polymer-pbs nanocrystal blend for efficient hybrid solar cells. Journal of Materials Chemistry 2012, 22 (46), MacLachlan, A. J.; Rath, T.; Cappel, U. B.; Dowland, S. A.; Amenitsch, H.; Knall, A.-C.; Buchmaier, C.; Trimmel, G.; Nelson, J.; Haque, S. A., Polymer/Nanocrystal Hybrid Solar Cells: Influence of Molecular Precursor Design on Film Nanomorphology, Charge Generation and Device Performance. Advanced Functional Materials 2015, 25 (3), Fradler, C.; Rath, T.; Dunst, S.; Letofsky-Papst, I.; Saf, R.; Kunert, B.; Hofer, F.; Resel, R.; Trimmel, G., Flexible polymer/copper indium sulfide hybrid solar cells and modules based on the 217

218 metal xanthate route and low temperature annealing. Solar Energy Materials and Solar Cells 2014, 124 (0), Rath, T.; Edler, M.; Haas, W.; Fischereder, A.; Moscher, S.; Schenk, A.; Trattnig, R.; Sezen, M.; Mauthner, G.; Pein, A.; Meischler, D.; Bartl, K.; Saf, R.; Bansal, N.; Haque, S. A.; Hofer, F.; List, E. J. W.; Trimmel, G., A Direct Route Towards Polymer/Copper Indium Sulfide Nanocomposite Solar Cells. Advanced Energy Materials 2011, 1 (6), Bansal, N.; O'Mahony, F. T. F.; Lutz, T.; Haque, S. A., Solution Processed Polymer Inorganic Semiconductor Solar Cells Employing Sb2S3 as a Light Harvesting and Electron Transporting Material. Adv. Energy Mater. 2013, 3 (8), Leventis, H. C.; King, S. P.; Sudlow, A.; Hill, M. S.; Molloy, K. C.; Haque, S. A., Nanostructured Hybrid Polymer Inorganic Solar Cell Active Layers Formed by Controllable in Situ Growth of Semiconducting Sulfide Networks. Nano Letters 2010, 10 (4), Kaltenhauser, V.; Rath, T.; Haas, W.; Torvisco, A.; Muller, S. K.; Friedel, B.; Kunert, B.; Saf, R.; Hofer, F.; Trimmel, G., Bismuth sulphide-polymer nanocomposites from a highly soluble bismuth xanthate precursor. Journal of Materials Chemistry C 2013, 1 (47), Lewis, E. A.; McNaughter, P. D.; Yin, Z.; Chen, Y.; Brent, J. R.; Saah, S. A.; Raftery, J.; Awudza, J. A. M.; Malik, M. A.; O Brien, P.; Haigh, S. J., In Situ Synthesis of PbS Nanocrystals in Polymer Thin Films from Lead(II) Xanthate and Dithiocarbamate Complexes: Evidence for Size and Morphology Control. Chemistry of Materials 2015, 27 (6), Lazell, M.; O'Brien, P.; Otway, D. J.; Park, J.-H., Single source molecular precursors for the deposition of III/VI chalcogenide semiconductors by MOCVD and related techniques. Dalton Trans. 2000, (24), Obrien, P.; Walsh, J. R.; Watson, I. M.; Hart, L.; Silva, S. R. P., Properties of cadmium sulphide films grown by single-source metalorganic chemical vapour deposition with dithiocarbamate precursors. Journal of Crystal Growth 1996, 167 (1-2), Ramasamy, K.; Malik, M. A.; O'Brien, P., The chemical vapor deposition of Cu2ZnSnS4 thin films. Chemical Science 2011, 2 (6), Lewis, E.; Haigh, S.; O'Brien, P., The synthesis of metallic and semiconducting nanoparticles from reactive melts of precursors. J. Mater. Chem. A 2014, 2 (3), Sreekumari Nair, P.; Radhakrishnan, T.; Revaprasadu, N.; Kolawole, G. A.; O Brien, P., Cd(NH2CSNHNHCSNH2)Cl2: a new single-source precursor for the preparation of CdS nanoparticles. Polyhedron 2003, 22 (23), Trindade, T.; O'Brien, P.; Zhang, X.-m., Synthesis of CdS and CdSe Nanocrystallites Using a Novel Single-Molecule Precursors Approach. Chemistry of Materials 1997, 9 (2), Tang, Y.; Zhang, H.; Fan, Z.; Li, M.; Han, J.; Dong, F.; Yang, B., Polymer-mediated growth of fluorescent semiconductor nanoparticles in preformed nanocomposites. Physical Chemistry Chemical Physics 2010, 12 (38), Li, D.; Xia, Y., Electrospinning of Nanofibers: Reinventing the Wheel? Advanced Materials 2004, 16 (14),

219 27. Greiner, A.; Wendorff, J. H., Electrospinning: A Fascinating Method for the Preparation of Ultrathin Fibers. Angewandte Chemie International Edition 2007, 46 (30), Richard-Lacroix, M.; Pellerin, C., Molecular Orientation in Electrospun Fibers: From Mats to Single Fibers. Macromolecules 2013, 46 (24), Pham, Q. P.; Sharma, U.; Mikos, A. G., Electrospinning of polymeric nanofibers for tissue engineering applications: a review. Tissue engineering 2006, 12 (5), Wang, P.; Wang, Y.; Tong, L., Functionalized polymer nanofibers: a versatile platform for manipulating light at the nanoscale. Light Sci Appl 2013, 2, e Botes, M.; Eugene Cloete, T., The potential of nanofibers and nanobiocides in water purification. Critical reviews in microbiology 2010, 36 (1), Dong, H.; Wang, D.; Sun, G.; Hinestroza, J. P., Assembly of Metal Nanoparticles on Electrospun Nylon 6 Nanofibers by Control of Interfacial Hydrogen-Bonding Interactions. Chemistry of Materials 2008, 20 (21), Ding, B.; Wang, M.; Wang, X.; Yu, J.; Sun, G., Electrospun nanomaterials for ultrasensitive sensors. Materials Today 2010, 13 (11), He, X.; Tan, L.; Wu, X.; Yan, C.; Chen, D.; Meng, X.; Tang, F., Electrospun quantum dots/polymer composite porous fibers for turn-on fluorescent detection of lactate dehydrogenase. Journal of Materials Chemistry 2012, 22 (35), Ren, S.; Chang, L.-Y.; Lim, S.-K.; Zhao, J.; Smith, M.; Zhao, N.; Bulović, V.; Bawendi, M.; Gradečak, S., Inorganic Organic Hybrid Solar Cell: Bridging Quantum Dots to Conjugated Polymer Nanowires. Nano Letters 2011, 11 (9), Fang, X.; Ma, H.; Xiao, S.; Shen, M.; Guo, R.; Cao, X.; Shi, X., Facile immobilization of gold nanoparticles into electrospun polyethyleneimine/polyvinyl alcohol nanofibers for catalytic applications. Journal of Materials Chemistry 2011, 21 (12), Ye, J.; Chen, Y.; Zhou, W.; Wang, X.; Guo, Z.; Hu, Y., Preparation of Polymer@PbS hybrid nanofibers by surface-initiated atom transfer radical polymerization and acidolysis by H2S. Materials Letters 2009, 63 (16), Schlecht, S.; Tan, S.; Yosef, M.; Dersch, R.; Wendorff, J. H.; Jia, Z.; Schaper, A., Toward Linear Arrays of Quantum Dots via Polymer Nanofibers and Nanorods. Chemistry of Materials 2005, 17 (4), Xu, J.; Cui, X.; Zhang, J.; Liang, H.; Wang, H.; Li, J., Preparation of CuS nanoparticles embedded in poly(vinyl alcohol) nanofibre via electrospinning. Bulletin of Materials Science 2008, 31 (2), Lu, X.; Zhao, Y.; Wang, C., Fabrication of PbS Nanoparticles in Polymer-Fiber Matrices by Electrospinning. Advanced Materials 2005, 17 (20), Zhang, J.; Maurer, F. H. J.; Yang, M., In situ Formation of TiO2 in Electrospun Poly(methyl methacrylate) Nanohybrids. The Journal of Physical Chemistry C 2011, 115 (21),

220 42. Di Benedetto, F.; Camposeo, A.; Persano, L.; Laera, A. M.; Piscopiello, E.; Cingolani, R.; Tapfer, L.; Pisignano, D., Light-emitting nanocomposite CdS-polymer electrospun fibres via in situ nanoparticle generation. Nanoscale 2011, 3 (10), Zhang, C.-L.; Yu, S.-H., Nanoparticles meet electrospinning: recent advances and future prospects. Chemical Society Reviews 2014, 43 (13), Zuo, J. M.; Mabon, J. C., Web-Based Electron Microscopy Application Software: Web- EMAPS. Microscopy and Microanalysis 2004, 10 (SupplementS02), Stadelmann, P., Image Analysis and Simulation Software in Transmission Electron Microscopy. Microscopy and Microanalysis 2003, 9 (SupplementS03), Meng, C.; Xiao, Y.; Wang, P.; Zhang, L.; Liu, Y.; Tong, L., Quantum-Dot-Doped Polymer Nanofibers for Optical Sensing. Advanced Materials 2011, 23 (33), Zhang, R.; Yu, H.; Li, B., Active nanowaveguides in polymer doped with CdSe-ZnS core-shell quantum dots. Nanoscale 2012, 4 (19), Liu, H.; Edel, J. B.; Bellan, L. M.; Craighead, H. G., Electrospun Polymer Nanofibers as Subwavelength Optical Waveguides Incorporating Quantum Dots. Small 2006, 2 (4), Cocoyer, C.; Rocha, L.; Sicot, L.; Geffroy, B.; de Bettignies, R.; Sentein, C.; Fiorini- Debuisschert, C.; Raimond, P., Implementation of submicrometric periodic surface structures toward improvement of organic-solar-cell performances. Applied Physics Letters 2006, 88 (13), Na, S.-I.; Kim, S.-S.; Jo, J.; Oh, S.-H.; Kim, J.; Kim, D.-Y., Efficient Polymer Solar Cells with Surface Relief Gratings Fabricated by Simple Soft Lithography. Advanced Functional Materials 2008, 18 (24), Pagliaro, M.; Ciriminna, R.; Palmisano, G., Flexible Solar Cells. ChemSusChem 2008, 1 (11), Akhtar, J.; Afzaal, M.; Vincent, M. A.; Burton, N. A.; Hillier, I. H.; O'Brien, P., Low temperature CVD growth of PbS films on plastic substrates. Chemical Communications 2011, 47 (7), Acharya, S.; Gautam, U. K.; Sasaki, T.; Bando, Y.; Golan, Y.; Ariga, K., Ultra Narrow PbS Nanorods with Intense Fluorescence. Journal of the American Chemical Society 2008, 130 (14), Khan, A. H.; Ji, Q.; Ariga, K.; Thupakula, U.; Acharya, S., Size controlled ultranarrow PbS nanorods: spectroscopy and robust stability. Journal of Materials Chemistry 2011, 21 (15), Fan, D.; Thomas, P. J.; O Brien, P., Pyramidal Lead Sulfide Crystallites With High Energy {113} Facets. Journal of the American Chemical Society 2008, 130 (33), Laera, A.; Resta, V.; Piscopiello, E.; Miceli, V.; Schioppa, M.; Scalone, A.; Benedetto, F.; Tapfer, L., In situ growth of well-dispersed CdS nanocrystals in semiconducting polymers. Nanoscale Research Letters 2013, 8 (1),

221 57. Carotenuto, G.; Martorana, B.; Perlo, P.; Nicolais, L., A universal method for the synthesis of metal and metal sulfide clusters embedded in polymer matrices. Journal of Materials Chemistry 2003, 13 (12), Antolini, F.; Burresi, E.; Stroea, L.; Morandi, V.; Ortolani, L.; Accorsi, G.; Blosi, M., Time and Temperature Dependence of CdS Nanoparticles Grown in a Polystyrene Matrix. Journal of Nanomaterials Antolini, F.; Di Luccio, T.; Laera, A. M.; Mirenghi, L.; Piscopiello, E.; Re, M.; Tapfer, L., Direct synthesis of II VI compound nanocrystals in polymer matrix. physica status solidi (b) 2007, 244 (8), Hu, J.; Bando, Y.; Zhan, J.; Zhi, C.; Golberg, D., Carbon Nanotubes as Nanoreactors for Fabrication of Single-Crystalline Mg3N2 Nanowires. Nano Letters 2006, 6 (6), Ersen, O.; Bégin, S.; Houllé, M.; Amadou, J.; Janowska, I.; Grenèche, J.-M.; Crucifix, C.; Pham- Huu, C., Microstructural Investigation of Magnetic CoFe2O4 Nanowires inside Carbon Nanotubes by Electron Tomography. Nano Letters 2008, 8 (4), Kakade, M. V.; Givens, S.; Gardner, K.; Lee, K. H.; Chase, D. B.; Rabolt, J. F., Electric Field Induced Orientation of Polymer Chains in Macroscopically Aligned Electrospun Polymer Nanofibers. Journal of the American Chemical Society 2007, 129 (10), Bellan, L. M.; Craighead, H. G., Molecular orientation in individual electrospun nanofibers measured via polarized Raman spectroscopy. Polymer 2008, 49 (13 14), Richard-Lacroix, M.; Pellerin, C., Orientation and Partial Disentanglement in Individual Electrospun Fibers: Diameter Dependence and Correlation with Mechanical Properties. Macromolecules 2015, 48 (13), Richard-Lacroix, M.; Pellerin, C., Partial Disentanglement in Continuous Polystyrene Electrospun Fibers. Macromolecules 2015, 48 (1), Gururajan, G.; Giller, C. B.; Snively, C. M.; Chase, D. B.; Rabolt, J. F., Molecular Orientation Evolution and Solvent Evaporation During Electrospinning of Atactic Polystyrene Using Real-Time Raman Spectroscopy. Applied Spectroscopy 2011, 65 (8), Liu, J.; Wang, S.; Bian, Z.; Shan, M.; Huang, C., Organic/inorganic hybrid solar cells with vertically oriented ZnO nanowires. Applied Physics Letters 2009, 94 (17), Park, H.; Chang, S.; Jean, J.; Cheng, J. J.; Araujo, P. T.; Wang, M.; Bawendi, M. G.; Dresselhaus, M. S.; Bulović, V.; Kong, J.; Gradečak, S., Graphene Cathode-Based ZnO Nanowire Hybrid Solar Cells. Nano Letters 2013, 13 (1), Ren, S.; Zhao, N.; Crawford, S. C.; Tambe, M.; Bulović, V.; Gradečak, S., Heterojunction Photovoltaics Using GaAs Nanowires and Conjugated Polymers. Nano Letters 2011, 11 (2), Qin, Y.; Wang, X.; Wang, Z. L., Microfibre-nanowire hybrid structure for energy scavenging. Nature 2008, 451 (7180),

222 Supporting information: Direct growth of ultra-narrow PbS nanowires in electrospun polymer nanofibres: a low temperature ligand-free synthesis. SI Figure 1. Thermogravimetric analysis of lead(ii)octylxanthate performed at two different heating rates: 1 C/ min (red) and 10 C/ min (blue). The apparent onset of decomposition is lower at the slower heating rate, suggesting that the precursor can decompose slowly at temperatures lower than previously expected. In both experiments a two-step decomposition is observed. 222

223 SI Figure 2. Additional SEM images of nanofibres heated for 8 hours at 90 C. Red arrows in the upper left image highlight beading. A Histogram, based on a representative sample of 60 fibres, shows the distribution of fibre diameters, the mean diameter is found to be 1.8 ± 0.5 μm. 223

224 SI Figure 3. Low magnification bright field TEM image of nanowires extracted and selected area diffraction pattern from the same region. The ring pattern is characteristic of a polycrystalline sample (as a large number of randomly oriented nanowires are present in the selected area) and indexes to PbS, confirming that the nanowires produced in the polymer fibres are crystalline PbS. SI Figure 4. EDX spectrum imaging of a bundle of nanowires. (a) HAADF STEM image of the region and Pb (b) and S (c) elemental maps. The spectrum image shows clear colocalisation of Pb and S. Cliff Lorimer quantification confirms that the elements are present in close to the expected 1:1 stoichiometry. 224

225 SI Figure 5. Low magnification bright field TEM (a) and HAADF STEM (b-d) images of nanowires extracted from nanofibre mat after heating at 90 C. The PbS nanostructures grown are predominantly narrow, high aspect ratio, PbS nanowires (a-c). However, some large nanocubes are also found (d). Based on analysis of typical regions of the grid we estimate that less than 4% of nanocrystals formed are nanocubes. 225

226 SI Figure 6. High resolution HAADF STEM images and corresponding Fourier transforms of PbS nanocrystals grown in polymer fibres. (a-d) show nanowires which are all found to be elongated in the [110] direction. (e-f) shows a PbS nanocube with {100} faces. 226

227 SI Figure 7. Histograms showing size distributions of nanowire width and length, measurements were made from HAADF STEM images using representative populations of 100 nanowires. The average width is 3.4 ± 1.8 nm and the average length is 46.2 ± 20.1 nm. SI Figure 8. Data form figure 4 presented with additional Fourier transform analysis. (b) shows the Fourier transform of the entire image in (a), (c) is taken from broad wire on the left of (a), while (d) is taken is taken form the bright narrow wire on the right of (a). 227

228 SI Figure 9. Another example of the coexistence of narrow and broad wires, as in figure 4 (and SI figure 7) the broader wire (top of the image) is viewed down the [001] zone axis, while the narrow wires (below it) are viewed down the [110] axis. SI Figure 10. High resolution HAADF STEM images of PbS nanocubes grown in polymer thin films with (a) 5 wt. % PbS and (b) 30 wt. % PbS. Fourier transforms and enlarged regions of the atomic resolution image are also shown. In both cases the cubes are single crystalline PbS with {100} faces. 228

229 SI Figure 11. HAADF STEM images of cross sectioned nanofibres. (a) shows 3 cross sectioned fibres, the fibre in the top right hand corner is shown in higher magnification in (b) and the area indicated by the red box in (b) is shown in (c). 229