Compositional mapping of semiconductor quantum dots by X-ray photoemission electron microscopy

Transcription

1 Compositional mapping of semiconductor quantum dots by X-ray photoemission electron microscopy Stefan Heun CNR-INFM, Italy, Laboratorio Nazionale TASC, Trieste and NEST-SNS, Pisa

2 Outline A brief introduction to x-ray photoemission electron microscopy Compositional mapping of semiconductor quantum dots Motivation InAs / GaAs Ge / Si Ge / Au / Si

3 Why XPS? XPS = X-ray Photoelectron Spectroscopy chemical state information G. Mori, S. Heun et al.: NIM B 246 (2006) 39.

4 Why XPS? chemical state information surface sensitive S. Hüfner et al.: NIM A 547 (2005) 8.

5 Why XPS? chemical state information surface sensitive ease of quantification (in general) nondestructive

6 Why spectromicroscopy? semicond. nanostructures: selforganized islands (dots) S. Heun et al.: Phys. Rev. B 63 (2001)

7 Why spectromicroscopy? semicond. nanostructures: selforganized islands (dots) surface faceting (wires) F.-J. Meyer zu Heringdorf, S. Heun et al.: Phys. Rev. Lett. 86 (2001) 5088.

8 Why spectromicroscopy? semicond. nanostructures: selforganized islands (dots) surface faceting (wires) carbon nanotubes S. Suzuki, S. Heun et al.: Appl. Phys. Lett. 85 (2004) 127.

9 Why spectromicroscopy? semicond. nanostructures: selforganized islands (dots) surface faceting (wires) carbon nanotubes catalysis, chemical waves A. Locatelli, S. Heun et al.: J. Am. Chem. Soc. 127 (2005) 2351.

10 Why spectromicroscopy? semicond. nanostructures: selforganized islands (dots) surface faceting (wires) carbon nanotubes catalysis, chemical waves surface magnetism (XMCD) A. M. Mulders, S. Heun et al.: Phys. Rev. B 71 (2005)

11 The SPELEEM at ELETTRA Best energy resolution: 250 mev Best lateral resolution: 25 nm Variable polarization ev Photon flux ph/s Small spot (2µm x 25µm)

12 The SPELEEM instrument Spectroscopic Photo-Emission and Low Energy Electron Microscope Monochromatic images

13 XPEEM: Core Level Spectroscopy Pb/W(110), Pb 5d core level, hv = 80 ev Best energy resolution: 250 mev

14 Outline A brief introduction to x-ray photoemission electron microscopy Compositional mapping of semiconductor quantum dots Motivation InAs / GaAs Ge / Si Ge / Au / Si

15 Motivation Quantum Dot Applications based on their particular electronic properties (confinement) Strain-driven self-assembly (SK-growth) Model systems: InAs/GaAs, Ge/Si Intermixing and alloying allow for partial strain relaxation Control of composition of individual QD, which determines their physical properties Controlled positioning of QDs on a suitable substrate

16 Concentration Profiles Concentration maps in cross-section: TEM Walther et al., PRL 86 (2001) 2381

17 Concentration Profiles Concentration maps in cross-section: TEM, STM Liu et al., PRL 84 (2000) 334

18 Concentration Profiles Concentration maps in cross-section: TEM, STM, XRD Kegel et al., PRL 85 (2000) 1694

19 Concentration Profiles Concentration maps in cross-section: TEM, STM, XRD in top-view: etching (Ge > 65%) Complementary views Full 3D mapping Denker et al., PRL 90 (2003)

20 Concentration Profiles Concentration maps in cross-section: TEM, STM, XRD in top-view: etching, XRM F. Ratto, S. Heun et al.: Small 2 (2006) 401.

21 Outline A brief introduction to x-ray photoemission electron microscopy Compositional mapping of semiconductor quantum dots Motivation InAs / GaAs Ge / Si Ge / Au / Si

22 InAs/GaAs Islands (LEEM) 1 µm Electron Microscopy (LEEM) 5 µm FOV E kin = 7.6 ev 500nm S. Heun et al.: J. Nanosci. Nanotech., in press.

23 Integral Core Level Spectra hv = 99.0 ev Spectra taken from a 1 µm x 1 µm sample area. G. Biasiol, S. Heun et al.: Appl. Phys. Lett. 87 (2005)

24 XPEEM Core Level Imaging hv (a) 500nm (b) 500nm In 4d XPEEM image hv = 99.0 ev, E kin = ev Ga 3d XPEEM image hv = 99.0 ev, E kin = ev G. Biasiol, S. Heun et al.: Appl. Phys. Lett. 87 (2005)

25 XPEEM Local Spectra Integration area 25 nm x 25 nm, energy resolution 1 ev G. Biasiol, S. Heun et al.: Appl. Phys. Lett. 87 (2005)

26 Core Level Line Profile Analysis hv = 99.0 ev Spectrum from Wetting Layer, Shirley Background subtracted Gauss 1 ev, Lor 0.16 ev, BR 1.5, SO: Ga 3d 0.45 ev, In 4d 0.85 ev G. Biasiol, S. Heun et al.: Appl. Phys. Lett. 87 (2005)

27 2D Fit of XPEEM Data 500nm In 4d peak area Min: 220, Max: nm Ga 3d peak area Min: 270, Max: 470 Ratio of Number of Atoms: n n In Ga = I I In Ga σ σ Ga In G. Biasiol, S. Heun et al.: Appl. Phys. Lett. 87 (2005)

28 Indium Surface Concentration Map n n In tot = I In I σ Ga In σ + I Ga Ga σ In nm 0.76 G. Biasiol, S. Heun et al.: Appl. Phys. Lett. 87 (2005)

29 Wetting Layer Composition Segregation models predict the following In concentration profile: Measured composition is average across topmost layers: d i d i λ λ < x >= xie e Shown profile would be measured as x ~ 0.75, in agreement with our data. O. Dehaese et al.: Appl. Phys. Lett. 66 (1995) 52.

30 Dot Composition from TEM At surface: x ~ 0.6 Max of x ~ 0.8 at 10 ML below the surface We would measure this profile as x ~ 0.65 Our data: x ~ 0.85 A. Rosenauer et al.: Phys. Rev. B 64 (2001)

31 Indium depth concentration profiles Strong In segregation also on surface of dots. Add double layer with x ~ (like WL) to surface. We would measure this profile as x ~ 0.85, in good agreement with our data. S. Heun et al.: J. Nanosci. Nanotech., in press.

32 Outline A brief introduction to x-ray photoemission electron microscopy Compositional mapping of semiconductor quantum dots Motivation InAs / GaAs Ge / Si Ge / Au / Si

33 Ge/Si(111) growth by LEEM LEEM Movie FOV 15 µm MBE growth T = 550ºC 3 to 8 ML Ge on Si(111) Has been used to study diffusion dynamics during the nucleation and growth of Ge/Si nanostructures on Si(111) F. Ratto, S. Heun et al.: Phys. Rev. Lett. 96 (2006)

34 Chemical contrast by XPEEM 10 ML Ge on Si(111), T = 560 C, hv = ev island island shadow F. Ratto, S. Heun et al.: Appl. Phys. Lett. 84 (2004) 4526.

35 Comp. mapping of Ge/Si islands Island height: about 25 nm Relative Si surface concentration in a Ge(Si) island on Si(111). The composition mapping is obtained by combining sequences of Si2p and Ge3d XPEEM micrographs with a lateral resolution of ~30 nm. Inset: LEEM image of the same 3D structure (~10 nm lateral resolution). 10 MLs Ge Rate: 0.2 ML/s T = 450 C F. Ratto, S. Heun et al.: Small 2 (2006) 401.

36 Si conc. vs. island morphology Si surface concentration as a function of island base area. At each deposition temperature, the stoichiometry is uniquely determined by the island s lateral dimensions. F. Ratto, S. Heun et al.: J. Appl. Phys. 97 (2005)

37 Outline A brief introduction to x-ray photoemission electron microscopy Compositional mapping of semiconductor quantum dots Motivation InAs / GaAs Ge / Si Ge / Au / Si

38 Metal-induced island ordering Au deposition Stencil mask HF -etched Si(100) Transfer in air to MBE chamber Ge deposition Au (1nm thick) J. T. Robinson et al.: Nano Lett. 5 (2005) 2070.

39 Diffusion barrier model Experiment Kinetic Monte Carlo simulations Objective: Understanding the origin of this diffusion barrier Strategy: Probing surface chemistry and growth dynamics using PEEM and LEEM.

40 As-deposited Au-patterned Si(001) Stencil Mask 200 nm x 200 nm square holes of various pitch LEEM (3 um FOV) Inset: LEED (23.4 ev) shows 1x1 Sharp Au-pattern with 400 nm spacing

41 As-deposited Au-patterned Si(001) Local spectra (hv = 200 ev): Si bulk XPEEM image at Au 4f Au-site Center-site SiO 2 Au-site In addition to Au and Si, another chemical species is present on the surface only at each Au-site, where Si 2p shows two peaks at KE of 92.5 ev (Si bulk) and at 88.5 ev (SiO 2 )

42 As-deposited Au-patterned Si(001) Composite PEEM image of the Si2p and Au4f Au 4f Negligible amount of Au is found inbetween Au sites Good correlation between the Au 4f and SiO 2 maxima Si bulk 1000 nm spacing SiO 2 SiO 2 component extends radially slightly beyond each Au-site

43 Annealed Au-patterned Si(001) LEED (53 ev) after annealing at 600ºC for 30 min. Reconstruction double domain 2x1. No Auinduced reconstruction! Cross-sectional Au 4f intensity profile before and after annealing. Au-sites do not spread during annealing. AFM image of Aupatterned sample after annealing. Several Au-Si islands have formed within each Au-site.

44 Phenomenological Model Si HF Si H-Si(001) Surface Si Au deposition SiO 2 Si Diffusion of Si atoms to the top of Au layers To MBE (air) Si Local oxidation of Si During pre-growth annealing, Au atoms are stabilized by SiO 2 even at substrate temperature higher than 363 o C (Au-Si eutectic point).

45 Ge on Au-patterned Si(001) LEEM images of the surface topography before and after Ge deposition at 450 o C(FOV 3 µm). Ge islands Au pattern Island growth occurs away from the Au-sites at (½, ½) type position of the original pattern.

46 Ge/Au-Si Surface Stoichiometry Highest Ge concentration at Ge islands, Gedepleted zone around original Au-patches Au distribution is similar to Ge: 3D islands are rich in Au Si-rich zone around original Au-sites

47 Phenomenological Model Ge deposition Si Small islands grow directly around Au-sites Etching with hydrogen peroxide (H 2 O 2 ) which removes Ge x Si 1-x for x > 0.65 Islands are Ge-rich Growth occurs only by direct impingement

48 Phenomenological Model Ge deposition Ge & Au Si XPEEM confirms the existence of Ge-denuded zone around Au-sites. Physical origin might be from oxide, which would likely produce compressive strain and promote Ge diffusion away from here. Si Ge wets both SiO 2 and Si. Ge atoms cover the oxide around the original Au-sites. Au can diffuse away. Si Ge/Au-Si island Ge-depleted region

49 Conclusions Surface concentration maps of quantum dot systems by photoemission microscopy. Dot composition neither pure InAs (Ge) nor homogeneous In x Ga 1-x As (SiGe). In (Ge) concentration decreases from center (high) to borders (low) of dots. In segregation (x ~ 0.9) on surface of InAs dots and WL. The Si content in Ge/Si dots increases with island lateral dimensions and deposition temperature. The interplay between Au-induced Si local oxidation and SiO 2 -enhanced Au stability plays an important role in the ordering process of Ge islands on Au-patterned Si.

50 Coworkers InAs / GaAs: G. Biasiol, G. B. Golinelli, L. Sorba, TASC, Trieste, Italy F. Z. Guo, SPring-8, Japan A. Locatelli, T. O. Mentes, Sincrotrone Trieste, Italy Ge / Si: F. Ratto, F. Rosei, University of Quebec, Canada S. Cherifi, S. Fontana, A. Locatelli, Sincrotrone Trieste, Italy Ge / Au / Si J. T. Robinson, O. D. Dubon, University of California, Berkeley O. Moutanabbir, Keio University, Japan F. Ratto, F. Rosei, INRS-EMT, Université du Québec, Canada A. Locatelli, O. T. Mentes, L. Aballe, Sincrotrone Trieste, Italy