ConceptGraphene New Electronics Concept: Wafer-Scale Epitaxial Graphene Small or medium-scale focused research project WP2 Characterization and integration Deliverable 2.2 Report on the influence of environment on the properties of graphene Main Authors: A. Tzalenchuk, S. Kubatkin Nature of deliverable: R = Report Dissemination level: PU = Public Due date of deliverable: M18 Actual submission date: M18 LIST OF CONTRIBUTORS Project funded by the European Commission under grant agreement n 257829
ConceptGraphene D2.2 31 March 2012 2 / 13 Partner Acronym Laboratory Name Name of the contact 1(coordinator) CTHA Chalmers Tekniska Hoegskola Tomas Löfwander AB 3 NPLML NPL Management Limited Alexander Tzalenchuk
ConceptGraphene D2.2 31 March 2012 3 / 13 TABLE OF CONTENTS Deliverable Summary... 4 1. The substrate... 5 2. Disorder... 6 3. Adsorbates... 6 4. Resist residues... 9 5. Environment engineering... 11
ConceptGraphene D2.2 31 March 2012 4 / 13 Deliverable Summary Deliverable 2.2 Report on the influence of SiC environment and correlated adsorbents and disorder on the properties of graphene Electronic properties of graphene are strongly influenced by its environment natural or engineered, such as the substrate, overlayers, atmosphere, precipitates, contaminants, etc. Often the environment is difficult to control, leading to deterioration of graphene parameters, but when properly engineered, the environment can play an important positive role. Either way, understanding the immediate graphene environment is important for the progress of graphene technology.
ConceptGraphene D2.2 31 March 2012 5 / 13 1. The substrate As-grown epitaxial graphene is always strongly n-doped. In 1 we explored the origin of this doping and presented a theoretical model of charge transfer from the SiC substrate to monolayer or bilayer graphene at zero magnetic field. We analysed doping of graphene grown on SiC in two models, which differ by the source of charge transferred to graphene, namely, from SiC surface and from bulk donors. For each of the two models, we found the maximum electron density induced in monolayer and bilayer graphene, which is determined by the difference between the work function for electrons in pristine graphene and donor states on/in SiC, and analysed the responsivity of graphene to the density variation by means of electrostatic gates. The model gives the saturation carrier density of 10 13 cm -2 for realistic values of the workfunction difference close to the densities normally observed in as-grown graphene on SiC. Such initial carrier densities impose constraints on various applications of graphene, such as in transistors or metrology. For the former, the tunability transport with electrostatic gating becomes impossible, whereas for the latter application, the range of quantising magnetic field becomes impractical. However, we noted that particular growth processes developed by Linköping and Erlangen produce SiC/G with a much lower doping level 2, 3 indicating that efficient annealing of donors on and near the SiC surface is possible. Figure 1. Charge transfer between SiC and bare/gated graphene
ConceptGraphene D2.2 31 March 2012 6 / 13 2. Disorder Disorder in the substrate plays a crucial role in the transport properties of graphene. In 4 we presented a comprehensive analysis of two quantum transport phenomena in epitaxial graphene grown on the Si-terminated face of SiC: the Aronov-Altshuler (AA) e-e interaction correction and weak localisation (WL) contribution to resistivity. The analysis of the AA correction has confirmed that electrons in epitaxial graphene display all the attributes of a disordered Fermi liquid. The analysis of WL has enabled us to determine that the symmetry-breaking disorder and intervalley scattering play a lesser role in determining epitaxial graphene resistivity than scattering from donors on SiC surface. Moreover, the analysis of the temperature dependence of the decoherence rate extracted from the WL magnetoresistance has enabled us to establish spin relaxation of electrons, at the time scale of ~50 ps, which we attribute to the presence of local magnetic moments of defects in or under graphene. The latter conclusion is driven from both the tendency of the decoherence rate to saturate at the lowest temperatures and also the absence of crossover from WL to WAL specific to conductors with strong spinorbit scattering. Based on this observation, we conclude that spin memory of electrons in a field-effect transistors based on epitaxial graphene with electron density n~10 11 cm -2 can exceed a micrometre scale, which proves the suitability of this system for spintronics applications. 3. Adsorbates The substrate-induced carrier density in graphene can be modified by adsorbates. One such adsorbate, always present in the atmosphere is water, which is a well-known p-dopant. For example in 5 (Supporting Information) the transport properties of epitaxial graphene had been measured by magnetotransport measurements at room temperature in vacuum and ambient atmosphere at room temperature. It was shown that the carrier density was significantly lower in the latter case. Furthermore the results showed a greater spread in parameters in this case, indicating nonhomogeneous doping. To get further quantitative insight into the water doping
ConceptGraphene D2.2 31 March 2012 7 / 13 on the local scale, NPL and Linköping have measured the surface potential of a deliberately inhomogeneous sample, partially covered with monolayer graphene (1LG) and partially bilayer (2LG). The sample was kept in atmosphere with controlled humidity, while the surface potential distribution was measured using a scanning Kelvin probe microscope (SKPM). The sequence of images shows that while 2LG is not affected by humidity, the 1LG is indeed progressively compensated as humidity increases. To quantify this compensation, we have plotted the change in the carrier density dn calculated as dn = 2d r n ~v F 2.0x10 12 1.8x10 12 Figure 2. Change in the carrier density of 1LG as a function of humidity. n, cm -2 1.6x10 12 1.4x10 12 1.2x10 12 1.0x10 12-2 0 2 4 6 8 10 12 14 1 Measurement no. (increasing humidity) Small-scale inhomogeneity of water doping in 1LG has also been studied, see Figure 3 below. Interestingly, the tapping phase AFM images of 1LG reveal ordering of water in ambient conditions, which gradually disappears as the temperature is increased. The ordering is not resolved in SKPM.
ConceptGraphene D2.2 2LG 31 March 2012 8 / 13 1L G Figure 2. From top left to bottom right: Kelvin Force Microscopy image of surface potential of 1LG and 2LG. Starting from vacuum and 200 C cleaned graphene in dry nitrogen with increasing humidity to ~30% RH. 2LG potential remains constant whereas the potential of 1LG drops as humidity increases.
ConceptGraphene D2.2 31 March 2012 9 / 13 Figure 3. Ordering of water nano-droplets on 1LG - tapping phase AFM images. 4. Resist residues From the technological perspective, another contaminant is particularly important the residue of the lithographic resists. Although every care is taken to remove these resists after processing, the question is how much remains. To verify this issue Linköping has prepared two samples of singlelayer graphene: at Chalmers one was initially covered with the usual combination of resists, and then reconditioned by chemically removing the resist as thoroughly as possible. The other sample remained completely untreated. Both samples travelled between partners in the same conditions. They were eventually investigated at NPL using secondary ion mass spectroscopy SIMS.
ConceptGraphene D2.2 31 March 2012 10 / 13 6 x10 6 x10 2.5 2.5 Intensity (counts) 2.0 1.5 1.0 0.5 Intensity (counts) 2.0 1.5 1.0 0.5 10 20 30 40 50 60 70 80 90 100 Mass (u) 10 20 30 40 50 60 70 80 90 100 Mass (u) 5 x10 5 x10 5.0 5.0 Intensity (counts) 4.0 3.0 2.0 1.0 Intensity (counts) 4.0 3.0 2.0 1.0 120 140 160 180 200 220 240 260 280 300 Mass (u) 120 140 160 180 200 220 240 260 280 300 Mass (u) 3 x10 3 x10 8.0 8.0 Intensity (counts) 6.0 4.0 2.0 Intensity (counts) 6.0 4.0 2.0 350 400 450 500 550 600 650 Mass (u) 350 400 450 500 550 600 650 Mass (u) Figure 4. Secondary ion mass spectra of two G-SiC samples (reconditioned #266, left, and untreated, #277, right). For the reconditioned sample the low mass range of the spectrum includes ions which primarily consists of hydrogen, carbon, oxygen and nitrogen. Peaks at masses 91, 105, 115 and 128 are part of a peak pattern which is seen through the spectrum to higher masses. These peaks are characteristic for carbon-rich polymers. Another series of peaks, which is observed in the + mass range 400-600 u, can generally be assigned to ions of the form C x H y O z and are likely to originate from oils, fatty acids or phthalates. The polymer additive Irgafos168 gives rise to strong peaks around m/z 650 u. For the untreated sample the low mass range in the spectrum from G267 is highly similar to the spectrum from the treated sample, though unsurprisingly the polymer related peaks are not observed. A number of peaks in the higher + mass range, e.g. at 149, 256, 282 and 419 u, can be identified as C x H y O z ions, some characteristic for phthalates and others for fatty acids (oleamide). Again, Irgafos 168 is seen in the spectrum it is most likely the result of contamination from the plastic box and the sticky tape holding the samples in place. The recommendation is to use special Fluoroware boxes for storage and shipping.
ConceptGraphene D2.2 31 March 2012 11 / 13 Peaks of low intensity can be assigned to C + x ions and are indicative of graphene. Finally, CH 3 O + ion maps were made of both samples. For the treated sample they show an uneven lateral distribution of low mass ions containing nitrogen and oxygen (only low mass ions were sufficiently intense to give useful chemical maps within the experiment time). The utreated sample shows an almost featureless map. Figure 5. Contamination maps of the two G-SiC samples (reconditioned #266, left, and untreated, #277, right) obtained with SIMS. 5. Environment engineering Once the carrier density in the as-grown graphene is made relatively low (compared to the saturation value of 10 13 cm -2 ), it can be efficiently tuned by gating. Normally this is done by application of a voltage between an isolated metal gate and graphene. In 6 we demonstrated a novel technology photochemical gating. The technology is based on generation of potent acceptors in a specially chosen polymer heterostructure covering graphene subjected to deep UV light. With this method we were able to reduce the electron density in graphene by a factor of 50. Importantly, this change in the carrier density resulted in a fivefold increase in carrier mobility up to 16 000 cm 2 /Vs at the liquid helium temperature. At room temperature we have demonstrated the mobility of 5100 cm 2 /Vs, thus achieving milestone MS2.
ConceptGraphene D2.2 31 March 2012 12 / 13 Figure 6. Photochemical gating of graphene. The donor states, which are responsible for the strong initial n-doping of graphene, are probably localised in the dead layer of carbon atoms, just underneath graphene. This layer is characterized by a 6 3 6 3 supercell of the reconstructed surface of sublimated SiC. Missing or substituted carbon atoms in various positions of such a huge supercell in the dead layer create localized surface states with a broad distribution of energies within the bandgap of SiC. In 7 we have shown that that the carrier density in graphene varies with magnetic field due to the charge transfer between surface-donor states in SiC and graphene. In high magnetic fields the carrier density saturates at a value up to 30% higher than the zero-field carrier density. Most importantly, we found magnetic field intervals of several Tesla, where the carrier density in graphene increases linearly with the magnetic field, resulting in pinning of the ν = 2 quantum Hall state with electrons at the chemical potential occupying SiC surface donor states half-way between the N = 0 and N = 1 LLs in graphene. The pinned filling factor manifests itself in a continuously increasing breakdown current toward the upper magnetic field end of the ν = 2 state far beyond the nominal value of B ν=2 calculated from the zero-field carrier density. Facilitated by the high breakdown current in excess of 500 µa at 14 T, we have achieved a precision of 3 parts in 10 10 in the Hall resistance quantization measurements of photochemically gated graphene heterostructures.
ConceptGraphene D2.2 31 March 2012 13 / 13 Figure 7. Pinning of the ν = 2 filling factor. Ic breakdown current the maximum non-dissipative current, which the quantum Hall state can sustain. 1. S. Kopylov, A. Tzalenchuk, S. Kubatkin and V. I. Fal'ko, Appl Phys Lett 97 (11), 112109 (2010). 2. A. Tzalenchuk, S. Lara-Avila, A. Kalaboukhov, S. Paolillo, M. Syvajarvi, R. Yakimova, O. Kazakova, T. J. B. M. Janssen, V. Fal'ko and S. Kubatkin, Nat Nanotechnol 5 (3), 186-189 (2010). 3. S. Weingart, C. Bock, U. Kunze, F. Speck, T. Seyller and L. Ley, Appl Phys Lett 95 (26), 262101 (2009). 4. S. Lara-Avila, A. Tzalenchuk, S. Kubatkin, R. Yakimova, T. J. B. M. Janssen, K. Cedergren, T. Bergsten and V. Fal'ko, Phys Rev Lett 107 (16), 166602 (2011). 5. J. A. Robinson, M. Hollander, M. LaBella, K. A. Trumbull, R. Cavalero and D. W. Snyder, Nano Lett 11 (9), 3875-3880 (2011). 6. S. Lara-Avila, K. Moth-Poulsen, R. Yakimova, T. Bjornholm, V. Fal'ko, A. Tzalenchuk and S. Kubatkin, Adv. Mater. 23 (7), 878-882 (2011). 7. T. J. B. M. Janssen, A. Tzalenchuk, R. Yakimova, S. Kubatkin, S. Lara- Avila, S. Kopylov and V. I. Fal'ko, Phys Rev B 83 (23), 233402 (2011).