Structural Mapping of Functional Ge Layers Grown on Graded SiGe

Nov 6, 2015 - *E-mail: [email protected]. ... Ivan Prieto , Roksolana Kozak , Oliver Skibitzki , Marta D. Rossell , Thomas Schroeder , ...
0 downloads 0 Views 4MB Size
Research Article www.acsami.org

Structural Mapping of Functional Ge Layers Grown on Graded SiGe Buffers for sub-10 nm CMOS Applications Using Advanced X‑ray Nanodiffraction Marie-Ingrid Richard,*,†,‡ Marvin H. Zoellner,¶ Gilbert A. Chahine,† Peter Zaumseil,¶ Giovanni Capellini,¶ Maik Hab̈ erlen,§ Peter Storck,§ Tobias U. Schülli,† and Thomas Schroeder¶,∥ †

ID01/ESRF, The European Synchrotron, 71 rue des Martyrs, 38043 Grenoble Cedex, France Aix-Marseille Université, CNRS, Université de Toulon, IM2NP UMR 7334, 13397 Marseille Cedex 20, France ¶ IHP, Im Technologiepark 25, 15236 Frankfurt (Oder), Germany § Siltronic AG, Hans-Seidel-Platz 4, 81737 München, Germany ∥ Institute of Physics and Chemistry, Brandenburg University of Technology, 03046 Cottbus, Germany ‡

S Supporting Information *

ABSTRACT: We report a detailed advanced materials characterization study on 40 nm thick strained germanium (Ge) layers integrated on 300 mm Si(001) wafers via strainrelaxed silicon−germanium (SiGe) buffer layers. Fast-scanning X-ray microscopy is used to directly image structural inhomogeneities, lattice tilt, thickness, and strain of a functional Ge layer down to the sub-micrometer scale with a real space step size of 750 μm. The structural study shows that the metastable Ge layer, pseudomorphically grown on Si0.3Ge0.7, exhibits an average compressive biaxial strain of −1.27%. By applying a scan area of 100 × 100 μm2, we observe microfluctuations of strain, lattice tilt, and thickness of ca. ±0.03%, ±0.05°, and ±0.8 nm, respectively. This study confirms the high materials homogeneity of the compressively strained Ge layer realized by the step-graded SiGe buffer approach on 300 mm Si wafers. This presents thus a promising materials science approach for advanced sub-10 nm complementary metal oxide−semiconductor applications based on strain-engineered Ge transistors to outperform current Si channel technologies. KEYWORDS: nanodiffraction, structural mapping, Ge layer, strain, CMOS



INTRODUCTION The integration of Ge materials is on the roadmap of advanced CMOS technology as key element for performance improvement beyond the 22 nm device node.1−3 Furthermore, Ge virtual wafers can be used as a technology platform for Ge photonics application to realize true group IV or III−V/Si hybrid electronic−photonic integrated circuits4,5 as well as a cost-effective substrate platform for high-efficiency concentrator photovoltaics.6 However, because of the lattice and thermal mismatch between the active Ge layer and the Si substrate, special techniques are required to achieve low-defect Ge epitaxial layers on Si. One solution is the growth of a SiGe strain relaxed buffer layer (SRB) between the Ge layer and the substrate, which is closer lattice matched to Ge and could help obtaining low-defect Ge thin film heterostructures. In the last decades, many efforts have been expanded to improve the quality of the SiGe surface by controlling the density and propagation of dislocations such as the use of a backside stressor7 or by pit-patterning of the Si(001) substrate.8 Numerous works have been performed to characterize the structural properties of the SiGe relaxed buffer layer.7,9−11 However, the correlated, nondestructive, model-free character© 2015 American Chemical Society

ization of the structure, strain, composition, and the micro- and nanofluctuations of a functional Ge layer on a SiGe buffer has not yet been addressed due to the lack of experimental methods. However, information on crystallographic structure and defects are required to understand the engineering processes applied to evaluate and control the future device performances. Recent advances in X-ray nanodiffraction experimental methods12 now allow nanostructures to be studied at the relevant local scale in device samples.13 In this article, we show that the strain and lattice orientation can be determined at the sub-micrometer length scale to analyze the quality of the functional Ge layer.



EXPERIMENTAL SECTION

The sample consists of a Ge layer deposited on a graded Si0.3Ge0.7 buffer. The epitaxial growth of the SiGe(001) layer is realized by a step-graded buffer on a polished 300 mm Si(001) substrate with a backside stressor, as outlined in more detail in refs 7 and 14. Prior to Received: September 14, 2015 Accepted: November 6, 2015 Published: November 6, 2015 26696

DOI: 10.1021/acsami.5b08645 ACS Appl. Mater. Interfaces 2015, 7, 26696−26700

Research Article

ACS Applied Materials & Interfaces the SiGe deposition the Si wafer received an in situ hydrogen bake. A flow of 75 standard liter per minute (slm) hydrogen carrier gas at atmospheric pressure and different flows of dichlorosilane and germaniumtetrachloride (GeCl4) were used. The SiGe buffer layers with a final concentration of 70% Ge were additionally capped with 1 μm of a constant Si0.3Ge0.7 composition. After a chemical mechanical polishing (CMP) step, which results in a flattening of the buffer layer surface (see ref 14 for more details), a metastable pseudomorphic and thus compressively strained Ge layer of targeted 40 nm thickness was deposited on top of the Si0.3Ge0.7 buffer. Diffraction experiments were performed using a nanofocused X-ray beam at beamline ID01 of the European Synchrotron Facility in Grenoble (France). The nanobeam was focused down to a 340 × 1100 nm2 (vertical × horizontal) spot size using a Fresnel Zone-plate (FZP) of 270 μm diameter and 80 nm outermost zone width. The nanodiffraction experiments were performed at a beam energy of 8 keV (wavelength λ of 1.55 Å), so that the 004 Ge Bragg peak was accessible at a scattering angle 2θ of 66.44°. The diffracted beam was recorded with a two-dimensional (2D) MAXIPIX photon-counting detector,15 characterized by 516 × 516 pixels of 55 μm pixel size and positioned at 1.12 m from the sample. The sample was mounted on a fast xyz scanning piezoelectric stage, with a lateral stroke of 100 μm, a resolution of 2 nm, and reproducibility in the range of 10 nm in our working conditions. It was itself mounted on a hexapod. An optical telescope was fixed on the marble supporting the Fresnel zone-plate, so that the interesting region could be found and positioned in the path of the X-ray beam (see Figure 1a). The experimental setup is shown in Figure 1a. By simultaneously combining high-speed continuous motion of the xyz scanning piezoelectric stage with highfrequency MAXIPIX image recording, the recently developed quicK continuous Mapping (K-Map) technique12 allows 2D diffraction maps to be obtained extremely quickly: a 100 × 100 μm2 area is mapped in 12 min with a 750 nm step size. A hardware module synchronizes the detector and the scanning stage, eliminating the holding and settling time, which includes data transfer and connection between the control program, the motors, and the detectors.12

Figure 1. Experimental setup. (a) Schematic overview of the experimental setup in coplanar diffraction geometry. The incoming X-ray beam is focused using an FZP and an order-sorting aperture (OSA). The incidence angle ω and the scattering angle 2θ are shown. ν denotes the deviation angle of the signal from the coplanar geometry. A microscope is positioned above the sample. A 30 × 30 μm2 AFM image of the Ge layer is shown at the sample position. (b) Radial X-ray diffraction scan in the vicinity of the 004 Ge, SiGe, and Si Bragg reflections as a function of 2θ and qz. The scan was measured with the nanobeam at a given (x, y) position.



RESULTS AND DISCUSSION Figure 1b shows a radial ω/2θ scan of the sample around the 004 Ge and 004 Si Bragg peaks and along the [001] direction. The curve is plotted as a function of the scattering angle 2θ and the out-of-plane scattering vector, qz = 4π sin(2θ/2)/λ. The Bragg peak from the Si substrate corresponds to the sharp intense peak. Besides the various peaks of the SiGe buffer in the angular range of 2θ = [67−69.5°], which will not be discussed further in this paper (see, e.g., refs 11 and 14), one observes the peak from the Ge layer at lower qz values (larger lattice parameters). It is characterized by some well-defined Kiessig fringes, whose spacing (Δqz = 0.0159 Å−1) is inversely proportional to the thickness of the Ge layer, leading to a value of 40 nm in line with the targeted thickness. The Kiessig fringes are an important indication of a high coherent structure quality of the smooth Ge layer on the Si0.3Ge0.7 buffer layer, pointing thus to its pseudomorphic growth; it is well-known that the onset of plastic relaxation eliminates these Kiessig fringes even on an early level of plastic relaxation.16 In addition, the Ge layer exhibits an intensity distribution centered at qz = 4.4 Å−1, leading to a⊥ = 5.712 Å and to an average out-of-plane strain εz of 0.954% with respect to bulk Ge. The average inplane lattice constant a// can be inferred by applying a biaxial strain model: a// = (aGe × (1 + P) − a⊥)/P

Figure 2a−c displays intensity maps of the same 100 × 100 μm2 region. The maps represent different 2D real-space maps of the total diffracted intensity of the 004 Ge Bragg peak at different incident angles, ω: at the left (Figure 2a, Δω = −0.08°) and right (Figure 2c, Δω = 0.07°) tails of the ω rocking-curve and at its maximum (Figure 2b, Δω = 0°), corresponding to 004 Ge diffraction condition. The intensity was integrated over the whole 2D detector. Vertical lines are observed, revealing inhomogeneous lattice orientation inside the Ge layer. No horizontal lines are observed despite a crosshatch pattern that is visible in the AFM image in Figure 1a. As the intensity was integrated in the detector plane (i.e., along the in-plane y and out-of-plane z directions; the beam being along the x-direction), these maps are insensitive to intensity fluctuations along the y-direction. The inhomogeneous strain field and bending of the Ge layer induced by the SRB SiGe buffer can be quantitatively measured by the fast-scanning X-ray nanodiffraction measurement. X-ray intensity maps were collected at 31 different incidence angles, ω, within an angular range of 0.01°, covering a total area of 100 × 100 μm2 (see Figure 2a−c as examples) with a step size of 750 nm. The measurement was performed at the 004 Ge Bragg peak, collecting 31 maps, yielding a total number of 737 769 detector frames (frames also collected when one of the

(1)

with P = 2ν/(1 − ν), ν being the Poisson’s ratio. This leads to a// = 5.586 Å. The Ge layer undergoes a biaxial compressive strain of ε// = −1.27% with respect to bulk Ge. 14

26697

DOI: 10.1021/acsami.5b08645 ACS Appl. Mater. Interfaces 2015, 7, 26696−26700

Research Article

ACS Applied Materials & Interfaces

Figure 2. Real vs reciprocal space. (a−c) 2D real-space maps of the logarithmic integrated intensity from a 100 × 100 μm2 surface area at different incident angles, Δω = −0.8° (a), 0° (b), and 0.7° (c) with respect to the 004 Ge Bragg condition. The letters indicate three different regions of the Ge layer, where 3D reciprocal space maps were measured. The sum of the measured intensity is displayed along Qy and Qz (d−f). (f) Thickness fringes. The dashed line is a guide for the eyes.

also displays the out-of-plane strain εz, which is related to the variation of the d004 spacing of the atomic planes. It can be retrieved for each position in real space: εz = (d004,meas − d004,ref)/d004,ref, where d004,meas = 2π/Q004. Q004 and d004,ref are the modulus of the out-of-plane scattering vector related to the 004 Bragg reflection and its reference d-spacing of bulk Ge, respectively. The out-of-plane lattice parameter c can also be retrieved from c = d004,meas/4. Figure 3a displays the calculated in-plane biaxial strain ε// of the Ge layer. The statistics for ε// are given in Figure 3b, where ε// = (−1.27 ± 0.03)%. In Figure 3b, the discrete sampling leads to artificial holes in the histogram. Figure 3c displays the magnitude and direction of the local orientation of the Ge planes, corresponding to the tilt, η, of the {001} Ge atomic planes. Assuming that the mosaic spread is homogeneously distributed in all directions, the lattice tilt of the (004)Ge planes is calculated from the layer peak position as η = cos−1(qz/Q004). The growth sequence leads to variations in strain and orientation in the Ge functional layer mainly because of the spatial distribution of defects associated with the plastic relaxation of the strain-relaxed, CMP-treated Si0.3Ge0.7 buffer layer.13 The crystallographic tilts within the 100 × 100 μm2 area vary up to 0.11°. A crosshatch pattern is visible in the atomic force microscopy (AFM) image in Figure 1a as

scanning motors is going back to its initial position; see ref 12). The total recording time was 6 h. The obtained detector frames at each spatial (X, Y) position are then reconstructed into threedimensional (3D) reciprocal space maps by converting the detector pixels and the angular coordinates into the reciprocal space coordinates, Qx, Qy, and Qz. Qy−Qz reciprocal space maps are displayed in Figure 2d−f at three different positions of the Ge layer, as indicated for ω = 0 in Figure 2b. Thickness fringes can be observed along Qz. At certain positions (for instance, Figure 2e), diffuse scattering appears along Qy. Threading dislocation defects, passing from the SRB Si0.3Ge0.7 layer into the functional Ge film, may induce it. The small deviations of Bragg peak obtained from the (004) maps allow determining the strain and tilt distribution inside the Ge layer. The average Bragg position is obtained by fitting the intensity of the Bragg peak by a Gaussian function. The reciprocal space coordinates of the scattering vector, Qx (a), Qy (b), and Qz (c) are displayed in Figure S1 (Supporting Information). Interestingly, only horizontal (vertical) lines are visible in the Qx (Qy) map. As the crosshatch pattern is aligned with the x- and y-directions, the Qx and Qy positions of the 004 Ge Bragg peak is only sensitive to tilts along the x- and ydirections, respectively. Figure S1d (Supporting Information) 26698

DOI: 10.1021/acsami.5b08645 ACS Appl. Mater. Interfaces 2015, 7, 26696−26700

Research Article

ACS Applied Materials & Interfaces

Figure 3. In-plane strain and tilt. (a) In-plane strain ε// of the Ge layer. The dashed lines are a guide for the eyes and are similar to the ones plotted in (c). (b) The statistics for ε//. The discrete sampling leads to artificial holes in the histogram. (c) 2D map distribution of tilt amplitude η (deg) with directional representation (arrows) of tilted {001} Ge atomic planes with respect to the sample surface normal. The dashed lines, at the same position as in (a), indicate tilt lines. (d) Absolute strain gradient along x-direction horizontally flipped. (e) Nanoscale thickness variation determined from the Qz-profile of the 004 Ge diffraction peak (hot wireframe: regions of high (low) thickness are colored yellow (red).

graded buffer layer for future sub-10 nm CMOS applications using model-free and nondestructive quantization with respect to the scanned 100 × 100 μm2 area. We were able to simultaneously measure its out-of-plane lattice parameter, the associated strain (ca. −1.27 ± 0.03%), its lattice tilt (±0.05°) and thickness variation (∼40.0 ± 0.8 nm) at the submicrometer scale. The main result is thus given by the fact that a compressive strain value well above a magnitude of 1% can be stabilized in the metastable ∼40 nm thick Ge film (pseudomorphically grown on the SRB Si0.3Ge0.7 layer), making this 300 mm virtual Ge wafer approach via SiGe grading of high relevance for sub-10 nm strain engineered Ge CMOS technologies. On the basis of our previously reported X-ray nanodiffraction experiment exclusively devoted to the Si0.3Ge0.7 buffer layer,13 it is evident that the observed strain and tilt fluctuations of the pseudomorphic Ge layer result from the relaxed SiGe buffer layer (only an approximation to a perfect crystal) underneath, as the strain and tilt fluctuation values of the latter are identical to the values reported here for the functional Ge film. This presents thus a promising materials science approach for advanced sub-10 nm CMOS applications based on strain-engineered Ge transistors to outperform current Si channel technologies. Advanced X-ray nanodiffraction experiments will be particularly valuable in the future to understand and control the deleterious structural effects in a fast and nondestructive manner and improve the performance of semiconductor heterostructures for not only

well as in the strain and lattice orientation maps (see Figure 3a,c). It arises from dislocation formation and relaxation in the step-graded buffer and leads to a nonuniform distribution of crystallographic orientations and results in localized strain gradients in the functional Ge layer. These tilt and strain fluctuations correlate with each other: higher in-plane strain gradient (see Figure 3d) leads to higher tilt. Accordingly, the full-width at half-maximum of the diffraction peak along the Qx (ΔQx) and Qy (ΔQy) directions, respectively, is broader in regions with low strain (see Figure S2 of Supporting Information), since the relaxation by defects causes a spreading of tilt. More interestingly, tilt lines appear next to strain lines, meaning that the strain gradient is directly correlated to the tilt. Indeed, it is shown that the tilt along x (Figure 3c) displays a fairly good correlation with the absolute strain gradient along the x-direction (Figure 3d). Further nanoscale structural information can be obtained by fitting the Qz profile of the diffraction peak (central peak and its thickness fringes) using a cardinal sine function. Figure 3e displays a map of the thickness of the functional Ge layer over an area of 100 × 100 μm2. A thickness variation of ±0.8 nm is observed over the displayed area, in agreement with AFM measurements.



CONCLUSION In summary, we have reported here X-ray nanodiffraction experiments on a very thin functional Ge layer grown on a SiGe 26699

DOI: 10.1021/acsami.5b08645 ACS Appl. Mater. Interfaces 2015, 7, 26696−26700

Research Article

ACS Applied Materials & Interfaces

(8) Mondiali, V.; Bollani, M.; Cecchi, S.; Richard, M.-I.; Schülli, T.; Chahine, G.; Chrastina, D. Dislocation Engineering in SiGe on Periodic and Aperiodic Si(001) Templates Studied by Fast Scanning X-Ray Nanodiffraction. Appl. Phys. Lett. 2014, 104, 021918. (9) Fitzgerald, E. A.; Xie, Y.-H.; Green, M. L.; Brasen, D.; Kortan, A. R.; Michel, J.; Mii, Y.-J.; Weir, B. E. Totally Relaxed GexSi1- X Layers with Low Threading Dislocation Densities Grown on Si Substrates. Appl. Phys. Lett. 1991, 59, 811−813. (10) Myrberg, T.; Jacob, A. P.; Nur, O.; Friesel, M.; Willander, M.; Patel, C. J.; Campidelli, Y.; Hernandez, C.; Kermarrec, O.; Bensahel, D. Structural Properties of Relaxed Ge Buffer Layers on Si (0 0 1): Effect of Layer Thickness and Low Temperature Si Initial Buffer. J. Mater. Sci.: Mater. Electron. 2004, 15, 411−417. (11) Kozlowski, G.; Zaumseil, P.; Schubert, M. A.; Yamamoto, Y.; Bauer, J.; Matejova, J.; Schulli, T.; Tillack, B.; Schroeder, T. Compliant Substrate versus Plastic Relaxation Effects in Ge Nanoheteroepitaxy on Free-Standing Si(001) Nanopillars. Appl. Phys. Lett. 2011, 99, 141901. (12) Chahine, G. A.; Richard, M.-I.; Homs-Regojo, R. A.; TranCaliste, T. N.; Carbone, D.; Jacques, V. L. R.; Grifone, R.; Boesecke, P.; Katzer, J.; Costina, I.; Djazouli, H.; Schroeder, T.; Schülli, T. U. Imaging of Strain and Lattice Orientation by Quick Scanning X-Ray Microscopy Combined with Three-Dimensional Reciprocal Space Mapping. J. Appl. Crystallogr. 2014, 47, 762−769. (13) Evans, P. G.; Savage, D. E.; Prance, J. R.; Simmons, C. B.; Lagally, M. G.; Coppersmith, S. N.; Eriksson, M. A.; Schülli, T. U. Nanoscale Distortions of Si Quantum Wells in Si/SiGe QuantumElectronic Heterostructures. Adv. Mater. 2012, 24, 5217−5221. (14) Zoellner, M.; Richard, M.-I.; Chahine, G.; Zaumseil, P.; Reich, C.; Capellini, G.; Montalenti, F.; Marzegalli, A.; Xie, Y.-H.; Schülli, T. U.; Häberlen, M.; Storck, P.; Schroeder, T. Imaging Structure and Composition Homogeneity of 300 Mm SiGe Virtual Substrates for Advanced CMOS Applications by Scanning X-Ray Diffraction Microscopy. ACS Appl. Mater. Interfaces 2015, 7, 9031. (15) Ponchut, C.; Rigal, J. M.; Clément, J.; Papillon, E.; Homs, A.; Petitdemange, S. MAXIPIX, a Fast Readout Photon-Counting X-Ray Area Detector for Synchrotron Applications. J. Instrum. 2011, 6, C01069. (16) Bhagavannarayana, G.; Zaumseil, P. Diffuse X-Ray Scattering of Misfit Dislocations at Si1- xGex/Si Interfaces by Triple Crystal Diffractometry. J. Appl. Phys. 1997, 82, 1172−1177. (17) Chahine, G. A.; Zoellner, M. H.; Richard, M.-I.; Guha, S.; Reich, C.; Zaumseil, P.; Capellini, G.; Schroeder, T.; Schülli, T. U. Strain and Lattice Orientation Distribution in SiN/Ge Complementary Metal− oxide−semiconductor Compatible Light Emitting Microstructures by Quick X-Ray Nano-Diffraction Microscopy. Appl. Phys. Lett. 2015, 106, 071902.

CMOS technologies (like those in ref 17) but also quantum well devices in optoelectronics and functional layer for energy materials like photovoltaics. With the availability of detectors that can operate in the kilohertz regime, the scanning speed is ultimately limited by the brilliance of the source. The technique will benefit from the ESRF Extremely Brilliant Source (EBS) Upgrade and is likely to become a routine analysis technique with in situ and operando capabilities to study processed semiconductor devices.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.5b08645. Two-dimensional maps of the three components of the scattering vector Q and of the out-of-plane strain of the Ge layer. Figures of the full-width at half-maximum of the 3D diffraction peak along the Qx and Qy directions. (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Funding

Financial support to this work by Deutsche Forschungsgemeinschaft is gratefully acknowledged by IHP. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors are grateful to ESRF Synchrotron for allocating beamtime. The measurement was performed on the ID01 beamline. We thank ID01 beamline staff for excellent support during the experiment.



REFERENCES

(1) Shifren, L.; Aitken, R.; Brown, A. R.; Chandra, V.; Cheng, B.; Riddet, C.; Alexandar, C. L.; Cline, B.; Millar, C.; Sinha, S.; Yeric, G.; Asenov, A. Predictive Simulation and Benchmarking of Si and Ge pMOS FinFETs for Future CMOS Technology. IEEE Trans. Electron Devices 2014, 61, 2271−2277. (2) Collaert, N.; Alian, A.; Arimura, H.; Boccardi, G.; Eneman, G.; Franco, J.; Ivanov, T.; Lin, D.; Loo, R.; Merckling, C.; Mitard, J.; Pourghaderi, M. A.; Rooyackers, R.; Sioncke, S.; Sun, J. W.; Vandooren, A.; Veloso, A.; Verhulst, A.; Waldron, N.; Witters, L.; Zhou, D.; Barla, K.; Thean, A.V.-Y. Ultimate Nano-Electronics: New Materials and Device Concepts for Scaling Nano-Electronics beyond the Si Roadmap. Microelectron. Eng. 2015, 132, 218−225. (3) Feng, J.; Woo, R.; Chen, S.; Liu, Y.; Griffin, P. B.; Plummer, J. D. P-Channel Germanium FinFET Based on Rapid Melt Growth. IEEE Electron Device Lett. 2007, 28, 637−639. (4) Liu, H.; Wang, T.; Jiang, Q.; Hogg, R.; Tutu, F.; Pozzi, F.; Seeds, A. Long-Wavelength InAs/GaAs Quantum-Dot Laser Diode Monolithically Grown on Ge Substrate. Nat. Photonics 2011, 5, 416−419. (5) Tsaur, B.-Y.; Geis, M. W.; Fan, J. C. C.; Gale, R. P. Heteroepitaxy of Vacuum-evaporated Ge Films on Single-crystal Si. Appl. Phys. Lett. 1981, 38, 779−781. (6) Sheldon, P.; Yacobi, B. G.; Jones, K. M.; Dunlavy, D. J. Growth and Characterization of GaAs/Ge Epilayers Grown on Si Substrates by Molecular Beam Epitaxy. J. Appl. Phys. 1985, 58, 4186−4193. (7) Kozlowski, G.; Fursenko, O.; Zaumseil, P.; Schroeder, T.; Vorderwestner, M.; Storck, P. (Invited) Epitaxial Growth of Low Defect SiGe Buffer Layers for Integration of New Materials on 300 Mm Silicon Wafers. ECS Trans. 2013, 50, 613−621. 26700

DOI: 10.1021/acsami.5b08645 ACS Appl. Mater. Interfaces 2015, 7, 26696−26700