Ge Virtual Substrates Research Articles

(Invited) Future Applications of SiGeSn and GeSn

Silicon photonics is considered to be an attractive pathway to drastically reduce power consumption in information technology based on classical Si microelectronics. Group IV alloys, namely Ge and SiGe on Si, are heavily investigated for the integration of modulators and detectors on Si [1,2]. However, higher integration levels of optical data transfer on the chip will require the integration of low power laser. The latter is approached by a wide range of technologies to deposit III/V structures on Si as well as by bonding III/V lasers Si (see ref. 3-7 in ref. [3]). Recently, a direct bandgap as well as optically pumped lasing was demonstrated for GeSn alloys deposited on Ge virtual substrates on Si (100) 8” wafers [3]. Here, efforts are reported of this ongoing endeavor towards an electrically pumped group IV laser. Moreover due to the predicted low effective mass for electrons in the Г-valley, resulting in high electron mobility as well as high carrier injection velocity, the material bears notable potential for high speed, low power circuitry. A novel epitaxial growth technology, namely, reactive gas source epitaxy has been applied to grow GeSn and SiGeSn film and heterostructures in a commercial 8” LPCVD tool, using Ge2H6, Si2H6 and SnCl4 has source gases in an N2 ambient. PH3 and B2H6 are used as sources for n- and p-type doping. Hetero- and quantum well structures containing SiGeSn/GeSn and strained Ge / GeSn layer, respectively, have been deposited at low growth temperatures (down to 325°C) and Sn concentrations up to 14.5%. Careful structural investigation were performed using X-ray diffractometry and transmission electron microscopy. Thick GeSn films grown on Ge will release about 80% of the strain. Interestingly, misfit dislocations are found typically only at the lowest interface towards the virtual Ge substrate, the interface structure within Hetero- and quantum wells structures appears to be free from extended defects. Optical pumping was performed using a ND-YAG laser with pulse duration of 5 ns. PL spectra are taken using a FTIR spectrometer equipped with a InSb detector. To improve the performance of optically pumped laser obtained from GeSn Fabry-Perot devices fabricated using simple mesa stripes [3], the GeSn mesa lines were undercut by RIE processing. This yields efficient strain relaxation in the undercut area. We have then an increase in the energy splitting between Г and L valley for electrons and a mode overlap of ~94% leading to an improved performance of these laser in comparison to the simple Fabry Perot devices previously reported [3]. These laser operate up to 135K, additionally the lasing efficiency is doubled and threshold pump power is cut by half. Similarly, micro-disc laser have been fabricated. Multi-mode lasing involving several TE and TM modes is observed up to 140K. The threshold is found at a pump power of ~225 kW/cm2. A minimum Sn concentration of ~ 8.5% is found to be required to obtain lasing in these relaxed GeSn structures. Light emitting diodes have been processed from p-i-n GeSn homojunctions. Light emission was observed up to 300K. First SiGeSn layers exhibiting a direct band gap, verified by comparing PL data and a joint density of state model are discussed. SiGeSn/GeSn p-i-n double heterostructures were grown, labelling the next milestone on the path towards an electrically pumped group IV laser. Moreover, the potential of GeSn alloys for the fabrication of low power, high frequency electronics is discussed. In particular, Bandstructure calculation reveal low effective masses, high mobilities and a high carrier injection velocity for electrons in the Gamma valley of the conduction band, making this material most interesting for the monolithic integration of optical and electronic devices. Acknowledgement The staff at PGI-9 and the clean room team at FZ-Jülich is acknowledged for their support. Special thanks to continuous input of Jeremy Witzens (RWTH) and Zoran Ikonic (Univ. Leeds) in analysing the data and modelling GeSn devices. References S. Assefa, F. Xia, F. and Y.A. Vlasov, “Reinventing germanium avalanche photodetector for nano-photonic on-chip optical interconnects. Nature 464, 80–84 (2010). Xu, Q., Schmidt, B., Pradhan, S. & Lipson, M. Micrometre-scale silicon electro-optic modulator. Nature 435, 325–327 (2005). S. Wirths, R. Geiger, N. von den Driesch, G. Mussler, T. Stoica, S. Mantl, Z. Ikonic, M. Luysberg, S. Chiussi, J. M. Hartmann, H. Sigg, J. Faist, D. Buca and D. Grützmacher, Nature Photonics 9, 88–92 (2015)

Doping Effects in Direct Bandgap GeSn Light Emitter

N-type doping is necessary for many applications in Ge and GeSn optoelectronics. In recent years the need for energy efficient electronic integrated circuits has led to the exploration of different approaches in order to bring light emitters on chips. Group-IV lasers are being considered, along with heterogeneous integration and flip-chip bonding. To overcome the lack of direct bandgap in Ge, options like injecting high amounts of tensile strain and/or n-type doping were investigated, aiming to achieve a sufficiently large electron population in the Γ-valley to support laser action. Investigations of highly n-doped Ge resulted in strong luminescence enhancement and in an optically pumped laser action. Alternatively, it has been shown that alloying Sn and Ge changes the electronic band structure so that the conduction band at Γ decreases below the L-valley, resulting in a transition from an indirect to a fundamental direct bandgap semiconductor. Optically pumped GeSn lasers were demonstrated, but their operation is still limited to temperatures of about 130 K. The naturally appearing question is whether n-doping can improve the laser performance of these direct bandgap materials, which still have a low directness (the separation between Γ and L valleys is typically 0-50 meV). Calculations indicate an increase of net gain up to a certain doping concentration, beyond which it saturates and then decreases. Here, we concentrate on the experimental study of n-type doping in direct bandgap Ge0.875Sn0.125 alloys. A series of samples with different doping concentrations, in the order of 1018 cm-3, were grown via reactive gas source epitaxy on Si(100) wafers buffered by a Ge virtual substrate. Digermane (Ge2H6), Tin tetrachloride (SnCl4) and phosphine (PH3) were used as precursors. The material is analyzed, regarding crystalline quality, strain and composition, by transmission electron microscopy (TEM), X-ray diffraction reciprocal space mapping (XRD-RSM) and Rutherford Backscattering Spectrometry (RBS). Photoluminescence under a chopped 532 nm laser excitation shows the influence of increased doping concentration on the material quality, regarding carrier recombination due to defects. The optical analysis of the bulk material also shows a very slight change in the bandgap, which can be attributed to bandgap narrowing caused by increasing doping concentration. The impact of the doping concentration is further studied in microdisk resonator lasers. The microdisks are defined by electron beam lithography, and mesas are etched using reactive ion etching (RIE) by Cl2/Ar plasma. To relax the strain of the GeSn layer, and to increase the optical mode confinement, the disks are under-etched by an isotropic CF4 plasma. Finally the structures are passivated with a 10 nm thick Al2O3layer deposited via atomic layer deposition (ALD). The advantage of using microdisk structures is that the deep under-etching of the GeSn disks results in the full strain relaxation of the active layer, eliminating the adverse effect of compressive strain on the band structure. Photoluminescence/laser studies were performed using an excitation with a pulsed Nd:YAG laser at 1064 nm wavelength, and a pulse width of 5 ns. Power and temperature dependence of the laser emission were studied. Figure 1 shows normalized photoluminescence from bulk GeSn and from an under-etched microdisk with a diameter of 8 µm made of n-doped GeSn. The lasing spectrum shows the contribution of multiple modes inside the disk. Undercutting the disks relaxes the strain, resulting in a decreased bandgap leading to a redshift of the laser peak. The photoluminescence peak energies depend on various effects: the strain in the material, the band filling due to high excitation densities and doping, and finally the doping-induced bandgap narrowing, while the laser emission peak position is also influenced by the free-carrier contribution to the refractive index. All these effects will be discussed, based on the systematic analysis of the above described material and laser structures. Additionally, calculations are used to investigate the influence of doping on optical gain. With L-valleys sitting only slightly above, and having a much larger density of states than the Γ valley, the majority of the photo-generated electrons will populate the L valleys if no n-type doping is employed. Only a fraction of electrons will therefore contribute to gain. This can be improved by n-doping, (e.g. by filling the L valleys), as already used in Ge lasers (mandatory due to the indirect bandgap, but also useful in cases of moderate directness). However, additional electrons also increase the free-carrier absorption, and the optimal doping will be a tradeoff between the two effects. Figure 1

(Invited) On the Manipulation of Phosphorus Diffusion as Well as the Reduction of Specific Contact Resistivity in Ge by Carbon Co-Doping

Thanks to its intrinsic high mobilities for carriers i.e. 3900 and 1900 cm2V-1s-1 for holes and electrons respectively, Ge has resurged as an alternative channel material for metal-oxide-semiconductor field-effect transistors (MOSFETs) [1]. To realize the-state-of-the-art high-performance Ge devices, the formation of highly activated shallow junctions especially n-type junction as well as the large contact specific contact resistivity (ρc), however, are still great challenges remained to be resolved, in spite of numerous efforts spent in the past decade [2, 3]. Impurity species such as carbon (C), nitrogen (N) and fluorine (F) are utilized to suppress the rapid diffusion of phosphorus in Ge by trapping vacancies [4, 5], while at the expense of a reduced activation level usually. Specifically, in this paper, the impact of C implantation parameters (energy and dose) on the diffusion and activation of P in Ge pre-amorphized (PAI) germanium epitaxial layers on Si (100), is systematically explored. All samples received a Ge PAI at 20 keV, 6×1014 at/cm2, followed by C implantation at different energies ranging from 3 keV to 20 keV and doses varying from 5×1014 to 2×101 5 at/cm2. P was implanted with a fixed condition using an energy of 5 keV and dose of 2×101 5 at/cm2 and was completely contained in the amorphous Ge layer. During annealing, to prevent out-diffusion of the implanted dopants, all samples were capped with a 20 nm thick SiO2 deposited by plasma enhanced chemical vapor deposition (PECVD) at 280 oC. The annealing process was carried out in N2 ambient at 600 oC for 60 s. It is illustrated in Fig. 1a that the shallowest N-type junction occurs for C implantation at 8 keV with a dose of 1×1015 cm-2. This optimum condition can be ascribed to the carbon distribution shown in Fig. 1b that carbon atoms straddles the amorphous/crystal (a/c) interface. As a result, both the diffusion of P in the amorphous layer and in the Ge virtual substrate beyond the a/c interface can be effectively suppressed. Except the control of P diffusion, the specific contact resistivity (ρc) between NiGe and n-Ge by the presence of C is also investigated. As illustrated in Fig. 2, the thermal stability of NiGe films is obviously enhanced by the presence of C and this can be inferred by the decreasing sheet resistance till 600 oC and the pretty smooth NiGe/Ge interface for samples with C. At 500 oC germanidation temperature, the rc values are reduced from 1.1×10-4 Ω-cm2 and 2.9×10-5 Ω-cm2 for NiGe/n- and p-Ge contacts without carbon to 7.3×10-5 Ω-cm2 and 1.4×10-5 Ω-cm2 for their counterparts with carbon, respectively, as illustrated in Fig. 3 and 4.

(Invited) Graphene Synthesis and Processing on Ge Substrates

Graphene is considered as a material which can enable new functionalities and performance improvements in a large variety of applications, among them in microelectronics [1, 2]. In microelectronics, techniques required for commercial large scale fabrication of graphene devices are not yet in place and further progress towards wafer-scale processing is required. Development of a wafer-scale Si technology-compatible graphene synthesis method and a toolbox of processes dedicated to handling, cleaning, patterning as well as integration of graphene with semiconductors, insulators, and metals is viewed as a prerequisite to practical applications of this material in electronic and photonic devices [3]. Semiconducting Ge surfaces appeared recently as attractive substrates for chemical vapor deposition (CVD) of graphene. There are several advantages of using them instead of metals. Firstly, the risk of metallic contaminations [4] is eliminated, ensuring the front-end-of-line compatibility in the device manufacturing process. Secondly, the thermal expansion coefficient mismatch between the Ge substrate and graphene is significantly lower than in the Cu-graphene system. This provides an opportunity to reduce the density of wrinkles which originate from thermal stress release and cause scattering of charge carriers in graphene. Furthermore, as a result of very low carbon solubility in Ge, the graphene growth on Ge can proceed, similarly to Cu substrates, in a “self-limiting” manner enabling more straightforward process control than for example on Ni. Finally, synthesis on patterned semiconducting Ge may enable direct use of graphene in some device concepts without the need of transfer which is inevitable in case of metallic substrates. In this talk, we will review the recent progress in graphene synthesis on Ge achieved by IHP and our collaborators starting from proof-of-concept molecular beam growth experiments [5], through low- [6] and high-pressure [7] CVD on Ge single crystals and Ge epi-layers on Si, to 200 mm wafer-scale CVD synthesis on Ge(100)/Si(100) virtual substrates. As an optimization approach towards improved rotational alignment of graphene on Ge [8,9] we will present results from the preparation of Ge(110)/Si(110) wafers and graphene growth on them. Beyond growth we will look at selected aspects of subsequent processing in a 200 mm Si wafer pilot line including patterning and interaction with photoresists as well as the deposition of other materials on graphene. Among the challenges ahead, re-usability of the Ge virtual substrates and wafer-scale graphene transfer procedures eliminating disadvantages of currently existing noble metal-assisted techniques [8] will be addressed. [1] A.C. Ferrari et al., “Science and technology roadmap for graphene, related two-dimensional crystals, and hybrid systems”, Nanoscale, 2015, 7, 4598-4810 [2] G. Fiori et al., “Electronics based on two-dimensional materials”, Nature Nanotechnology 2014, 9, 768–779 [3] L. Colombo et al., “Graphene growth and device integration”, IEEE Proceedings, 2013, 101, 1536-1556 [4] G. Lupina et al., “Residual metallic contamination of transferred chemical vapor deposited graphene”, ACS Nano, 2015, 9, 4776-4785 [5] G. Lippert et al., “Graphene grown on Ge(001) from atomic source”, Carbon, 2014, 75, 104-112 [6] J. Dabrowski et al., “Direct growth of low-doped graphene on Ge/Si(100) surfaces”, 2016, arXiv:1604.02315 [7] I. Pasternak et al., “Graphene growth on Ge(100)/Si(100) substrates by CVD method”, Scientific Reports, 2016, 6, 21773 [8] J.-H. Lee et al., “Wafer scale growth of single-crystal monolayer graphene on reusable hydrogen-terminated germanium”, Science, 2014, 344, 286-289 [9] B. Kiraly et al., “Electronic and mechanical properties of graphene-germanium interfaces grown by chemical vapor deposition”, Nano Lett., 2015, 15, 7414-7420

(Invited) On the Manipulation of Phosphorus Diffusion as Well as the Reduction of Specific Contact Resistivity in Ge by Carbon Co-Doping

In this work, the manipulation of phosphorus diffusion as well as the reduction of the specific contact resistivity in Ge by carbon co-doping is explored. It is found that there is an optimum condition for carbon implantation to suppress the rapid P diffusion effectively. For such a condition, the implanted carbon sits at the half range of the pre-amorphized Ge layer. In this case, both the diffusion of P in the amorphous layer and in the Ge virtual substrate beyond the amorphous/crystal (a/c) interface can be effectively manipulated. Except the control of P diffusion, the specific contact resistivity (ρc) between as-formed NiGe and the Ge substrate in the presence of C can be also reduced, compared with the reference without C. At 500°C germanidation temperature,the ρc values are reduced from 1.1x10-4 Ω-cm2 and 2.9x10-5 Ω-cm2 for NiGe/n- and p-Ge contacts without carbon to 7.3x10-5 Ω-cm2 and 1.4x10-5 Ω-cm2 for their counterparts with carbon, respectively.

The low temperature epitaxy of Ge on Si (1 0 0) substrate using two different precursors of GeH4 and Ge2H6

We have investigated the initial stage of low temperature epitaxy (LTE) of Ge on 8″-dia. Si (100) substrate using a rapid thermal chemical vapor deposition (RTCVD) with two different precursors of GeH4 and Ge2H6. The quality of LTE Ge films such as surface morphology, defects and crystallinity were analyzed using SEM, AFM and TEM. Experimental results confirmed that the LTE Ge using Ge2H6 precursor was much more beneficial than the LTE using GeH4 in terms of growth rate (×10), stress relaxation (85% at surface), and crystal quality (low TDDs). The discrepancy looks originated from the weak GeGe bonds requiring their dissociation energy small compared to the GeH bonds in GeH4 precursors, and the abundant supply of GeH3 molecules should stimulate chemical reactions at free surface sites. Our LTE technology would be promising for very thin Ge virtual substrate as well as be beneficial for nano-micro electronic devices in need of low temperature processes below 300–500°C.

Density and Capture Cross-Section of Interface Traps in GeSnO2 and GeO2 Grown on Heteroepitaxial GeSn.

An imperative factor in adapting GeSn as the channel material in CMOS technology, is the gate-oxide stack. The performance of GeSn transistors is degraded due to the high density of traps at the oxide-semiconductor interface. Several oxide-gate stacks have been pursued, and a midgap Dit obtained using the ac conductance method, is found in literature. However, a detailed signature of oxide traps like capture cross-section, donor/acceptor behavior and profile in the bandgap, is not yet available. We investigate the transition region between stoichiometric insulators and strained GeSn epitaxially grown on virtual Ge substrates. Al2O3 is used as high-κ oxide and either Ge1-xSnxO2 or GeO2 as interfacial layer oxide. The interface trap density (Dit) profile in the lower half of the bandgap is measured using deep level transient spectroscopy, and the importance of this technique for small bandgap materials like GeSn, is explained. Our results provide evidence for two conclusions. First, an interface traps density of 1.7 × 10(13) cm(-2)eV(-1) close to the valence band edge (Ev + 0.024 eV) and a capture cross-section (σp) of 1.7 × 10(-18) cm(2) is revealed for GeSnO2. These traps are associated with donor states. Second, it is shown that interfacial layer passivation of GeSn using GeO2 reduces the Dit by 1 order of magnitude (2.6 × 10(12) cm(-2)eV(-1)), in comparison to GeSnO2. The results are cross-verified using conductance method and saturation photovoltage technique. The Dit difference is associated with the presence of oxidized (Sn(4+)) and elemental Sn in the interfacial layer oxide.

Temperature varying photoconductivity of GeSn alloys grown by chemical vapor deposition with Sn concentrations from 4% to 11%

Pseudomorphic GeSn layers with Sn atomic percentages between 4.5% and 11.3% were grown by chemical vapor deposition using digermane and SnCl4 precursors on Ge virtual substrates grown on Si. The layers were characterized by x-ray diffraction rocking curves and reciprocal space maps. Photoconductive devices were fabricated, and the dark current was found to increase with Sn concentration. The responsivity of the photoconductors was measured at a wavelength of 1.55 μm using calibrated laser illumination at room temperature and a maximum value of 2.7 mA/W was measured for a 4.5% Sn device. Moreover, the responsivity for higher Sn concentration was found to increase with decreasing temperature. Spectral photoconductivity was measured using Fourier transform infrared spectroscopy. The photoconductive absorption edge continually increased in wavelength with increasing tin percentage, out to approximately 2.4 μm for an 11.3% Sn device. The direct band gap was extracted using Tauc plots and was fit to a bandgap model accounting for layer strain and Sn concentration. This direct bandgap was attributed to absorption from the heavy-hole band to the conduction band. Higher energy absorption was also observed, which was thought to be likely from absorption in the light-hole band. The band gaps for these alloys were plotted as a function of temperature. These experiments show the promise of GeSn alloys for CMOS compatible short wave infrared detectors.

Growth mechanism of an aluminium-induced solid phase epitaxial (AI-SPE) Si0.5Ge0.5layer using in situ heating transmission electron microscopy

The mechanism of growth of an epitaxial Si0.5Ge0.5 layer on a single crystalline (sc) Si (100) substrate by aluminum-induced solid phase epitaxy (AI-SPE) at a relatively low temperature (450 °C) has been revealed using in situ heating transmission electron microscopy (TEM). The analysis of the thermodynamics exactly supports the finding from in situ TEM. It evidences that the Si0.5Ge0.5 prefers to nucleate at the interface of the Al layer and the sc-Si (100) substrate due to the lowest critical thickness for nucleation. Based on the results from in situ TEM and thermodynamic analysis, the germanium (Ge) virtual substrate of the compositional gradient can be successfully prepared via a multi-run AI-SPE process at low-temperature.

Fabrication of GeSn-multiple quantum wells by overgrowth of Sn on Ge by using molecular beam epitaxy

We report on the fabrication and structural characterization of epitaxially grown ultra-thin layers of Sn on Ge virtual substrates (Si buffer layer overgrown by a 50 nm thick Ge epilayer followed by an annealing step). Samples with 1 to 5 monolayers of Sn on Ge virtual substrates were grown using solid source molecular beam epitaxy and characterized by atomic force microscopy. We determined the critical thickness at which the transition from two-dimensional to three-dimensional growth occurs. This transition is due to the large lattice mismatch between Ge and Sn (≈14.7%). By depositing Ge on top of Sn layers, which have thicknesses at or just below the critical thickness, we were able to fabricate ultra-narrow GeSn multi-quantum-well structures that are fully embedded in Ge. We report results on samples with one and ten GeSn wells separated by 5 and 10 nm thick Ge spacer layers that were characterized by high resolution transmission electron microscopy and X-ray diffraction. We discuss the structure and material intermixing observed in the samples.

Doping of Direct Gap Ge1 -y Sn y Alloys to Attain Electroluminescence and Enhanced Photoluminescence

Alloying of elemental Ge with Sn permits tuning of the band gap, as well as a transition towards direct gap behavior. Epitaxial deposition of direct gap Ge 1-y Sn y alloys (y≥0.09) requires ultra-low temperatures of ~300°C.1–3 In order to fully realize the electroluminescence and photoluminescence (PL) potential of this class of materials, strategies for p- and n-type doping that are compatible with this temperature range are needed. In our previous work, combination of both traditional and novel dopant sources were utilized for doping of materials ranging from pure Ge to Ge 1-y Sn y alloy at the direct to indirect crossover composition of y=0.09.4,5 Presented here are the extensions of these methods to include alloys beyond the crossover composition. The n-type doped alloys were grown on virtual Ge substrates in a hot wall UHV-CVD reactor using trigermane (Ge3H8) and deuterostannane (SnD4) sources to provide the respective elements. Trisilylphosphine (P(SiH3)3) was used as a P source for n-type doping alloys in the composition range y=0.09-0.13, which required growth temperatures of 300°C-275°C. Choice of this doping agent was due to its greater stability (which facilitates long term storage) and higher vapor pressure (which allows straightforward use in mixture preparation) when compared to trigermylphosphine (P(GeH3)3), another agent capable of delivering donor P atoms at low temperatures. The use of these precursor combinations has enabled the deposition of direct gap Ge 1-y Sn y alloy films with 300-500 nm thicknesseswith dopant concentrations 5x1018-3x1019 cm-3, as determined by IR spectroscopic ellipsometry. RBS channeling spectra and (004) rocking curves show these materials to have excellent crystallinity. Furthermore, analysis using SIMS shows that the incorporated donor atoms are fully activated, and the amount of Si incorporated into the materials from using P(SiH3)3 as the doping precursor is at impurity levels which does not affect the band gap properties. These observations are in agreement with previous experiments in which P(SiH3)3 was used as a doping agent, albeit at higher temperatures.4,5 Using identical protocols but removing P(SiH3)3 from the precursor mixtures, it was possible to obtain intrinsic Ge 1-y Sn y alloys with concentrations up to y=0.14. By choosing n-type doped virtual Ge substrates for these depositions, it was possible to investigate electroluminescence, after growth of appropriate p-type layers to form n-i-p diodes, as described below. The p-type doped alloys were grown on either intrinsic or n-type doped Ge­1-y Sn y alloys with y=0.09-0.14 in order to fabricate n-p or n-i-p device stacks. The depositions were carried out in a hot wall UHV-CVD reactor with a temperature profile that enables the pre-activation of the precursor molecules prior to reaching the growth surface. The precursors used for the depositions were digermane (Ge2H6), SnD4 and diborane (B2H6). The films obtained were 100-300 nm thick, with active dopant concentrations in the 2-3×1019 cm-3 range, and compositions of y=0.08-0.12. As shown in Fig.1, n-type doping of the Ge 1-y Sn y alloys enhances PL in comparison to intrinsic counterparts. Furthermore, prototype LEDs produced with the aid of above doping strategies exhibit strong electroluminescence,2 paving the way for obtaining emission for a broad spectral range, from pure Ge at 1550 nm up to 2700 nm and beyond. References (1) Gallagher, J. D.; Senaratne, C. L.; Kouvetakis, J.; Menéndez, J. Appl. Phys. Lett. 2014, 105, 142102. (2) Gallagher, J. D.; Senaratne, C. L.; Sims, P.; Aoki, T.; Menéndez, J.; Kouvetakis, J. Appl. Phys. Lett. 2015, 106, 091103. (3) Senaratne, C. L.; Gallagher, J. D.; Jiang, L.; Aoki, T.; Menéndez, J.; Kouvetakis, J. J. Appl. Phys. 2014, 116 (13), 133509. (4) Senaratne, C. L.; Gallagher, J. D.; Aoki, T.; Kouvetakis, J.; Menéndez, J. Chem. Mater. 2014, 26 (20), 6033–6041. (5) Xu, C.; Gallagher, J. D.; Sims, P.; Smith, D. J.; Menéndez, J.; Kouvetakis, J. Semicond. Sci. Technol. 2015, 30 (4), 045007. Fig. 1 – PL spectra obtained from a liquid nitrogen cooled InGaAs detector for n-type doped (solid line) and intrinsic (dashed line) Ge 1-y Sn y samples with y=0.09. Both samples show strong luminescence as expected for direct gap materials. However the doped sample, which has an active carrier concentration of 1×1019 cm-3, shows clear enhancement of PL compared to the intrinsic sample. The inset shows the same spectra obtained from a thermoelectrically cooled PbS detector with extended spectral range beyond the cutoff point of the InGaAs detector, albeit at expense of poorer signal-to-noise ratio. Figure 1

Ge growth on porous silicon: The effect of buffer porosity on the epilayer crystalline quality

We report on the epitaxial growth of Ge virtual substrates directly on Si (001) and on different porosity porous silicon (pSi) buffers. Obtained results indicate that Ge grown on low porosity (22%) pSi buffer has a better crystalline quality compared to Ge grown on bulk Si and on higher porosity buffers. This result is attributed to the compliant nature of pSi and to its reduced Young's modulus, which leads to plastic tensile deformation of the 22% porosity buffer under the in-plane tensile stress introduced by Ge lattice. The same result is not observed for higher porosity buffers, this effect being attributed to the higher buffer fragility. A low porosity pSi layer can hence be used as buffer for the growth of Ge on Si virtual substrates with reduced dislocation content and for the growth of Ge based devices or the successive integration of III-V semiconductors on Si.

Metastability and relaxation in tensile SiGe on Ge(001) virtual substrates

We systematically study the heteroepitaxy of SiGe alloys on Ge virtual substrates in order to understand strain relaxation processes and maximize the tensile strain in the SiGe layer. The degree of relaxation is measured by high-resolution x-ray diffraction, and surface morphology is characterized by atomic force microscopy. The results are analyzed in terms of a numerical model, which considers dislocation nucleation, multiplication, thermally activated glide, and strain-dependent blocking. Relaxation is found to be sensitive to growth rate and substrate temperature as well as epilayer misfit and thickness, and growth parameters are found which allow a SiGe film with over 4 GPa of tensile stress to be obtained.

GeSn/Ge multiquantum well photodetectors on Si substrates.

Vertical incidence GeSn/Ge multiquantum well (MQW) pin photodetectors on Si substrates were fabricated with a Sn concentration of 7%. The epitaxial structure was grown with a special low temperature molecular beam epitaxy process. The Ge barrier in the GeSn/Ge MQW was kept constant at 10 nm. The well width was varied between 6 and 12 nm. The GeSn/Ge MQW structures were grown pseudomorphically with the in-plane lattice constant of the Ge virtual substrate. The absorption edge shifts to longer wavelengths with thicker QWs in agreement with expectations from smaller quantization energies for the thicker QWs.

Virtual Substrate Technology for Ge1-XSnX Heteroepitaxy on Si Substrates

The fabrication of GeSn layers is one of the worldwide current areas of research in material science. The driving force behind it, is the possibility to realize a group IV direct bandgap semiconductor. Theoretical calculations show values between 10 % for relaxed GeSn crystals and 6 % for tensile-strained GeSn crystals. However, only its synthesis on Si substrates is of interest for the photonic components applications. For the manufacture of GeSn based components on Si substrates, a specific Ge VS (virtual substrate) has to be first realized in order to overcome the large lattice mismatch (≥ 4.2 %). This method has been developed by Kasper et al. for SiGe heterostructures on Si and its here used to provide, without epitaxial break, the desired lattice constant for the grow of Sn based alloys. The use of such VS minimize the number of the threading dislocations as required for high performance optical devices.The significant progress made in recent years allowed the demonstration of GeSn based electro-optical components integrated on Si.A compressively strained heterostructure can by relaxed by various methods. These are, for example, special annealing steps, increasing the layer thickness over the critical thickness tcrit or by means of point defects introduced during low temperature growth. If a Ge or GeSn film is deposited directly on Si it relaxes very quickly since the critical thickness is only a few nanometer thick. At the same time the threading dislocation density is extremely high. An alternative is an ultra-thin Ge VS developed in the last few years. The threading dislocation density is significantly reduced by high temperature annealing steps close to the melting point of Ge. This technology is used here as a basis for GeSn structure growth.The optical properties of the GeSn layers with 2 % and 4 % Sn content are addressed via pin photodetectors fabrication (see Fig. 1). As a reference the GeSn pin detectors were manufactured without VS directly on Si. An overview of the layer structures are shown in figure 2.The strain statuses of the individual samples were analyzed by RSM (reciprocal space mapping). These show that the samples on Si are 100 % strain relaxed and the samples on Ge VS are still under 50 % rest compressive strain.The dark current of the individual samples are shown in the figure 3. The GeSn samples grown directly on Si show a much higher dark current due to the significantly higher threading dislocation densities compared to the samples on Ge-VS. At a bias of -1 V, the dark current is increased by more than one order of magnitude. In Figure 4, the optical responsivity of each sample is shown as a function of wavelength.

Keynote; Epitaxy-Based Strain-Engineering Methods for Advanced Devices

In Si microtechnology with the economy of scale slackening its pace, energy efficiency is the technology driver of the future. The replacement of metallic by optical interconnects and the rigorous reduction of the supply voltage (VDD) of logic circuitry are suitable options to drastically reduce power consumption. Thus the integration of optoelectronic components and steep slope devices, like the T-FET, will meet this demands. However, both elements will either need or benefit from a direct band gap material, which is not supplied by Si. Possible solutions are the given by III/V compound or by Ge based semiconductors, such as GeSn alloys grown on Si and combinations thereof. These are challenging tasks and many obstacles have to be overcome, like large lattice mismatch, alloys with compositions far beyond the solid solubility limit, antiphase domains and interface and surface states. In this paper we will discuss several epitaxial methods to address these challenges. Furthermore, basing on profound Si technology, possible fabrication routes for electronic and optoelectronic devices will be discussed. Nanotechnology in combination with advanced selective area epitaxy offer unique paths to fabricate high quality III/V nanostructures on Si. In particular templated selfassembly of vertical core/ shell and lateral nanowires on nanopatterned Si substrates will be discussed. The structural properties have been investigated using high resolution transmission microscopy. Fig. 1 depicts a lateral InAs nanowires grown on Si (110) substrates revealing (111) facets at the InAs/Si interface and rotational twins in the InAs nanowire.Regarding photonic devices the most promising group IV material is Ge, since the conduction band minimum at the G-point of the Brillouin-zone (referred to as G-valley) is placed only ca. 140 meV above the fourfold degenerate indirect L-valley. To band engineer Ge towards a direct bandgap semiconductor, two approaches are very promising, Ge with a large tensile strain (> 1.5%) and relaxed GeSn alloys with Sn concentrations above 10% [1,2]. Here, both systems are discussed, however, special emphasis is put on the growth of GeSn and SiGeSn alloys. These are predicted to have a direct band gap energy of less than 0.5 eV. Thus GeSn provides advantages for optoelectronics as well as for ultralow power T-FETs, since the small direct band gap is beneficial for the realization of large tunnel currents. However, due to the large lattice mismatch to Si, the small solubility of Sn in Ge and the necessity to strain engineer the layers, epitaxy is extremely demanding. Here we introduce an epitaxial growth method using enhanced surface kinetical reactions to provide high quality SiGeSn/GeSn heterostructures on Ge virtual substrates on Si with Sn concentrations above 10%.Fig. 2 shows a TEM cross section of a strained Ge layer embedded in relaxed SiGeSn cladding layers grown on a relaxed GeSn buffer layer. The double heterostructure provides a type I band alignment and includes doping levels for active optoelectronic devices. Structural characterization using X-ray diffraction and RBS to determine strain and crystallographic perfection have been performed. Photoluminescence and absorption spectroscopy give insights into the indirect to direct band gap transition in these materials. Finally, novel concepts for Tunnel-FET´s using III/V and group IV heterostructures will be discussed. The recently achieved strained-Si nanowire TFET exhibiting a subthreshold swing of less than 60 mV and large Ion/Ioff ratios are base for the integration of these heterostructure concepts. SiGeSn alloys potentially promise an energy efficient platform by the integration of ultralow power and optoelectronic devices. [1] Camacho-Aguilera, R. E. et al. An electrically pum-ped germanium laser. Opt. Express 20, 11316–20 (2012). [2] Bauer, M. et al. Ge–Sn semiconductors for band-gap and lattice engineering. Appl. Phys. Lett. 81, 2992 (2002).

Ni(SiGeSn) Metal Contact Formation on Low Bandgap Strained (Si)Ge(Sn) Semiconductors

GeSn has gained a lot of attention as alternative high mobility and low bandgap material to boost the performance of MOSFETs and Tunnel-FETs. Self-aligned contacts with low sheet resistance and good thermal stability are desired for GeSn FETs. Preliminary works on Ni-GeSn contacts indicate a low thermal stability. However, to date only GeSn with low Sn concentrations below 6 % have been reported. Here we investigate the formation of metallic contacts for high mobility Sn-based group IV semiconductors. The formation of ternary NiGeSn alloys for a large range of Sn compositions up to 12 at.% and quaternary NiSiGeSn alloys, formed on SiGeSn ternaries with Si/Sn compositions ranging from 20%/3% to 3%/12% is presented.(Si)GeSn layers with thicknesses of about 30-60 nm were pseudomorphically grown on high quality Ge virtual substrates (VS) on Si(100) by RP-CVD. The growth temperature was varied between 350°C and 475°C to adjust the Sn and Si concentration within the epilayers.For the layers under investigation Sn and Si concentrations between 3 – 12 at.% and 3 – 18 at.% have been obtained. All GeSn layers exhibit high compressive strain while for SiGeSn ternaries biaxial compressive and biaxial tensile strain is observed according to the Si/Sn ratio. For the metallization studies 10 nm Ni were evaporated followed by a forming gas annealing step for 10 or 30 s. Here, the temperature was varied between 250°C and 400°C.Specular XRD, taken in off-orientation relative to the Si substrate was performed for phase identification of differently annealed (275°C – 350°C) Ni-Ge0.91Sn0.09 samples (Fig. 1a). The temperature dependence of the stanogermanide phase formation shows that the Ni5X3 phase (x = Ge0.88Sn0.12, Ge0.91Sn0.09 Ge0.94Sn0.06) appears already at 275°C. The epitaxial Ni5X3 phase transforms into a [Ni1X1 + Ni5X3] multi-phase at 300°C while at 325°C only poly crystalline grains of the Ni1X1 phase are observed. The morphology of the stanogermanides was investigated by TEM. For all investigated Sn concentrations at a formation temperature of 300°C, the formed metallic layers are continuous with a mean thickness of about 25 nm. High resolution TEM images taken on NiGe0.94Sn0.06 and NiGe0.88Sn0.12 are shown in Fig. 1b,c. The rough Ni1X1/GeSn interface is typical for poly-crystalline layers containing large and non-substrate oriented grains. The TEM micrograph also proves that the grains are crystalline with well-defined grain boundaries and an unconsumed single crystalline GeSn layer. The well-oriented Ni5X3 phase possess a flat Ni5X3/GeSn interface (not shown here).Compared to the stanogermanides, we obtained an extended thermal stability window for the Ni5(SiGeSn)3 phase. This epitaxial phase, demonstrated by pole figures, is observed for all Si/Sn ratios, from high Si content (20% Si and 3% Sn) to high Sn content (3% Si and 12% Sn) alloys. The sheet resistance of NiSiGeSn alloys versus temperature is shown in Fig. 1d for different compositions. For Sn concentrations up to 8 at.% low sheet resistances (< 10 W/sq) are obtained for the complete temperature range. The thermal stability for 12 at.% Sn is very limited.

In Situ Studies of Germanium-Tin and Silicon-Germanium-Tin Thermal Stability

GeSn and SiGeSn semiconductors provide a wealth of opportunities to enable the realization of mid-infrared photonics in group-IV semiconductors will enable on-chip CMOS optoelectronic systems with a potential impact on chemical and biological sensing, spectroscopy, and free-space communication. Moreover, the fact that these alloys are silicon-compatible will make possible the integration of group-IV-based photonics and optoelectronics with CMOS technology. Efficient group-IV-based light emitting devices and photodetectors can now be implemented using band gap engineering in GeSn and SiGeSn semiconductors, which show indirect-to-direct transition at a particular composition and strain. Besides the potential applications in photonics and optoelectronics, Sn-containing group IV semiconductor alloys and heterostructures are also highly relevant for high-mobility and low-power electronics. Additionally, the control of the composition and structure of SiGeSn alloys and heterostructures are also crucial to implement carbon-free energy conversion devices such thermoelectrics and high-efficiency solar cells.A deep understanding of the structural and morphological stability of GeSn and SiGeSn metastable alloys is of utmost importance in order to achieve the aforementioned technologies. With this perspective, we present in this contribution detailed in situ studies of the evolution throughout thermal processing of both composition and structure of set of monocrystalline binary and ternary Sn-containing group-IV alloys. The investigated layers were grown using an industry compatible metal cold-wall Reduced Pressure AIXTRON TRICENT reactor (RP-CVD) with a showerhead for 200/300mm wafers. The epitaxial layers were grown using Si2H6, Ge2H6 (10% diluted in H2) and SnCl4 precursors, and N2 carrier gas, which warrant reasonable growth rates at growth temperatures in the 350-475 °C range. The growth of GeSn and SiGeSn layers was performed on Si(100) wafers using a low-defect density Ge virtual substrate. Figure 1 displays a representative image of cross-sectional scanning transmission electron microscopy of Si0.04Ge0.84Sn0.12layer. The composition and strain of the grown layers were investigated using a variety of experimental techniques including Raman spectroscopy, Rutherford backscattering spectrometry, x-ray reciprocal space mapping, and energy dispersive x-ray spectroscopy.Subsequently, the as-grown layers were subjected to in situ investigations of their structural and elemental properties as a function of annealing temperatures using low energy electron microscopy (LEEM), photoelectron emission microscopy (PEEM), and nano-Auger spectroscopy. These investigations have unraveled unprecedented insights into the stability of these layers as well as into the dynamics of phase separation in Sn-rich alloys. In the latter, we have traced the formation, evolution, surface diffusion of Sn-rich droplets and clusters. We have also indentified the interplay between Sn concentration and the critical temperature that triggers the alloy instabilities in both binary and ternary layers. The effects of dislocations on the dynamics of phase separation were also identified and elucidated. A theoretical treatment including both thermodynamic and kinetic considerations was developed to discuss the observed phenomena.

Comparison between experimental and simulated strain profiles in Ge channels with embedded source/drain stressors

AbstractTwo planar Ge‐based MOSFET structures were analysed for n‐ and p‐type transistors with tensile and compressive strain implemented in the &lt;110&gt; channel direction using source/drain stressors. Strain profiles measured by nano‐beam diffraction (NBD) have been compared with Sentaurus process simulations. The TCAD simulations were tuned with the actual process using NBD measurements, resulting in a more realistic simulation and, therefore, a more reliable interpretation is now possible. dark‐field scanning transmission electron microscopy images added to the strain profiles returned interesting results about the Ge virtual substrate condition and how it disturbs the NBD readings due to the presence of extended defects in strain relaxed Ge buffer layers. (© 2014 WILEY‐VCH Verlag GmbH &amp; Co. KGaA, Weinheim)

GeSn-on-Si normal incidence photodetectors with bandwidths more than 40 GHz

GeSn (Sn content up to 4.2%) photodiodes with vertical pin structures were grown on thin Ge virtual substrates on Si by a low temperature (160 °C) molecular beam epitaxy. Vertical detectors were fabricated by a double mesa process with mesa radii between 5 µm and 80 µm. The nominal intrinsic absorber contains carrier densities from below 1 · 10(16) cm(-3) to 1 · 10(17) cm(-3) for Ge reference detectors and GeSn detectors with 4.2% Sn, respectively. The photodetectors were investigated with electrical and optoelectrical methods from direct current up to high frequencies (40 GHz). For a laser wavelength of 1550 nm an increasing of the optical responsivities (84 mA/W -218 mA/W) for vertical incidence detectors with thin (300 nm) absorbers as function of the Sn content were found. Most important from an application perspective all detectors had bandwidth above 40 GHz at enough reverse voltage which increased from zero to -5 V within the given Sn range. Increasing carrier densities (up to 1 · 10(17) cm(-3)) with Sn contents caused the depletion of the nominal intrinsic absorber at increasing reverse voltages.