Open Access
Add to BibSonomy
Abstract
The persistent interest of the epitaxy of group IV alloy GeSn is mainly driven by the demand for an efficient light source that could be monolithically integrated on Si for mid-infrared Si photonics. For chemical vapor deposition of GeSn, the exploration of the growth window is difficult from the beginning due to the metastable nature of the material requiring non-equilibrium growth condition. In this work, we demonstrated an effective pathway to achieve high quality GeSn with high levels of Sn incorporation. The GeSn films were grown on Ge-buffered Si via ultra-high vacuum chemical vapor deposition using GeH4 and SnCl4 as the precursor. The influence of both SnCl4 flow fraction and growth temperature on the Sn incorporation and material quality were investigated. Different growth regimes were explored leading to an optimized regime at low temperature which suppressed the Sn precipitation allowing for increased Sn incorporation. The prototype GeSn photoconductors were fabricated with typical samples, which show the promising device applications.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Progress in group IV photonics over recent years has elevated the GeSn material system to the forefront of photonic integration on Si substrates [1–3]. Demonstrations of true direct bandgap GeSn in LED and lasing devices show the growing potential of the GeSn material system [4–9]. Applications based on this technology could flood the photonics market with inexpensive and efficient light emission/detection devices.
Growth of GeSn on Ge is difficult due to low solubility (<1%) of Sn in Ge and the instability of α-Sn above 13°C. In order to grow GeSn material, growth techniques were developed under non-equilibrium growth conditions such as low temperature MBE [10–16] or CVD [17–24]. The CVD growth of GeSn has been increasingly investigated over the past decade. Various Sn and Ge precursors in conjunction with carrier gases were used attempting to achieve high Sn incorporation and high material quality. Early growths were carried out using deuterium-stabilized stannane (SnD4) as the Sn precursor [25], whose high cost and instability drove the motivation to seek other Sn precursors. It has been reported that tin-tetrachloride (SnCl4) is a low cost, stable, and commercially available precursor and the GeSn material growth was initially demonstrated by Vincent at el [26]. On the other hand, various Ge hydrides have also been explored by Kouvetakis et al [27]. Higher order germane were commonly used due to their favorable decomposition at low temperatures [28–31]. Our recent progress has allowed for low cost germane (GeH4) to be used for Ge and GeSn growth via CVD [19,21,22,32] and plasma enhanced-CVD [33] in UHV chamber.
Previous reports of GeSn material growth focus on the successful deposition conditions and the final product. While providing useful guidance, they have not provided an effective route to produce high quality GeSn as optimized growth conditions vary with reactor designs and configurations. The growth of GeSn material detailed in the previous reports showed the material properties of the grown films with limited description of growth conditions, however no result has yet to demonstrate the entire development pathway to high quality GeSn material. The goal of this work was to detail an optimization methodology for the development of high-quality GeSn material on Ge buffered Si that could be potentially transferable across reactor design and configuration. The results from optimization of GeSn using a home-built UHV-CVD are presented as demonstration of the optimization process. The two dominating factors were growth temperature and SnCl4 flow fraction both of which were extensively explored across a wide range. Specifically, the GeSn growth window exploration began with the variation of temperature while the SnCl4 flow fraction was kept at a constant overpressure regime. A cloudy surface was observed for all GeSn depositions conducted under these conditions, which was due to Sn segregation on the surface. Two distinct Sn incorporation regions were determined. The next step was to decrease SnCl4 flow fraction while keeping the temperature constant, driving the GeSn growth window from SnCl4 overpressure regime to the optimized regime. With the decrease of SnCl4 flow fraction a mirror-like GeSn surface and substitutional Sn incorporation were achieved. After that, the growth temperature was further decreased while keeping the SnCl4 flow fraction at the optimized region. As a result, the Sn incorporation was further enhanced. Once the high-quality and high-Sn-content GeSn was obtained, photoconductor devices were fabricated with infrared imaging also being accomplished. It is noteworthy that, growth pressure is also an effective factor in reducing adatom mobility as well as enhancing GeH4 decomposition both of which are partially responsible for the formation of Sn droplets. However, in this paper, the growth was conducted at constant pressure and the role of pressure in Sn incorporation is not discussed. The effect of higher pressure growth towards atmospheric and sub-atmospheric growth regimes is a separate topic that would be investigated independently.
2. Experimental methods
It has been acknowledged that the combination of Ge hydrides and SnCl4 in the incorrect proportions can result in the severe etching effect, thus limiting the success of GeSn growth [21] The etching effect originates from the generation of chlorinated species during the reaction between Ge hydrides and SnCl4. The overall growth of GeSn is the competition between deposition and etching, both of which heavily depend on the supply of SnCl4 and growth temperature. In order to explore the optimal epitaxy window of GeSn, we designed the experiments by following three logical sequences. First of all, we grew the well-known low/high temperature two-step Ge buffer layer to accommodate the lattice mismatch between GeSn and Si [34,35]. Second, we demonstrated our previously reported GeSn growth [21] directly on Si substrate based on the initial exploration of low temperature Ge growth. The epitaxy of GeSn film on Si could be easily identified by visual inspection and subsequent material characterization. It also eases the efforts to determine the epitaxy window where deposition dominates over etching by visual contrast between GeSn and Si; Third, we grew high quality GeSn on a two-step Ge buffer layer. Based on the knowledge of GeSn growth on Si at the second step, we mainly focus on the optimization of two dominating factors for the effective Sn incorporation: SnCl4 supply and growth temperature. In this paper the growth was conducted at constant pressure and the role of pressure in Sn incorporation is not discussed. The effect of higher pressure growth towards atmospheric and sub-atmospheric growth regimes is a separate topic that would be investigated independently.
All material growth work was done using p-type Si (001) 4-inch wafers as substrates. The wafers were processed using standard piranha etch and HF dip prior to growth as described in our previous work [21]. Material growth was carried out in a cold wall UHV-CVD chamber with base pressures below 1×10−9 torr. The growth was conducted using GeH4, SnCl4 as precursors and Ar as the carrier gas. The Ge buffer layer with approximately 1 µm thickness was grown by low/high temperature two-step growth [34,35]. The first 100-nm Ge layer was grown as a seeding layer at 325°C temperature while the second 900-nm Ge layer was grown at 600°C. The GeSn epitaxy was initiated after Ge buffer was cooled down. During the GeSn growth the process pressure was fixed at 2 torr. A wide range of growth temperature and SnCl4 flow fraction were explored, as shown in Fig. 1. The flow fraction of SnCl4 was defined as $F_{SnCl_4}/\left( {F_{SnCl_4} + F_{GeH_4} + F_{Ar}} \right)$, where the ${F_{SnCl_4}}$, ${F_{GeH_4}}$ and ${F_{Ar}}$ indicate the flow rate of SnCl4, GeH4 and Ar, respectively. Three groups of GeSn were grown on Ge buffer:
Group 1: From sample A to G the growth temperature decreases from 325 to 240°C while the SnCl4 flow fraction was fixed at 2.9×10−3. The growth time of samples A, B, C, D, E1, F, and G was 30 mins while the growth time of sample E2 was 60 mins.
Group 2: From sample E2, H, I, and J, the SnCl4 flow fraction decreased from 2.9×10−3 to 2.3×10−4 while the growth temperature was fixed at 270°C. The corresponding growth time for these samples was 60 mins.
Group 3: From sample J, K and L, the growth temperature decreased from 270 to 250°C while the SnCl4 flow fraction was fixed at 2.3×10−4. The corresponding growth time was 60 mins.
The initial characterization of GeSn films is visual inspection to observe the mirror like or milky surface, showing possible Sn segregation. Following the visual inspection, the films were characterized by three techniques: 1) Raman Spectroscopy to determine the crystallization of GeSn; 2) Spectroscopic Ellipsometry to measure the thickness and absorption coefficient of GeSn; and 3) Photoluminescence (PL) to find out the optical properties of the GeSn. The PL emission was analyzed by using standard off-axis configuration and lock-in techniques with a chopping frequency of 40 Hz. The PL signal was collected using a grating-based spectrometer supplemented with a thermoelectrically cooled PbS detector which has a cutoff at 3.0 µm. A 1064 nm pulsed laser was initially used as the pumping source with 45 kHz repetition rate and 6 ns pulse duration. The laser spot size was measured as 52 µm in diameter. Using a high carrier injection per pulse ensures sufficient PL emission from GeSn. GeSn samples exhibiting strong PL were further characterized using a 532 nm continuous wavelength (CW) laser. The 1064 nm pulse pump laser had an average power density of 6 kW/cm2 and a peak power density of 2.7 × 104 kW/cm2. The 532 nm pump laser had a power density of 15 kW/cm2. The film strain, Sn composition, and crystallinity were then evaluated using X-ray diffraction (XRD) rocking curves and reciprocal space mapping (RSM). The XRD was performed using a Philips X’pert MRD system equipped with a standard four-bounce Ge (220) monochromator and a three bounce (022) channel cut Ge analyzer crystal along with the 1.6 kW Cu Kα1 X-ray tube with vertical line focus. Table 1 summarized the Sn compositions and compressive strains extracted from the XRD for all the samples.
Table 1. The summary of Sn compositions, strains and PL peak positions for all the samples.
The PL peak positions were also extracted from PL spectra as a comparison between different techniques. The surface morphology of GeSn was further inspected by scanning electron microscopy (SEM). Film thicknesses of the typical samples were also cross-checked by the cross-sectional transmission electron microscopy (TEM). The TEM images were generated using an aberration corrected Titan 80-300 with a Schottky field emission gun (FEG) operating at 300 kV.
3. Results and discussion
3.1 GeSn growth in the SnCl4 overpressure regime
Visual inspection of the wafer after growth provides useful information about the film surface. Diffusion of light reflecting off the surface is quite noticeable by the human eye and can serve as an indicator of surface roughness. For this work, the film surfaces were divided into three categories: i) the sample surface was milky, in which visual reflections were not easily visible, ii) the sample surface was hazy, such that, visual reflections were easily visible but not clear, and iii) the sample surface was clear or mirror-like shining. Based on the subsequent SEM measurement, different surfaces correspond to the size of Sn droplets seen on the surface (milky: ∼ 3 µm, hazy: < 1 µm, and mirror-like: no detectable droplets. The droplet density on the milky and hazy surfaces were less than 0.5 and more than 1 droplet per square micrometer, respectively. Optical and SEM images from these classifications are given in section C. Successive growths were conducted to verify the dependence of Sn incorporation as a function of deposition temperature [35]. The Sn droplet formation could be best explained by the Ostwald ripening mechanism [36] where inhomogeneous small particles dissolve over time and redeposit into larger particles. The formation of the Sn droplets thus leads to suppressing the effective Sn incorporation into Ge. These temperatures were varied from our starting point determined by the growth of previous materials described in Reference 22. In this step of the optimization process, all the film surfaces were milky as Sn droplets dominated the surface of the film. Room temperature PL from the first group of films (A-G) is shown in Fig. 2.
Fig. 2. PL spectra for samples of the growth temperature dependent experiment. The spectra are normalized and stacked to easily show the shifting peaks. For the Ge sample, the spectra was fitted with two Gaussian peaks for direct transitions (green line) and the indirect transition (red line). For GeSn samples (Samples A-G), two Gaussian peaks were fitted for two layers: Low-Sn-content GeSn (red line) and high-Sn-content GeSn (green line). The overall peak fit (blue line) was shown for eye guidance.
In Fig. 2, two peaks were evident in the PL spectra below a growth temperature of 325 °C. The PL signal was weak and very noisy for all samples excluding the Ge bulk reference. On the PL spectra only direct bandgap emission was observed while the indirect bandgap emission was not detected. The inefficient indirect bandgap recombination in GeSn is suppressed by the fast non-radiative recombination induced by defects. The majority of photon-pumped electrons accumulated at the indirect L valley are recombined with holes via the non-radiative Shockley-Read-Hall (SRH) process. Two peaks were fitted on PL spectra with Gaussian, suggesting the direct emissions from two different GeSn layers. The long wavelength peak (>1800nm) red shifted and reduced in intensity as the growth temperature was reduced to 260°C, corresponding to the direct emission of high-Sn-content GeSn. There was no PL evident at temperatures below 260°C. The extension of the wavelength indicated higher Sn incorporation, thus confirming the growth behavior matched what was expected. The short wavelength peak (∼1600 nm) shifts with the decreased growth temperature and does not match with Ge buffer emission, suggesting a separate layer of low composition Sn could be present, which is further confirmed by other subsequent characterizations.
Figure 3 shows a XRD comparison of the different samples grown in the SnCl4 overpressure regime. The peaks observed at ∼69°, 66° and those less than 66° correspond to the Si substrate, Ge buffer and GeSn respectively. The shift of GeSn peaks to lower angles with reduction of growth temperature agreed with the extension of PL peak wavelength, indicating increased Sn incorporation. The sample grown at 325°C showed only two layers present in the material however, below 325°C another peak appeared and remained throughout the rest of the temperature reduction. The presence and approximate position of this peak did not change with growth temperature reduction. The shift in the low Sn content peak, shown in Fig. 3 in the shaded area, from Ge suggested the layer contained 1-2 at. % Sn. At 325 °C the GeSn peak bordered along the shaded region and maintained intensity and linewidth like the Ge buffer. Upon reduction of the growth temperature below 325 °C, the peak corresponding to high Sn content reduced in intensity and broadened to only a shoulder by 270 °C and was no longer visible at 240 °C. Selected samples from the first batch were investigated further by XRD-RSM and TEM to better understand the origin of the low Sn peak seen in XRD. The characterization of a selected sample that was representative of all samples measured is shown in Fig. 4.
Three areas were marked on the XRD RSM ($\bar{2}\bar{2}4$) of sample E1 (grown at 270 °C for 30 min), in Fig. 4(a), 1) a Ge buffer peak displaying tensile strain, that was measured at 0.02%, 2) a psuedomorphic layer peak with low Sn composition, (1.6% Sn), 3) A broad high Sn composition peak (7.9% Sn). The two layers of GeSn were grown psuedomorphically on Ge buffer, corresponding the high compressive strain of 0.057% and 0.84%, respectively. The Sn compositions of two-layer GeSn were then extracted from the rocking curve of XRD (004) shown in Fig. 3(c). The broad peak of 7.9%-Sn-content GeSn is likely due to the disordered crystal structure introduced by excessive Sn incorporation into Ge in the SnCl4 overpressure regime.
Fig. 3. XRD (004) patterns for selected samples of group 1 from growth-temperature dependent experiment. Arrows indicate the growth temperature dependent region of GeSn while the shaded area corresponds to 1-2% Sn layer.
Fig. 4. (a) XRD RSM ($\bar{2}\bar{2}4$) of Sample E1 (grown at 270 °C for 30 min) showing the Ge buffer and GeSn peaks. Dashed line show the relaxation for psuedomorphic growth (R = 0), (b) Corresponding dark field TEM image of an area with a Sn droplet. (c) EDS point measurement from Sn droplet shows no Ge incorporation. The inset shows the position of the EDS measurement on the droplet.
TEM imaging of this sample is shown in Fig. 4(b). A Ge buffer thickness was observed with a thickness of 1035 nm which was near the 1 µm targeted buffer thickness demonstrating good growth control. Also visible in Fig. 4(b) is the GeSn layer with a large Sn droplet on the surface which had a diameter of > 1 µm. The Energy Dispersive X-Ray Spectroscopy (EDS) scan of the droplet shows pure Sn composition [Fig. 4(c)]. Further EDS scan near the droplet shows slight increase in the Sn composition before the sudden jump to 100% in the droplet. The mechanism through which the Sn droplets form from GeSn films could be found elsewhere [37]. The Sn droplet is expected to have a β-Sn crystal structure as discussed in Ref. [38]. However, further XRD scan did not show an extra peak around 64 degrees for Sn. This is attributed to formation of polycrystalline structure, which does not contribute to a high-count peak in comparison with single crystal GeSn peak.
PL experiments on these films consisted of two peaks in the spectra with a low Sn peak and a high Sn peak. The (004) XRD rocking curve characterization showed that the layer consisted of a low Sn layer with a high Sn tail supporting that two different compositions made up the films. One possible explanation for these results is that the bulk of the film has 1% Sn content, which is covered with Sn droplets and a small amount of Sn has inter-diffused into the layer near the droplet. The RSM clearly shows the 1% Sn layer with some extension that could correspond to the high Sn tail from the rocking curves and the TEM clearly shows the pure Sn droplet on the surface. These pieces of evidence further support the high Sn peak in the PL and rocking curves were localized near the Sn droplet.
3.2 Role of SnCl4 flow ratio in enhancing the material quality
To fully understand the effect of SnCl4 in the growth of GeSn, the ratio of SnCl4 gas flow was calculated and showed that the first batch was grown under a SnCl4 flow fraction of 2.9× 10−3. This value was two orders of magnitude above that which has been reported in the growth of high quality GeSn [24]. Therefore, a series of growths were conducted with reduced SnCl4 gas supply fraction. These growths were accomplished using the same temperature, pressure, and time (samples H, I, and J). The precursor and carrier gas flow rates were adjusted to reduce the SnCl4 partial pressure. The SnCl4 gas supply ratio was reduced by an order of magnitude from the starting condition until a condition that produced a clear mirror like surface was obtained. The initial growth condition (Sample E2) showed Sn surface segregation and by SnCl4 reduction resulted in the elimination of surface Sn and clear films. From sample E2 to J, the growth condition changes from SnCl4 overpressure regime (sample H) to the SnCl4 underpressure regime (sample J). As a result, the sample surface evolves from milky surface (sample E2) to mirror like surface (sample J) with the decrease of Sn segregation on the surface. The same characterization philosophy used in the first batch of samples was applied to this group with visual inspection occurring upon removal from the growth chamber. Optical characterizations including PL and XRD were also accomplished. Visual imaging, SEM imaging, and PL results are contained in Fig. 5.
Fig. 5. Visual images of samples E2, H, I and J are shown in (a1), (b1), (c1) and (d1), respectively. From sample E2 to J the flow fraction of SnCl4 decreases from 2.9×10−3 to 2.3×10−4 while the temperature was fixed at 270°C. Room-temperature PL of samples E2, H, I and J are shown in (a2), (b2), (c2) and (d2), respectively. The inset: The SEM images exhibit the cloudy surface (sample E2), hazy surface (sample H) and mirror surface (sample J), respectively.
Visual observation of the sample surface shown in Fig. 5(a1)-5(d1) gave evidence of successful Sn droplet reduction. Pieces were taken of Samples E2, H, I, and J, for other measurements before imaging was accomplished. It can be seen in the initial state the sample surface (E2) was milky and sample H wafer surface is still hazy indicating Sn droplet formation on the surface which covers the entire wafer. As the SnCl4 molar flow fraction reduced, the cloudiness of the sample surface cleared up and sample optical quality improved. Visual observation of the samples showed that by sample I the wafer center was clear with only the outer edge being hazy. Sample I is under the transition mode between SnCl4 over- and under-pressure regime. Sample J showed mirror-like sample surface across the full wafer. This showed that a window exists for the growth with the upper limit on SnCl4 molar flow fraction, from sample I, was 4.5 × 10−4. Sample J showed a clearer sample surface; so, the upper limit SnCl4 molar flow fraction was set to show where the surface Sn was removed, and the wafer surface began to clear. This did not verify that all the Sn was removed from the surface. A more in-depth look at the sample surface was accomplished using a scanning electron microscope (SEM). The inset of Fig. 5(a2)-5(d2) shows the SEM imaging of the sample surface of selected samples with separate levels of cloudiness.
In the inset of Fig. 5(a2), it is shown that the droplet sizes ranged up to 3 µm in diameter. Another interesting feature of the surface in the inset of Fig. 5(a2) was that there appeared to be pits in the surface where Sn droplets looked to have been located. This could be explained by the agglomeration of surface Sn which attracted neighboring Sn thus, forming the large clumps on the surface and leaving behind the pits. The hazy surface shown in the inset of 5(b2) shows an accumulation of Sn droplets on the surface much like the surface from the initial state. However, the size of the droplet had decreased in size to be < 1 µm for the largest of droplets. It can also be seen in this figure that the pits left over were also reduced in size, supporting that it was left behind after Sn agglomerates on the sample surface. The inset of Fig. 5(d2) shows a clear surface with only a single droplet to be seen in the image. To further gauge improvement of optical quality, PL on the samples was completed and shown in Fig. 5(a2)-5(d2).
In Fig. 5(a2) and 5(b2), the first two samples (E2 and H) have PL signal similar in intensity and line shape indicating very little change in the quality of the grown films. The extra peak seen in Fig. 5 (a2) from sample E2 was due to the second order diffraction of the 1064 nm pumping laser through the grating based. This peak is less obvious and fades away as the intensity of the PL increases in samples H-J. Sample I in Fig. 5(c2) showed 4 times increase over sample E2 in PL intensity and red-shift of PL peak wavelength. This suggests increased incorporation with the Sn surface droplet reduction and higher optical quality of the sample. Sample J in Fig. 5(d2) showed a 6-time increase in PL intensity over that of sample E2, indicating higher optical quality than sample I; However, the PL peak is shifted back toward shorter wavelengths. This was attributed to the combination of Sn composition and compressive strain in sample J which was further discussed in XRD section. Additionally, the reduction of defects causes the defect states that are lower than bandgap (contributed for longer wavelength emission) to reduce and that causes a sharper peak and slightly blue shift of the peak.
The (004) XRD rocking curves contained in Fig. 6(a), show similar peak position for each sample regardless of the flow fraction of SnCl4 or GeH4. Sample E2 had a similar shoulder in the XRD located at the same angle as samples I and J as it was explained in Section B. The XRD peaks of sample I and J were extracted as 65.1° and 65.14°, respectively. More intense peaks from sample I and J and the XRD peak disappearance of the 1% Sn-content GeSn shows high material quality and suggests that the Sn has incorporated into the growing film instead of agglomerating on the surface. From the analysis of XRD data, Composition of the sample J was determined from XRD-RSM ($\bar{2}\bar{2}4$) shown in Fig. 6(b), with the GeSn film comprising 5.6% Sn and -0.51% strain. The Ge buffer was found to be under 0.02% tensile strain. The Sn composition and strain of sample I were also obtained as 6.0% and -0.58%, respectively. The Sn composition of sample I exhibits higher Sn incorporation than sample J, which could be explained as follows. From sample I to J the growth condition evolves from SnCl4 overpressure regime to underpressure regime. In the SnCl4 underpressure regime, the Sn incorporation was heavily affected by both SnCl4 supply and growth temperature. At the fixed growth temperature, the Sn incorporation was primarily limited by the SnCl4 supply. Comparing with sample I, the SnCl4 flow fraction of sample J decreases in half. As a result, the Sn incorporation of sample J is smaller than sample I. The dark field TEM image of sample J was shown in Fig. 6(c), in which no Sn segregation was observed on the surface. The GeSn film also exhibited no extended threading dislocations penetrating through the film. Improvement in the growth conditions allowing for high-quality growth sets the stage for continued GeSn growth to further increase the Sn incorporation.
Fig. 6. (a) (004) XRD rocking curves of Sn reduction tests (samples I and J). The GeSn film thickness of 228 and 310 nm for samples I and J, respectively, was derived using spectroscopic ellipsometry (not shown); (b) XRD-RSM ($\bar{2}\bar{2}4$) for sample J. Dashed lines show the relaxation when R = 0 for psuedomorphic growth, while R = 1 is for relaxed growth; (c) Dark field TEM image of sample J. No Sn droplets were observed on the GeSn surface.
3.3 Role of temperature in increasing Sn incorporation
The prior section showed improvements in the Sn incorporation and a reduction/elimination of Sn droplets on the surface by optimizing the SnCl4 flow. The GeSn films with effective Sn incorporation and mirror-like surfaces have been achieved by decreasing the Sn dilution ratio down to the gas flow limit of mass flow controller. In order to further increase the Sn incorporation, the growth temperature was decreased while maintaining the Sn dilution ratio at 2.3×10−4. By visual inspection after GeSn epitaxy, the surface morphology of GeSn film changes with decreasing growth temperature from 270 to 250°C. As shown in Fig. 7(b) inset, the majority of GeSn surface was mirror-like without the significant Sn droplet segregation on the surface at the temperature of 270°C. When the temperature drops to 260 °C, the hazy area started growing from the edge to center as the ring on the surface while the GeSn surface at the center is clear as shown in Fig. 7(c) inset. The hazing area is attributed to the aggregated Sn droplet on the GeSn surface. As the temperature continues decreases to 250 °C, the hazy area continued growing and covered the majority of the surface, leaving the clear surface only sparsely at the center of the surface as shown in Fig. 7(d) inset. The segregation of Sn and droplet formation was attributed to both increased Sn incorporation and reduction in the reactivity of GeH4 at lower temperatures. Lower GeH4 reactivity resulted in less Ge crystal formation to host further incorporated Sn atoms and the excessive Sn would precipitate on the surface.
Fig. 7. Room-temperature PL spectra of (a) reference sample, (b) sample J, (c) sample K and (d) sample L, respectively. The 1064 pulsed laser was used as the pumping source for reference sample and samples J, K and L while the 532 nm continuous wave laser was used for sample K as comparison. The samples J, K and L were grown at the growth temperature of 270, 260 and 250°C, respectively. The inset: visual images of sample J, K and L.
Room-temperature PL spectra was conducted in order to study the optical properties of GeSn films. As shown in Fig. 7, the PL spectrum of a GeSn sample grown by a commercial reactor is shown as a reference for comparison with our CVD-grown GeSn samples from J to L. The GeSn reference sample was grown on Ge buffered Si substrate by using the ASM Epsilon 2000 Plus reduced pressure CVD (RPCVD) system with commercially available precursors of GeH4 and SnCl4. The film thickness of GeSn film is 83 nm. The corresponding Sn composition and compressive strain of commercial sample were 5% and -0.52%, respectively. On the PL spectra of Fig. 7, the pumping power of the 1064 nm laser for PL emission was 300 mW and all other PL setup conditions between GeSn reference and samples J to K were kept the same. From Fig. 7(a) to 7(d) the direct bandgap emissions were observed for all the samples. From Fig. 7(a) of GeSn reference, the wavelength of PL peak was 1916nm. From Fig. 7(b) to 7(d) of sample J to L, the PL peak wavelength shifted from 1924 to 2072nm when the growth temperature decreased from 270 to 250 °C. The pronounced red shift from sample J to K suggested increased Sn incorporation into Ge matrix when the temperature was decreased. The GeSn reference sample and the samples of this study both showed similar PL characteristics. The FWHMs of reference sample and sample J are 88.5 meV (264 nm), 96 meV (288 nm), respectively. The 532 nm continuous-wave (CW) pumping laser was also used for the PL characterization of sample K, as shown in Fig. 7(c). The pumping power of 532 nm laser was 500 mW and the beam size was 100 µm in diameter. The peak power density per pulse of the 1064 nm pumping laser was calculated as four orders of magnitude higher than that of the 532 CW laser and the observation of PL from 532 nm laser pumping indicated achieving higher quality material. The PL comparison study further confirmed the significant improvement of optical quality of our grown GeSn samples as PL emission was not observed for other samples with 532 nm laser.
The XRD was conducted in order to investigate the Sn incorporation and compressive strain of the GeSn films. The rocking curve of the reference GeSn is shown in Fig. 8 as well. The peaks of Si substrate, Ge buffer, and GeSn film were clearly resolved in the rocking curves of GeSn reference and samples J, K, and L. From Fig. 8(a), the GeSn peak of reference was located at 65.08°. Figure 8(b) to 8(d) shows that GeSn peaks of sample J to L were shifted from 65.14° to 64.66° which indicated the increase of Sn incorporation when temperature dropped from 270 to 250 °C. The XRD results were consistent with the PL analysis showing the peak shift toward to longer wavelengths as the temperature was reduced. Comparing the GeSn reference and sample J, the GeSn peaks were very close at 65.08° and 65.14°, respectively. The FWHMs of GeSn reference and sample J was Gaussian fitted at 0.19° and 0.20°, suggesting the material quality of our grown GeSn was getting close to that of the GeSn reference. It is worth mentioning that at the growth temperature of 250°C a new peak between the main peaks of GeSn and Ge emerged and was attributed to the GeSn peak with ∼1% Sn composition. As the Sn droplets continued growing over the surface at the growth temperature of 250°C, the GeSn with ∼ 1% Sn composition reappeared, which is in agreement with the analysis in section B.
Fig. 8. The XRD rocking curves (004) of (a) reference sample, (b) sample J, (c) sample K and (d) sample L, respectively. The Si, Ge and GeSn peaks were marked in XRD rocking curves, respectively.
3.4 Prototype GeSn photoconductors
To further evaluate the optical properties of the GeSn samples grown in this study, GeSn, photoconductors were fabricated from selected films. Device characteristics including, I-V, responsivity (not shown), detectivity, and infrared imaging were obtained with selected results shown in Fig. 9.
Fig. 9. (a) I-V curves for interdigitated GeSn photoconductors. Inset shows measured device; (b) Calculated D* for the GeSn photoconductors. Upper inset is a visible image of the object. Lower inset is infrared image generated using the GeSn photoconductor.
The GeSn films were fabricated into photoconductor devices such as the one shown in Fig. 9(a). Photoconductors were chosen due to the simplicity of the device and ease of fabrication. Interdigitated structures were fabricated to allow higher collection efficiency over that of coplanar devices, with an image shows as inset of Fig. 9(a). The I-V characteristics measured at both 77 and 300 K are shown in Fig. 9(a). At each temperature the I-V showed linear I-V characteristics suggesting good ohmic contacts. The dark current at 300 K was 40 mA but reduced to below 10 mA at 77 K.
The specific detectivity (D*) was calculated for the fabricated GeSn photoconductor and plotted, in Fig. 9(b), against commercial extended InGaAs detectors, and a 10% GeSn photoconductor that has been previously reported with material grown by commercial reactor. [39]. It was seen that at 300 K the D* of sample K had similar detectivity to our previously reported GeSn photoconductor with shorter cutoff wavelength, 2.25 and 2.5 µm respectively. However, at 77 K the 10% GeSn device outperformed the sample K by just over an order of magnitude showing there is still room to improve the material quality. In Fig. 9(b) insets, a visible image (upper inset) and an infrared image of an object (golden star) was demonstrated using the photoconductor fabricated from sample K. Imaging was accomplished by using a single photoconductor being moved in an array pattern to generate the image. White light was used to illuminate the 6 mm star and a 1600 nm long pass filter was used to filter out non-infrared light. Generating an image from the grown films shows the potential of GeSn to be used for future low cost infrared imaging applications
4. Summary and conclusion
The experimental results discussed in this work indicate that during GeSn epitaxy the SnCl4 flow fraction has the supply limit for each growth temperature. The high growth temperature corresponds to the high supply limit of SnCl4 and vice versa. Exceeding the SnCl4 supply limit leads to the Sn precipitation on the surface and the deterioration of GeSn material. Additionally, at low growth temperature GeH4 reactivity decreases and the bottom line of the growth window is also limited to Ge crystal formation to host incorporated Sn atoms. The key to achieve optimized growth regime in which high quality film and high Sn incorporation are achieved through proper parameter matching between SnCl4 flow fraction and growth temperature. In the optimized regime, the Sn precipitation on the surface is significantly suppressed and Sn atoms are effectively incorporated into Ge lattice matrix.
For Group 1 samples, the GeSn growth was conducted in a SnCl4 overpressure regime with variable growth temperatures. The excess Sn segregated on the surface and developed as droplets due to their low surface free energy. The subsequent Sn atoms supplied by the thermal decomposition of SnCl4 tend to be attracted and absorbed by the high-volume Sn droplets instead of being buried by Ge atoms. The Sn droplets continue growing on the surface via the Ostwald ripening mechanism [36]. In this case, only 1-2% Sn atoms could be incorporated into the Ge lattice matrix, which is close to the Sn solubility (1%) into Ge at the equilibrium condition. High Sn incorporation only occurred in the vicinity of Sn-rich GeSn surface with limited material quality.
For group 2 samples, the SnCl4 flow fraction was gradually reduced while keeping the growth temperature at constant. With decreased SnCl4 supply, the sizes of Sn droplets on the surface became smaller. Eventually the GeSn growth entered into the optimized growth regime, where the flow fraction of SnCl4 and growth temperature reached a good match. Therefore, no appreciable Sn droplets developed on the surface, leading to more Sn atoms buried into the Ge. Meanwhile the material quality of GeSn was significantly enhanced.
For group 3 samples, the SnCl4 flow fraction was fixed while decreasing the growth temperature. It has been reported that in the optimized growth regime, the decrease of growth temperature has led to increased Sn incorporation, which was further verified in this work. However, the continuous decrease of temperature broke the good match between SnCl4 supply and growth temperature and shifted the growth condition into SnCl4 overpressure regime. As a result, the Sn atoms started segregating on the surface, which again jeopardized the effective Sn incorporation and material quality. Additionally, the low temperature window was only available until GeH4 reactivity was appropriate for crystal formation in order to accommodate the incorporated Sn in the crystal sites.
In conclusion, this paper has discussed a systematic methodology to pursue growth of high quality GeSn on Ge buffered Si using GeH4 in a home-built UHV-CVD system while most growth techniques use either Ge2H6 precursor or a commercial CVD system. The study started in a growth regime in which SnCl4 overpressure dominated the overall growth of the GeSn material. The detrimental effects of SnCl4 overpressure lead to reduction of the SnCl4 molar flow fraction. This reduction in SnCl4 molar flow fraction provided better growth conditions to allow Sn to incorporate into the growing film and less agglomeration on the sample surface. Optimal growth conditions were achieved for the 270 °C growth temperature, producing high quality GeSn with mirror-like surface. Furthermore, temperature reduction below 270 °C resulted in increasing surface Sn with decreasing GeH4 breakdown. The GeH4 and SnCl4 ratio used for Samples I, J and K was 40, however the carrier gas diluted the overall ratio of SnCl4 in the system to 2.3 × 10−4. Prototype GeSn photoconductors were fabricated and used to obtain an infrared image showing the great potential for materials to be used for future device applications.
Funding
Air Force Office of Scientific Research (AFOSR) (FA9550-14-1-0205, FA9550-16-C-0016); National Aeronautics and Space Administration (NASA) (NNX15AN18A); U.S. Department of Defense (DOD) Basic Research Funding (W911NF1910004).
Acknowledgments
The authors acknowledge Dr. M. Benamara’s assistance in TEM imaging and Dr. A. Kuchuk’s assistance in XRD measurement from Institute for Nanoscience & Engineering, University of Arkansas.
References
1. R. Soref, D. Buca, and S.-Q. Yu, “Group IV Photonics: Driving Integrated Optoelectronics,” Opt. Photonics News 27(1), 32 (2016). [CrossRef]
2. S.Q. Yu, S.A. Ghetmiri, W. Du, J. Margetis, Y. Zhou, A. Mosleh, S. Al-Kabi, A. Nazzal, G. Sun, R.A. Soref, J. Tolle, B. Li, and H. A. Naseem, “Si based GeSn light emitter: mid-infrared devices in Si photonics,” Proc. SPIE 9367, Silicon Photonics X 9367, 93670R (2015).
3. S. Gupta, R. Chen, B. Magyari-Kope, H. Lin, B. Yang, A. Nainani, Y. Nishi, J.S. Harris, and K.C. Saraswat, “GeSn technology: Extending the Ge electronics roadmap,” Tech. Dig. - Int. Electron Devices Meet. IEDM 398 (2011).
4. S. Al-Kabi, S. A. Ghetmiri, J. Margetis, T. Pham, Y. Zhou, W. Dou, B. Collier, R. Quinde, W. Du, A. Mosleh, J. Liu, G. Sun, R. A. Soref, J. Tolle, B. Li, M. Mortazavi, H. A. Naseem, and S. Q. Yu, “Systematic study of GeSn heterostructure-based light-emitting diodes towards mid-infrared applications,” Appl. Phys. Lett. 109(17), 171105 (2016). [CrossRef]
5. 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, “Lasing in direct-bandgap GeSn alloy grown on Si,” Nat. Photonics 9(2), 88–92 (2015). [CrossRef]
6. W. Du, Y. Zhou, S. A. Ghetmiri, A. Mosleh, B. R. Conley, A. Nazzal, R. A. Soref, G. Sun, J. Tolle, J. Margetis, H. A. Naseem, and S. Q. Yu, “Competition of optical transitions between direct and indirect bandgaps in Ge1−xSnx,” Appl. Phys. Lett. 105(5), 051104 (2014). [CrossRef]
7. S. Stange, R. Wirths, C. Geiger, B. Schulte-Braucks, N.V. Marzban, G. Den Driesch, T. Mussler, T. Zabel, J.M. Stoica, S. Hartmann, Z. Mantl, D. Ikonic, H. Grützmacher, J. Sigg, D. Witzens, and Buca, “Optically Pumped GeSn Microdisk Lasers on Si,” ACS Photonics 3(7), 1279–1285 (2016). [CrossRef]
8. W. Dou, Y. Zhou, J. Margetis, S. Ghetmiri, W. Du, J. Liu, G. Sun, R. Soref, J. Tolle, B. Li, M. Mortazavi, and S.Q. Yu, “Optically Pumped GeSn-edge-emitting Laser with Emission at 3 µm for Si Photonics,” Conference on Lasers and Electro-Optics (CLEO) , San Jose, CA (2018).
9. J. Margetis, S. Al-Kabi, W. Du, W. Dou, Y. Zhou, T. Pham, P. Grant, S. Ghetmiri, A. Mosleh, B. Li, J. Liu, G. Sun, R. Soref, J. Tolle, M. Mortazavi, and S.-Q. Yu, “Si-based GeSn lasers with wavelength coverage of 2–3 µm and operating temperatures up to 180 K,” ACS Photonics 5(3), 827–833 (2018). [CrossRef]
10. O. Gurdal, P. Desjardins, J. R. A. Carlsson, N. Taylor, H. H. Radamson, J. E. Sundgren, and J. E. Greene, “Low-temperature growth and critical epitaxial thicknesses of fully strained metastable Ge1−xSnx (x≲0.26) alloys on Ge(001)2×1,” J. Appl. Phys. 83(1), 162–170 (1998). [CrossRef]
11. R. Chen, H. Lin, Y. Huo, C. Hitzman, T. I. Kamins, and J. S. Harris, “Increased photoluminescence of strain-reduced, high-Sn composition Ge1−xSnx alloys grown by molecular beam epitaxy,” Appl. Phys. Lett. 99(18), 181125 (2011). [CrossRef]
12. E. Kasper, J. Werner, M. Oehme, S. Escoubas, N. Burle, and J. Schulze, “Growth of silicon based germanium tin alloys,” Thin Solid Films 520(8), 3195–3200 (2012). [CrossRef]
13. M. Coppinger, J. Hart, N. Bhargava, S. Kim, and J. Kolodzey, “Photoconductivity of germanium tin alloys grown by molecular beam epitaxy,” Appl. Phys. Lett. 102(14), 141101 (2013). [CrossRef]
14. P. Tao, L. Huang, H. H. Cheng, H.-H. Wang, and X.-S. Wu, “Epitaxial growth of Ge1–xSnx films with x up to 0.14 grown on Ge (00 l) at low temperature,” Chin. Phys. B 23(8), 088112 (2014). [CrossRef]
15. S. Kim, N. Bhargava, J. Gupta, M. Coppinger, and J. Kolodzey, “Infrared photoresponse of GeSn/n-Ge heterojunctions grown by molecular beam epitaxy,” Opt. Express 22(9), 11029 (2014). [CrossRef]
16. Y. G. Sadofyev, V. P. Martovitsky, A. V. Klekovkin, V. V. Saraikin, and I. S. Vasil’Evskii, “Thermal Stability of Ge/GeSn Nanostructures Grown by MBE on (001)Si/Ge Virtual Wafers,” Phys. Procedia 72, 411–418 (2015). [CrossRef]
17. J. Taraci, S. Zollner, M. R. McCartney, J. Menendez, M. A. Santana-Aranda, D. J. Smith, A. Haaland, A. V. Tutukin, G. Gundersen, G. Wolf, and J. Kouvetakis, “Synthesis of Silicon-Based Infrared Semiconductors in the Ge-Sn System Using Molecular Chemistry Methods,” J. Am. Chem. Soc. 123(44), 10980–10987 (2001). [CrossRef]
18. S. Wirths, D. Buca, A. T. Tiedemann, B. Hollander, P. Bernardy, T. Stoica, D. Grutzmacher, and S. Mantl, “Reduced Pressure CVD Growth of Ge and Ge1−xSnx Alloys,” ECS Trans. 50(9), 885–893 (2013). [CrossRef]
19. A. Mosleh, M. Alher, L. Cousar, H. Abu-safe, W. Dou, P. C. Grant, S. Al-Kabi, S.A. Ghetmiri, H. Tran, W. Du, M. Benamara, B. Li, M. Mortazavi, S.-Q. Yu, and H.A. Naseem, “Enhancement of Material Quality of (Si) GeSn Films Grown by SnCl4 Precursor,” ECS Trans. 69(5), 279–286 (2015). [CrossRef]
20. S. A. Ghetmiri, W. Du, J. Margetis, A. Mosleh, L. Couser, B. R. Conley, L. Domulevicz, A. Nazzal, G. Sun, R. A. Soref, J. Tolle, B. Li, H. A. Naseem, and S. Yu, “Direct-bandgap GeSn grown on Silicon with 2230 nm photoluminescence,” Appl. Phys. Lett. 105(15), 151109 (2014). [CrossRef]
21. A. Mosleh, M. A. Alher, L. C. Cousar, W. Du, S. A. Ghetmiri, T. Pham, J. M. Grant, G. Sun, R. A. Soref, B. Li, H. A. Naseem, and S.-Q. Yu, “Direct growth of Ge1− xSnx films on Si using a cold-wall ultra-high vacuum chemical-vapor-deposition system,” Front. Mater. 2, 30 (2015). [CrossRef]
22. A. Mosleh, M. Alher, L. C. Cousar, W. Du, S. A. Ghetmiri, S. Al-Kabi, W. Dou, P. C. Grant, G. Sun, R. A. Soref, B. Li, H. A. Naseem, and S. Q. Yu, “Buffer-Free GeSn and SiGeSn Growth on Si Substrate Using In Situ SnD4 Gas Mixing,” J. Electron. Mater. 45(4), 2051–2058 (2016). [CrossRef]
23. J. Margetis, S.-Q. Yu, N. Bhargava, B. Li, W. Du, and J. Tolle, “Strain engineering in epitaxial Ge1− x Sn x: a path towards low-defect and high Sn-content layers,” Semicond. Sci. Technol. 32(12), 124006 (2017). [CrossRef]
24. J. Margetis, A. Mosleh, S. Al-Kabi, S. A. Ghetmiri, W. Du, W. Dou, M. Benamara, B. Li, M. Mortazavi, H. A. Naseem, S. Q. Yu, and J. Tolle, “Study of low-defect and strain-relaxed GeSn growth via reduced pressure CVD in H2 and N2 carrier gas,” J. Cryst. Growth 463, 128–133 (2017). [CrossRef]
25. J. Tolle, A. V. G. Chizmeshya, Y.-Y. Fang, J. Kouvetakis, V. R. D’Costa, C.-W. Hu, J. Menéndez, and I. S. T. Tsong, “Low temperature chemical vapor deposition of Si-based compounds via SiH3 SiH2 SiH3: Metastable SiSn/GeSn/Si(100) heteroepitaxial structures,” Appl. Phys. Lett. 89(23), 231924 (2006). [CrossRef]
26. B. Vincent, F. Gencarelli, H. Bender, C. Merckling, B. Douhard, D. H. Petersen, O. Hansen, H. H. Henrichsen, J. Meersschaut, W. Vandervorst, M. Heyns, R. Loo, and M. Caymax, “Undoped and in-situ B doped GeSn epitaxial growth on Ge by atmospheric pressure-chemical vapor deposition,” Appl. Phys. Lett. 99(15), 152103 (2011). [CrossRef]
27. J. Kouvetakis and A. V. G. Chizmeshya, “New classes of Si-based photonic materials and device architectures via designer molecular routes,” J. Mater. Chem. 17(17), 1649 (2007). [CrossRef]
28. S. Wirths, D. Buca, G. Mussler, A. T. Tiedemann, B. Hollander, P. Bernardy, T. Stoica, D. Grutzmacher, and S. Mantl, “Reduced Pressure CVD Growth of Ge and Ge1−xSnx Alloys,” ECS J. Solid State Sci. Technol. 2(5), N99–N102 (2013). [CrossRef]
29. V. Reboud, A. Gassenq, N. Pauc, J. Aubin, L. Milord, Q. M. Thai, M. Bertrand, K. Guilloy, D. Rouchon, J. Rothman, T. Zabel, F. Armand Pilon, H. Sigg, A. Chelnokov, J. M. Hartmann, and V. Calvo, “Optically pumped GeSn micro-disks with 16% Sn lasing at 3.1 µm up to 180 K featured,” Appl. Phys. Lett. 111(9), 092101 (2017). [CrossRef]
30. J. Aubin, J. M. Hartmann, A. Gassenq, L. Milord, N. Pauc, V. Reboud, and V. Calvo, “Impact of thickness on the structural properties of high tin content GeSn layers,” J. Cryst. Growth 473, 20–27 (2017). [CrossRef]
31. C. L. Senaratne, J. D. Gallagher, L. Jiang, T. Aoki, D. J. Smith, J. Menéndez, and J. Kouvetakis, “Ge1- y Sn y (y = 0.01-0.10) alloys on Ge-buffered Si: Synthesis, microstructure, and optical properties,” J. Appl. Phys. 116(13), 133509 (2014). [CrossRef]
32. P. C. Grant, W. Dou, B. Alharthi, J. M. Grant, A. Mosleh, W. Du, B. Li, M. Mortazavi, H. A. Naseem, and S. Q. Yu, “Comparison study of the low temperature growth of dilute GeSn and Ge,” J. Vac. Sci. Technol., B: Nanotechnol. Microelectron.: Mater., Process., Meas., Phenom. 35(6), 061204 (2017). [CrossRef]
33. B. Alharthi, W. Dou, J. M. Grant, P. C. Grant, T. Morgan, A. Mosleh, W. Du, M. Mortazavi, B. Li, H. Naseem, and S.-Q. Yu, “Low temperature epitaxy of high-quality Ge buffer using plasma enhancement via UHV-CVD system for photonic device applications,” Appl. Surf. Sci. 481, 246–254 (2019). [CrossRef]
34. H.-C. Luan, D. R. Lim, K. K. Lee, K. M. Chen, J. G. Sandland, K. Wada, and L. C. Kimerling, “High-quality Ge epilayers on Si with low threading-dislocation densities,” Appl. Phys. Lett. 75(19), 2909–2911 (1999). [CrossRef]
35. J. Margetis, S. A. Ghetmiri, W. Du, B. R. Conley, A. Mosleh, R. A. Soref, G. Sun, L. Domulevica, H. A. Naseem, S.-Q. Yu, and J. Tolle, “Comparison study of the low temperature growth of dilute GeSn and Ge,” ECS Trans. 64(6), 711–720 (2014). [CrossRef]
36. R. D. Vengrenovitch, “On the ostwald ripening theory,” Acta Metall. 30(6), 1079–1086 (1982). [CrossRef]
37. H. Groiss, M. Glaser, M. Schatzl, M. Brehm, D. Gerthsen, D. Roth, P. Bauer, and F. Schäffler, “Free-running Sn precipitates: an efficient phase separation mechanism for metastable Ge 1− x Sn x epilayers,” Sci. Rep. 7(1), 16114 (2017). [CrossRef]
38. N. Oehl, G. Schmuelling, M. Knipper, R. Kloepsch, T. Placke, J. Kolny-Olesiak, T. Plaggenborg, M. Winter, and J. Parisi, “In situ X-ray diffraction study on the formation of α-Sn in nanocrystalline Sn-based electrodes for lithium-ion batteries,” CrystEngComm 17(44), 8500–8504 (2015). [CrossRef]
39. B. R. Conley, J. Margetis, W. Du, H. Tran, A. Mosleh, S. A. Ghetmiri, J. Tolle, G. Sun, R. Soref, B. Li, H. A. Naseem, and S. Q. Yu, “Si based GeSn photoconductors with a 1.63 A/W peak responsivity and a 2.4 µm long-wavelength cutoff,” Appl. Phys. Lett. 105(22), 221117 (2014). [CrossRef]
