High-quality GeSn Layer with Sn Composition up to 7% Grown by Low-Temperature Magnetron Sputtering for Optoelectronic Application

In this paper, a high-quality sputtered-GeSn layer on Ge (100) with a Sn composition up to 7% was demonstrated. The crystallinity of the GeSn layer was investigated via high-resolution X-ray diffraction (HR-XRD) and the strain relaxation degree of the GeSn layer was evaluated to be approximately 50%. A novel method was also proposed to evaluate the averaged threading dislocation densities (TDDs) in the GeSn layer, which was obtained from the rocking curve of GeSn layer along the (004) plane. The photoluminescence (PL) measurement result shows the significant optical emission (1870 nm) from the deposited high-quality GeSn layer. To verify whether our deposited GeSn can be used for optoelectronic devices, we fabricated the simple vertical p-i-n diode, and the room temperature current–voltage (I–V) characteristic was obtained. Our work paves the way for future sputtered-GeSn optimization, which is critical for optoelectronic applications.


Introduction
Recently, GeSn alloys have been shown the capability to become a real direct bandgap material with a Sn composition of 10% [1][2][3], which makes it a promising optical gain medium for group IV light sources [4][5][6]. Moreover, GeSn is also an attractive material owing to its compatibility with the mature Si complementary metal-oxide-semiconductor (CMOS) process. The epitaxial growth of high-Sn-composition GeSn with high material quality is challenging because the solid solubility of Sn in Ge or Ge in Sn is less than 1%. The smaller surface free energy of Sn compared to that of Ge makes Sn more likely to immigrate to the surface of the GeSn film during epitaxial growth and thermal treatment [7][8][9][10][11]. So far, several techniques such as chemical vapor deposition (CVD) [12][13][14], molecular beam epitaxy (MBE) [15][16][17], and magnetron sputtering [18][19][20][21][22] have been employed to achieve the crystalline GeSn layers.
Reduced pressure chemical vapor deposition (RPCVD) has grown high-quality GeSn with a Sn composition up to 12.6% using Ge 2 H 6 and SnCl 4 as precursors, thus achieving the first demonstration of an optically pumped GeSn laser [23]. Lasing from GeSn micro disks with a Sn composition up to 16% has also been achieved by using the same precursors [24]. At the same time, GeH 4 and SnCl 4 have been utilized to grow high-quality GeSn with a Sn composition up to 17%, also contributing to

Fabrication of Simple GeSn Diode
After the deposition of the GeSn layer, a B implant (dose of 4 × 10 15 cm −2 and energy of 40 keV) was performed to dope the 400 nm i-GeSn, followed by rapid thermal annealing (RTA) in N 2 to activate the p-type dopant concentration. Then, the p-i-n diode was processed via metallization and RTA treatment. The schematic cross-section of the layer structure of the sample is shown in Figure 1a. In order to form a simple GeSn p-i-n diode, we cut the sample into small pieces of 0.8 × 0.8 mm square by laser slicing. Then, the typical current-voltage (I-V) characteristic of the sample was measured by a Keithley 4200 semiconductor characterization system parameter analyzer (Tektronix, Beaverton, OR, USA). After the deposition of the GeSn layer, a B implant (dose of 4 × 10 15 cm −2 and energy of 40 keV) was performed to dope the 400 nm i-GeSn, followed by rapid thermal annealing (RTA) in N2 to activate the p-type dopant concentration. Then, the p-i-n diode was processed via metallization and RTA treatment. The schematic cross-section of the layer structure of the sample is shown in Figure  1a. In order to form a simple GeSn p-i-n diode, we cut the sample into small pieces of 0.8 × 0.8 mm square by laser slicing. Then, the typical current-voltage (I-V) characteristic of the sample was measured by a Keithley 4200 semiconductor characterization system parameter analyzer (Tektronix, Beaverton, Oregon, USA).

Characterization Method for GeSn Layer
The crystallinity quality of the GeSn layer was determined by high-resolution X-ray diffraction (HR-XRD) measurement. The Sn composition of the GeSn layer was verified by energy dispersive spectrometry (EDS). The XRD 2θ-ω scans along the (004) and (224) planes were accomplished. The rocking curve of GeSn (004) was also found to determine the averaged threading dislocation densities (TDDs) in the GeSn layer. PL was used to ascertain the band gap energy. The photoluminescence (PL) setup consisted of a Fourier transform infrared spectroscopy, liquid nitrogen (LN2)-cooled InGaAs detector, a 532 nm continuous wave (CW) laser, and a grating monochromator. Figure 2a shows the XRD 2θ-ω (004) scan of the sample, in which the diffraction peak of GeSn and Ge can be seen. The peaks at 66° and 65.18° correspond to the Ge substrate and GeSn layer, respectively. The out-of-plane lattice constant of the GeSn layer was extracted from the following expression:

Crystallinity, Sn Content, and Strain of GeSn Layer
where λ represents the wavelength of Cu K1 (λ = 1.5406 nm) and θ004 is the diffraction peak of the GeSn layer along the (004) plane. Hence, the out-of-plane lattice constant of the GeSn layer is calculated to be 0.5721 nm. Figure 2b shows the XRD 2θ-ω (224) scan of the sample in which the diffraction peaks of GeSn and Ge can also be clearly seen. The peaks at 83.1° and 82.98° correspond to the Ge substrate and the GeSn layer, respectively. The in-plane lattice constant of the GeSn layer can be reckoned as:

Characterization Method for GeSn Layer
The crystallinity quality of the GeSn layer was determined by high-resolution X-ray diffraction (HR-XRD) measurement. The Sn composition of the GeSn layer was verified by energy dispersive spectrometry (EDS). The XRD 2θ-ω scans along the (004) and (224) planes were accomplished. The rocking curve of GeSn (004) was also found to determine the averaged threading dislocation densities (TDDs) in the GeSn layer. PL was used to ascertain the band gap energy. The photoluminescence (PL) setup consisted of a Fourier transform infrared spectroscopy, liquid nitrogen (LN 2 )-cooled InGaAs detector, a 532 nm continuous wave (CW) laser, and a grating monochromator. Figure 2a shows the XRD 2θ-ω (004) scan of the sample, in which the diffraction peak of GeSn and Ge can be seen. The peaks at 66 • and 65.18 • correspond to the Ge substrate and GeSn layer, respectively. The out-of-plane lattice constant of the GeSn layer was extracted from the following expression:

Crystallinity, Sn Content, and Strain of GeSn Layer
where λ represents the wavelength of Cu K1 (λ = 1.5406 nm) and θ 004 is the diffraction peak of the GeSn layer along the (004) plane. Hence, the out-of-plane lattice constant of the GeSn layer is calculated to be 0.5721 nm. Figure 2b shows the XRD 2θ-ω (224) scan of the sample in which the diffraction peaks of GeSn and Ge can also be clearly seen. The peaks at 83.1 • and 82.98 • correspond to the Ge substrate and the GeSn layer, respectively. The in-plane lattice constant of the GeSn layer can be reckoned as: where d 224 is the crystal spacing. For the XRD scan along the (224) plane, d 224 can be calculated using the following equation: Therefore, d 224 of the GeSn layer is estimated to be 0.1162 nm. Substituting the value of d 224 into Equation (2), the in-plane lattice constant of the GeSn layer is calculated to be 0.5634 nm.
where d224 is the crystal spacing. For the XRD scan along the (224) plane, d224 can be calculated using the following equation: (3) Therefore, d224 of the GeSn layer is estimated to be 0.1162 nm. Substituting the value of d224 into Equation (2), the in-plane lattice constant of the GeSn layer is calculated to be 0.5634 nm. To confirm the Sn composition for the GeSn layer, we performed the surface EDS spectra of the GeSn layer. The spectra image was obtained from typical EDS analysis. Figure 3 shows that seven primary peaks were formed at 0.4, 1.3, 3, 3.4, 3.6, 4.4, and 9.86 keV. The peaks at 0.4, 1.3, 3, 3.6 and 4.4 keV match the spectral lines of Sn. Furthermore, the peaks at 1.3 and 9.86 keV match the spectral lines of Ge. Thus, the Sn composition of the GeSn layer is estimated to be 7%. Based on the lattice elastic theory, the bulk lattice constant of GeSn ( bulk a ) can be obtained using the following equation: We can first substitute the Sn composition (7) obtained from EDS into Equation (5), and the value of C11/C12 is calculated to be 0.0966. bulk a can also be calculated using the in-plane lattice To confirm the Sn composition for the GeSn layer, we performed the surface EDS spectra of the GeSn layer. The spectra image was obtained from typical EDS analysis. Figure 3 shows that seven primary peaks were formed at 0.4, 1.3, 3, 3.4, 3.6, 4.4, and 9.86 keV. The peaks at 0.4, 1.3, 3, 3.6 and 4.4 keV match the spectral lines of Sn. Furthermore, the peaks at 1.3 and 9.86 keV match the spectral lines of Ge. Thus, the Sn composition of the GeSn layer is estimated to be 7%. Based on the lattice elastic theory, the bulk lattice constant of GeSn (a bulk ) can be obtained using the following equation: where a ⊥ is the out-of-plane lattice constant of the GeSn layer, a // is the in-plane lattice constant of the GeSn layer, C 11 and C 12 are the elastic coefficients. Moreover, there is a relationship between C 11 , C 12 , and the Sn composition of the GeSn [39]: We can first substitute the Sn composition (7) obtained from EDS into Equation (5), and the value of C 11 /C 12 is calculated to be 0.0966. a bulk can also be calculated using the in-plane lattice constant of the GeSn layer, the out-of-plane lattice constant of the GeSn layer, and the value of C 11 /C 12 (0.0966). The bulk lattice constant of the GeSn is calculated to be 0.5707 nm. Therefore, the relaxation degree can be demonstrated as: where a Ge is the lattice constant of the Ge (a Ge = 0.5658 Å), a ⊥ is the out-of-plane lattice constant of the GeSn, a // is the in-plane lattice constant of the GeSn, and a bulk is the bulk lattice constant of the GeSn. Finally, the relaxation degree of the GeSn layer is evaluated to be approximately 50%.
constant of the GeSn layer, the out-of-plane lattice constant of the GeSn layer, and the value of C11/C12 (0.0966). The bulk lattice constant of the GeSn is calculated to be 0.5707 nm. Therefore, the relaxation degree can be demonstrated as: where Ge a is the lattice constant of the Ge ( Ge a = 0.5658 Å), a ⊥ is the out-of-plane lattice constant of the GeSn, // a is the in-plane lattice constant of the GeSn, and bulk a is the bulk lattice constant of the GeSn. Finally, the relaxation degree of the GeSn layer is evaluated to be approximately 50%. In order to determine the surface morphology of the GeSn layer, atomic force microscopy (AFM) measurement was performed. Figure 4 shows the typical 5 µm × 5 µm AFM image of the GeSn layer and the RMS (root mean square roughness) value of the GeSn sample was extracted from AFM scans. The as-grown GeSn layer showed a smooth surface, and the root mean square roughness (Rq) value and average root mean square (Ra) for the GeSn sample were found to be 0.8 and 0.6 nm, respectively. Compared with other low-Sn-composition GeSn layer deposited by magnetron sputtering (Table 1), our result is better than other RMS values of the crystalline GeSn layer.  In order to determine the surface morphology of the GeSn layer, atomic force microscopy (AFM) measurement was performed. Figure 4 shows the typical 5 µm × 5 µm AFM image of the GeSn layer and the RMS (root mean square roughness) value of the GeSn sample was extracted from AFM scans. The as-grown GeSn layer showed a smooth surface, and the root mean square roughness (Rq) value and average root mean square (Ra) for the GeSn sample were found to be 0.8 and 0.6 nm, respectively. Compared with other low-Sn-composition GeSn layer deposited by magnetron sputtering (Table 1), our result is better than other RMS values of the crystalline GeSn layer. constant of the GeSn layer, the out-of-plane lattice constant of the GeSn layer, and the value of C11/C12 (0.0966). The bulk lattice constant of the GeSn is calculated to be 0.5707 nm. Therefore, the relaxation degree can be demonstrated as: where Ge a is the lattice constant of the Ge ( Ge a = 0.5658 Å), a ⊥ is the out-of-plane lattice constant of the GeSn, // a is the in-plane lattice constant of the GeSn, and bulk a is the bulk lattice constant of the GeSn. Finally, the relaxation degree of the GeSn layer is evaluated to be approximately 50%. In order to determine the surface morphology of the GeSn layer, atomic force microscopy (AFM) measurement was performed. Figure 4 shows the typical 5 µm × 5 µm AFM image of the GeSn layer and the RMS (root mean square roughness) value of the GeSn sample was extracted from AFM scans. The as-grown GeSn layer showed a smooth surface, and the root mean square roughness (Rq) value and average root mean square (Ra) for the GeSn sample were found to be 0.8 and 0.6 nm, respectively. Compared with other low-Sn-composition GeSn layer deposited by magnetron sputtering (Table 1), our result is better than other RMS values of the crystalline GeSn layer.

Averaged TDD in GeSn Layer
Transmission electron microscopy (TEM) and etching pit density (EPD) are utilized to evaluate the TDDs in GeSn layers. Meanwhile, it is very difficult to ascertain the TDDs for GeSn using EPD when the TDDs are beyond 10 6 cm −2 . Moreover, TEM is also limited by its small affected region because TEM can only obtain the TDDs of the material in a small region. So, we evaluated the TDDs of the GeSn layer using the rocking curve. From our previous report [41], the intrinsic FWHM (full width at half maxima) for the strained GeSn can be described as: where r c is the radius of the electron, θ is the Bragg angle for the GeSn, λ is the X-ray wavelength, |F hkl | is the reflection structure factor of the GeSn (hkl), a 0 is the lattice constant of the bulk GeSn, and φ is the angle between the crystal surface and the diffracting planes. The reflection structure factors for the GeSn (004) are all 64f 2 , where f is the dispersion factor for the atom. The FWHM broadening by the TDDs in GeSn layer can represented as: where β m is the measured FWHM of the strained GeSn, β 0 is the intrinsic FWHM of the strained GeSn, and β d is the FWHM broadening by incident beam difference of the HR-XRD equipment. Owing to the large lattice constant mismatch between Ge and GeSn, the majority of the dislocation locates at the interface between Ge and GeSn. The TDDs in GeSn decreased with the increase of thickness and the previously deposited GeSn layer can be regarded as a buffer layer for the top GeSn layer. So, we believe that the TDDs in the whole GeSn layer are not uniform. However, the XRD rocking curve measurement ( Figure 5) demonstrates that the FWHM of the whole GeSn layer and the TDDs obtained from the XRD result reflect the TDDs from the whole GeSn layer. For this reason, we can conclude that the TDDs obtained from the XRD result can be used to present the averaged TDD of the whole GeSn layer. The photoluminescence (PL) characterization was undertaken to examine the luminescence property of the GeSn layer deposited by magnetron sputtering. The PL measurement was excited by a 532 nm green laser. During the excitation, laser beam was focused to be 100 µm spot and its power was intercalated as 500 mW. The PL emission was collected by FTIR analysis, equipped with a liquid nitrogen-cooled (LN2) InGaAs detector. Figure 6 shows the room temperature PL spectra of the GeSn The averaged TDD in the strained GeSn layer can be expressed as an empirical formula: The value of b is 0.4. The calculated values of β 2 m , β 2 0 , β 2 d , β 2 TDD , and the TDDs of the GeSn layer are outlined in Table 2. In addition, we compare the calculated TDDs with other results, and it was found that the calculated value agrees well with the reference TDDs. The photoluminescence (PL) characterization was undertaken to examine the luminescence property of the GeSn layer deposited by magnetron sputtering. The PL measurement was excited by a 532 nm green laser. During the excitation, laser beam was focused to be 100 µm spot and its power was intercalated as 500 mW. The PL emission was collected by FTIR analysis, equipped with a liquid nitrogen-cooled (LN 2 ) InGaAs detector. Figure 6 shows the room temperature PL spectra of the GeSn layer and the luminescence peak was located at 1870 nm. Comparing the PL peak position of GeSn with the results of another work [1], our result is smaller than that of a direct bandgap GeSn with a 2230 nm PL peak. From the HR-XRD result, the relaxation degree of the GeSn layer was evaluated to be approximately 50%. Therefore, we conclude that the GeSn layer is an indirect bandgap material. This finding can be attributed to the large lattice mismatch between Ge and GeSn (the lattice constant of GeSn is larger than Ge), which indicates that GeSn layer is under compressive strain and it is adverse to the bandgap transformation of Ge from indirect GeSn material to direct GeSn material. The photoluminescence (PL) characterization was undertaken to examine the luminescence property of the GeSn layer deposited by magnetron sputtering. The PL measurement was excited by a 532 nm green laser. During the excitation, laser beam was focused to be 100 µm spot and its power was intercalated as 500 mW. The PL emission was collected by FTIR analysis, equipped with a liquid nitrogen-cooled (LN2) InGaAs detector. Figure 6 shows the room temperature PL spectra of the GeSn layer and the luminescence peak was located at 1870 nm. Comparing the PL peak position of GeSn with the results of another work [1], our result is smaller than that of a direct bandgap GeSn with a 2230 nm PL peak. From the HR-XRD result, the relaxation degree of the GeSn layer was evaluated to be approximately 50%. Therefore, we conclude that the GeSn layer is an indirect bandgap material. This finding can be attributed to the large lattice mismatch between Ge and GeSn (the lattice constant of GeSn is larger than Ge), which indicates that GeSn layer is under compressive strain and it is adverse to the bandgap transformation of Ge from indirect GeSn material to direct GeSn material. Finally, the current-voltage (I-V) characteristic was carried out at room temperature. The electrical property of the p-i-n diode was performed using a Keithley 4200 semiconductor characterization system parameter analyzer (Figure 7a). Figure 7b shows the typical I-V characteristic of an 8 mm × 8 mm square p-i-n diode with the GeSn layer deposited by magnetron sputtering. The very high dynamic resistance may be attributed mainly to the fact that the Si wafer and Ge substrate have a high series resistor. Ultimately, we conclude that the sputter-deposited Finally, the current-voltage (I-V) characteristic was carried out at room temperature. The electrical property of the p-i-n diode was performed using a Keithley 4200 semiconductor characterization system parameter analyzer (Figure 7a). Figure 7b shows the typical I-V characteristic of an 8 mm × 8 mm square p-i-n diode with the GeSn layer deposited by magnetron sputtering. The very high dynamic resistance may be attributed mainly to the fact that the Si wafer and Ge substrate have a high series resistor. Ultimately, we conclude that the sputter-deposited GeSn layer has the great potential to be used for the fabrication of optoelectronic devices.

Conclusions
In summary, a high-quality GeSn layer with a Sn content up to 7% was successfully grown on a Ge (100) substrate via magnetron co-sputtering. The HR-XRD result shows the remarkable single-crystalline property of the GeSn layer. Due to the fact that the TDDs in the whole GeSn layer are not uniform, we conclude that the TDDs obtained from the XRD result can give the averaged TDD of the whole GeSn layer (8.7 × 10 9 cm −2 ) along the (004) plane. The PL measurement result shows the optical emission from the deposited high-quality GeSn layer. Furthermore, the fabricated vertical p-i-n device exhibits a good room temperature current-voltage (I-V) characteristic. From the results, we predict that the sputter-deposited GeSn will have great potential to achieve high-Sn-composition GeSn layers with proper design, which is critical for optoelectronic applications.