Open Access
Add to BibSonomy
Abstract
Herein we utilize polarized photoluminescence (PL) microscopy and spectral analysis to locate and characterize many different types of µm-scale extended defects present in melt-grown bulk crystals and metal-organic vapor-phase epitaxy (MOVPE)-grown epitaxial thin films of β-Ga2O3 and β-(Al,Ga)2O3. These include pits, divots, mounds, scratches, rotation domain boundaries, stacking faults, cracks, and other defect categories. Some types of µm-scale defects simply decrease overall PL yield, while others emit different spectra than single crystal regions. We combine PL microscopy with atomic force microscopy (AFM) and scanning electron microscopy (SEM) to provide detailed characteristics of these different types of features which can arise from both bulk crystal growth, surface preparation, and epitaxial growth processes. We show that sample quality (in terms of extended defects) can be determined by using PL and that attributing spectral features to isolated point defects is invalid unless the sample is proven to not contain extended defects.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Gallium oxide (Ga2O3) is an attractive material for various electronic and optoelectronic applications. Ga2O3 is estimated to have a Baliga’s figure of merit near 3200 times that of Si [1]. The most thermochemically-stable phase of Ga2O3 is monoclinic β-Ga2O3. It has an indirect (or dipole-forbidden direct depending on direction [2]), ultra-wide fundamental bandgap around 276 nm (4.5 eV), but anisotropy makes the optically-determined bandgap vary from 282 to 253 nm (4.5 to 4.9 eV) depending on the light polarization [1,3]. In order to develop and improve the performance and yield of β-Ga2O3 devices, defects (point and extended defects) need to be better understood and characterized to improve growth processes during research phases. Furthermore, noninvasive and quick characterization methods will be necessary for in-between sample quality checks during fabrication of devices. To this end, we recently reported systematic photoluminescence (PL) characterization showing that the presence of extended defects in epitaxial thin films is correlated with changes in spatially-averaged PL yield and shifts in the luminescence spectrum [3]. In this work, we directly correlate modified emission spectra and intensities with individual structural defects using PL microscopy combined with atomic force (AFM) and scanning electron (SEM) microscopies. These results, first and perhaps most importantly, underscore the point that spectral features in spatially-averaged luminescence (and by extension, other defect spectroscopies) should not be assumed to arise only from point defects and complexes. They also provide a basis for understanding trends in structural quality of epitaxially- and melt-grown samples.
As was discussed previously, PL spectra from samples of β-Ga2O3, which except for very rare circumstances exhibits significant Stokes shifts, can be (somewhat naively) decomposed into multiple Gaussians or more complex asymmetric vibronically-broadened line shapes for self-trapped hole and defect emissions [3–6]. The strong electron-phonon coupling for self-trapped hole and defect emissions results in broadening of order of magnitude hundreds of meV, even at cryogenic temperatures, making meaningful deconvolution of luminescence spectra challenging [3–6]. As such, there has been a lot of uncertainty in the literature regarding attribution of spectral features, and everything was attributed, without much basis, to isolated point defects. A comprehensive review of that literature was provided in the Supplement 1 in our previous work [3]. We found that all previous literature attributed changes in PL spectra to various point defects. The general consensus (which appears undisputed) attributes the UV band luminescence to recombination of conduction band electrons with self-trapped holes. Visible emission (blue, green, etc.) is likely of extrinsic origin with a host of native defects and dopants predicted to emit in strongly-overlapping bands, rendering attempts to infer the presence of any one or few particular defect(s) from such spectral features impossible [7–26]. Further literature that focused on PL microscopy to study β-Ga2O3 used similar reasoning. At that time, red emission peaks were attributed to Cr3+ defects, as well as Si-related contamination for three other different localized red emissions [27]. Recently Huso, et al. discussed surface defects and showed that hydrogen-annealed single crystals of unintentionally doped (UID) β-Ga2O3 have bright, localized emission at 3.27eV (∼380 nm), of which the brightest occurred near surface pits created by the hydrogenation [23].
Our previous study [3] revealed that extended defects can also play a critical role in the PL spectrum. In that study, we showed that a single crystal sample, with few extended defects, will have a UV dominant emission. In contrast, a sample with a large density of extended defects will have a shifted dominant peak with a now blue emission [3]. To further support those conclusions from our previous study, PL microscopy is performed here on a series of samples, verifying that extended defects emit away from the UV band (conduction band to self-trapped hole band) and are the specific cause of blue emission. This implies that it is erroneous to attribute spectral features to isolated point defects before first verifying that extended defects are not present in the sample. Single crystals and high-quality epitaxial films, on the other hand, tend to be UV dominant. Furthermore, characterization and identification of different types of extended defects are discussed using PL, AFM, and SEM. They show that previous literature, which attributed PL shifting to a type of point defect, are misleading as extended defects play a large role in the PL spectrum. PL microscopy is also of use for growers during research phases, as well as for fabrication of devices, to noninvasively check sample quality and identify potential killer defects.
2. Experiment
2.1 Sample preparation
All thin film samples discussed herein were grown epitaxially by metal-organic vapor-phase epitaxy (MOVPE) in an Agnitron Agilis reactor using triethylgallium (TEGa) as the gallium precursor and O2 as the oxygen precursor. Growth details of all samples are listed in Cooke et al. [3], and are presented in Table 1. They include for all, unless stated, a 15.53 µmol/min total molar flow, 1100 sccm argon flow rate, 500 sccm oxygen flow rate, and 15 Torr chamber pressure. The substrates used for growth were either Fe-doped (-201) or (010) oriented β-Ga2O3 grown by Novel Crystal Technology, Japan. C-plane sapphire purchased from Cryscore was also used. The substrates were pre-cleaned using 5 min acetone, 5 min methanol, and 5 min DI water in an ultrasonic cleaner followed by a 30 min dip in HF (49%) bath before epitaxial growth. Furthermore, an edge-defined film fed (EFG) grown (-201) oriented unintentionally doped (UID) β-Ga2O3 bulk sample was purchased from Novel Crystal Technology for comparison with the epitaxial films. Lastly, a Czochralski grown (100) 10% bulk aluminum-gallium oxide (AGO) sample, β-Al0.2Ga1.8O3, was produced at Washington State University [28].
Table 1. Growth parameters for the samples [3]
In the prior work, we analyzed three series of samples using PL [3], and the same samples were examined using PL microscopy herein, with representative samples discussed. The first series consisted of epitaxial films of varying Si-doping grown on Fe-doped β-Ga2O3 (010). The second series compared samples of the same (−201) orientation grown through different methods, including a bare (−201) UID wafer, UID films grown on a (−201) Fe-doped Ga2O3 wafer, and UID films grown on c-plane sapphire such that (−201) Ga2O3 || (0001) sapphire (but with rotation domains). Details on x-ray diffraction (XRD) for Ga2O3 grown on sapphire can be found in the report by Ghadbeigi et al. [29]. The 3rd series compared (AlxGa1-x)2O3 films, i.e., AGO of varying Al concentrations, grown on Fe-doped β-Ga2O3 (010) and on c-plane sapphire. Finally, we compare the 10% AGO film grown on Fe-doped (010) β-Ga2O3 from the 3rd series to a more-recently grown 10% AGO sample having fewer extended defects due to improved growth conditions that reduce precipitates that generate nodular defects, as well as growing a thinner film to reduce the size and density of defects from heteroepitaxy stress and strain. Growth details of the two 10% AGO films are listed and compared in Table 1. The aluminum composition values were obtained from the precursor molar ratio (PMR) during growth, which has been calibrated previously by our group [3]. PMR or transmission electron microscopy energy dispersive X-ray spectroscopy (TEM EDS) is used because strain and relaxation in the homoepitaxial film causes XRD values to shift giving a larger percent concentration than expected as discussed in our previous report [3].
2.2 Characterization
Scanning electron microscopy (SEM) was done using a FEI TENEO with a trinity detection system. Images were taken using 5 kV and a spot size of 8. Atomic Force Microscopy (AFM) was performed using a Bruker dimension icon AFM. All AFM roughness is given as the root mean square average of the height deviation taken from the mean image data plane (Rq) over a 255 µm2 (15 × 15 µm) area. A schematic of the µ-photoluminescence microscopy system is provided in the Supplement 1 as Fig. S1. Ultrafast (fs) pulses from a wavelength-tunable (690 -1040 nm or 1.8 - 1.2 eV) Ti:Sapphire (Coherent Chameleon Vision Ultra) laser were used. The laser then passed through a third-harmonic generator (Coherent Harmonics) and was polarized using a linear polarizer (Glan-Laser alpha-BBO polarizer prism, 210 - 450 nm or 5.9 - 2.76 eV). This was followed by a zero-order half-wave plate to control the linear polarization angle of the laser. Two lenses were then used to expand the beam before entering a Picoquant MicroTime 200 Time-resolved Fluorescence Microscope designed to operate in the UV. Two dichroic mirrors reflect the beam into an Olympus 1X71 inverted microscope with a 40x and 0.6 NA Ultrafluar UV objective, which can scan an image using a nano-positioning stage. An image size of 384 × 384 pixels was used for measurements with either an 80 × 80 µm (0.208 µm/px) or 15 × 15µm (0.039 µm/px) scanning range. Dwell time was 5 ms with a learning time of 20 s. A quartz plate was used to hold the sample on the microscope stage with a drop of glycerin (Immersol G) used as a medium between the objective and quartz to prevent reflections. Fluorescence generated by the sample is collected by the objective, passes through the dichroic followed by a 325 nm EdgeBasic long-pass edge filter, and is focused through a 30 µm pinhole. The emission beam is then passed through either a 357 ± 44 nm Newport bandpass filter, 450 ± 40 nm bandpass filter, or no filter before being collected by a UV-sensitive photomultiplier tube (PMT) photon-counting detector (PMA 175-M Ultra). SymPhoTime 64 was used for data analysis [30].
3. Results
Figure 1 shows results of characterizing 1 µm thick UID β-Ga2O3 grown on (010) Fe-doped β-Ga2O3 at 600°C. The results include PL microscopy, AFM, and SEM. The results of the other two samples from the Si-doped series are in the Supplement 1 and show similar results (Fig. S2 and S3). In Fig. 1(a) and 1(b), it is observed that emission intensity depends on the polarization of the excitation. This is because both PL maps were excited at 266 nm (4.66 eV) with polarization parallel to either the E||c* (Fig. 1(a)) or E||a (Fig. 1(b)). The optical transition threshold energies for β-Ga2O3 are ordered as E||c* < E||a < E||b [31]. Because the excitation energy falls further above the E||c*, threshold excited carrier densities for that orientation will be larger than for E||a, resulting in higher PL emission. This is especially true for epitaxial films of thickness comparable to the absorption depth grown on Fe-doped substrates, which have much lower PL yield than epitaxial MOVPE films. In Fig. 1(a) it is noted that the emission from the sample is very constant, except for spots seen across the sample surface which have drastically lower PL yield and appear black. Figure 1(b) taken with E||a shows the opposite contrast, with several of the spots now appearing brighter than the constant ‘background’ from the rest of the single-crystalline films [3,32,33]. AFM revealed a roughness of 0.8 ± 0.2 nm for the background seen in Fig. 1(c). The spots, seen as bright white and black spots in AFM, match sympetalous defects which have been characterized in previous literature [30]. The sympetalous defects are polarization-dependent in PL microscopy. AFM shows they have a height difference of 100 nm between the lowest and highest points located near the center of the defect and the top surface view, showing the defect to be polygon shaped. The SEM images of Fig. 1(d) and 1(e) also shows sympetalous defects in both secondary electron and backscattering modes. Dark spots were also seen throughout the secondary electron SEM images, which are surface divots found in AFM, and are 4 nm deep. These were likely caused by contamination during growth causing crater-like defects from flake contamination [34] and have drastically lower PL yield, appearing black. Lastly, the rougher background texture is likely not real. It is caused when AFM runs over a large area causing bowing which must be corrected by the software. But the large (in height) defects cause some error when collecting and correcting this bowing. Since the sample is β-Ga2O3 grown on (010) Fe-doped β-Ga2O3, which generates the most smooth and single crystalline film, this background texture is not real. The other samples in the series, Si-doped β-Ga2O3 grown on (010) Fe-doped β-Ga2O3, shows the same uniform emission in the Supplement 1 as Fig. S2 and S3 with sympetalous defects on the films.
Fig. 1. UID β-Ga2O3 grown on (010) Fe-doped β-Ga2O3. All parts use roughly the same scale. (a) PL microscopy at 266 nm (4.66 eV) excitation E||c*, (b) PL microscopy of the same area at 266 nm with E||a, (c) AFM of a different location on the sample (3D image in Fig. S9a), (d) secondary electron SEM of an area of the sample different from PL and AFM areas, (e) backscattered SEM of the same area as secondary electron SEM.
A (−201) bulk β-Ga2O3 sample is imaged in Fig. 2 using PL microscopy, AFM, and SEM. PL microscopy displays straight lines and pits running across the sample. These same lines and pits are also seen in AFM and SEM and are likely due to scratching caused during the polishing process in which mirror polishing was used. The AFM surface roughness is 6.2 ± 0.2 nm, with the largest scratches being around 40 nm deep in Fig. 2(c). These deep scratches and pits generate the largest difference in emission contrast from the bulk and are what are most easily seen in the PL microscopy of Fig. 2(a) and 2(b). Any β-Ga2O3 single crystal with homogeneous point defect concentrations would be expected to have uniform PL emission (at this length scale of the microscopy system resolution) except at extended defects. However, topography also has a role in emission intensity. Like SEM, there is an angle of acceptance for how much emission is collected, and edges/bumps on a sample surface can cause the amount of emission collected at a point to change based on refraction and diffraction. The detector will in general collect fewer photons from holes and grooves but more from hills and ridges. The number of counts can also depend on the optical anisotropy of Ga2O3. Figure 2(a) has more emission compared to Fig. 2(b). This is due to the excitation energy being greater than the absorption edge for E⊥b in Fig. 2(a), while the excitation energy is less than the absorption edge for E||b in Fig. 2(b). The shorter absorption depth for E⊥b also makes it more surface sensitive. Therefore, emission will come from the surface causing surface flaws to provide contrast. On the other hand, Fig. 2(b) shows more topography contrast as β-Ga2O3 emits weakly along E||b due to the excitation energy being less than the absorption edge along E||b. Therefore, the emission is more dependent on the path (i.e., refraction and diffraction from the surface). The 10% bulk AGO sample, shown in the Supplement 1, shows similar results in Fig. S4. Since it was cleaved from a boule, the face of the sample is uneven in areas shown in PL microscopy.
Fig. 2. Bulk (−201) β-Ga2O3. All parts use roughly the same scale. (a) PL microscopy at 266 nm (4.66 eV) excitation E⊥b, (b) PL microscopy of the same area at 266 nm (4.66 eV) excitation E||b, (c) AFM of a different location on the sample (3D image in Fig. S9b), (d) secondary electron SEM of an area of the sample different from PL and AFM areas, (e) backscattered SEM of the same area as secondary electron SEM.
Figure 3 shows the characterization of 400 nm thick (−201) β-Ga2O3 grown on sapphire at 810°C. When β-Ga2O3 is grown on sapphire, it forces a 2-fold symmetric crystal to grow on a 3-fold (technically rhombohedral which is 3-fold, but the slight deviations allow it to be considered hexagonal which is 6-fold) symmetric substrate plane, which causes the film to develop rotational domains [3,29,35,36]. As such, the rotationally misaligned crystal domains each have their in-plane directions oriented differently from one another. In spatially-integrated PL, this causes the polarization-dependent intensity seen for single crystals to rotationally average out. However, as seen in Fig. 3(a) and 3(b), PL microscopy from such samples shows lateral contrast associated with the size and orientations of individual rotation domains. The rotational domains are around 1 × 1 µm in size. The brightest spots within the two images have a strong or weak emission dependent on the excitation polarization angle. Likely, the b-direction crystal domains aligned perpendicular to the excitation polarization angle when they are brightest (E⊥b). This is because perpendicular to b-direction is a mix of a and c-direction. The weaker, textured background is likely a combination of in-plane orientation directions, though topography may also contribute to the contrast as well. As seen in AFM, there are higher and lower spots all over the sample with a roughness of 14.8 ± 0.2 nm. SEM verifies that rotational domains are present throughout the sample as it matches the rotational domains seen in Ghadbeigi et al. [29].
Fig. 3. β-Ga2O3 grown on sapphire. All parts use roughly the same scale. Inserts are of the same image zoomed in. (a) PL microscopy at 266 nm (4.66 eV) excitation with the insert of images brightened, (b) PL microscopy of the same area at 266 nm (4.66 eV) excitation polarized 90° from (a) with the insert of images brightened, (c) AFM of a different location on the sample (3D image in Fig. S9c), (d) secondary electron SEM of an area of the sample different from PL and AFM areas, (e) backscattered SEM of the same area as secondary electron SEM.
The 400 nm thick 25% AGO grown on (010) Fe-doped β-Ga2O3 at 650°C also shows a PL microscopy that matches AFM and SEM. The PL microscopy seen in Fig. 4(a) and 4(b) shows a pattern caused by extended structural crystalline defects riddled throughout the film as well as potentially also caused by topography. TEM showed, in our previous report, that there was a high density of extended structural crystalline defects [3]. EDS taken from TEM also showed constant aluminum concentration throughout the film so phase segregation was not seen. Extended structural defects are also shown in AFM, where surface roughness is 21.8 ± 0.2 nm. In fact, the extended structural crystalline defects that make up this pattern are so large that they can be easily seen with AFM or SEM. Similar observations are seen in the Supplement 1, Fig. S5, S6, and S7, which shows AGO grown on sapphire. In those cases, the PL microscopy matches AFM and SEM due to extended defects riddled throughout. Those samples have rotational domains from growing on sapphire and other potential extended defects caused by growing an alloy heteroepitaxially [3].
Fig. 4. 25% AGO grown on (010) Fe-doped β-Ga2O3. All parts use roughly the same scale. (a) PL microscopy at 235 nm (5.28 eV) excitation E||c*, (b) PL microscopy of the same area at 235 nm (5.28 eV) excitation E||a (c) AFM of a different location on the sample (3D image in Fig. S9d), (d) secondary electron SEM of an area of the sample different from PL and AFM areas, (e) backscattered SEM of the same area as secondary electron SEM.
The 400 nm thick (−201) β-Ga2O3 grown on a Fe-doped β-Ga2O3 at 650°C sample shows very different results for the PL microscopy in Fig. 5. The pattern from PL microscopy does not match AFM or SEM. AFM and SEM show long sliver-like features, all extending in one direction, with a roughness of 11.4 ± 0.2 nm. However, PL microscopy shows round-like spots, more similar to what is seen for β-Ga2O3 grown on sapphire. As previously reported, this sample is riddled with extended defects [3,37]. We hypothesize that some extended defects may emit more brightly than other defects causing this discrepancy. For example, Eisner et al. found that most defects in (−201) β-Ga2O3 grown on Fe-doped β-Ga2O3 are incoherent stacking faults [37], which likely causes the texture seen in SEM and AFM. But other defects, such as nodular defects, may also be present and emitting more brightly than the stacking faults, thus overwhelming the stacking faults’ luminescence. A similar situation is seen in the Supplement 1 of Fig. S8 with another (−201) β-Ga2O3 film grown on a Fe-doped β-Ga2O3 wafer. In this case, the PL microscopy shows some of same long sliver-like features also seen in AFM and SEM but the PL microscopy also has excitation polarization angle dependent spots similar to sympetalous defects or rotational domains.
Fig. 5. (−201) β-Ga2O3 grown on Fe-doped β-Ga2O3. All parts use roughly the same scale. (a) PL microscopy at 266 nm (4.66 eV) excitation E⊥b, (b) PL microscopy of the same area at 266 nm (4.66 eV) excitation E||b, (c) AFM of a different location on the sample (3D image in Fig. S9e), (d) secondary electron SEM of an area of the sample different from PL and AFM areas, (e) backscattered SEM of the same area as secondary electron SEM.
This hypothesis was verified by comparing two 10% AGO grown on (010) β-Ga2O3 samples of different thicknesses. The first is a 400 nm film sample characterized in Fig. 6. The sample shows multiple extended defects, including lines running along the [001] and crystal clusters jutting out of the film. AFM shows an overall roughness of 10.4 ± 0.2 nm with a 0.6 ± 0.2 nm smooth film between crystal clusters. SEM shows similar results of lines which are likely cracks running along the [001], which are produced to relax strain caused by lattice mismatch of heteroepitaxial growth [38,39]. The crystal clusters are likely nodular defects caused due to contamination during growth [34] or due to sympetalous defects [3]. PL microscopy shows all of these defects. The crystal clusters are polarized and emit most intensely when excited E||a. The lines never emit even though the AFM topography shows the lines are one of the highest topography points on the sample. Other than that, two different background textures were seen when exciting E||a or E||c*.
Fig. 6. 400 nm thick 10% AGO grown on (010) Fe-doped β-Ga2O3. All parts use roughly the same scale. Inserts are of the same image zoomed in. (a) PL microscopy at 235 nm (5.28 eV) excitation E||c*, (b) PL microscopy of the same area at 235 nm (5.28 eV) excitation E||a, (c) AFM of a different location on the sample (3D image in Fig. S9f), (d) secondary electron SEM of an area of the sample different from PL and AFM areas, (e) backscattered SEM of the same area as secondary electron SEM.
A second 10% AGO grown on (010) β-Ga2O3 sample using improved growth conditions, including a lower TEGa flow rate and thinner film, was made to reduce the extended defects in the film and compare with the results from the thicker sample. This sample was only 150 nm thick and showed far less defects throughout the sample, as seen in Fig. 7. In our previous work, we proved that extended defects emit blue (430–460 nm, 2.7–2.9 eV), and single-crystal emits in the UV (∼390 nm, 3.2 eV) [3]. In this case, the thinner sample has a dominant UV emission, while the thicker sample has a dominant blue emission. There are also highlighted regions in Fig. 7 showing where the bandpass filters used for PL microscopy allow emission to pass through for the UV bandpass (purple highlighted) and blue bandpass (blue highlighted).
Fig. 7. Normalized PL of two 10% AGO films grown on Fe-doped β-Ga2O3 excited at 235 nm (5.28 eV). Colored regions show where the filters used for PL microscopy pass-through for UV bandpass (purple highlighted region) and blue bandpass (blue highlighted region).
For the thicker 10% AGO sample, the two textures seen in PL microscopy in Fig. 8 come from the excitation polarization angle used, which generates an emission that is dependent on the sample structure. This can be seen by using bandpass filters where the spectral shapes between samples is compared. As seen in Fig. 8, E||c* excitation shows a lot more constant-like intensity from the crystal using the UV filter, with weaker intensity coming from defects with the blue filter. On the other hand, E||a excitation highlights the extended defects throughout the sample for both the UV and blue filter. This data also suggests that there could be some types of defects that emit UV and not just blue as seen on the E||a using the UV filter. The background is not constant with brighter spots seen throughout the image, though this could also be partly due to the height location of defects in the film.
Fig. 8. 400 nm thick 10% AGO grown on (010) Fe-doped β-Ga2O3. PL microscopy at 248 nm excitation using linear polarization at E||a and E||c*, with different filters used to filter the emission taken of the same spot on the sample. Inserts depict the same images brightened.
A second 10% AGO grown on (010) β-Ga2O3 sample using improved growth conditions, including a lower TEGa flow rate and thinner film, was made to reduce the extended defects in the film and compare with the results from the thicker sample. This sample was only 150 nm thick and showed far less defects throughout the sample, as seen in Fig. 9. In this case, the background is constant and appears more like results from β-Ga2O3 grown on (010) β-Ga2O3. The constant background is only seen when using a UV filter. The defects are all coming from the nodular defect clusters and only show emission using a blue filter. Though, for the E||c* excited image, some UV could be leaking through the blue filter due to the weaker emission from the defects.
Fig. 9. 150 nm thick 10% AGO grown on (010) Fe-doped β-Ga2O3. PL microscopy at 248 nm excitation using linear polarization at E||a and E||c*, with different filters used to filter the emission taken of the same spot on the sample.
4. Conclusions
Using the samples characterized through PL in our previous work [3], we elaborate on how extended defects change the appearance of samples utilizing PL microscopy. Single crystal samples are expected to have uniform UV emission. Any topography can alter the emission intensity; and any topography that is altered due to defects can also cause contrast in emission intensity. As such, PL microscopy tends to match AFM and SEM images. But, there are cases where PL microscopy can diverge from AFM and SEM due to other extended defects within the film, which are more luminescent than surface defects/topography. This was seen in the background patterns for 10% AGO which were dependent with the excitation polarization angle. It was seen that defect clusters within the AGO film emit brightly blue. The cracks running along [001] emit weakly even though the topography shows the cracks generate peaks, and some defects emit both UV and blue.
This demonstrates that characterization of samples and extended defects is possible with PL microscopy. PL allows for sample quality (density of extended defects) to be determined quickly and noninvasively. Characterization and identification of different types of extended defects are also possible. Furthermore, attributing spectral features to isolated point defects is invalid unless the sample is proven to not contain extended defects. This is of use for growers to characterize growth quality of samples and identify defects caused by their growth, as well as fabricators of devices who can also use this to noninvasively check sample quality and identify potential ‘device killer’ defects during the fabrication process.
Funding
Air Force Office of Scientific Research (FA9550-21-1-0507).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
Supplemental document
See Supplement 1 for supporting content.
References
1. S. J. Pearton, J. Yang, P. H. Cary, F. Ren, J. Kim, M. J. Tadjer, and M. A. Mastro, “A review of Ga2O3 materials, processing, and devices,” Appl. Phys. Rev. 5(1), 011301 (2018). [CrossRef]
2. H. Peelaers and C. G. Van De Walle, “Sub-band-gap absorption in Ga2O3,” Appl. Phys. Lett. 111(18), 182104 (2017). [CrossRef]
3. J. Cooke, P. Ranga, J. Jesenovec, J. S. McCloy, S. Krishnamoorthy, M. A. Scarpulla, and B. Sensale-Rodriguez, “Effect of extended defects on photoluminescence of gallium oxide and aluminum gallium oxide epitaxial films,” Sci. Rep. 12(1), 3243 (2022). [CrossRef]
4. A. Alkauskas, M. D. McCluskey, and C. G. Van De Walle, “Tutorial: Defects in semiconductors - combining experiment and theory,” J. Appl. Phys. 119(18), 181101 (2016). [CrossRef]
5. Y. Zhang, “Applications of Huang-Rhys theory in semiconductor optical spectroscopy,” J. Semicond. 40(9), 091102 (2019). [CrossRef]
6. S. L. Li and D. G. Truhlar, “Franck-Condon models for simulating the band shape of electronic absorption spectra,” J. Chem. Theory Comput. 13(6), 2823–2830 (2017). [CrossRef]
7. W. C. Herbert, H. B. Minnier, and J. J. J. Brown, “Self-Activated Luminescence of β-Ga2O3,” J. Electrochem. Soc. 116(7), 1019 (1969). [CrossRef]
8. Y. Tomm, P. Reiche, D. Klimm, and T. Fukuda, “Czochralski grown Ga2O3 crystals,” J. Cryst. Growth 220(4), 510–514 (2000). [CrossRef]
9. K. Shimamura, E. G. Víllora, T. Ujiie, and K. Aoki, “Excitation and photoluminescence of pure and Si-doped β-Ga2O3 single crystals,” Appl. Phys. Lett. 92(20), 201914–30 (2008). [CrossRef]
10. H. Wakai, Y. Sinya, and A. Yamanaka, “Effect of Cr3 + ions on optical properties in β-Ga2O3 semiconductor,” Phys. Status Solidi C 8, 537–539 (2011). [CrossRef]
11. M. Yamaga, T. Ishikawa, M. Yoshida, T. Hasegawa, E. G. Villora, and K. Shimamura, “Polarization of optical spectra in transparent conductive oxide β-Ga2O3,” Phys. Status Solidi C 8, 2621–2624 (2011). [CrossRef]
12. T. Onuma, S. Fujioka, T. Yamaguchi, M. Higashiwaki, K. Sasaki, T. Masui, and T. Honda, “Correlation between blue luminescence intensity and resistivity in β-Ga2O3 single crystals,” Appl. Phys. Lett. 103(4), 041910–6 (2013). [CrossRef]
13. S. Yamaoka and M. Nakayama, “Evidence for formation of self-trapped excitons in a β-Ga2O3 single crystal,” Phys. Status Solidi C 13, 93–96 (2016). [CrossRef]
14. O. M. Bordun, B. O. Bordun, I. Y. Kukharskyy, and I. I. Medvid, “Photoluminescence Properties of β-Ga2O3 Thin Films Produced by Ion-Plasma Sputtering,” J. Appl. Spectrosc. 84(1), 46–51 (2017). [CrossRef]
15. H. Oda, N. Kimura, D. Yasukawa, H. Wakai, and A. Yamanaka, “Time-resolved spectroscopy of luminescence in a wide gap Si-doped β-Ga2O3,” Phys. Status Solidi A 214(3), 1600670 (2017). [CrossRef]
16. Q. D. Ho, T. Frauenheim, and P. Deák, “Origin of photoluminescence in β-Ga2O3,” Phys. Rev. B 97(11), 115163 (2018). [CrossRef]
17. T. T. Huynh, L. L. C. Lem, A. Kuramata, M. R. Phillips, and C. Ton-That, “Kinetics of charge carrier recombination in β- Ga2O3 crystals,” Phys. Rev. Mater. 2, 1–6 (2018). [CrossRef]
18. C. Liu, Y. Berencén, J. Yang, Y. Wei, M. Wang, Y. Yuan, C. Xu, Y. Xie, X. Li, and S. Zhou, “Irradiation effects on the structural and optical properties of single crystal β-Ga2O3,” Semicond. Sci. Technol. 33(9), 095022 (2018). [CrossRef]
19. Y. Wang, P. T. Dickens, J. B. Varley, X. Ni, E. Lotubai, S. Sprawls, F. Liu, V. Lordi, S. Krishnamoorthy, S. Blair, K. G. Lynn, M. Scarpulla, and B. Sensale-Rodriguez, “Incident wavelength and polarization dependence of spectral shifts in β-Ga2O3 UV photoluminescence,” Sci. Rep. 8(1), 18075 (2018). [CrossRef]
20. Y. Berencén, Y. Xie, M. Wang, S. Prucnal, L. Rebohle, and S. Zhou, “Structural and optical properties of pulsed-laser deposited crystalline β-Ga2O3 thin films on silicon,” Semicond. Sci. Technol. 34(3), 035001 (2019). [CrossRef]
21. H. Cui, Q. Sai, H. Qi, J. Zhao, J. Si, and M. Pan, “Analysis on the electronic trap of β-Ga2O3 single crystal,” J. Mater. Sci. 54(19), 12643–12649 (2019). [CrossRef]
22. I. Hany, G. Yang, C. E. Zhou, C. Sun, K. Gundogdu, D. Seyitliyev, E. O. Danilov, F. N. Castellano, D. Sun, and E. Vetter, “Low temperature cathodoluminescence study of Fe-doped β-Ga2O3,” Mater. Lett. 257, 126744 (2019). [CrossRef]
23. J. Huso, M. D. McCluskey, Y. Yu, M. M. Islam, and F. Selim, “Localized UV emitters on the surface of β-Ga2O3,” Sci. Rep. 10(1), 21022–7 (2020). [CrossRef]
24. J. Bin Cho, G. Jung, K. Kim, J. Kim, S. K. Hong, J. H. Song, and J. I. Jang, “Highly asymmetric optical properties of β-Ga2O3 as probed by linear and nonlinear optical excitation spectroscopy,” J. Phys. Chem. C 125(2), 1432–1440 (2021). [CrossRef]
25. J. Jesenovec, M. H. Weber, C. Pansegrau, M. D. McCluskey, K. G. Lynn, and J. S. McCloy, “Gallium vacancy formation in oxygen annealed β-Ga2O3,” J. Appl. Phys. 129(24), 245701 (2021). [CrossRef]
26. N. Zhang, I. M. Kislyakov, C. Xia, H. Qi, J. Wang, and H. F. Mohamed, “Anisotropic luminescence and third-order electric susceptibility of Mg-doped gallium oxide under the half-bandgap edge,” Opt. Express 29(12), 18587 (2021). [CrossRef]
27. C. Remple, J. Huso, and M. D. McCluskey, “Photoluminescence and Raman mapping of β-Ga2O3,” AIP Adv. 11(10), 105006 (2021). [CrossRef]
28. J. Jesenovec, B. Dutton, N. Stone-Weiss, A. Chmielewski, M. Saleh, C. Peterson, N. Alem, S. Krishnamoorthy, and J. S. McCloy, “Alloyed β-(AlxGa1-x)2O3 bulk Czochralski single β-(Al0.1Ga0.9)2O3 and polycrystals β-(Al0.33Ga0.66)2O3, β-(Al0.5Ga0.5)2O3), and property trends,” J. Appl. Phys. 131(15), 155702 (2022). [CrossRef]
29. L. Ghadbeigi, J. Cooke, G. T. Dang, T. Kawaharamura, T. Yasuoka, R. Sun, P. Ranga, S. Krishnamoorthy, M. A. Scarpulla, and B. Sensale-Rodriguez, “Optical characterization of gallium oxide α and β polymorph thin-films grown on c-plane sapphire,” J. Electron. Mater. 50(6), 2990–2998 (2021). [CrossRef]
30. J. Cooke, P. Ranga, A. Bhattacharyya, X. Cheng, Y. Wang, S. Krishnamoorthy, M. A. Scarpulla, and B. Sensale-Rodriguez, “Sympetalous defect,” in Metalorganic Vapor Phase Epitaxy (MOVPE)-Grown Homoepitaxial β-Ga2O3 Films (2022).
31. F. Ricci, F. Boschi, A. Baraldi, A. Filippetti, M. Higashiwaki, A. Kuramata, V. Fiorentini, and R. Fornari, “Theoretical and experimental investigation of optical absorption anisotropy in β-Ga2O3,” J. Phys.: Condens. Matter 28(22), 224005 (2016). [CrossRef]
32. S. Rafique, L. Han, M. J. Tadjer, J. A. Freitas, N. A. Mahadik, and H. Zhao, “Homoepitaxial growth of β-Ga2O3 thin films by low pressure chemical vapor deposition,” Appl. Phys. Lett. 108(18), 182105–8 (2016). [CrossRef]
33. Y. An, L. Dai, Y. Wu, B. Wu, Y. Zhao, T. Liu, H. Hao, Z. Li, G. Niu, J. Zhang, Z. Quan, and S. Ding, “Epitaxial growth of β-Ga2O3 thin films on Ga2O3 and Al2O3 substrates by using pulsed laser deposition,” J. Adv. Dielectr. 11(3), 035319 (2021). [CrossRef]
34. P. Panjan, A. Drnovšek, P. Gselman, M. Čekada, and M. Panjan, “Review of growth defects in thin films prepared by PVD techniques,” Coatings 10(5), 447 (2020). [CrossRef]
35. R. Schewski, G. Wagner, M. Baldini, D. Gogova, Z. Galazka, T. Schulz, T. Remmele, T. Markurt, H. Von Wenckstern, M. Grundmann, O. Bierwagen, P. Vogt, and M. Albrecht, “Epitaxial stabilization of pseudomorphic α-Ga2O3 on sapphire (0001),” Appl. Phys. Express 8(1), 011101 (2015). [CrossRef]
36. S. Rafique, L. Han, A. T. Neal, S. Mou, J. Boeckl, and H. Zhao, “Towards High-Mobility Heteroepitaxial β-Ga2O3 on sapphire − dependence on the substrate off-axis angle,” Phys. Status Solidi A 215, 1700467 (2018). [CrossRef]
37. B. A. Eisner, P. Ranga, A. Bhattacharyya, S. Krishnamoorthy, and M. A. Scarpulla, “Compensation in (-201) homoepitaxial β-Ga2O3 thin films grown by metalorganic vapor-phase epitaxy,” J. Appl. Phys. 128(19), 195703 (2020). [CrossRef]
38. S. Mu, M. Wang, H. Peelaers, and C. G. Van De Walle, “First-principles surface energies for monoclinic Ga2O3 and Al2O3 and consequences for cracking of (AlxGa1-x)2O3,” APL Mater. 8(9), 091105 (2020). [CrossRef]
39. A. F. M. A. U. Bhuiyan, Z. Feng, L. Meng, A. Fiedler, H. L. Huang, A. T. Neal, E. Steinbrunner, S. Mou, J. Hwang, S. Rajan, and H. Zhao, “Si doping in MOCVD grown (010) β-(AlxGa1-x)2O3 thin films,” J. Appl. Phys. 131(14), 145301 (2022). [CrossRef]
