Study of Nitridation Effect on Structural, Morphological, and Optical Properties of GaAs Film Growth on Silicon Substrates via Close Space Vapor Transport Technique
, Antonio Coyopol 1,*,†, Godofredo GarcÃa-Salgado 1,†, Román Romano-Trujillo 1, Crisóforo Morales-Ruiz 1, Enrique Rosendo-Andrés 1, Marco Antonio Vásquez-AgustÃn 2, Justo Miguel Gracia-Jiménez 3, Reina Galeazzi-Isasmendi 1 and Francisco Morales-Morales 4
Round 1
Reviewer 1 Report
Within the paper Study of nitridation effect on structural, morphological and optical properties of GaAs films growth on silicon substrates via close space vapor transport technique authors presented GaAs deposition on Si substrate and then nitridation of GaAs film. Presented topic is interesting, however the paper needs some improvement before the publication. Below I listed my detailed remarks:
1. lines 22, 27 and many more - include citation in the sentence as it is a part of it
2. lines 44-45 - GaN is a III-V semiconductor material NOT II-IV!
3. line 67 - what was conductivity type of GaAs substrate used as a target?
4. what was a distance between Si substrate and GaAs target? Is it adjustable?
5. what about EDS/EDX results for sample M1 and M3 (table 1)?
6. Did authors consider XPS measurements to determine bonding configuration of the prepared GaN and GaAs surface?
7. line 110 - missing space in the 8mW
8. Please add for comparison purpose the JPDF data for GaAs in Figure 4 and for GaN and Ga2O3 in Figure 7
9. What is a thickness of M1-M3 GaAs film, and later what is a thickness of M4-M6 GaN film?
10. What is an origin of oxide in GaAs layers, as the deposition process is performed in hydrogen atmosphere?
11. Figure 6 - I am not sure that this feature can be assigned as GaxOy mode. Please note that frequenceis of 2*261 and 2*285 are exactly in this range what meant that you observed 2TO and 2LO modes of GaAs. See for instance: https://doi.org/10.1016/j.jpcs.2005.09.061 and https://doi.org/10.1103/PhysRevB.17.1865
12. Why authors do not observe Si signal at 519 cm-1 for GaAs films, while for nitrided sample the Si feature is visible in Figure 9?
13. How, based on simple XRD scan authors determined three lattice parameters of GaN unit cell?
14. The a=b unit cell of GaN is significantly smaller than expected (2.73 vs 3.89 A) what does it mean? Is the GaN layer relaxed?
15. line 172 - what are the lattice parameters for Ga2O3?
16. Lines 180-182 - authors claims that oxygen concentration decrease because of high temperature nitridation at 900 C. What about M2 and M3 fabricated at 900 and 1000C? In case of these two samples oxygen groups did not volatize during deposition?
17. Do authors have cross section SEM images of fabricated structures?
18. GaN band gap is 3.4 eV NOT 3.5 eV - line 233
Author Response
Response to Reviewer 1
Point 1: Lines 22, 27 and many more - include citation in the sentence as it is a part of it
Response 1: The reviewer's suggestion is appropriate, relevant changes were made throughout the document. We add new references; [1], [3], [4], [5], [6], [7], [9], [10], [11], [12], [13], [14], [15], [16] and [22] in order to complement the appropriate reference.
Point 2: Lines 44-45 - GaN is a III-V semiconductor material NOT II-IV!
Response 2: The reviewer's suggestion is appropriate. In fact line 48 was rewritten.
Point 3: Line 67 - what was conductivity type of GaAs substrate used as a target?
Response 3: The n-type conductivity of solid source (GaAs) was added in line 70.
Point 4: What was a distance between Si substrate and GaAs target? Is it adjustable?
Response 4: It is worth mentioning that the distance between the solid source (GaAs) and the tungsten filament is kept constant at 7mm. On the other hand, the distance between the solid source (GaAs) and silicon substrate is adjusted by the thickness of the quartz ring; 1 mm, 2 mm and 3mm, to obtain a temperature around ~1000, ~900 and ~800ºC respectively. Said information was added on lines 88-91.
Point 5: What about EDS/EDX results for sample M1 and M3 (table 1)?
Response 5: We only made the measurements on sample M2 to demonstrate the reduction of Arsenic composition and Nitrogen replacement after nitriding. The EDS values are qualitative, so it was not considered to measure the sample M1 and M3. In addition to this observation, Table 1 was modified, leaving only the data from sample M2.
Point 6: Did authors consider XPS measurements to determine bonding configuration of the prepared GaN and GaAs surface?
Response 6: In this work we focus on the study of the nitridation effect of GaAs films on silicon substrates, analyzing the morphological, structural and optical properties. As it is well mentioned, XPS could give us greater precision in the atomic compositions of the elements present in films. In fact, XPS analysis in-depth profile could give us better arguments respect to compositions; however, this technique is outside the scope of this work. We will plan to do these characterizations and other experiments for future work.
Point 7: Line 110 - missing space in the 8mW
Response 7: The reviewer suggestion was corrected in line 120. In fact, was corrected incident wavelength (633 nm) in Raman measurements.
Point 8: Please add for comparison purpose the JPDF data for GaAs in Figure 4 and for GaN and Ga2O3 in Figure 7
Response 8: The reviewer's suggestion is appropriate; their corresponding diffraction patterns were added in Figure 2 and Figure 5. For this, PDFs 32-0389 and 50-0792 were added.
Point 9: What is a thickness of M1-M3 GaAs film, and later what is a thickness of M4-M6 GaN film?
Response 9: The thicknesses of GaAs films with and without nitriding were measured by profilometry and cross section in SEM measurements as was suggested by the reviewer. In the case of GaAs film (M2), the thickness (~641 nm) was added to table 1. Likewise, information on the measurement instrumentation is written in line 123.
The thicknesses of M1 and M3 do not generate a trend with M2 and are of the order of M2 (M1 thickness = 659 nm M3 thickness = 630 nm)
Respects to M4-M6 thickness (323-342 nm) also were added in table 1. It is worth mentioning that the decrease in the thicknesses of nitrided films is almost double with respect to the non-nitrided M1-M3. This reduction is due to the rearrangement of the film structure due to the effect of annealing in nitriding process (900 ºC) as well as arsenic desorption and low nitrogen incorporation as can be seen in the EDS information (Table 1).
Point 10: What is an origin of oxide in GaAs layers, as the deposition process is performed in hydrogen atmosphere?
Response 10: The GaAs films growth on silicon substrate is carried out in a system in which there is a tungsten filament that reaches high temperatures (2000 °C). We assume that high temperature in the tungsten filament causes reactive hydrogen (H°) that attacks the quartz ring. Thus could be explains the possible source of oxygen that gives rise to the formation of Ga2O3 present in the X-ray diffraction spectra and EDS measurements. This effect is supported in different works, where the attack of a quartz solid source (SiO2) placed 3 mm from the filament tungsten generates a large amount of SiO(g) compounds for the formation of SRO or SiOx films. [https://doi.org/10.1016/j.jlumin.2013.07.013, https://doi.org/10.1155/2012/368268]. In this case, the distance between the filament and the quartz ring is around 7 mm, however, we believe that it still attack the ring quartz responsible for the oxygen present in the GaAs samples. Additionally, the presence of oxygen in the solid source (GaAs) and in the silicon substrate due to native oxide could also be pointed out as a possibility. Such information was added in lines 152-163.
Point 11: Figure 6 - I am not sure that this feature can be assigned as GaxOy mode. Please note that frequencies of 2*261 and 2*285 are exactly in this range what meant that you observed 2TO and 2LO modes of GaAs. See for instance: https://doi.org/10.1016/j.jpcs.2005.09.061 and https://doi.org/10.1103/PhysRevB.17.1865
Response 11: The reviewer's suggestion is appropriate. Lines 173-176 were rewritten as follows.
Furthermore, a band at 576 cm-1 corresponding to the 2LO mode of GaAs appears for sample M1. On the other hand, no signals or modes were observed regarding the Silicon substrate. Due to the average thickness of the GaAs samples (~ 643 nm), the Raman penetration depth is not enough for generate interaction with the Silicon substrate.
Point 12: Why authors do not observe Si signal at 519 cm-1 for GaAs films, while for nitrided sample the Si feature is visible in Figure 9?
Response 12: Because the GaAs (M2) sample has an average thickness of 641 nm, the penetration depth is not sufficient for it to interact with the silicon substrate. On the other hand, the M4-M6 layers have a thickness of less than 342 nm, enough to penetration depth. Pulzara-Mora, et al reports a penetration depth equal to 250nm for a laser line of 632.8 nm at 20 W. (https://www.redalyc.org/articulo.oa?id=94251122008 ).
Such information is discussed in lines 173-176.
Point 13: How, based on simple XRD scan authors determined three lattice parameters of GaN unit cell?
Response 13: For a hexagonal structure the parameters "a=b" and "c" can be calculated from interplanar spacings formula of diffractions planes (100) and (002). We add the important relationships to the manuscript to determine the lattice parameters and interplanar distances in lines 186-196.
It is worth mentioning that lattice parameters calculated in this second revision are close to those reported in the literature for GaN in bulk, therefore Table 3 was also updated. We appreciate your observation as it allows us to add important information.
Point 14: The a=b unit cell of GaN is significantly smaller than expected (2.73 vs 3.89 A) what does it mean? Is the GaN layer relaxed?
Response 14: The answer was added in line 195.
Point 15: Line 172 - what are the lattice parameters for Ga2O3?
Response 15: The new revision of lattice parameters calculated are close to GaN in bulk, so the idea written in the first version of the manuscript in line 172 was deleted. Line 195 in present manuscript explains the effect in GaN lattice.
Respect lattice parameters for Ga2O3, Kohn et al reported monoclinic crystal cell dimensions with the values a=12.23, b=3.04 and c=5.8 Ǻ and β=103.7º (https://doi.org/10.1063/1.1731237). We have a close correlation with two GaN parameters that are b=3.17 and c=5.18.
Point 16: Lines 180-182 - authors claims that oxygen concentration decrease because of high temperature nitridation at 900 C. What about M2 and M3 fabricated at 900 and 1000C? In case of these two samples oxygen groups did not volatize during deposition?
Response 16: Answering the question about deposition of GaAs films in samples M2 and M3, the interaction of the reactive hydrogen with the quartz ring promotes the formation of oxygen species, so there is no volatilization of oxygen in the film, but there is an incorporation of oxygen in the deposits of GaAs. XRD results in figure 2, confirm the presence of Ga2O3 in the GaAs samples.
Point 17: Do authors have cross section SEM images of fabricated structures?
Response 17: The SEM measurements were made in cross section, although it only served to corroborate the thickness of the samples measured by profilometry. Said thicknesses measured by both techniques are similar and are of the order of ~335 nm for the GaN films and ~643 nm for the GaAs samples.
Point 18: GaN band gap is 3.4 eV NOT 3.5 eV - line 233
Response 18: The reviewer suggestion was corrected in line 260.
Reviewer 2 Report
The authors suggested a method to synthesis GaN from GaAs films via CSVT. In order to characterize their resultant samples, the authors performed a set of surface measurements and kept strict comparison between different growth parameters, and both the qualitative and quantitative experiments are designed with scientific considerations. However, before considering to be published, the authors should address the following questions.
1. Please combine Figures 1 & 2, Figures 10 & 11, and Figures 12, 13 & 14, in order to save figure indices and merge results from the same techniques. This helps the manuscript be more organized.
2. Are Figures 5 & 8 directly screenshots from the tool-driving softwares? For formal research journal articles, please redraw the figures using Origin, et al.
3. It can be seen from Figure 6 that the characteristic Raman peaks are generally wide and low in intensity. This is probably a result of the combination of different GaxAsy compositions in the resultant film samples that caused the dispersion of the peaks. The question is, will these different compositions affect the overall properties/performances of the material?
4. Also in Figure 6, the characteristic peak for Si at ~ 520 cm-1 is not observed. Why?
Author Response
Response to Reviewer 2
Point 1: Please combine Figures 1 & 2, Figures 10 & 11, and Figures 12, 13 & 14, in order to save figure indices and merge results from the same techniques. This helps the manuscript be more organized.
Response 1: The reviewer's suggestion is appropriate. The changes suggested were made and added to the new version of the manuscript.
Point 2: Are Figures 5 & 8 directly screenshots from the tool-driving softwares? For formal research journal articles, please redraw the figures using Origin, et al.
Response 2: Unfortunately the EDS data provided to us are only images, so it would be difficult to extract data from an image. On the other hand, in many publications EDS results are reported as an image.
Point 3: It can be seen from Figure 6 that the characteristic Raman peaks are generally wide and low in intensity. This is probably a result of the combination of different GaxAsy compositions in the resultant film samples that caused the dispersion of the peaks. The question is, will these different compositions affect the overall properties/performances of the material?
Response 3: We agree that the degree of impurity causes tensions or stress in the crystal lattice. In our case, the GaxOy groups present in the samples with and without nitriding can affect the performance of the material.
Point 4: Also in Figure 6, the characteristic peak for Si at ~ 520 cm-1 is not observed. Why?
Response 4: Because the GaAs (M2) sample has an average thickness of 641 nm, the penetration depth is not sufficient for it to interact with the silicon substrate. On the other hand, the M4-M6 layers have a thickness of less than 342 nm, enough to penetration depth. Pulzara-Mora, et al reports a penetration depth equal to 250nm for a laser line of 632.8 nm at 20 W. (https://www.redalyc.org/articulo.oa?id=94251122008 ).
Such information is discussed in lines 173-176.
Round 2
Reviewer 1 Report
I appreciate the corrections made by authors. However, there are still some issues which should be adressed before the publications:
- missing space: line 91
- citation is a part of the sentence: lines 172, 183, 185 and many moore
-
Author Response
Response to Reviewer 1
Response 1 and 2: We thank each of your comments for which we have made the pertinent changes.
Reviewer 2 Report
Response to:
2. We do feel it inappropriate to only post a screenshot instead of re-drawn formal plots. Even if it were from collaborators, the authors are requested to manage to obtain the original data and draw the plot using the data. The key point is, the EDS data is not images, but spectra. If the authors (and in turn, the readers) cannot get precise intensity information for each emission energy, how do we discover any potential new phenomenon that might only differ from common sense by tiny amounts? We also ask attention from the editors on this issue. Raw data is even more important than the conclusions themselves.
3. The authors are requested to elucidate this effect in the context.
Author Response
Response to Reviewer 2
Response: We thank each of your indications, we inform you that we obtained the data. The EDS spectra were plotted in figures 3 and 6 respectively.
