Growth of Low-Temperature Epitaxial Lithium Niobate Thin Films and Guided-Wave Optical Properties

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Bui et al reported growth LiNbO₃ thin films using Chemical Beam Vapor Deposition and measured the morphological properties using XRD and Raman measurements. In addition, the optical properties in terms of bi-refringence (Dn). Authors aimed in this work to investigate the crystallinity and optical properties of LiNbO₃ films deposited on sapphire by CBVD at a lower deposition temperature than previously reported. The work seems interesting.  However, I have one major comment on this manuscript is that “what are the main application and added advantage of this thin films grown by CBVD compared over the bulk LinbO₃”. In the manuscript they mentioned about applications “many applications in photonic and acoustic industry, such as waveguide-based optical devices, surface acoustic wave devices (SAW), memory units, and neuromorphic systems”, but in the manuscript I did not see the practical application or optical measurements except the bi-refringence. How these thin films show better performance over the bulk materials?

Some typo mistakes:

(1)   Line 33, truncation of line,

(2)   Line 66, double time Figure 1,

(3)   In Figure .3, 152 cm-1, 870 cm-1 should be 152 cm^-1, 870 cm^-1

(4)   In Figure 8 legend, no and ne should be n_o and n_e

Comments for author File: Comments.pdf

Author Response

Dear,

Thank your for your review.

In this manuscript we focus only on LN deposition by CBVD and examine its properties compared to other methods and bulk LN. Currently, thin-film LN (TFLN) fabricated by smart-cut or crystal ion slicing (CIS) is commonly used in practical application. The growth of LN is still challenging with film quality, thickness. However, we are working on investigate photonics and acoustic application, and get some promising results on SAW by using the same sample in this manuscript, we will publish soon. The better performance of thin film compared to bulk is device miniaturization and enhanced functional properties. For example in RF filter, the thinner thickness is the higher frequency is. As I know the TFLN by CIS from bulk LN can not go lower than 300 nm. However, we cannot ignore the difficulty of manufacturing technology. In addition, the crystallinity of TFLN by CIS is almost equivalent to us (S. Huang, DOI:10.3390/electronics12132964).

If you have any question, if you have any question please feel free to ask me.

Best Regards,

Thanh BUI

 

Reviewer 2 Report

Comments and Suggestions for Authors  

After carefully reading the article "Growth of Low Temperature Epitaxial Lithium Niobate Thin Films and Guided-Wave Optical Properties," I find it to be interesting and sufficiently novel. However, a few corrections are required before acceptance, as listed below:

1. Line 33 needs revision.
2. Please discuss why the sapphire peaks in the Raman signal intensified.
3. What is the rationale behind presenting the XRD peak in Figure 4 on a log scale?
4. The lower X-ray diffraction peak of LN is not observed in Figure 4. This should be discussed in the revised manuscript.

Author Response

Dear,

Thank you for your review.

Please find my answer below:

1. Line 33 needs revision.

Answer: Done

2. Please discuss why the sapphire peaks in the Raman signal intensified.

Answer: Raman is a light scattering technique. If the thickness of thin film is too thin, the light can penetrate the substrate, leading to the appearance of substrate peaks on the Raman spectrum. About the limitation of film thickness to avoid or reduce the intensity of substrate, at the moment I cannot give the answer. Because it also depends on the transparency of thin film, we have not investigated the value yet. According to literature, LN transparent from 0.32 to 5 micron

3. What is the rationale behind presenting the XRD peak in Figure 4 on a log scale?

Answer: The log scale can show almost lower intensity peaks. In this manuscript, the intensity of (006) orientation is approximately 10⁴, if the other orientations of LiNbO₃ or second phases is about 10¹ to 10², it is very difficult to see in linear scale and more difficult to see it on figure of manuscript.

4. The lower X-ray diffraction peak of LN is not observed in Figure 4. This should be discussed in the revised manuscript.

Answer: Do you mean we need to list other orientations of LN? According to JCPDS sheets of LN (#00-020-0631), there are many characteristic peaks. In most of publication that focus on growth of LN film, they focus on c-axis (006) orientation of LN on c-plane sapphire. Other orientation can observed on other kind of substrate.

I hope my answers will satisfy you. if you have any question please feel free to ask me.

Best Regards,

Thanh BUI

Reviewer 3 Report

Comments and Suggestions for Authors

In this paper, Bui et al. show a novel approach to grow single-crystalline thin film lithium niobate on a c-plane sapphire wafer at a reduced temperature. They carried out extensive analysis on the crystallinity and optical properties of the deposited film. The growth technique and metrology studies of material properties are sounded and well explained. The Raman spectroscopy shows that the material composition is predominantly LiNbO₃ and the XRD data confirms the material purity and crystallinity. These findings are furthered strengthened by optical measurements, where the refractive index and bi-refringence properties are measured and compared to literature.

 

While the presented results clearly support the finding of an epitaxial growth of single-phase c-axis oriented LiNbO₃ layer on sapphire wafer, I have several doubts regarding the novelty and significance of the research results before it can be published on Photonics. My major concern is the novelty and significance of the reported epitaxial growth technique. What is the significant advantages of growing TFLN at a reduced temperature? In the introduction, the authors mentioned that the high temperature method will lead to material cracks/defects and increase manufacturing costs. However they didn't highlight the advantages of their low temperature approach in this regard. Does it help at improving the quality and purity of LiNbO₃ layer or it is just an alternative approach with a lower temperature so the power consumption is lower?  In the conclusion part, the authors stated that the film quality is the same, which doesn't seem to be convincing enough as a novel and unique growth technique. Beside the above comment, I also have a few questions regarding the results in this paper:

 

1. The deposited layer thickness is ~357 nm, is this the maximal layer thickness with this technique? Due to the lattice mismatch and different thermal expansion rates, growing high quality TFLN via epitaxy with a layer thickness for photonic and acoustic applications always remain challenging. The authors are encouraged to address the layer thickness range with their technique. 

2. As pointed out by the authors, the grown LiNbO₃ film shows several cracks under SEM (line 136), which will limit its performance for photonic applications. Have the authors tried anything to reduced this phenomenon?  The reported rms from AFM is 2.0 nm, whereas in figure 6 the scale bar shows variation up to 0.02 um (20 nm). The authors should verify these values. If the scale bar is correct, then a 20 nm height change in a 357 nm thick layer is quite significant for any real applications.

3. The extracted FWHM from XRD measurements is 0.04°, which is still one order of magnitude higher than the bulk one. How is this value compared to other epitaxial TFLN? 

4. The investigation of refractive index and bi-refringence is very extensive, however some studies or estimations on optical loss are encouraged since this is what differentiate a high quality LN film from an average one.

5. In the introduction, the authors mentioned that optical measurements support index mapping at any area of the wafer. Therefore, a summary of the layer uniformity should be provided. 

 

 

 

Comments on the Quality of English Language

There is no general concern about the fluency of this paper. However, the authors should avoid using vague descriptions such as "probably compete with" in line 129.  The  benefits and novelty of growing epitaxy LN at a reduced temperature should also be emphasized both in the introduction and conclusion parts.  The authors should also correct formatting mistakes such as the disconnection at line 33. 

Author Response

Dear,

Thank you for your review, it gives me idea to modify my manuscript better.

Please find my answers in blue:

1. The deposited layer thickness is ~357 nm, is this the maximal layer thickness with this technique? Due to the lattice mismatch and different thermal expansion rates, growing high quality TFLN via epitaxy with a layer thickness for photonic and acoustic applications always remain challenging. The authors are encouraged to address the layer thickness range with their technique.

In our technique, the growth rate can go higher than 100 nm/h, it is possible to grow higher thickness. As you said, due to the lattice mismatch and different thermal expansion coefficient, our main challenge is crack at higher thickness. In our technique, we can deposit a thickness gradient sample on 8” wafer, we have done it, and at thickness higher than approx. 270 nm crack appears. However, we have some ideas to improve it, I will answer question 2.

We are investigating our layers on photonic and acoustic applications.

• On photonic applications, we need higher thickness, at least 500 nm based on our partner demand. The second approach is used bilayer, it means that another layer will deposit on top of LN layer.
• On acoustic applications, it depends on what kind of devices, BAW or SAW. We have achieved some results on SAW, and we will submit it soon. 350 nm of TFLN (same sample on this manuscript) is suitable for SAW applications. In addition, S. Huang et al, 2023 fabricated SAW on 275 nm of TFLN. However, the shorter period of IDTs is a challenge for technology due to the lower thickness. On BAW, the fabrication technology is more complex than SAW, but it is suitable for 400 nm of TFLN (Plessky, V. et al, 2019, DOI: 10.1109/MWSYM.2019.8700876)

2. As pointed out by the authors, the grown LiNbO3 film shows several cracks under SEM (line 136), which will limit its performance for photonic applications. Have the authors tried anything to reduced this phenomenon? The reported rms from AFM is 2.0 nm, whereas in figure 6 the scale bar shows variation up to 0.02 um (20 nm). The authors should verify these values. If the scale bar is correct, then a 20 nm height change in a 357 nm thick layer is quite significant for any real applications.

We are optimizing the process to avoid cracks while maintaining the properties of TFLN (buffer layer, step deposition…). For AFM, I've contacted the person who have responsibility on it.

3. The extracted FWHM from XRD measurements is 0.04°, which is still one order of magnitude higher than the bulk one. How is this value compared to other epitaxial TFLN?

In my bibliography, the lowest FWHM of TFLN deposited on C-plane sapphire is 0.0024°(narrow peak) and 0.03° (full peak) by MBE, reported by M.Tellekamp et al (DOI : 10.1007/s11664-016-4986-3). Other technique, such as MOCVD and PLD have achieved lowest FWHM at 0.04°. I updated these information in new version.

4. The investigation of refractive index and bi-refringence is very extensive, however some studies or estimations on optical loss are encouraged since this is what differentiate a high quality LN film from an average one.

In M-line system from Metricon, we can measure the optical loss of thin film. However, we are crack on the surface, It is not easy to make this measurement. 

5. In the introduction, the authors mentioned that optical measurements support index mapping at any area of the wafer. Therefore, a summary of the layer uniformity should be provided.

It is my mistake, index mapping for other publication. I remove this sentence in new version.

Thanks again. If you have any question please feel free to ask me.

Best Regards,

Thanh BUI

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The paper can be accept in the present form.