Novel High-Contrast Photoacoustic Imaging Method for Cancer Cell Monitoring Based on Dual-Wavelength Confocal Metalenses

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript by Zhang et al. reports a high-contrast photoacoustic (PA) imaging methodology based on a dual-wavelength confocal metalens for monitoring the dissemination of cancer cells and informing subsequent cancer treatment strategies. In general, the paper is scientific sound. However, there are some comments which need to address clearly before I can recommend its publication in Photonics.

 

1. The numerical aperture of the designed metalens is about 0.25. Some discussions about the reason for choosing such numerical aperture need to be added in the revised manuscript.
2. The metalens is composed of two metasurfaces that perform filtering and focusing functions. Some details about how to design the metasurface for filtering need to be provided.
3. For the section 3.3, the consumed time of Monte Carlo simulation is better to be provided.
4. In the figure caption of Figure 3, the spaces between the 532/785 and “nm” are neglected.

Comments for author File: Comments.pdf

Author Response

Dear reviewers,

Thank you very much for your comments and professional advice. These points help to improve the academic rigor of our article. Based on your suggestions and requests, we have corrected and revised the original manuscript. We first made some notes on the revised manuscript and then responded to the reviewers' comments one by one.

Here are some notes on the response:

Three colors of text can be seen in the revised manuscript. The black text is unmodified. The red text is modified according to the suggestions of Reviewer #1. The green text correspond to the suggestions of Reviewer #2. The details of the text and image rectification will be mentioned in our specific replies to reviewers.

 

The following are the specific responses to the two reviewers:

Reviewer #1

Comments:

The manuscript by Zhang et al. reports a high-contrast photoacoustic (PA) imaging methodology based on a dual-wavelength confocal metalens for monitoring the dissemination of cancer cells and informing subsequent cancer treatment strategies. In general, the paper is scientific sound. However, there are some comments which need to address clearly before I can recommend its publication in Photonics.

1. Comment: The numerical aperture of the designed metalens is about 0.25. Some discussions about the reason for choosing such numerical aperture need to be added in the revised manuscript.

Reply: The design with NA=0.25 balances lateral resolution (approximately λ/2NA ≈ 1.06–1.57 μm) and imaging depth (low NA reduces scattering, making it suitable for deep tissue imaging), meeting the requirements of photoacoustic imaging in skin models. The revisions are shown in red in paragraph 1 of the Section 3.2.

2. Comment: The metalens is composed of two metasurfaces that perform filtering and focusing functions. Some details about how to design the metasurface for filtering need to be provided.

Reply: Thanks for your comments. The unit structure and scanning results of the metasurface for filtering have been incorporated into the manuscript (They are shown in Figure 5). We believe that it can facilitate a more comprehensive understanding of its operational mechanism.

3. Comment: For the section 3.3, the consumed time of Monte Carlo simulation is better to be provided.

Reply: We sincerely appreciate the reviewer’s insightful comment regarding the computational time of the Monte Carlo simulation in Section 3.3. The reviewer raises a valid point—the execution time of such simulations is indeed highly dependent on the hardware configuration of the computer used (e.g., CPU, RAM, parallelization capabilities). For this calculation, our configuration is as follows: the CPU is an Intel Core-i9-14900K, and the GPU is an NVIDIA GeForce RTX 3060. The average time for the Monte Carlo simulation is approximately 5 hours. As a result, reporting absolute time values may have limited reference value for readers with different computational resources.

 

4. Comment: In the figure caption of Figure 3, the spaces between the 532/785 and “nm” are neglected.

Reply: We sincerely appreciate the reviewer’s careful attention to detail. The spaces between "532/785" and "nm" in the caption of Figure 3 have now been corrected as suggested. Thank you for catching this oversight—we have revised the manuscript accordingly.

We would like to thank the referee again for taking the time to review our manuscript.

Yours sincerely,

Bin Wang

 

 

 

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors This manuscript presents and simulates a photoacoustic  imaging based on a dual-wavelength confocal metalens. It may hold practical value for cancel cell monitoring. However, regarding the metalens design, it's suggested to further enhance its innovation to meet experimental requirements. Detailed comments are as follows:

1. If this proposal is to be experimentally demonstrated, what's the size of the metalens to be fabricated? Is it much larger than the simulated one? There maybe difference between the focusing propertise of small-area metalens in simulation and large-area ones in experiment.
2. The designed metalens with spatial multiplexing technology uses different materials and nanopillar heights for different wavelengths, increasing fabrication difficulty. It's recommended to adopt a more experimentally feasible design.
3. The author should explain how the "total power of the focused spot at the focal plane" is calculated for focusing efficiency. Also, after combining into a dual-wavelength metalens, the focusing efficiencies of 532nm and 785nm differ greatly despite similar efficiencies for single-wavelength ones. A more specific explanation is needed.
4. Using spatial multiplexing technology may deform the focus spot in the xy-plane. Please provide the PSF of the dual-wavelength metalens at the focal plane and discuss its impact on subsequent detection.
5. For Figure 7's filtered metasurface, a more detailed description of its structure and transmittance is needed. Furthermore, can the filter and metalens be integrated into a single layer metasurface to better reflect the metalens's integration advantage?

Author Response

Dear reviewers,

Thank you very much for your comments and professional advice. These points help to improve the academic rigor of our article. Based on your suggestions and requests, we have corrected and revised the original manuscript. We first made some notes on the revised manuscript and then responded to the reviewers' comments one by one.

Here are some notes on the response:

Three colors of text can be seen in the revised manuscript. The black text is unmodified. The red text is modified according to the suggestions of Reviewer #1. The green text correspond to the suggestions of Reviewer #2. The details of the text and image rectification will be mentioned in our specific replies to reviewers.

The following are the specific responses to the two reviewers:

Reviewer #2

Comments:

       This manuscript presents and simulates a photoacoustic imaging based on a dual-wavelength confocal metalens. It may hold practical value for cancel cell monitoring. However, regarding the metalens design, it's suggested to further enhance its innovation to meet experimental requirements. Detailed comments are as follows:

1. Comment: If this proposal is to be experimentally demonstrated, what's the size of the metalens to be fabricated? Is it much larger than the simulated one? There maybe difference between the focusing propertise of small-area metalens in simulation and large-area ones in experiment.

Reply: We sincerely appreciate the reviewer's insightful comments regarding the experimental validation of the metalens design. The concerns raised about potential discrepancies between simulation and experimental performance due to size differences are indeed valuable and well-noted. In our proposed approach for photoacoustic imaging, the system does not rely on a single large-area metalens but rather employs an integrated array of microscale metalenses (with a designed radius of 25 μm and each unit radius approximately several tens of nanometers).  This design allows for sequential scanning of the target tissue area, thereby efficiently delivering optical energy to subsurface cells while maintaining the focusing characteristics observed in simulations.  By leveraging this strategy, we aim to bridge the gap between small-scale simulation results and practical implementation without compromising performance. However, it must be acknowledged that the fabrication process is highly complex and prohibitively expensive. The unit structure radius ranges from 10 nm to 90 nm, with a typical processing error of approximately 30%, which can be expected to induce significant variations in related performance and properties. Therefore, the reviewer's point on this matter is valid.

Thank you again for this constructive feedback, which has helped us clarify this important aspect of our methodology.

2. Comment: The designed metalens with spatial multiplexing technology uses different materials and nanopillar heights for different wavelengths, increasing fabrication difficulty. It's recommended to adopt a more experimentally feasible design.

Reply: Thank you for your valuable suggestions. Regarding the potential fabrication process of the metalens designed using spatial multiplexing technology, which involves the use of different materials and nanopillar heights at different wavelengths, we have provided additional explanations in the final paragraph of Section 3.2 and Fig.9.

To address this challenge, a possible solution is to deposit Si and TiO₂ films with different thicknesses by physical vapor deposition (PVD), where the film thickness can be controlled by adjusting the deposition time. Subsequently, both Si and TiO₂ can be etched simultaneously using inductively coupled plasma (ICP) etching with appropriate gas mixtures.We sincerely appreciate your insightful comments, which have helped improve the technical feasibility of our manuscript.

3. Comment: The author should explain how the "total power of the focused spot at the focal plane" is calculated for focusing efficiency. Also, after combining into a dual-wavelength metalens, the focusing efficiencies of 532nm and 785nm differ greatly despite similar efficiencies for single-wavelength ones. A more specific explanation is needed.

Reply: Thank you for your valuable comments. In this study, the total power of the focused spot is quantified as the cumulative optical intensity encompassed within the 2×FWHM boundary. Regarding The added contents are shown in green in paragraph 3 of the Section 3.1.

The difference in focusing efficiency between the two wavelengths, we attribute this primarily to the higher absorption of near-infrared light (785 nm) compared to visible light (532 nm) in the spatially multiplexed and filtering surface design. Additionally, there may be unintended modulation of 785 nm light by the phase profile optimized for 532 nm, leading to energy dispersion. We have supplemented these contents to our paper in green in paragraph 2 of the Section 3.2.

To address this issue, potential solutions could include exploring materials with lower absorption in the infrared regime or optimizing the geometric parameters of the unit structures. Your insightful suggestions have significantly enhanced the rigor of our study, and we sincerely appreciate your guidance.

4. Comment: Using spatial multiplexing technology may deform the focus spot in the xy-plane. Please provide the PSF of the dual-wavelength metalens at the focal plane and discuss its impact on subsequent detection.

Reply: Thank you for your valuable suggestion. We have now included the focal-plane intensity distributions for dual-wavelength focusing in Figure 8. While the focal spot exhibits some distortion, the majority of the optical energy remains concentrated within twice the FWHM of the focal point. Since the primary function of the metalens in photoacoustic imaging is to deliver optical energy into the skin, this performance still meets the requirements for effective photoacoustic excitation.

We greatly appreciate your insightful feedback, which has helped strengthen the clarity and rigor of our manuscript.

5. Comment: For Figure 7's filtered metasurface, a more detailed description of its structure and transmittance is needed. Furthermore, can the filter and metalens be integrated into a single layer metasurface to better reflect the metalens's integration advantage?.

Reply: Thanks for your comments. The unit structure and scanning results of the metasurface for filtering have been shown in Figure 5. This enables readers to gain a clearer understanding of the operational principles of filtering metasurfaces. However, with current technology, integrating both filtering and focusing functions onto a single metasurface remains highly impractical. Even with various phase modulation approaches, achieving this functionality effectively is still challenging. Therefore, we employ a multilayer structure to realize these two functions.

We would like to thank the referee again for taking the time to review our manuscript.

 

Yours sincerely,

Bin Wang

 

 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have not fully addressed my previous comments. The specific issues are outlined below:

Comment 1:
The authors mentioned that they plan to adopt an integrated array of microscale metalenses for practical applications. However, this point is not discussed in detail in the manuscript. In particular, Section 3.3 lacks any explanation or comparison related to this approach. The figures and results presented in that section only reflect the performance of a single large-area metalens, without discussing the potential differences and trade-offs between the two configurations.

Comment 2:
The authors proposed using a PVD method to deposit Si and TiO₂ at different locations on the same plane. Firstly, this technique is inherently challenging. Even assuming it is technically feasible, PVD can only deposit amorphous silicon, which has a high absorption coefficient at the wavelength of 786 nm. Based on the refractive index values cited in the manuscript, it appears that the simulations are based on crystalline silicon, which indeed exhibits low absorption at 786 nm but cannot be deposited using PVD. The authors should clarify which material is actually used in the simulations and whether such a structure can realistically be fabricated. In fact, I recommend the authors consider using only one material for focusing both wavelengths to avoid the issues raised in this comment.

Comment 3:
The explanation given by the authors for the low focusing efficiency at 786 nm is not convincing. First, both the standalone metalens and the color filter for 785 nm light exhibit high efficiency individually. Therefore, attributing the reduced efficiency to integration alone is not reasonable. Second, even if there is crosstalk between different wavelengths, it should mainly affect the focusing point spread function (PSF), not cause a significant drop in overall efficiency.

Author Response

Dear reviewer,

Thank you very much for your comments and professional advice. These points help to improve the academic rigor of our article. Based on your suggestions and requests, we have corrected and revised the original manuscript. We first made some notes on the revised manuscript and then responded to the reviewer's comments one by one.

 Here are some notes on the response:

Three colors of text can be seen in the revised manuscript. The black text is unmodified. The green text correspond to the suggestions of Reviewer #2. The details of the text and image rectification will be mentioned in our specific replies to reviewers.

The following are the specific responses to the reviewer:

 Reviewer #2

1. Comment: The authors mentioned that they plan to adopt an integrated array of microscale metalenses for practical applications. However, this point is not discussed in detail in the manuscript. In particular, Section 3.3 lacks any explanation or comparison related to this approach. The figures and results presented in that section only reflect the performance of a single large-area metalens without discussing the potential differences and trade-offs between the two configurations.

Reply: In our preliminary design, the practical photoacoustic imaging system should consist of a large-scale metasurface lens array composed of numerous individual metalenses (radius = 25 μm). The primary function of this array is to achieve dual-wavelength optical confocal focusing and efficiently couple light energy into biological tissues (e.g., skin). Given the required high uniformity in optical performance across all constituent metalens units within the array, this study focuses on characterizing the focusing properties of a single metalens element, which effectively represents the collective performance of the entire array system. This approach is justified by the periodic nature and fabrication consistency of the array architecture.

2. Comment: The authors proposed using a PVD method to deposit Si and TiO2 at different locations on the same plane. Firstly, this technique is inherently challenging. Even assuming it is technically feasible, PVD can only deposit amorphous silicon, which has a high absorption coefficient at the wavelength of 786 nm. Based on the refractive index values cited in the manuscript, it appears that the simulations are based on crystalline silicon, which indeed exhibits low absorption at 786nm but cannot be deposited using PVD. The authors should clarify which material is actually used in the simulations and whether such a structure can realistically be fabricated. In fact, I recommend the authors consider using only one material for focusing both wavelengths to avoid the issues raised in this comment.

Reply: We sincerely thank the reviewer for pointing out the potential fabrication issues of the proposed metalens. Regarding the choice of materials used in the simulation, we have added updates marked in green in Section 2.2. As for the potential fabrication process of the metalens designed using spatial multiplexing techniques, we have provided new explanations in the green text at the end of Section 3.2 and in Figure 9.

To address this challenge, one possible solution is to first deposit Si using chemical vapor deposition (CVD), followed by another CVD process to deposit TiO₂. Compared to using TiO₂ alone, the combination of Si and TiO₂ provides greater flexibility in phase control. In contrast to using Si alone—despite the relatively high absorption of crystalline Si at 785 nm—the combination with TiO₂ helps to reduce absorption losses and improve overall efficiency. From the fabrication perspective, single-material systems are indeed advantageous in terms of simplicity. However, such designs may not fully meet the performance requirements or may only partially satisfy the design goals. Therefore, we have ultimately chosen to integrate both Si and TiO₂ in our design. We have tried a lot of chances to make this design good.  Please understand.

We greatly appreciate your insightful feedback, which has helped strengthen the rigor of our manuscript.

 

3. Comment: The explanation given by the authors for the low focusing efficiency at 786 nm is not convincing. First, both the standalone metalens and the color filter for 785 nm light exhibit high efficiency individually. Therefore, attributing the reduced efficiency to integration alone is not reasonable. Second, even if there is crosstalk between different wavelengths, it should mainly affect the focusing point spread function (PSF), not cause a significant drop in overall efficiency.

Reply: Possible explanation is that the dual-metasurface structure introduces additional interfacial reflections. When the 532 nm light satisfies the λ/4 anti-reflection condition, the 785 nm light may undergo destructive interference due to deviations in optical thickness from the optimal value. It may lead to lower efficiency under multi-wavelength conditions compared to single-wavelength measurements of unit structures.

Additionally, the silicon unit height of 1.2 μm is close to the 785 nm wavelength, potentially exciting higher-order waveguide modes. A portion of the energy may dissipate as evanescent waves in the SiO₂ substrate rather than contributing to far-field focusing

Author Response File: Author Response.docx

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

The authors did not fully address my previous comments, resulting in continued ambiguity and even incorrect statements in the manuscript. Therefore, I still cannot recommend this paper for publication at this stage.

Regarding the first comment, I agree with the authors’ explanation and suggest that it be clearly incorporated into the revised manuscript.

Regarding the second comment, the reference cited by the authors (Ref. 35) describes single-crystal silicon films grown over silica-covered regions of a single-crystal silicon wafer. In essence, this still involves depositing single-crystal silicon onto a single-crystal silicon substrate. The cited work does not provide evidence that single-crystal silicon can be directly deposited onto a silica wafer. Therefore, the authors’ proposed fabrication approach for their designed structure remains incorrect. To avoid confusion, I also recommend that all instances of "silicon" in the revised manuscript be explicitly clarified as "single-crystal silicon" where appropriate.

Regarding the third comment, I do not agree with the authors’ explanation. Their response is entirely speculative, relying on vague language such as “may,”. I suggest the authors should support simulations to explain this point. Furthermore, the proposed physical mechanism is unconvincing. For instance:

• What is the thickness of the silica layer used in the current simulations?

• If the effect is due to an anti-reflection condition, does varying the thickness of the silica layer enhance the focusing efficiency at 785 nm?

• If evanescent waves are responsible for the reduced efficiency, why is this issue absent in the case of single-wavelength metalenses?

Author Response

Dear reviewers,

 Thank you very much for your comments and professional advice. Major revision has been made. These points help to improve the academic rigor of our article. Based on your suggestions and requests, we have corrected and revised the original manuscript. We first made some notes on the revised manuscript and then responded to the reviewers' comments one by one.

 Here are some notes on the response:

 

Three colors of text can be seen in the revised manuscript. The black text is unmodified. The red text correspond to the suggestions of Reviewer. The details of the text and image rectification will be mentioned in our specific replies to reviewers.

 

The following are the specific responses to the two reviewers:

 Reviewer

1. Comment: The authors mentioned that they plan to adopt an integrated array of microscale metalenses for practical applications. However, this point is not discussed in detail in the manuscript. In particular, Section 3.3 lacks any explanation or comparison related to this approach. The figures and results presented in that section only reflect the performance of a single large-area metalens without discussing the potential differences and trade-offs between the two configurations.

Reply: Thank you very much for your constructive feedback.

Compared to conventional lenses, metasurfaces offer significant advantages in terms of thickness, weight, ease of integration, and capability for complex optical functions. For the first time, this study proposes a dual-wavelength confocal metalens structure for photoacoustic imaging to distinguish cancer cells from normal cells. Throughout the design process of the entire structure, extensive optimization and verification efforts were conducted. The final results demonstrate successful differentiation between cancerous and normal cells while maintaining anti-interference capabilities. The optical properties of individual metalenses and metalens arrays show high consistency, meaning the performance of a single metalens is representative of the relevant characteristics of the array. Due to technical complexities, the practical implementation of the metalens array was not realized. This addition would make clear that, although these evaluations were beyond the current scope, they are recognized as important and will be addressed in future work.

Additional clarifications are provided in red text in the final part of Section 3.3.

2. Comment: The authors proposed using a PVD method to deposit Si and TiO2 at different locations on the same plane. Firstly, this technique is inherently challenging. Even assuming it is technically feasible, PVD can only deposit amorphous silicon, which has a high absorption coefficient at the wavelength of 786 nm. Based on the refractive index values cited in the manuscript, it appears that the simulations are based on crystalline silicon, which indeed exhibits low absorption at 786nm but cannot be deposited using PVD. The authors should clarify which material is actually used in the simulations and whether such a structure can realistically be fabricated. In fact, I recommend the authors consider using only one material for focusing both wavelengths to avoid the issues raised in this comment.

Reply: We sincerely thank the reviewer for highlighting the potential fabrication challenges associated with the proposed metalens. Regarding the material selection used in our simulations, we have carefully considered the comments and have made corresponding revisions marked in red in Section 2.2. As for the potential fabrication process of the metalens designed using spatial multiplexing techniques, we have provided additional explanations in red text at the end of Section 3.2 and in Figure 9. We have accordingly unified the materials and adopted a single material, TiO₂, for the entire design.

We greatly appreciate your insightful feedback, which has helped strengthen the rigor of our manuscript.

 

3. Comment: The explanation given by the authors for the low focusing efficiency at 786 nm is not convincing. First, both the standalone metalens and the color filter for 785 nm light exhibit high efficiency individually. Therefore, attributing the reduced efficiency to integration alone is not reasonable. Second, even if there is crosstalk between different wavelengths, it should mainly affect the focusing point spread function (PSF), not cause a significant drop in overall efficiency.

Reply: We sincerely thank the reviewer for pointing out the issues with the proposed metalens. The filtering structure plays a critical role in controlling the transmission of light at specific wavelengths. Due to the adoption of spatial multiplexing technology, different filtering structures are integrated on the same side of the lens. These distinct configurations exhibit varying transmittance for 532 nm and 785 nm light, resulting in reduced overall transmittance of the structure and consequently leading to a decline in the focusing efficiency of the entire system. Figure 5 in Section 2.3 illustrates the transmittance of different structures across various wavelengths.

Yours sincerely

 

Author Response File: Author Response.docx

Round 4

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have answer all my questions and improved the manuscript. It can be published in Photonics