Finite Element Optimization of 3D Abiotic Glucose Fuel Cells for Implantable Medical Devices
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors-
The methodology section lacks rigorous mathematical formulation. The calculation of the rate constant K, the treatment of reaction–diffusion coupling, and the boundary assumptions are presented in an overly simplified manner and require clearer justification.
-
The study relies exclusively on COMSOL simulations without any experimental validation. To strengthen the reliability of the findings, the authors should provide sensitivity analyses, convergence tests, or comparisons with experimental results reported in the literature.
-
The discussion section primarily reiterates simulation trends rather than providing deeper mechanistic insights. The authors are encouraged to analyze the underlying causes of diffusion limitations, the trade-offs in electrode spacing, and the practical manufacturability of the proposed designs.
-
The comparison with existing implantable power sources is overly simplistic. A more quantitative evaluation of lifetime, power density, and device size is necessary to highlight the practical advantages and limitations of the proposed approach.
Author Response
Comment 1 :[The methodology section lacks rigorous mathematical formulation. The calculation of the rate constant K, the treatment of reaction–diffusion coupling, and the boundary assumptions are presented in an overly simplified manner and require clearer justification.]
Response 1: [We appreciate this comment. We have derived the formula for the reaction constant k and provided supplementary explanations. A more detailed explanation has been provided for the boundary assumptions and the formation of concentration gradients due to diffusion-reaction coupling effects.]
Comment 2 :[The study relies exclusively on COMSOL simulations without any experimental validation. To strengthen the reliability of the findings, the authors should provide sensitivity analyses, convergence tests, or comparisons with experimental results reported in the literature.]
Response 2:[Reaction rate dynamics have a complex set of influencing factors: Geometric surface, materials, nano-structuring, etc.
We agree that a typical simulation paper would connect simulations to experimental results. But given the complexities above, we thought that there was a danger of presenting potentially over-optimistic experimental data.
As such, we decided to review the entire field. We present the prior data in Figure 3, create 3 scenarios of levels of optimization for a realistic long-term implant, and then work backwards from there to calculate the reaction rate.
Diffusion dynamics are very well established, and we don’t feel the need for further calibration. Hence, we believe our approach to match our simulations to prior data is the best way to get realistic conclusions based on where the field currently is.]
Comment 3:[The discussion section primarily reiterates simulation trends rather than providing deeper mechanistic insights. The authors are encouraged to analyze the underlying causes of diffusion limitations, the trade-offs in electrode spacing, and the practical manufacturability of the proposed designs.]
Response 3:[We appreciate this comment. We have updated the discussion accordingly]
Comment 4:[The comparison with existing implantable power sources is overly simplistic. A more quantitative evaluation of lifetime, power density, and device size is necessary to highlight the practical advantages and limitations of the proposed approach.]
Response 4:[A more detailed explanation of existing implantable batteries has been provided, clarifying the advantages of fuel cells.]
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsIn this MS, the optimization of abiotic and biotic fuel cells and its application clinical implants were aimed in the development of 3D architecture design. This will help in finding out design parameters for future implantable medical devices. It is very important piece of work in the development of fuel cells in the field of medicine. Based on the following comments, the MS should be revised
Title looks like a review paper, it should give the details about the current research investigation rather than a general title. The title of the MS should be modified by the incorporation of the aim and objectives and application.
Introduction Section: The information given in this section is very much extensive and Too lengthy. It has nearly 43 references cited with many results of the earlier works. It has 3 figures, reactions and 1 table cited. This kind of references should be included in the discussion. And should be compared with the results of the present investigation. As a research paper, the introduction should be rewritten based on the background of the focused research field in a brief manner.
In this MS, all data recorded were developed on hypothetical (mathematically simulation models) basis and how this will be exactly implement in development of real time devices. This should be discussed in detail in the discussion section.
All cited references in this MS has represented at Introduction and methods. The total discussion were not compared and justified with literature. A complete revision in the discussion is very much essential. The hypothesis discussed in this MS should need comparative statement. The results obtained in this investigations are not validated in the discussion. In this MS a total of 88 references used only in the introduction and methodology. I suggest that the MS should be completely revised.
Author Response
Comment 1:[Title looks like a review paper, it should give the details about the current research investigation rather than a general title. The title of the MS should be modified by the incorporation of the aim and objectives and application.]
Response 1:[The title has been changed to :“Finite Element Optimization of 3D Abiotic Glucose Fuel Cells for Implantable Medical Devices”]
Comment 2 :[Introduction Section: The information given in this section is very much extensive and Too lengthy. It has nearly 43 references cited with many results of the earlier works. It has 3 figures, reactions and 1 table cited. This kind of references should be included in the discussion. And should be compared with the results of the present investigation. As a research paper, the introduction should be rewritten based on the background of the focused research field in a brief manner. ]
Response 2:[We appreciate that in a typical paper, the literature review would be more concise. However, see our response to point 2 from Reviewer 1.
In summary we have used the past literature to create Fig 3, which essentially calibrates our simulation data.
Nevertheless, in the interest of brevity, we maintain figure 3, but have removed some of the literature review and relocated them to the discussion section for comparison with the simulation results.]
Comment 3:[In this MS, all data recorded were developed on hypothetical (mathematically simulation models) basis and how this will be exactly implement in development of real time devices. This should be discussed in detail in the discussion section.]
Response 3:[The discussion has been added to cover trade-offs during the simulation and manufacturing stages, and the conclusion section has already touched upon the potential benefits for future equipment manufacturing processes. I believe there is no need for excessive repetition.]
Comment 4:[All cited references in this MS has represented at Introduction and methods. The total discussion were not compared and justified with literature. A complete revision in the discussion is very much essential. The hypothesis discussed in this MS should need comparative statement. The results obtained in this investigations are not validated in the discussion. In this MS a total of 88 references used only in the introduction and methodology. I suggest that the MS should be completely revised. ]
Response 4:[The discussion section has been expanded to include comparisons with previous literature.]
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for Authors- The manuscript uses a simplified rate law (Eq.4) and then assumes a fixed cell voltage (~0.4V) to compute power, effectively treating voltage as independent of current density. This neglects electrochemical polarization (activation, ohmic, concentration overpotentials) and can substantially overestimate power. See Methods and integration sections.
- The manuscript sets boundary concentrations to 5mM glucose and 4.5mM oxygen (and states venous Oâ‚‚ ~6mM), which appears to conflate bound oxygen and dissolved oxygen. Dissolved Oâ‚‚ (the species available for diffusion to the electrode) is much lower — the conversion/justification is not shown.
- A serious mistake was discovered. The diffusion coefficients listed in Table II have incorrect unit conversions. Example in the manuscript: 0.96×10^-9 m²/s (= 0.96×10^-15 mm²/s. The author must carefully review the entire text and handle the manuscript with care.
- The manuscript only states the mesh was set to “extremely fine” (COMSOL v6.0) without mesh-independence tests, DOFs, solver tolerances, or model files. This prevents assessment of numerical reliability and reproducibility.
- While the manuscript discusses series/parallel connections (Fig.15) and leakage through tissue, the simulation does not explicitly model electronic/ionic transport, internal resistance, or leakage paths. The current approach integrates reaction rate to current and multiplies by a fixed voltage, ignoring circuit/load behavior.
Author should consider these:
(1) Include internal resistance and leakage modeling (either via COMSOL Electric Currents module or an equivalent circuit). Simulate typical load cases and show how output power/efficiency depends on series/parallel configuration and insulation properties.
(2) Introduce an explicit “charge-collection efficiency” parameter (or model electron recombination/leakage) and present sensitivity.
- The manuscript states that abiotic glucose fuel cells could operate indefinitely if biofouling is excluded, but provides no quantitative treatment of biofouling rates or how coatings affect long-term performance. Comparison with literature (e.g., Gonzalez-Solino et al.) is conceptual but lacks matched-condition validation.
Author Response
Comment 1:[The manuscript uses a simplified rate law (Eq.4) and then assumes a fixed cell voltage (~0.4V) to compute power, effectively treating voltage as independent of current density. This neglects electrochemical polarization (activation, ohmic, concentration overpotentials) and can substantially overestimate power. See Methods and integration sections.]
Response 1:[We thank the reviewer for pointing this out. We fully agree that in practice, the cell voltage is current-dependent, with activation, ohmic, and concentration overpotentials lowering the terminal voltage. It is also worth noting that subsequently, the load voltage be dependent on the ratio of the Thevenin output resistance of the cell to the load times that fundamental cell voltage.
We see 3 different levels of understanding for the modelling: The Mesoscopic (mm) scale, microscopic scale (um) and nanoscopic scale (nm). This paper is essentially exploring effects at the mescopic scale, where we explore the effect of multi-reaction plate architecture. We therefore treat individual reaction plates as having fixed properties with 3 different scales of power generation efficiency.
Simulating the drift-diffusion equations properly requires detailed microscopic-scale simulations to explore the effect of microscopic distance and thus electric field distribution between interdigitated anode/cathode sections of the reaction plates. The different scale means these are separate simulations and would require a lot more investigation, making the paper perhaps too long. As such, this is the current focus of our follow-up work.Therefore, in our model we fixed the operating voltage at 0.4 V, based on literature reports for abiotic glucose fuel cells using nanoporous gold electrodes [Gonzalez-Solino et al., 2020]. We are fully aware that the practical voltage range could deviate +/- 50% from this value, depending various micros-scale design decisions. So this is why we provided 3 scenarios for power output.
We have made this more clear in the methodology and discussion.]
Comment 2:[The manuscript sets boundary concentrations to 5mM glucose and 4.5mM oxygen (and states venous Oâ‚‚ ~6mM), which appears to conflate bound oxygen and dissolved oxygen. Dissolved Oâ‚‚ (the species available for diffusion to the electrode) is much lower — the conversion/justification is not shown.]
Response 2:[We thank the reviewer for highlighting this important point. It is not possible to put these fuel cells directly into blood vessels. Rather we envisage them in special capsules in the interstitial space. The capsules will be to prevent diffusion of cells onto the reaction plates.
As such, we concur that the concentration of free dissolved oxygen in plasma is only ~0.05–0.1 mM, with the majority of oxygen bound to haemoglobin. Our aim was not to equate 4.5 mM with the dissolved fraction, but rather to represent an effective perfusion support boundary. In highly vascularised regions, the unloading of haemoglobin continuously replenishes the dissolved fraction at the capillary-tissue interface, thereby preventing severe oxygen limitation at the implant boundary.
We therefore selected 4.5 mM as an optimistic upper bound to ensure our simulation isolates the geometric and diffusion characteristics of the fuel cell itself, rather than systemic oxygen delivery.
We have supplemented this in our Methods section.]
Comment 3:[A serious mistake was discovered. The diffusion coefficients listed in Table II have incorrect unit conversions. Example in the manuscript: 0.96×10^-9 m²/s (= 0.96×10^-15 mm²/s. The author must carefully review the entire text and handle the manuscript with care.]
Response 3:[Thanks very much for pointing out the error in the unit conversion; This was a typo on the paper and has been rectified.
I have rechecked the units and derivations in the simulation files, which are correct, so the final results remain valid.]
Comment 4:[The manuscript only states the mesh was set to “extremely fine” (COMSOL v6.0) without mesh-independence tests, DOFs, solver tolerances, or model files. This prevents assessment of numerical reliability and reproducibility.]
Response 4:[We appreciate the request for an independent study of the formal mesh independence.
Our COMSOL model considers two factors – concentration gradients and surface reactions. The former should be independent of mesh density. i.e. the concentration between two boundary conditions remains the same regardless of the number of sampling points between them. Though of course the mesh density determines quantization error in the gradient profile. In contrast for the latter case, if a reaction surface is divided into mesh points, then careful consideration is needed.
In the COMSOL model, concentration gradients determine the local availability of O2 and glucose. These concentrations determine the local reaction rate. In order to ensure independence of mesh density, we integrated the results in MATLAB and calibrated across the whole surface to achieve the overall reaction rate and current densities. Multiplication by the assumed operating voltage then yields the power output. This workflow is described in detail in the manuscript.
It takes several hours to perform each of the simulations, of which there were many. So we would prefer not to spend 3 months re-simulating given that we believe our results to be independent of mesh density.
We fully agree with regards open science. we will provide our COMSOL + Matlab models along with precise solver settings onto our Github page: https://github.com/Neuroprosthetics-Lab]
Comment 5:[While the manuscript discusses series/parallel connections (Fig.15) and leakage through tissue, the simulation does not explicitly model electronic/ionic transport, internal resistance, or leakage paths. The current approach integrates reaction rate to current and multiplies by a fixed voltage, ignoring circuit/load behavior.
Author should consider these:
(1) Include internal resistance and leakage modeling (either via COMSOL Electric Currents module or an equivalent circuit). Simulate typical load cases and show how output power/efficiency depends on series/parallel configuration and insulation properties.
(2) Introduce an explicit “charge-collection efficiency” parameter (or model electron recombination/leakage) and present sensitivity.]
Response 5:[We acknowledge this valuable point.
Please see our comments above with regards the difference between meso, micro and nano-scale simulations.
The present study focuses on meso-scale structural aspects of fuel cells, including cell dimensions, geometry, inter-cell separation, and diffusion–reaction coupling. Electronic/ionic transport, internal resistance, and leakage pathways were therefore not explicitly modelled.
We fully agree that these factors need to be taken into account, though this falls within the subsequent micrometre-scale simulations. Future work will extend the finite element analysis framework by coupling the Dilute Species Transport Module with the Electrical Module in COMSOL, thereby explicitly accounting for ion/electron transport and leakage pathways. However, these aspects are beyond the scope of this paper.]
Comment 6:[The manuscript states that abiotic glucose fuel cells could operate indefinitely if biofouling is excluded, but provides no quantitative treatment of biofouling rates or how coatings affect long-term performance. Comparison with literature (e.g., Gonzalez-Solino et al.) is conceptual but lacks matched-condition validation.]
Response 6:[We thank the reviewer for highlighting this important limitation.
We agree that biofouling is one of the most critical barriers to long-term operation. Another factor could be electrolytic degradation of nano-electrodes, which may time diminishes the performance of abiotic glucose fuel cells.
However, research on electrode stability and long-term performance remains limited, with few animal studies or longevity tests available and little strong data to substantiate reported findings. These areas clearly need further investigation.
Therefore, premature simulation without empirical support would risk producing unreliable results.
We have incorporated comparisons with past literature into our discussion and clarified the conditions more explicitly.]
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe author should give much attention to improve the discussion section. As i stated in my previous report about the validation of the current results with earlier reports. In this revision the author not fulfilled the issue in the discussion section. In this revision the authors just added a paragraph (358-369). The discussion section should deals the results in comparison with the earlier reports. In this MS, the discussion section should be rewritten with comparative statements (with earlier reports) in each and every statement of the results discussed.
Author Response
Comment:[The author should give much attention to improve the discussion section. As i stated in my previous report about the validation of the current results with earlier reports. In this revision the author not fulfilled the issue in the discussion section. In this revision the authors just added a paragraph (358-369). The discussion section should deals the results in comparison with the earlier reports. In this MS, the discussion section should be rewritten with comparative statements (with earlier reports) in each and every statement of the results discussed. ]
Response :[We appreciate this comment and have expanded the discussion to include further literature comparisons.
Most prior research on abiotic glucose fuel cells has focused on electrode materials and 2D configurations, with few studies of diffusion-reaction behaviour within 3D structures. This makes many of the meso-scale observations presented in this paper difficult to compare with existing literature.]
Author Response File: Author Response.pdf
Round 3
Reviewer 1 Report
Comments and Suggestions for AuthorsThank you for the effort.
Reviewer 2 Report
Comments and Suggestions for AuthorsAccept in present form
