Kinetic Modelling for Hydrothermal Conversion of Food Wastes

Round 1

Reviewer 1 Report

This work presents the kinetic modelling of hydrothermal liquefaction for food wastes. The authors used more datasets from various sources to develop the model.  The product yields from the developed model are compared with the experimental data. The model prediction are satisfactory with the experimental data. However, the model needs to be improved for better prediction of all products yields. The paper is written well. It can be considered for possible publication with the following comments

1.       In page 2, expand the abbreviation of SOCOR and CAPINOV

2.       In Table 3, the initial distribution factors of model compounds in different phase needs to justified.

3.       In page 8, for the equation (5), the reference needs to be provided.

4.       Could author justify the validity of Eqn.(5)  for different solvents

5.       In page 9, provide the space for Figure4

6.       Could the authors provide the reason for low R² value   of aqueous phase predicted by new model?

7.       The predicted activation energy for various reactions from the new models needs to compare with the reported value in the literature.

8.       The authors need to be discussed the improvement of the present model in the manuscript.

Author Response

This work presents the kinetic modelling of hydrothermal liquefaction for food wastes. The authors used more datasets from various sources to develop the model.  The product yields from the developed model are compared with the experimental data. The model prediction are satisfactory with the experimental data. However, the model needs to be improved for better prediction of all products yields. The paper is written well. It can be considered for possible publication with the following comments

1. In page 2, expand the abbreviation of SOCOR and CAPINOV

> These are the name of companies as mentioned in the text, no need to further explain.

2. In Table 3, the initial distribution factors of model compounds in different phase needs to justified.

> They cannot really be justified. This table is a proposition that works but can be adjusted. It is part of the proposed model.

3. In page 8, for the equation (5), the reference needs to be provided.

> This equation cannot really be justified. This table is a proposition that works but can be adjusted. It is part of the proposed model.

4. Could author justify the validity of Eqn.(5)  for different solvents

> This equation cannot really be justified. This table is a proposition that works but can be adjusted. It is part of the proposed model. The idea was to propose a simple model with data readily available for all solvents.

5. In page 9, provide the space for Figure4

> OK well spotted

6. Could the authors provide the reason for low R²value of aqueous phase predicted by new model?

> As mentioned in the text, the aqueous phase experimental data was not used in the model fitting. There therefore a good explanation for the R2 value to be low.

7. The predicted activation energy for various reactions from the new models needs to compare with the reported value in the literature.

> Yes this could be done, readers can also find this data in the references. The problem is that the data from both the Valdez and Obeid schemes exist for a number of different resources. Lignin reactions were also added to the Valdez scheme so the results will numerically be different. A numerical comparison does not make a lot of sense. A comment was added to section 4 after table 5 to explain some differences with the other published models.

8. The authors need to be discussed the improvement of the present model in the manuscript.

> The conclusions were revised. It is difficult to claim that the new model really produce better results. The new model allows to take more phenomena into account and is proposed as a modelling basis to create new and better models and take into account intermediary data, not possible with the traditional models.

Reviewer 2 Report

In this work, a kinetic model was developed for the prediction of HTL products yields. This work needs a major revision before it can be published on Eng. Specific details are as follows:

1.      Figure 1 shows the experimental conditions in the HTL experiments on BCP and BSG. How about other feedstocks, like FW2, FFOM and DFOR?

2.      In section 2.3, please give the equation of bio-oil yield.

3.      Section 2.3, “Biocrudes are complex mixtures containing an oily and a solid fraction. This solid fraction is in fact an insoluble fraction in a determined solvent.”. Why does biocrude include solids? I cannot understand the Eq. 1, please explain it in detail.

4.      The author mentioned in the third paragraph of Section 3.1 that Valdez’s model does not give pre-exponential factors and activation energies. However, Sheehan and Savage from the same group as Valdez reported these data in the following literature.

Sheehan JD, Savage PE. Modeling the effects of microalga biochemical content on the kinetics and biocrude yields from hydrothermal liquefaction. Bioresource Technology 2017;239:144-50.

5.      Data Availability Statement. I cannot find data from BCP and BSG. Please summarize the dataset and clearly present it. For example, which data were used for determining the kinetic parameters and which data were used for validating the model.

6.      In formula (5), what is the meaning of EA?

7.      How to obtain the data in Table 3?

8.      Why is R² negative in Table 6? R² should be between 0 and 1. How to calculate the R2. Please list the equation.

9.      Why can’t this model accurately predict the yields of gas and aqueous phase?

10.  Did the author measure the actual heating profile. The heating profile in Figure 1 is too ideal.

Author Response

In this work, a kinetic model was developed for the prediction of HTL products yields. This work needs a major revision before it can be published on Eng. Specific details are as follows:

1. Figure 1 shows the experimental conditions in the HTL experiments on BCP and BSG. How about other feedstocks, like FW2, FFOM and DFOR?

> In the text : « For food wastes, the experimental conditions were less structured. The temperatures were 200, 250, 300, 325 and 350 °C with holding times 0 and 30 min. The same, programmed linear temperature ramp of 15 °C/min was applied in most cases, unless other-wise stated with the experiment. » The data file in the supplementary material gives all the details.

2. In section 2.3, please give the equation of bio-oil yield.

> OK this equation was added.

3. Section 2.3, “Biocrudes are complex mixtures containing an oily and a solid fraction. This solid fraction is in fact an insoluble fraction in a determined solvent.”. Why does biocrude include solids? I cannot understand the Eq. 1, please explain it in detail.

> HTL experiments produce a black material that is sometimes sticky, sometimes fluid and sometimes a brown powder. This largely depends on the resource and the conditions. We call this material “biocrude”. When en extraction with a solvent is applied, a certain amount of the biocrude can be dissolved in the solvent, this part is oil. The remaining is called char. This approach is quite common in the field. The nomenclature is not always the same.

4. The author mentioned in the third paragraph of Section 3.1 that Valdez’s model does not give pre-exponential factors and activation energies. However, Sheehan and Savage from the same group as Valdez reported these data in the following literature.

Sheehan JD, Savage PE. Modelling the effects of microalga biochemical content on the kinetics and biocrude yields from hydrothermal liquefaction. Bioresource Technology 2017;239:144-50.

> We have included this reference. The data presented is for micro algae and has no direct applicability on food wastes. The text in section 4 was modified to discuss the differences.  The results in this publication are numerically slightly different but they also present a low R2.

5. Data Availability Statement. I cannot find data from BCP and BSG. Please summarize the dataset and clearly present it. For example, which data were used for determining the kinetic parameters and which data were used for validating the model.

> Yes well spotted. You are absolutely right, this was an oversight and has been corrected. Sorry about this. The data is added as supplementary material.

6. In formula (5), what is the meaning of EA?

>Ethyl acetate, mention added to the text

7. How to obtain the data in Table 3?

> These values were chosen as part of the proposed model. They cannot be justified, other than that they are reasonable and correspond to qualitative observations.

8. Why is R²negative in Table 6? R² should be between 0 and 1. How to calculate the R2. Please list the equation.

> The equation was added in the text. It is pretty basic but worthwhile repeating. It is actually possible to have negative values. 1 is a perfect fit, 0 means a null model characterized by an intercept only (horizontal line). There is no limit on the negative side, as the model can be bad, very bad or catastrophically bad. An alternative formulation has no upper limit with a lower limit of zero, but this is rarely used.

9. Why can’t this model accurately predict the yields of gas and aqueous phase?

> As mentioned in the text, the aqueous phase experimental data was not used in the model fitting. There is therefore a good explanation for the R2 value to be low.

10. Did the author measure the actual heating profile. The heating profile in Figure 1 is too ideal.

> Yes we do of course measure the temperature. We have added the temperature measurement of one series of experiments in Figure 1.

Reviewer 3 Report

The paper submitted by Haarlemmer and Briand (2022) on title “Kinetic Modelling for Hydrothermal Conversion of Food Wastes” deals kinetic model for the prediction of HTL product yields based on a chemical mechanism. The writing of the manuscript too poor and the explanation of results very poor. After carefully reviewing the manuscript, I recommend it for rejection. Below are some of my serious comments.

Q.1. The graphical abstract need to be add for the better understanding of the present work. The GA should be clean and clear so that reader can understand the overall study.

Q. 2. The abstract need revision for the better understanding of the readers. The present form of the abstract did not showed interest of readers. Add few suitable data in the abstract.

Q. 3. The keywords have chemical mechanism but hardily able to find it throughout the, manuscript. Author should explain.

Q. 4. Author should add relevant highlight of the presented work. Make sure the highlight should be relevant and indicate major findings.

Q. 5. Reference 1 cited in the manuscript but the reference is unclear. Is reference still have the same information?

Q. 6. “The hydrothermal conversion takes place at temperatures between 300 and 400 °C and at pressures above the saturation pressure to ensure that water remains in the liquid phase, typically above 100 bar.” Author should add the suitable reference to support the stated lines.

Q. 7. Author should not add any personal word like our, we my mine etc. in the manuscript. Also, “Modelling with machine learning algorithms is becoming popular in the HTL com-munity [6–9].” Is it possible to add only single reference to support this lines?

Q. 8. The author did not mention the novelty and objective of the present study? Why this study becomes essential? How it is different than others? How this manuscript is good than other reported study in recent years?

Q. 9. What is SOCOR and CAPINOV? Did author explain it before? Also, I did not able to find Table 1 in the text. In addition, table 1 should go to results and discussion section. In table 1, there is no consistency. Somewhere added standard deviation, some where one point after decimal somewhere nothing after decimal point. Author should stick in one format.

Q. 10. How author selected temperature and holding time for the experiments?

Q. 11. All the fig. need to revise and replaced with higher resolution. At least 600 dpi. Experimental condition is repeated twice why? Author should correct it.

 Q. 12. The manuscript need substantial English revision. There are many mistakes throughout the manuscript.

Q. 13. Arrhenius expression need an equation number. Author should follow one trend through the manuscript. Also, section 3 need more explanation.

Q. 14. “Figure 2. Reaction mechanisms from Valdez et al. (a) and Obeid et al. (b).” title for the figure need to correct.

Q. 15. The data listed in table 3 is produced by the author or taken from the literature. If taken form the published literature must produce supporting reference.

Q. 16. Fig. 4 and 5 need more explanation. Also compare the results with other study. Results and discussion section need a through revision. The present form is very poor and hardly able to understand.

Q. 17. The present activation energy in table 5 is too high. What is the physical significance of this on the actual HTL process? What is the meani8ng of negative value in Table 6.

Q. 18. “Even if we see a larger scattering of the results with some others experimental data, globally all three models that were tested in this paper are able to reproduce results on a wide range of food waste and agro-industrial wastes, processed at different HTL conditions and with different experimental ways of products recovery. The solvent effect is taken into account with an extremely simple model that needs to be improved.” What is the meaning?

Q. 19. The conclusion is too long and place many unnecessary things that should be removed.

Author Response

The paper submitted by Haarlemmer and Briand (2022) on title “Kinetic Modelling for Hydrothermal Conversion of Food Wastes” deals kinetic model for the prediction of HTL product yields based on a chemical mechanism. The writing of the manuscript too poor and the explanation of results very poor. After carefully reviewing the manuscript, I recommend it for rejection. Below are some of my serious comments.

> We are sorry you did not like our paper. The document was reviewed for its English and improved. Even if we do not present a perfect model, we do believe that this paper makes a significant contribution to the field.

Q.1. The graphical abstract need to be add for the better understanding of the present work. The GA should be clean and clear so that reader can understand the overall study.

> Graphical abstracts are not used in Eng.

2. 2. The abstract need revision for the better understanding of the readers. The present form of the abstract did not showed interest of readers. Add few suitable data in the abstract.

> The abstract was revised

3. 3. The keywords have chemical mechanism but hardily able to find it throughout the, manuscript. Author should explain.

> We do not agree, the chemical reaction scheme should qualify as a mechanism. The chemical mechanism for HTL conversion of organic wastes is provided in figure 3

4. 4. Author should add relevant highlight of the presented work. Make sure the highlight should be relevant and indicate major findings.

> There are no highlights for papers in Eng.

5. 5. Reference 1 cited in the manuscript but the reference is unclear. Is reference still have the same information?

> We are not sure what you mean. The reference is valid. It is a link to a public report of a European project that can be found on the internet. However the actual numerical values cited in this report were indeed slightly different. This has been corrected.

6. 6. “The hydrothermal conversion takes place at temperatures between 300 and 400 °C and at pressures above the saturation pressure to ensure that water remains in the liquid phase, typically above 100 bar.” Author should add the suitable reference to support the stated lines.

> OK done

7. 7. Author should not add any personal word like our, we my mine etc. in the manuscript. Also, “Modelling with machine learning algorithms is becoming popular in the HTL com-munity [6–9].” Is it possible to add only single reference to support this lines?

> The language was revised.

> I am not sure how the number of references of ML papers is relevant. There are more references to kinetic studies than for ML studies.

8. 8. The author did not mention the novelty and objective of the present study? Why this study becomes essential? How it is different than others? How this manuscript is good than other reported study in recent years?

> There are quite a few simple kinetic models in the literature. While these models work for certain resources, they are limited. This work attempts to contribute to the field by proposing a more complex model that allows taking into account different laboratory practices and resources. This is not an easy task but very important. The idea is to propose a methodology to the scientific community.

9. 9. What is SOCOR and CAPINOV? Did author explain it before? Also, I did not able to find Table 1 in the text. In addition, table 1 should go to results and discussion section. In table 1, there is no consistency. Somewhere added standard deviation, some where one point after decimal somewhere nothing after decimal point. Author should stick in one format.

> SOCOR and CAPINOV are commercial laboratories, company names, as mentioned in the text. You are right about the decimal places in the standard deviations, this is corrected. Standard deviation was not always available.

10. 10. How author selected temperature and holding time for the experiments?

> Conditions were selected to be in the typical range to HTL experiments in the literature.

11. 11. All the fig. need to revise and replaced with higher resolution. At least 600 dpi. Experimental condition is repeated twice why? Author should correct it.

> We will check with the editor, the website mentions 300 dpi.

12. 12. The manuscript need substantial English revision. There are many mistakes throughout the manuscript.

> The language was revised.

13. 13. Arrhenius expression need an equation number. Author should follow one trend through the manuscript. Also, section 3 need more explanation.

> Equation number was added. More explanations were included.

14. 14. “Figure 2. Reaction mechanisms from Valdez et al. (a) and Obeid et al. (b).” title for the figure need to correct.

> OK, this was a mistake and has been corrected.

15. 15. The data listed in table 3 is produced by the author or taken from the literature. If taken form the published literature must produce supporting reference.

> This data constitutes the proposed model and a reference is therefore not provided. Some more explanation is provided.

16. 16. Fig. 4 and 5 need more explanation. Also compare the results with other study. Results and discussion section need a through revision. The present form is very poor and hardly able to understand.

> The graphs are pretty self-explanatory. More explanation is provided in the text.

17. 17. The present activation energy in table 5 is too high. What is the physical significance of this on the actual HTL process? What is the meani8ng of negative value in Table 6.

> The activation energies we calculated are actually lower that those calculated by Sheehan and Savage. The activation energy is just a numerical value, difficult to give a chemical explanation in cases like this, other than that this process depends moderately on the temperature, less than other reactions (Fischer-Tropsch for example). More explanation is provided in section 3

> A negative R2 means that the model is not very good on these results. 1 is perfect, the value can go to minus infinity for really bad models. More explanation is provided in section 3.

18. 18. “Even if we see a larger scattering of the results with some others experimental data, globally all three models that were tested in this paper are able to reproduce results on a wide range of food waste and agro-industrial wastes, processed at different HTL conditions and with different experimental ways of products recovery. The solvent effect is taken into account with an extremely simple model that needs to be improved.” What is the meaning?

> It means that this very simple approach gives interesting results. To our knowledge no models of this kind exist. We start with a simple model, share it and hope that the scientific community picks this up and collectively we improve.

19. 19. The conclusion is too long and place many unnecessary things that should be removed.

> The conclusions are revised and consensed.

Round 2

Reviewer 2 Report

After revision, this manuscript has been improved. However, this work still needs a major revision before it can be published on Eng. Some responses to the comments are not correct. Specific details are as follows:

2 In section 2.3, please give the equation of bio-oil yield.

The equation is incorrect. The equation of bio-oil yield should be equation 2.

3 Section 2.3, “Biocrudes are complex mixtures containing an oily and a solid fraction. This solid fraction is in fact an insoluble fraction in a determined solvent.”. Why does biocrude include solids? I cannot understand the Eq. 1, please explain it in detail.

The author only provided the response to this comment, but no revision in the manuscript.

7 How to obtain the data in Table 3?

The author did not answer how to obtain the data in Table 3. By experiments, by calculation or from other references?

Author Response

After revision, this manuscript has been improved. However, this work still needs a major revision before it can be published on Eng. Some responses to the comments are not correct. Specific details are as follows:

2 In section 2.3, please give the equation of bio-oil yield. The equation is incorrect. The equation of bio-oil yield should be equation 2.

• The numbering has been changed. We also added the missing equation for the yield in the aqueous phase.

3 Section 2.3, “Biocrudes are complex mixtures containing an oily and a solid fraction. This solid fraction is in fact an insoluble fraction in a determined solvent.”. Why does biocrude include solids? I cannot understand the Eq. 1, please explain it in detail. The author only provided the response to this comment, but no revision in the manuscript.

• This phrase should be taken litteraly. As be explain in the text, we consider the product recovered after the transformation « biocrude ». In the literature this product comes under many names, raw residue or even as bio-oil. Please understand that it can be a nice oil with little or no char but also a solid with little or no oil. This definition makes it easier to understand char and bio-oil. The text was clarified.

7 How to obtain the data in Table 3? The author did not answer how to obtain the data in Table 3. By experiments, by calculation or from other references?

• Initially these values were part of the optimisation problem. In practice the solutions found were erratique and did not correspond to anything physical. From many years of analysis we propose these values. They are just that propositions. They do correspond to observed behaviour but we do not pretend that they are perfect.

Reviewer 3 Report

Authour revised the manuscript and answerd all the rasied question. However, there is still few things that need to explain before get accpted.

Q. 1. The reviwe file showing track chnage mode that make many confusing while reading sentance. It would be bettr if author  send the corrected file wthoiut track chnge mode.

Q. 2. "In overallEurope, it can be estimated atapproximately 22 50 MT of dry matter for food waste is produced each year and 100 MT for agro-industrial residues [1]." the sentace is revised and data is modified but refrence is same. Authour must check it.

Q. 3. "he experimental database is based on our own HTL batch experiment following our usual experimental procedure in addition to data gathered from the literature for the same kind of resource, the food wastes. " authour must not use any persoonal word in manuscript such as our, my, us etc.

Q. 4. Authour must visit the below refrence for better understanding of hydtrothermal process. "https://www.sciencedirect.com/science/article/abs/pii/S0016236122002459".

Q. 5. In section 3, authour used eqation 45. How ? it should be equation 4 or 5. must correct it.

Q. 6. There is still english issue in the manuscript. The prsent form is ok but there is a cxhnage to improve it.

Q. 7.  In think fig. 6 should plot in line and symbol mode. it paper published in balck and while coloer then it becomes difficult to understand what line prsent what?

Author Response

Authour revised the manuscript and answerd all the rasied question. However, there is still few things that need to explain before get accpted.

1. 1. The reviwe file showing track chnage mode that make many confusing while reading sentance. It would be bettr if author  send the corrected file wthoiut track chnge mode.

> Track change is imposed by the editor. I cannot upload two versions of the document. I do agree with you that it sometimes makes the document more difficult to read. However it is a very useful tool and there is no way around it.

2. 2. "In overallEurope, it can be estimated atapproximately 22 50 MT of dry matter for food waste is produced each year and 100 MT for agro-industrial residues [1]." the sentace is revised and data is modified but refrence is same. Authour must check it.

> Yes we corrected the data to be in line with the reference. I am sorry, but the reference is correct and you will find the data in the referenced document, we added a reference to the project as well.

3. 3. "he experimental database is based on our own HTL batch experiment following our usual experimental procedure in addition to data gathered from the literature for the same kind of resource, the food wastes. " authour must not use any persoonal word in manuscript such as our, my, us etc.

> OK, we now have removed all prosonal words. The language was revised.

4. 4. Authour must visit the below refrence for better understanding of hydtrothermal process. "https://www.sciencedirect.com/science/article/abs/pii/S0016236122002459".

> Yes I had knowledge of this paper, I added it as a reference.

5. 5. In section 3, authour used eqation 45. How ? it should be equation 4 or 5. must correct it.

> The equation number changed from 4 to 5 and you mistook the track change for a double number. All is well. The 4 was barred after being deleted, but the barring sign is at the same level as the horizontal line in the 4. Sorry about this, but it comes out OK without track change.

6. 6. There is still english issue in the manuscript. The prsent form is ok but there is a cxhnage to improve it.

> We have reviewed the language once more.

7. 7.  In think fig. 6 should plot in line and symbol mode. it paper published in balck and while coloer then it becomes difficult to understand what line prsent what?

> Done. You were right.

Round 3

Reviewer 2 Report

7 How to obtain the data in Table 3? The author did not answer how to obtain the data in Table 3. By experiments, by calculation or from other references?

Why were these values part of the optimisation problem？

Author Response

7 How to obtain the data in Table 3? The author did not answer how to obtain the data in Table 3. By experiments, by calculation or from other references? Why were these values part of the optimisation problem？

• The text was even more clarified. These parameters are not part of the optimisation problems. They were initially, but this sometimes leads to un-physical solutions.