Thermal Pyrolysis of Polystyrene Aided by a Nitroxide End-Functionality. Experiments and Modeling

Round 1

Reviewer 1 Report

This article concerns the interesting topic of thermal pyrolysis of polystyrene to generate styrene monomers.  Although styrene pyrolysis experiments have been conducted in many previous studies, it seems that this might be the first study concerned with pyrolysis of polystyrene that contains nitroxide moieties.  I have several serious concerns about this article that need to be addressed:

1.  The main issue is that the data presented in Figures 3 to 6 don't really show any convincing influence of the nitroxide content of the polystyrene on the amount and types of the various products that are obtained.  There is a definite influence of temperature, which has been revealed in previous studies.  Based on the data presented, however, it is not at all clear to me that there really is any sort of significant effect of the nitroxide, except perhaps in Figure 6.  In the other figures, the FRP results are very similar to those obtained using the different nitroxide levels.  Without any error bars on the plots, it is not clear whether we are seeing any kind of important and reproducible response, or if the minor deviations between different experiments are just due to noise.  Did the authors perform any replicate experiments?  Given that it is hard to discern any important trends (except for the influence of temperature), it is hard to know whether the modeling efforts are justified and whether they really help the authors to better understand the system.  In my view, the issue of what is just noise and what is a reproducible trend needs to be addressed before this paper can be published.  My other concerns listed below are not as important.

2.  The first paragraph of the Introduction requires several references to back up the many claims made by the authors.  Also, the referencing elsewhere in the text needs fixing.  For example, on p. 4 at line 131, Ebert, Ederer and Schroder is missing the 4th author and should be changed to Ebert et al.  On p. 14, a reference to the moment-closure expression of Seidel and Katz should be provided.  Further down, near line 354, a reference should be provided to support the statement that "NMP is among ...".  The references provided for the patents listed as items 20 to 25 on p. 31 do not provide enough information for someone to easily locate these documents.  Are they US patents, World patents?

3.  The first two paragraphs in the Experimental results should be in the Introduction.  The information in these paragraphs is part of the literature review and not part of the results.  Also, the authors should make it clear that the reaction shown in Figure 2 is the same as the 4th reaction in Table 3.

 

Author Response

Answers to reviewer 1 comments are interspersed below:

This article concerns the interesting topic of thermal pyrolysis of polystyrene to generate styrene monomers.  Although styrene pyrolysis experiments have been conducted in many previous studies, it seems that this might be the first study concerned with pyrolysis of polystyrene that contains nitroxide moieties.  I have several serious concerns about this article that need to be addressed:

1. The main issue is that the data presented in Figures 3 to 6 don't really show any convincing influence of the nitroxide content of the polystyrene on the amount and types of the various products that are obtained.  There is a definite influence of temperature, which has been revealed in previous studies.  Based on the data presented, however, it is not at all clear to me that there really is any sort of significant effect of the nitroxide, except perhaps in Figure 6.  In the other figures, the FRP results are very similar to those obtained using the different nitroxide levels.  Without any error bars on the plots, it is not clear whether we are seeing any kind of important and reproducible response, or if the minor deviations between different experiments are just due to noise.  Did the authors perform any replicate experiments?  Given that it is hard to discern any important trends (except for the influence of temperature), it is hard to know whether the modeling efforts are justified and whether they really help the authors to better understand the system.  In my view, the issue of what is just noise and what is a reproducible trend needs to be addressed before this paper can be published.  My other concerns listed below are not as important.

 

Answer.

The reviewer is right in the sense that we should have included information about replicates, reproducibility and error bars. This is now fixed. Experiments were in fact run and analyzed by triplicate in some cases (the experimental design now shows the information about replicates). Specifically, for a given temperature, the experiments at the zero nitroxide level (blank or FRP) as well as the experiments at N/I ratio of 1.3 were run and analyzed by triplicate, and a pooled variance was calculated to include (new version) error bars of 1 standard deviation for liquid, solid and gas fractions, styrene, dimer and reaction time. The same pattern of replicates was used for all the temperatures tested. Figures 3, 5 and 6 were re-plotted to include the error bars. It was decided not to include error bars in Figure 4 since not all the values there are relevant and the absolute yields of the most important components are provided in Figure 5 (instead of yields relative to the liquid fraction as in Figure 4). It is now seen that some of the effects (at specific conditions) are indeed above the experimental error, while others represent noise. Additionally, by performing replicate experiments, we were able to observe clear qualitative differences in the presence and absence of nitroxide in the samples. These were not reported before, but are now included in the discussion. We were especially careful in generating reproducible results and we are convinced that the presence of nitroxide definitely influences the pyrolysis process. On the other hand, in the first version of the paper we did not supply enough information on this issue and therefore the concern of the reviewer is understandable. We believe that the additional information included in the revised version should fix this problem.

2. The first paragraph of the Introduction requires several references to back up the many claims made by the authors.  Also, the referencing elsewhere in the text needs fixing.  For example, on p. 4 at line 131, Ebert, Ederer and Schroder is missing the 4th author and should be changed to Ebert et al.  On p. 14, a reference to the moment-closure expression of Seidel and Katz should be provided.  Further down, near line 354, a reference should be provided to support the statement that "NMP is among ...".  The references provided for the patents listed as items 20 to 25 on p. 31 do not provide enough information for someone to easily locate these documents.  Are they US patents, World patents?

 

Answer.

Thank you for the corrections and recommendations.

Several references have been added in the first and second paragraphs of the introduction to back up the statements made.

The referencing has been fixed by using one last name and et al. for references with 3 or more authors, or the last names (1 or 2) for references with up to two authors.

The reference to Saidel and Katz was already included (ref 26 previous version, 33 in the new version) but the reference number was missing. This is now fixed.

A reference was added to support the importance of NMP.

The references for patents (20 to 25) are now complete. The company and the country of the patent have been added in all the cases.

3. The first two paragraphs in the Experimental results should be in the Introduction.  The information in these paragraphs is part of the literature review and not part of the results.  Also, the authors should make it clear that the reaction shown in Figure 2 is the same as the 4th reaction in Table 3.

Answer.

We agree with the reviewer, so the first 3 paragraphs in the Experimental results were moved to the Introduction. Also, a clarifying note was added in Table 3 regarding the reaction in Figure 2.

Author Response File: Author Response.docx

Reviewer 2 Report

This paper deals with the use of nitroxide end-capped polymers to aid in the recycling of polystyrene by pyrolysis. The paper is in general well written and clear and I recommend publication subject to the following changes:

The introduction is extensive but has some repitition and could be reduced.

The activation energy reported from ref 28 seems incorrect (activation energy of 266500 kcal/mol is incredibly high for example).

The starting point of the simulations is not clear. i.e. what is the initial molecular weight and assumed livingness.

It would be interesting to know if the differences observed experimentally are reproducible or whether they are within the experimental error. The authors should comment on this.

My main issue is with the model used for simulations. In previous work the depolymerization process has been followed and fit to a model (see for example work of Broadbelt and coworkers ref 30.) Why not just take this model as a base and add in the nitroxide association/dissociation reaction? It seems that the model used here is missing a number of reactions (propgation of double bonds from scission reactions, intramolecular transfer to polymer (backbiting) etc.) that would influence the results.

 

Author Response

The answers to reviewer 2 comments are interspersed below:

This paper deals with the use of nitroxide end-capped polymers to aid in the recycling of polystyrene by pyrolysis. The paper is in general well written and clear and I recommend publication subject to the following changes:

 

The introduction is extensive but has some repitition and could be reduced.

 

Answer

In the new version we tried to eliminate repeated information in the introduction.

 

The activation energy reported from ref 28 seems incorrect (activation energy of 266500 kcal/mol is incredibly high for example).

 

Answer

Thank you for the correction. There was a problem with the units, but the values have been fixed.

 

The starting point of the simulations is not clear. i.e. what is the initial molecular weight and assumed livingness.

 

Answer

The base case was assumed to have a Mn of 50,000 with dispersity of 1.9 (same as the FRP case in Table 1). For the dormant polymer a livingness of 100% was assumed. This information has been added in the Base case simulation section.

 

It would be interesting to know if the differences observed experimentally are reproducible or whether they are within the experimental error. The authors should comment on this.

 

Answer.

The reviewer is right regarding this issue and this was also pointed out by another reviewer, so our answer is similar to that for the other reviewer.  We should have included information about replicates, reproducibility and error bars, so now we have included this information. Experiments were run and analyzed by triplicate in some cases (the experimental design now shows the information about replicates). For each given temperature, the experiments at the zero nitroxide level (blank or FRP) as well as the experiments at N/I ratio of 1.3 were run and analyzed by triplicate, and a pooled variance was calculated to include (new version) error bars of 1 standard deviation for styrene, dimer and reaction time. Figures 3, 5 and 6 were re-plotted to include the error bars. It is now clear that some of the effects (at specific conditions) are indeed above the experimental error, while others represent noise. Additionally, by performing replicate experiments, we were able to observe clear qualitative differences in the presence and absence of nitroxide in the samples. These were not reported before, but are now included in the discussion. We were especially careful in generating reproducible results and we are convinced that the presence of nitroxide at the polymer ends definitely influences the pyrolysis process.

My main issue is with the model used for simulations. In previous work the depolymerization process has been followed and fit to a model (see for example work of Broadbelt and coworkers ref 30.) Why not just take this model as a base and add in the nitroxide association/dissociation reaction? It seems that the model used here is missing a number of reactions (propgation of double bonds from scission reactions, intramolecular transfer to polymer (backbiting) etc.) that would influence the results.

Answer.

 At the beginning of the modeling effort we considered several options including that proposed by the reviewer; however, in our view, existing models were either already too complex or too simplistic for our purposes. In particular, regarding the Broadbelt model, this was a very complete model for the pyrolysis of conventional PS, but a bit too complex as starting point for our purposes. The added complexity implied by the nitroxide-end functionality, which as shown here is significant, led us to the decision of using a simpler starting model for the conventional PS. In ongoing work we plan to introduce additional reactions missing in this preliminary model.

Author Response File: Author Response.docx

Reviewer 3 Report

The manuscript by Saldívar-Guerra  et al. reports a study on the thermal pyrolysis process of PS in the temperature range of 390-450°C. The authors compare two types of PS, the first one synthesized by free-radical polymerization and the second by nitroxide-mediated polymerization. The authors also propose a kinetic and mathematical model for the pyrolysis of dead and dormant polymer, able of explaining the main experimental results.

The paper is very well written, and although the topic is rather difficult for this reviewer, the interpretation of the experimental data given by the authors are clearly described.

So, I can recommend the paper for publication on Processes in the present form.

Author Response

The manuscript by Saldívar-Guerra  et al. reports a study on the thermal pyrolysis process of PS in the temperature range of 390-450°C. The authors compare two types of PS, the first one synthesized by free-radical polymerization and the second by nitroxide-mediated polymerization. The authors also propose a kinetic and mathematical model for the pyrolysis of dead and dormant polymer, able of explaining the main experimental results.

The paper is very well written, and although the topic is rather difficult for this reviewer, the interpretation of the experimental data given by the authors are clearly described.

So, I can recommend the paper for publication on Processes in the present form.

Answer.

We thank reviewer 3 for reading the submitted paper and for his/her comments.

Round 2

Reviewer 1 Report

I am satisfied that the authors have addressed my major concern about the reproducibility of the experimental results.  The error bars that are now shown provide valuable context for the experimental results.

Reviewer 2 Report

The authors have made the changes as suggested and responded to the comments I made on the previous version. I have no further issues with the paper.