From the journal Digital Discovery Peer review history

31-May-2022

Dear Dr Junkers:

Manuscript ID: DD-ART-04-2022-000035
TITLE: Operator-Independent High-Throughput Polymerization Screening Based on Automated Inline NMR and Online SEC

Thank you for your submission to Digital Discovery, published by the Royal Society of Chemistry. I sent your manuscript to reviewers and I have now received their reports which are copied below.

I have carefully evaluated your manuscript and the reviewers’ reports, and the reports indicate that major revisions are necessary.

Please submit a revised manuscript which addresses all of the reviewers’ comments. Further peer review of your revised manuscript may be needed. When you submit your revised manuscript please include a point by point response to the reviewers’ comments and highlight the changes you have made. Full details of the files you need to submit are listed at the end of this email.

Digital Discovery strongly encourages authors of research articles to include an ‘Author contributions’ section in their manuscript, for publication in the final article. This should appear immediately above the ‘Conflict of interest’ and ‘Acknowledgement’ sections. I strongly recommend you use CRediT (the Contributor Roles Taxonomy from CASRAI, https://casrai.org/credit/) for standardised contribution descriptions. All authors should have agreed to their individual contributions ahead of submission and these should accurately reflect contributions to the work. Please refer to our general author guidelines https://www.rsc.org/journals-books-databases/author-and-reviewer-hub/authors-information/responsibilities/ for more information.

Please submit your revised manuscript as soon as possible using this link:

*** PLEASE NOTE: This is a two-step process. After clicking on the link, you will be directed to a webpage to confirm. ***

https://mc.manuscriptcentral.com/dd?link_removed

(This link goes straight to your account, without the need to log on to the system. For your account security you should not share this link with others.)

Alternatively, you can login to your account (https://mc.manuscriptcentral.com/dd) where you will need your case-sensitive USER ID and password.

You should submit your revised manuscript as soon as possible; please note you will receive a series of automatic reminders. If your revisions will take a significant length of time, please contact me. If I do not hear from you, I may withdraw your manuscript from consideration and you will have to resubmit. Any resubmission will receive a new submission date.

The Royal Society of Chemistry requires all submitting authors to provide their ORCID iD when they submit a revised manuscript. This is quick and easy to do as part of the revised manuscript submission process.   We will publish this information with the article, and you may choose to have your ORCID record updated automatically with details of the publication.

Please also encourage your co-authors to sign up for their own ORCID account and associate it with their account on our manuscript submission system. For further information see: https://www.rsc.org/journals-books-databases/journal-authors-reviewers/processes-policies/#attribution-id

I look forward to receiving your revised manuscript.

Yours sincerely,
Professor Jason Hein
Associate Editor, Digital Discovery

************

Reviewer 1

There are minor English issues. For example,
p 9 "good average proofs" should be "good average, proves"

Also some issues with terminology,
Probably all instances of rate constant should be rate coefficient

Ref 1 should be formatted as a Book reference.

The NMR spectrum in Figure S13 does not look well-phased.It is indicated to be a raw NMR spectrum. What processing was carried out on the NMR spectrum? Add more details to SI.

I am wondering why AIBN was chosen as initiator. In AIBN initiated polymerization of acrylates the rate of 2-cyano-2-propyle radical adding to monomer can be a rate determining step in the initiation process as a consequence the initiator efficiency may vary with monomer,, monomer concentration, and monomer conversion.[see (1) Prog. Polym. Sci. 2019, 88, 130-188. https://doi.org/10.1016/j.progpolymsci.2018.08.003]. This does not affect the generic conclusion.

SI p5, A reference should be cited re the MHKS parameters for polystyrene. These are not the same as those in ref 3. The GPC operating temperature may also be different (30 vs 40 degrees) . Because MHKS parameters are strongly cross-correlated, and may be temperature dependent, it is usually best to use a pair from the same data source. This may be a factor in the small departure from theory seen in the Mn conversion plots.

Reviewer 2

Review attached.

Reviewer 3

Congratulations to the authors on this impressive work and the interesting way they have written it up for a journal outside their usual field of play. I happily recommend this work for publication with just some minor revisions.

To be honest the only thing about this manuscript that irritated me was the failure of the authors to proofread it adequately - I counted 35 errors such as incorrect prepositions, incorrect verb tense, singular nouns that should have been plural, and so on. This is far too many to list here, so I will just email the pdf to Tanja. The authors should be aware that some reviewers object to being used as proofreaders, and take unpalatable action as a consequence!

Some additional suggestions for improvement:

* The first 5.5 lines of the Abstract are content that belong in the Introduction, for they are just general background information, not work specifically done in this manuscript.

* The opening sentence of the Introduction is pointless - it states something axiomatic (that, as a physical science, chemistry involves experimentation), and furthermore it tries to imply that the work of this paper will take chemistry beyond experimentation, but in fact all this work does is provide a more efficient means of doing experiments. So yes, this paper changes the paradigm, but not in the way the opening sentence implies. (Sorry to nag on this, but if you want to be like the Rolling Stones, then your opening riff has to be powerful, sharp and on point!)

* Page 2, bottom column 1: Is it worth making absolutely clear what the difference between inline and online is in the present context? It never hurts to eliminate any ambiguity in terminology - one person's inline may be another's online.

* Page 2, last paragraph of column 2: Isn't a potential issue that systematic error gets hidden by higher precision from this increased reproducibility? I feel this should be discussed.

* Page 4, "From there on, the operator only needs to provide key information to the software (such as weighed in amount of monomer, initiator, etc) to start the experiments."
Wouldn't it be relatively straightforward to automate these aspects of the procedure too?

* Page 4, "An example for human bias is setting spectral integration limits, or various assumptions being made when analysing data."
These procedures are not special to automated, flow experiments - they can just as easily (and profitably!) be applied to manual and/or batch reactor operation (and they should be!).

* Page 6, "it must be noted that no individual calibrations of all monomers are available, and hence deviations between experiment and theory most likely stem from SEC calibration errors that have no root in the automation".
This is a classic example of my earlier point about systematic error.

* Page 6, refs. 32 and 33: Surely this is an appropriate place to reference the recent IUPAC work by Harrisson et al. in Polymer Chemistry, which very powerfully evidences what is written here (including for acrylates).

* Page 6, second column: I feel these regular claims about the method leading to mechanistic understanding are a bit overblown. What is done here is a very impressive piece of chemical engineering, i.e., reducing a complex situation to a single parameter. But the authors well know that genuine mechanistic understanding in radical polymerization kinetics requires far more sophisticated delving.

please see attached detailed reply letter

This text has been copied from the PDF response to reviewers and does not include any figures, images or special characters:

Dear Prof. Hein,

thank you very much for sending us the reviewer's comments for the above manuscript and the opportunity to respond to them in a revision of our manuscript. All comments that we had received have been very carefully considered. Below, our responses to the referees’ comments are outlined and the changes to the manuscript are described. For your convenience, the referee comments are set in green. All changes made to the manuscript text are highlighted in red.






Referee 1:

There are minor English issues. For example,
p 9 "good average proofs" should be "good average, proves"

We have carefully revised the manuscript and corrected spelling or other errors.

Also some issues with terminology,
Probably all instances of rate constant should be rate coefficient
We found two instances on page 6 of the manuscript, and these were changed to ‘rate coefficient’.

Ref 1 should be formatted as a Book reference.

The formatting of reference 1 has been changed in the bibliography of the manuscript.

The NMR spectrum in Figure S13 does not look well-phased.It is indicated to be a raw NMR spectrum. What processing was carried out on the NMR spectrum? Add more details to SI.

The NMR spectrum in Figure S13 is a screenshot of the Spinsolve software. Reactions are monitored using the ‘RMX’ protocol of the software. All spectra are auto-phased prior to integral calculated. This information is added to the SI (page 5). Please note that the device used is a low field benchtop NMR (60 MHz) in a flow setup hence resolution is inherently lower compared to offline analyses. It is part of the automation process that a manual correction cannot be carried out for each individual spectrum.

I am wondering why AIBN was chosen as initiator. In AIBN initiated polymerization of acrylates the rate of 2-cyano-2-propyle radical adding to monomer can be a rate determining step in the initiation process as a consequence the initiator efficiency may vary with monomer,, monomer concentration, and monomer conversion.[see (1) Prog. Polym. Sci. 2019, 88, 130-
188. https://doi.org/10.1016/j.progpolymsci.2018.08.003]. This does not affect the generic conclusion.

As AIBN is a well-studied thermal initiator for acrylate polymerizations, and probably used in the vast majority of all studies in radical polymerization. Hence, we opted to choose this initiator for the validation of the platform. Yet, the reviewer is of course correct that the choice of initiator can influence the overall kinetics of the reaction. In future studies, the initiator and its concentration will be studied consequently. This will, however, be focus of a future contribution.

SI p5, A reference should be cited re the MHKS parameters for polystyrene. These are not the same as those in ref 3. The GPC operating temperature may also be different (30 vs 40 degrees) . Because MHKS parameters are strongly cross-correlated, and may be temperature dependent, it is usually best to use a pair from the same data source. This may be a factor in the small departure from theory seen in the Mn conversion plots.

The reference for the MHKS parameters for polystyrene has been added to the SI. Concerning the different operating temperatures, we believe that this is within the acceptable error margin for the purpose of this study, and a deviation that is commonly accepted. For the deviation of experimental Mn from theory, it should be noted that we did not have MHKS parameters for all monomers, hence explaining the differences. Compared to not knowing MHKS parameters at all, a slight temperature change is inconsequential.
Referee 2:

Herck and coworkers present a fully automated, continuous polymer synthesis platform as a flexible screening tool for accelerating materials discovery of new macromolecules. In this work, the authors (1) describe the workflow between in-house software and instrumentation to orchestrate rapid screenings of controlled radical polymerizations, (2) perform model RAFT polymerizations from 8 monomers by 8 unique operators, and (3) demonstrate automated analyses of ~3600 NMR and ~400 GPC test datasets. The authors motivate this work by emphasizing the growing digitalization of science and engineering, with special emphasis on the lagging adoption of truly automated capabilities in polymer chemistry.

Overall, this review presents a comprehensive account of an interesting platform for high- throughput polymer screening. The software communication setup to run RAFT synthesis campaigns in line with routine instrumentation represents a powerful approach to save time and effort in new materials development. The interdisciplinary nature of this work and the effort it took to build the platform are clear.

Below I only note a few questions and suggestions for the authors to implement. Upon re- evaluation and approval of these major revisions by the editor, I would be supportive of this work to be published in Digital Discovery.

We are thankful for the evaluation of the reviewer and the constructive feedback. We hope that we can convince the reviewer of the quality of our work with the revisions made.

- Safeguards for automated polymer characterization as a black box: One of the concerns often brought up for democratizing GPC+NMR in this manner is transparency and understanding of the “cleaning” algorithm before showing the results to a user. The authors do an excellent job describing issues such as maintaining raw data records, manual baseline corrections, and outlier effects. I only have a few clarification questions:
o In the introduction, it would be useful for the authors to provide a better account of what currently exists today. There are numerous reports outlined in Oliver, S. et al. (“Living in the Fast Lane—High Throughput Controlled/Living Radical Polymerization” Macromolecules 2019, 52, 1, 3–23)—although not all the tabulated examples involve flow chemistry, highlighting a few select examples of synthesis tools (i.e., liquid handlers, parallel synthesizing reactors, 96-well plates, etc.) to compare/contrast would be beneficial for the broader audience of Digital Discovery to appreciate the state-of-the-art in accelerating polymer chemistry. Almost 1/3 of the total references refer to the corresponding author’s own work— more diverse representation from the field would be beneficial to the audience as a whole.
The reviewer is of course correct that our work does not represent the first advance towards high throughput experimentation for RDRP. Our aim is to describe a novel and expandable approach using flow chemistry automation, for which we believe significant advantages exist due to the possibility to incorporate directly online analysis. To account better for the previous work, we added citations to high-throughput screenings in batch as well as the recommended reference (Macromolecules 2019) to the manuscript (reference 11 to 14). As for the own citations, this is a consequence of this project being a culmination of previous projects from our group. We hope though that the additional references result in a better balance.

o Are crude samples not purified before direct injection into the GPC? Since the sampling residence time has been shortened, is there not concern about overlapping peaks or damage to the column from small molecules remnants? What if a RAFT reaction results in multimodal or high dispersity distributions— will the software pick up on this non-deal use case and correct the analysis in the running queue?
Crude samples are not purified before GPC injection. Previous work of this method showed no issues when using this method. Of course, overlapping of distributions must account for the overall broadness and molecular weight range of residual RAFT polymers, and the method has been optimized for products typically obtained. Broader distributions would result in lower time resolution.
At present, the cleaning algorithm can detect deviations, and will at present simply delete the affected data. Since the injection parameter is controlled by the GPC software and not the automation software, a dynamic control is currently not possible. The operator would, however, be able to pick up the deviation quite quickly and adjust experiments accordingly. We have added a sentence in the discussion to note on this.

o The authors only report detailed kinetics and characterization of homopolymers. A far more useful end use application of this system would be to analyse copolymers (N monomers) or block polymers (M blocks), where complexity quickly limits the number of samples a student can usually study in N and M. Have any efforts been pursued by the researchers to fully characterize a multi- component system? This would be extremely impactful to the utility of the platform as a whole.
This manuscript indeed only exemplifies the kinetic analysis of homopolymers. The options for studying various systems are endless. However, each new system requires a different workflow of data treatment. It is expected that the platform will be continuously updated with new software features to also enables the screening of copolymers or block copolymers. As we noted, the presented data in this work is only a starting point, not the end of the investigations.

o For this flow system, what is the upper limit in molecular weight that can be examined for different polymer systems? In other words, what viscosity would be prohibitive for its use?
The viscosity of polymer samples is affected by various parameters. Careful monitoring of viscosity is recommended to avoid reactor blocking. Typically, we find for acrylates at higher monomer concentrations limits are in the range of about 50 000 Da. Enough examples exist in literature though for materials up to millions of Daltons. We added a sentence in the discussion on this limitation.

o Do the researchers deoxygenate their stock solutions? What temperature is used for the reactions? I cannot find the important synthetic details of this in the manuscript.
Reaction protocols can be found in the supporting information. To avoid confusion, we also added the information now to the main manuscript as well (page 5).

- Toward ‘Big Data’ Use: The automation scripts that the team has developed are an important step in managing the large datasets that are being generated by the platform. However, it was unclear how well-structured the metadata is for the aggregated dataset. More broadly, there is no data management statement provided to accompany this platform. It would be useful for the authors to give their perspectives on how this work can benefit the field of materials informatics, both being pursued by the academic and industrial workforce, collectively as a whole.
This is an excellent comment. We made all code available via Github, so the data structure in the generated CSV files is accessible. Our vision is in the next step to store this data in an open database that follows the FAIR principles. This is now made clearer in the conclusion section.

- Do the authors plan on making their polymer synthesis and characterization data follow FAIR guiding principles (Findability, Accessibility, Interoperability, and Reusability; Wilkinson, M. et al. Sci. Data 2016, 3, 160018.)? In terms of novelty, automating polymer synthesis with modular tools and connecting samples to GPC/NMR analysis have been well described in the literature (Oliver, S. et al. “Living in the Fast Lane— High Throughput Controlled/Living Radical Polymerization” Macromolecules 2019, 52, 1, 3–23), but I have yet to see a comprehensive report that follows FAIR practices so that materials data can be leveraged for machine learning purposes for the broader informatics community.
Indeed, we will be using the FAIR principles. We believe only truly open science will allow to generate the amount of data required for true big data analysis in the long term. As mentioned above, this is commented on in the conclusion sections now.

Minor Points:
- The authors denote polymer dispersity as “D” throughout the manuscript and Supporting Information. Please correct the symbol of dispersity (Đ) in all text and figures.
Thank you for picking up this mistake. Dispersity is now consistently represented by Đ.

- The font size of Figures 3-5 is too small and difficult to read.
We have redrawn the figures in the manuscript using larger font sizes.

- The plots in Figure S-32 are blurry and cannot be read. Please include a higher quality image of this collage.

The figure in question is an overview Figure, and the individual plots are given beforehand. In some sense, this figure is not meant to be readable, and merely displays the wealth of data obtained. It is a feature of ‘big data’ that it cannot be clearly shown in its full extend in a single page/figure. Hence, we wish to make no change here.

- The TOC graphic appears to reproduce a PhD Comics graphic by Jorge Cham in 2007. This will probably run into copyright issues with RSC re-publishing this as part of the manuscript, so the authors need to generate a new TOC.
We are of course aware of the copyright issues. We have repeatedly tried to contact the copyright owner to ask for permission already before we received these comments, and unfortunately have not received a reply yet. Hence we have designed a new TOC and replaced the previous version.



Referee 3:

Congratulations to the authors on this impressive work and the interesting way they have written it up for a journal outside their usual field of play. I happily recommend this work for publication with just some minor revisions.

Many thanks for the positive and encouraging comment.

To be honest the only thing about this manuscript that irritated me was the failure of the authors to proofread it adequately - I counted 35 errors such as incorrect prepositions, incorrect verb tense, singular nouns that should have been plural, and so on. This is far too many to list here, so I will just email the pdf to Tanja. The authors should be aware that some reviewers object to being used as proofreaders, and take unpalatable action as a consequence!

We appreciate the given suggestions and corrections and changed the indicated errors throughout the manuscript.
Some additional suggestions for improvement:

* The first 5.5 lines of the Abstract are content that belong in the Introduction, for they are just general background information, not work specifically done in this manuscript.

We agree on this. The first sentences in the abstract were consequently removed as suggested.

* The opening sentence of the Introduction is pointless - it states something axiomatic (that, as a physical science, chemistry involves experimentation), and furthermore it tries to imply that the work of this paper will take chemistry beyond experimentation, but in fact all this work does is provide a more efficient means of doing experiments. So yes, this paper changes the paradigm, but not in the way the opening sentence implies. (Sorry to nag on this, but if you want to be like the Rolling Stones, then your opening riff has to be powerful, sharp and on point!)
We would like to thank the reviewer for this critical note. We removed the first sentence and start the manuscript from the statement that chemical space screening is a tedious task.

* Page 2, bottom column 1: Is it worth making absolutely clear what the difference between inline and online is in the present context? It never hurts to eliminate any ambiguity in terminology - one person's inline may be another's online.

A definition of the terms inline and online was added to the manuscript (page 2).

* Page 2, last paragraph of column 2: Isn't a potential issue that systematic error gets hidden by higher precision from this increased reproducibility? I feel this should be discussed.

On first glance, this is of course a correct observation. The increased data accuracy can potentially reveal systematic errors. However, the nature of machine learning and statistics (if applied proficiently) is able to account for systematic errors as much as signal noise. Ideally, this would involve creation of data with different instruments and laboratories, and long-term this is our goal. We have added a comment in the conclusion section to note on systematic errors, and how we plan to handle them in the future. We wish to not discuss this in detail though as this is not our current aim with this publication, and once ready will also not be the result of our own work only.

* Page 4, "From there on, the operator only needs to provide key information to the software (such as weighed in amount of monomer, initiator, etc) to start the experiments."
Wouldn't it be relatively straightforward to automate these aspects of the procedure too?

The current setup is based on syringe pumps. Solution preparation and filling the syringes are task that cannot be automated currently. In future work, we aim to expand our setup to be able to mix its own reaction solutions. This would go too far at the present stage though.

* Page 4, "An example for human bias is setting spectral integration limits, or various assumptions being made when analysing data." These procedures are not special to automated, flow experiments - they can just as easily (and profitably!) be applied to manual and/or batch reactor operation (and they should be!).

We agree that setting spectral integration limits should be standardized to obtain unbiased and correct results. The development of automated analysis algorithms is indeed relatively straightforward. However, such practises are not really performed consistently in the science community. We therefore believe that emphasising on this part is necessary to showcase the full potential of digital transformation.

* Page 6, "it must be noted that no individual calibrations of all monomers are available, and hence deviations between experiment and theory most likely stem from SEC calibration errors that have no root in the automation".
This is a classic example of my earlier point about systematic error.

Indeed. Yet, it is an issue that is in principle correctable by obtaining correct MHKS values for the individual monomers. No change is required here.

* Page 6, refs. 32 and 33: Surely this is an appropriate place to reference the recent IUPAC work by Harrisson et al. in Polymer Chemistry, which very powerfully evidences what is written here (including for acrylates).

The recommended reference has been added to the manuscript.

* Page 6, second column: I feel these regular claims about the method leading to mechanistic understanding are a bit overblown. What is done here is a very impressive piece of chemical engineering, i.e., reducing a complex situation to a single parameter. But the authors well know that genuine mechanistic understanding in radical polymerization kinetics requires far more sophisticated delving.

We disagree with that statement. Of course, we are well aware of the complexities of chain growth mechanisms, and the interplay of the many involved single reaction steps. Yet, past research has focused on isolating all individual reactions to study them in isolation. While this works for some reactions such as propagation, it is impossible for others. We believe that our method opens a pathway to a completely machine-learning based understanding of mechanisms. While more data will be required to achieve deep learning analysis of simple rate of polymerization data, we very much believe that such analysis will become feasible in the nearer future. We have added some comments on this on page 6 to make our statement clearer.

02-Jul-2022

Dear Dr Junkers:

Manuscript ID: DD-ART-04-2022-000035.R1
TITLE: Operator-Independent High-Throughput Polymerization Screening Based on Automated Inline NMR and Online SEC

Thank you for submitting your revised manuscript to Digital Discovery. I am pleased to accept your manuscript for publication in its current form.

You will shortly receive a separate email from us requesting you to submit a licence to publish for your article, so that we can proceed with the preparation and publication of your manuscript.

You can highlight your article and the work of your group on the back cover of Digital Discovery. If you are interested in this opportunity please contact the editorial office for more information.

Promote your research, accelerate its impact – find out more about our article promotion services here: https://rsc.li/promoteyourresearch.

If you would like us to promote your article on our Twitter account @digital_rsc please fill out this form: https://form.jotform.com/213544038469056.

By publishing your article in Digital Discovery, you are supporting the Royal Society of Chemistry to help the chemical science community make the world a better place.

With best wishes,

Professor Jason Hein
Associate Editor, Digital Discovery


******
******

Please contact the journal at digitaldiscovery@rsc.org

************************************

DISCLAIMER:

This communication is from The Royal Society of Chemistry, a company incorporated in England by Royal Charter (registered number RC000524) and a charity registered in England and Wales (charity number 207890). Registered office: Burlington House, Piccadilly, London W1J 0BA. Telephone: +44 (0) 20 7437 8656.

The content of this communication (including any attachments) is confidential, and may be privileged or contain copyright material. It may not be relied upon or disclosed to any person other than the intended recipient(s) without the consent of The Royal Society of Chemistry. If you are not the intended recipient(s), please (1) notify us immediately by replying to this email, (2) delete all copies from your system, and (3) note that disclosure, distribution, copying or use of this communication is strictly prohibited.

Any advice given by The Royal Society of Chemistry has been carefully formulated but is based on the information available to it. The Royal Society of Chemistry cannot be held responsible for accuracy or completeness of this communication or any attachment. Any views or opinions presented in this email are solely those of the author and do not represent those of The Royal Society of Chemistry. The views expressed in this communication are personal to the sender and unless specifically stated, this e-mail does not constitute any part of an offer or contract. The Royal Society of Chemistry shall not be liable for any resulting damage or loss as a result of the use of this email and/or attachments, or for the consequences of any actions taken on the basis of the information provided. The Royal Society of Chemistry does not warrant that its emails or attachments are Virus-free; The Royal Society of Chemistry has taken reasonable precautions to ensure that no viruses are contained in this email, but does not accept any responsibility once this email has been transmitted. Please rely on your own screening of electronic communication.

More information on The Royal Society of Chemistry can be found on our website: www.rsc.org