Direct Observation of Growth Rate Dispersion in the Enzymatic Reactive Crystallization of Ampicillin
, Andreas S. Bommarius 1,2,3, Martha A. Grover 1,2 and Ronald W. Rousseau 1,*Round 1
Reviewer 1 Report
Review of manuscript «Direct observation of GRD in the reactive crystallization of ampicillin” by McDonald, Bommarius, Grover and Rousseau, submitted to Processes
The manuscript studies experimentally the growth of individual crystals of ampicillin (needle-shaped) during reactive crystallization in a suspension kept between two glass surfaces, with crystals observed through a microscope. Since crystals do not move it is possible to follow and measure their growth for long times, e.g. two hours as in the experiment illustrated in Figure 1.
The manuscript is novel and original in its approach hence it is an important and stimulating contribution to the literature on growth rates and growth rate dispersion. It should definitely be published, but I believe that it requires a major revision to make sure that its impact is as large as the work deserves.
I am suggesting that the authors address the following specific points:
1. First of all, the whole analysis is based on the observation of 887 crystals. Although this appears to be a large number of crystals, it is obviously minuscule when compared with the number of crystals in a suspension of a typical lab experiment. Therefore, there is an issue of statistical representability of the measurements reported. I think that this is the major disadvantage of the “simple” technique illustrated in Figure 1 that the authors have selected (instead of the approach of working with the whole population as illustrated in Figure 2). I am not arguing that the approach proposed is not useful; I am arguing that the authors should discuss this point critically and carefully.
2. The statistical tests discussed in section 3.1, in section 3.2 and in section 4 are the key tools used to interpret the experimental data. I believe that such tests should be explained in more detail, so as the reader can appreciate not only the conclusion drawn from the outcome of the text, but also the details of the statistical test (why and how that specific test).
3. In the context of the previous point, I believe that it is extremely important that the authors discuss the significance of tests, which depend on a specific value of the confidence level. In certain cases, the outcome of the statistical test seems to be in contradiction with the large scatter observed in the data, e.g. in figure 5a and 5b. I recommend that the authors discuss this point more in detail and critically.
4. As far as the analysis around the use of the space partition illustrated in figure 6, I am puzzled by the fact that the authors assume that the growth in the width direction can be neglected when doing material balances around growth of needles in specific subdomains of the suspension volume. In fact, the change in volume V of a crystal is related to the change in length L and width W by: dV/V = dL/L + 2 dW/W. This shows – at least to me – that the growth in the width direction corresponds to a more significant volume change than the growth in length. I would like that the authors discuss this point and clarify their assumption.
Author Response
See the replies in the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
McDonald and co-workers present in this work a methodology to study and characterize the crystallization of ampicillin using the “reactive crystallization” method. In this study they show that the dispersion of the growth rate observed during the crystallization of ampicillin trihydrate is explained by the constant crystal growth model controlled by the crystals’ surface morphology and crystal imperfections, although the action of the precursor may also have a relevant influence. Although the theme and methodology are very interesting, in its current form this article is very hard to follow due to the weak contextualization and the lack of well-accepted terminology and methodology.
There are two relevant facts in this work. Firstly, the are using a seeding protocol, to produce their desired final product but surprisingly none of the words: seeding, seed, seeds, appear along the article proving a lack of knowledge of this methodology. Among other advantages, seeding protocols are used to obtain homogeneous out-put products which therefore have grown with low dispersion of the growth rate. Please see Encarnacion-Gomez et al., (2016) [1] for a similar study to the one presented in this manuscript.
I have to quote “reactive crystallization” since this terminology can be misunderstood. This is sometimes used to define just crystallization or precipitation but since in this work the reaction is enzymatic, why not use the word enzymatic in the title? Actually, the enzymatic reaction produces the ampicillin that crystallized and therefore if the rate of generation of ampicillin is slower than the crystal growth rate, it controls the supersaturation level and therefore the growth.
Other points that need to be addressed:
- Homogeneity of the seed. SEM, TEM, micrographs could work.
- Did the authors optimized when to seed the reaction solution? If the seeding was done when the concentration was not high enough, the seeds will be dissolved affecting the homogeneity of the system i.e. higher local supersaturation. In all experiments it seems that seeds were added at the beginning. If this is right, it will be difficult to understand the results and to compare among the different experiments since at the higher loaded seeds the amount of dissolved ampicillin will be much more.
- Please used acronyms only for names, not for process.
[1] Encarnación-Gómez LG, Bommarius AS, Rousseau RW. Reactive crystallization of β-lactam antibiotics: strategies to enhance productivity and purity of ampicillin. Reaction Chemistry & Engineering. 2016;1:321-9.
Author Response
Please see the replied in the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
The authors describe the growth rate dispersion of growing ampicillin fibers.
This is an extensive study of ampicillin crystal growth and a lot of measurements is presented, statistically evaluated, and discussed.
Nevertheless, I would like to see a more detailed technical description of the microscopy part (model, objective lens used, camera, frame size in pixels, pixel size, bit depth, etc.).
Author Response
Please see the replies in the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
I am very happy with the answers and the changes.
There are only two points in Appendix B:
(i) The authors refer to eqn S1, but now this is called eqn. B1 in figure B2.
(II) The third column (volume flux) in the table in fig B2 is not clear to me: the authors should provide the equation used to calculate it.
Author Response
(a) S1 is changed to B1.
(b) The following statement is added to the end of the last paragraph of Appendix B: The volume flux, defined in terms of crystal volume added (ΔV, μm3) per area of specific crystal face (A, μm2) per minute (Δt, min), occurs at the ends of the crystal. For the sample crystal in Figures B1 and B2, the volume flux to the different faces is summarized in the table in B2.
Reviewer 2 Report
I have no further comments.
