The manuscript by Lamaty, Giani and co-workers reports the use of several optimization strategies to maximize yield and minimize PMI for the acylation of amino acids with chloroacetyl chloride under mechanochemical conditions in a vibratory ball mill. The study presents a mechanochemical variant of a widely used amidation protocol and explores Design of Experiments (DoE) and Bayesian Optimization (BO) to improve performance. The topic fits the scope of RSC Mechanochemistry and will interest both process engineers and academic researchers in mechanochemical organic synthesis. The paper could also help popularize data-driven optimization methods in mechanochemistry.
However, before acceptance, the clarity and presentation should be improved. In its current form, the manuscript reads like a technical report rather than a critical study. A more critical, mechanochemistry-centered discussion of the optimization methods and their practical implications is needed. Specifically, the following points should be addressed:
1) The manuscript places strong emphasis on reagent stoichiometry, which is important but rather trivial. It would be valuable to discuss in greater detail the factors specific to mechanochemistry that should be included in optimization campaigns. The authors are encouraged to extend the discussion on the selection of these factors and the choice of suitable parameter ranges, since defining appropriate ranges is critical for DoE. Such recommendations would be highly useful for researchers aiming to apply BO and DoE to optimize mechanochemical reactions.
2) Is there a possibility to use stoichiometric reagent ratios by selecting a base that does not generate water as a by-product (unlike NaHCO3), since the water hydrolyses the acyl chloride? From a sustainability perspective, operating at stoichiometric ratios is essential, offering maximal reaction mass efficiency and optimal PMI.
3) While discussing BO and the importance of milling load, the authors note that the rheology of the reaction mixture and poor mixing efficiency compromise reproducibility. A general question therefore arises that should be properly discussed: how important is input-data quality for both DoE and BO models, and how well do the models tolerate erroneous results (e.g., caused by poor mixing)? Should such points be excluded, or do they not significantly affect the optimization outcome? If to exclude, how to identify such bad data points? In the presented data sets, several yields display large errors. If this was a consequence of poor and irreproducible mixing, the authors could demonstrate whether it is crucial or not to exclude such cases from the model.
4) Is it correct to start BO by using previously optimized results from DoE and OFAT? Shouldn’t BO be started from “scratch” for a fair comparison and a clearer demonstration of the advantages of BO, especially considering that the true maximum, or a point very close to it, might already be located in this a relatively simple system via OFAT and/or DoE?
5) When discussing the trade-off between yield and PMI, please clarify that these are not independent parameters (it is straightforward to derive an equation linking both values). They show opposite trends when excess reagents are applied, which necessitates identifying an appropriate trade-off.
6) In Scheme 1, the optimized conditions should be indicated below the diagrams displaying optimized yield and PMI value.
7) In the “comparison of the methods” section, please provide a clear summary comparing the three optimization methods, specifying the optimal conditions found, the yields achieved, and the number of experiments performed to reach the optimal result. Replace vague and low-informative statements, such as “huge number of experiments,” “a lot of experiments,” “little number of experiments,” with more informative ones, presenting specific numbers and relevant details.
8) Besides the cited works 17 and 30–32, please consider citing the recent study in which a DoE approach was essential to maximize yield in a mechanochemical synthesis: Suut-Tuule et al., Cell Reports Physical Science, 2024, 5, 102161 (DOI: 10.1016/j.xcrp.2024.102161).
9) Conclusion: “The results clearly demonstrate that mathematical and statistical tools can significantly enhance the optimization of chemical processes.” This statement is trivial and adds no value. Please revise the conclusion to be more specific, highlighting the most important outcomes that are directly relevant to mechanochemistry.

The manuscript has been improved after the revision. The following minor changes can be introduced to further improve clarity and correctness:
- When reporting yields together with uncertainty values, please follow the significant figure rules and accepted standards in error analysis: the uncertainty should normally be given with one (or, only in justified cases, two) significant figures, and the reported value should be rounded to the same decimal place as the uncertainty (e.g., 79 ± 9% but not 79 ± 8.9%). Please check this throughout the manuscript.
- On page 8, the sentence “Finally, noting the high standard deviation in entry 11, the addition of a liquid additive was studied.” could benefit from clarification. For example, please briefly explain that the high variability was attributed to unfavourable rheology and thus insufficient mixing, which motivated the use of a liquid additive.
- On page 11 (below Fig. 2), in the sentence “Moreover, since a lot of experiments were performed in the first part of the optimization (OFAT)…”, please specify the number of experiments.
- On page 11, in the sentence “Very good values were easily reached, with two sets of conditions providing the product 2a with a 93% NMR adjusted yield”, it would read more clearly as “High yield values were easily reached, with two sets of conditions providing product 2a in 93% NMR-adjusted yield.”
- On page 12, the expression “(i.e., yield x Molecular weight of the product x number of moles of 1a)” is currently not clearly interpreted as a calculation. Since this is an equation for determining the mass of product, please present it in a clearer and more explicit format. The use of “x” here is not immediately interpreted as multiplication.

The manuscript has been properly revised and can be accepted for publication