Process Monitoring of Antisolvent Based Crystallization in Low Conductivity Solutions Using Electrical Impedance Spectroscopy and 2-D Electrical Resistance Tomography

Round 1

Reviewer 1 Report

The manuscript "Process monitoring of antisolvent based crystallization in low conductivity solutions using electrical impedance spectroscopy and 2-D electrical resistance tomography" was reviewed. Summarily, this manuscript presents a method of analyzing antisolvent based crystallization of sucrose solutions via EIS and ERT. This manuscript is generally well written and likely of interest to the readership of Applied Sciences. However, the referee recommends minor revisions before acceptance for publication based on the comments below.

1. Importantly, the authors are performing electrical impedance tomography (EIT) not ERT. The distinction is that ERT uses DC interrogation whereas EIT makes use of AC interrogation, which the authors are clearly doing. This needs to be amended throughout the manuscript.

2. The authors state several times the difficulty of performing EIT/ERT on low-conductivity materials, but, to the best of the referee's knowledge, they never actually report on the conductivity of this solution. This should be provided for better context.

3. Speaking of EIT/ERT applied to low-conductivity materials, expansive work exists in this area with regard to materials imaging via EIT. This should be discussed in the introduction to provide better context of this challenge to the readers and to direct readers to cases where the low-conductivity problem has been effectively overcome. Potential references include:

-H. Ghaednia et al., 2020, "Interfacial load monitoring and failure detection in total joint replacements via piezoresistive bone cement and electrical impedance tomography," Smart Materials and Structures.

-Z. Yang, 2020, "Detection of impact damage for composite structure by electrical impedance tomography," ACMSM 25.

-M. Clausi et al., 2019, "Direct effects of UV irradiation on graphene-based nanocomposite films revealed by electrical resistance tomography," Composites Science and Technology.

4. Page 3: The statement "Electrical conductivity is an intensive property and its distribution cannot be measured directly; it must be calculated from a measurement of the corresponding extensive property such as resistance." should be clarified. Strictly speaking, no material property can be directly measured and all must be calculated from a discrete measurement (e.g. elastic properties are likewise determined from measured forces, displacements, and with knowledge of specimen dimensions). 

5. EIT/ERT is highly dependent on regularization. The authors should clarify how they regularized the inverse problem and their rationale for that particular regularization.

6. Labels and units for the color bars should be added to Figure 12.

7. The conclusion section should be expanded to provide more detail.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The paper presents a well-defined study of EIS and ERT in different solutions with an application to industrial process monitoring. Experimental setup, materials and methods used have been adequately addressed.

EIS provided information in 1 dimension. Could ERT be used with different alternating voltages, at different frequencies, to provide spectroscopy information?

Figure 12 shows the ABS phantom reconstructed for demineralized water and sucrose solutions. What about tap water? Units should be shown in the legend. A better explanation of these obtained results should be given (how are the density distribution of crystals and the non-conductive region related?, why that shape? can these results be compared to other experiments or references?)  

minor:

Figure 1 and Figure 11 legends should be enlarged, as they cannot be read easily.

The inverse imaging equation used could be reminded in the article, as well as the Bayesian reconstruction method used.

Some of the foreseen applications or future uses could be commented in the abstract, to improve the readers interest in the paper.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

In this work, electrical impedance spectroscopy (EIS) and 2-D electrical resistance tomography were used as initial study of the sucrose crystallization in water. Initially simple EIS spectra were recorded for a large range of sucrose-water-alchol concentrations. These spectra were fitted with up to a second order model. Subsequently, resistance tomography was used to investigate the crystallization of sucrose.

The work is nice and well-written. It contains some good ideas and results, but some of the discussion is left behind, for example about Figure 9 and 10. Clearly this is a first approach to the process analysis of crystallization, but it shows some promising outcomes.

Detailed suggestions and comments

1. Bode plots should be made with the frequency in logarithmic scale, otherwise only the last decade is visible and all the rest of the data appears squeezed at the origin.

2. A suggestion for further works, although it is not much, it appears that there is some noise in the impedance data. This can lower the quality of the fitting. A very easy way to decrease the noise is to increase the integration time for the measurements or the sampling rate or both.

3. In Figure 8 the lines are too thin to understand which line fits which one. Usually it is convenient to use symbols or markers for the experimental points so that they can be seen also below the fitting lines.

4. Figure 9 and 10 are only presented but not discussed. What about the results of Model 3? The author could rationalize the model-related results through principal component analysis so to decrease the variable space to a manageable one.

5. Is there a physical meaning for the parameters of the models? Please comment on this.

6. In Table 3 it appears that sometimes model 2 has lower quality than model 1, although it has higher degrees of freedom. Why?

7. The Bayesian reconstruction algorithm was neither explained nor discussed. What are the main features? Why is it important?

Author Response

Please see the attachment

Author Response File: Author Response.docx