Bulgarian Mavrud Wine Under Nanofiltration and Reverse Osmosis: Evaluating the Composition After the Process

Comment 1: The abstract could include more specific details, such as the percentage of ethanol reduction achieved and changes in the main components.

Response 1: Thank you for pointing this out. We agree with this comment. In the revised version of the text the abstract includes some information about the lowest achieved alcohol concentration and examples of the rejection for the carboxylic acids.

Comments 2:   The results section could be expanded to include a sensory evaluation of the wine after filtration (were the aromas and flavour profile preserved?) .

Response 2: Thank you for the proposing this. It fits perfectly to the research but the following was evaluated in a separate work, which was cited in the article. The citation and its discussion now is revised to underline the connection more obviously. in another research “Separation of ethanol and polyphenols from red wine by diananofiltration” by Tsibranska et. al (10.1016/j.cherd.2025.02.023), published this year presents the changes in the aromatics during the filtration process.

Comment 3: Regarding statistical data – the article lacks information on the repeatability of the analyses (number of repetitions) and basic statistical indicators (e.g., SD or RSD), or this has been omitted.

Response 3: Thank you for pointing this out and being patient. The statistical analysis initially wasn’t part of the initial concept of the paper. It was essential to present the results achieved during the nanofiltration process. When the HPLC method was developed the limits of detection and quantification were assessed for the ethanol and for the two major groups of compounds – carboxylic acids and sugars (including the glycerol). All standard solutions samples were triple-repeated. With the delivered data, the SD was calculated for all of the compounds, and a calibration plot was prepared for each component. The groups LoD and LoQ was defined using the SD and the calibration plot.

Comment 4: In Table 3, it would be advisable to present the data in a clearer way, e.g., by dividing the results for each membrane into separate tables.

Response 4: Dear Reviewer, thank you for pointing this out. Your recommendation was implemented to improve the reader's understanding of the tables’ content. This table is divided into three parts, each with its description.

Comment 5: In the conclusions, it would be worth emphasizing the practical prospects for applying NF/RO technology in commercial production of low-alcohol wines.

Response 5: Thank you for the valuable advice. A comment in the conclusions suggests that the application of NF/RO technology could be implemented to control the alcohol content and preserve the unique Mavrud content profile.

Comment 1: 1. The introduction is to long and the last part should be placed in the Discussion section. Please shorten it and the exceeding information add it to the Discussion part.

Response 1: Dear reviewer thank you for the attention given to our text. Your suggesttion has been implemented. In the revised version of the text the last part of the Introduction is now fitted in the Discussion part.

Comments 2: Materials and Methods section is missing so please write it.

Response 2: Dear reviewer thank you for pointing this. The Materials and methods part is now majorly revised and including all the necessary information.

Comment 3: 3. Section 2.1. does not present anything about the wine characteristics. Please remake it. Also lines 125-129 should be deleted.

Response 3: Dear reviewer, thank you for the attention given to the text. The following section gone under major revision. The lines you’ve pointed are deleted.

Comment 4: 4. You did not perform any statistical analyses for the measured parameters. Please explain why?.

Response 4: It wasn’t intentional to miss the statistical analysis. Since there has been a hurry during the preparation. The statistical analysis of the presented results was missed to be included, also there due to uncertainty of what method of analysis would be more appropriate. It was decided to use a sample standard deviation to asses the deviation in the components groups in the received rejection coefficients. The latter confirms the observed trends during the nanofiltration processes that have been carried.

In the revised version of the paper the following statistical data is included:

-           limits of detection and quantification for the used method of analysis;

-           standard deviation for the obtained results from the collected samples (a graphical representation for the experiments with the three different membranes in concentration mode of operation);

-           error bars (error in %) for each compound on the figures that represent the retention coefficient.

Comment 5: Lines 222-241 for Results and discussions should be part of a subsection of Materials and methods.

Response 5: Dear reviewer, we would like to inform you that the revision suggestion was accepted and the results and discussion part is revised.

Comment 6: Conclusions are too long. Part of it should be transferred to the discussion section.

Response 6: Dear reviewer thank you for the paid attention. Your suggestion has been implemented in the revised text of the paper.

5. Additional clarifications

Dear editors and reviewers, thank you for the spent time, the patience and attention to all the details that any paper submitted to the journal needs.

Comment 1: End of line 63 does not seem structurally or meaning wise correct to me. Please check.

Response 1: Dear reviewer, thank you for the attention to the text. We agree with this comment. In the revised version of the introduction the following line is revised to fit the main idea of the paragraph.

Comments 2:   What was the frequency of fouling removal?

Response 2: Dear reviewer thank you for your interesting question. The short answer – the membranes are cleaned after each experiment. While the more detailed answer to the question could be formulated in the following manner.

The used membranes are conditioned before work, and after the work, they are cleaned with distilled water. The cleaning procedure runs at specific filtration parameters (TMP; feed flow and permeate flow). The main parameter that have to be controlled, this is the permeate flow.  In case the normalized permeate flow drops about 10-15%, chemical cleaning has been applied with 0.01 to 0.05% aqueous sodium hydroxide solutions at the following conditions: pressure 5 bar; temperature 30-35⁰C; circulation time 10 - 30 minutes; followed by a 10 - 30 minutes soaking period and then a final 30 minutes’ recirculation before flushing. These recommendations are provided by the manufacturers of the membranes. This allows to keep the membranes in working condition.

Comment 3: Flow: 0.6 not 0,6 mL/min on line 167.

Response 3: Dear reviewer thank you for your comment. The given recommendation was accepted and revised in the text. Allow to explain briefly the use of the operational units - it was considered to give them since they are considered important information.

Comment 4: 4. 4.         How is the column temperature of 70 C decided? Did the authors check the RSD% of control at 70 C? If yes, what was the RSD%? If not then why not? Did the authors check other temperature?

Response 4: Dear reviewer, thank you for the interesting question.

The used temperature was based on the available literature about the used column (Schneider, Gerbi and Redoglia, 1987; Ding, Koizumi and Suzuki, 1995). The literature was provided by the producer of the analytical column – Agilent Technology, where applications of the column are discussed. One of the applications was wine analysis. Based on this, the named temperature was chosen for the method. Other temperatures weren’t tested for the method application. The following is now included in the text of the work.

The RSD% can be calculated from the S/N ratio of the samples using the equation RSD% ≈ 50/(S/N). A calibration procedure was carried out for the compounds presented in the work.

When the HPLC method was developed, the limits of detection and quantification were assessed for the ethanol and the two major groups of compounds – carboxylic acids and sugars (where glycerol was included). All standard solution samples were triple-repeated. With the delivered data, the SD was calculated for all of the compounds, and a calibration plot was prepared for each component. The components LoD and LoQ were defined using the SD and the calibration plot.

Comment 5: Were there any carryover?

Response 5: Dear reviewer, thank You for the interesting question. Briefly – no. This is because when the development of the method started a large injection volume was used (15 μL) and it was known that any higher volume leads to column carryover. Several trials with various injection volumes (5, 8, 10, 15 μL) of the standard solutions (maximum concentration) and the wine were tested initially. Based on the peak form and its asymmetry the 5 μL injection volume was chosen to work with. The last is included in the revised text of the work.

Comment 6:  Can the authors please briefly state what is meant by rejection for the greater audience? What percentage of rejection is expected? How does the obtained rejection differ to conventional wine filtration processes?

Response 6:

Rejection coefficient represents the ability of a membrane to selectively retain specific molecules and allow others to pass through it freely. For a compound, the rejection is defined as the ratio between the concentration of the component in the solution that has passed through the membrane (permeate) and its concentration in the solution that is left behind the membrane (retentate).

The membrane processes discussed in the article are nanofiltration and reverse osmosis which are widely used in the food and beverage industry. These processes are aimed to partially or fully remove specific compounds from the wine’s solution.  The obtained results on the rejection of the discussed components are in agreement with those published by Kumar et al. 2024 (https://doi.org/10.1016/j.lwt.2024.116228).

The conventional membranes for nanofiltration and reverse osmosis allow for achieving rejection coefficients of about 90% for the carboxylic acids and sugars, and between 2 and 8% for the ethanol. When the membrane processes are compared with other conventional techniques for lowering the alcohol content, their advantage clear to achieve a high degree in preservation of the valuable compounds and rejecting ethanol and other components that are not needed in the solution.

Comment 7: Line 213 has typo.

Response 7: Dear reviewer, thank you for the time and attention you spent on the text that we submitted. The typo was corrected, and the whole text of the paper is now revised.

Comment 8: I recommend the authors include representative chromatograms to demonstrate the peak for each of the marker compounds they tested and reported.

Response 8: Dear reviewer, thank you for the proposal that will contribute to the scientific value of the paper. The following is done, but a combined representative chromatogram with the peaks of EtOH, carboxylic acids and sugars (combined) is now included in the text. Since the it is considered that it will take too much space if all of the components were included. It is included in the “Analytical method” section.

Comment 9: How the compounds were quantitated? Did they use standards for each in a from of standard mix at a specific concentration? These details need to be added to the manuscript.

Response 9: Dear reviewer, thank you for the proposal that will contribute to the scientific value of the paper. The compounds were quantified using standards for each compound in the form of a standard mix. In the laboratory were prepared mixed solutions with known mass concentration – 1 g/L (+/- 0.0001 g) for each component. The solutions were diluted (10 and 100 times) to achieve various concentration, based on this a method calibration was performed. The components were split in two groups – ethanol and carboxylic acids plus sugars (were the glycerol was included). Then based on the standard deviation for the lowest concentration the lower limits of detection and quantitation were found. The last is presented in the revised text of the paper, section “Analytical method” in a form of table.

Comment 10: The bar graphs do not have any error bars. If they tested each samples in duplicates, they should calculate the standard deviation or standard errors and should include them.

Response 10: The error (in %) was added on the bar graph now, based on deviation between the duplicated samples. Statistical analysis is performed. The standard deviation extracted from the experimental data is used as a indicator when discussing the effectiveness of the operating modes and membranes used during the experiment for wine filtration.

5. Additional clarifications

Dear editors and reviewers, thank you for the spent time, the patience and attention to all the details that any paper submitted to the journal needs.

Comment 1: Comments on the Quality of English Language

The English language must be improved concerning the grammar errors noticed during the review

Response 1: Dear reviewer, we highly value your attention to the text and the time spent to give your recommendations in the PDF attachment. All recommendations were approved and implemented to bring up the scientific value of the paper. Major revision of the text, results representation, graphical representation, table information has been made. Also a statistical analysis was made in order to give an unambiguous interpretation of the results obtained and presented in the paper.