An Autopsy Study of a Fouled Reverse Osmosis Membrane Used for Ultrapure Water Production
Round 1
Reviewer 1 Report
1. Line 76: correct citation style, Pororov et al. [16]
2. Line 103: why was the RO membrane operated with such a low recovery rate?
3. Table 2: Specifications H+ OH-
4. Line 178: three-dimensional fluorescence excitation-emission matrix
5. Delete lines 189-200
6. Line 206, insignificant
7. Line 209, hydrophobic ?
8. Line 228, fluorophores were only slightly lower
9. Please present Figures 2-6 in colour (for better understandability), showing the different intensities in EEM, using different line colours for chromatograms
10. Line 272, membrane reversible(?) fouling
11. Line 273, (i) chemically-reversible fouling ?
12. Line 274, cleaning) and (ii) hydraulically-reversible fouling ?
13. Line 276, 10.9%; DOC fraction of RO-B = 30.0%; DOC fraction of RO-S = 42.5%)
14. Line 288, higher SUVA value of desorbed foulants from RO-B compared
15. Lines 303 and 304: Emission < 380 nm corresponds to tryptophan-like (or protein-like) fluorescence and Emission > 380 nm for humic-like
16. Line 375, (51.7 º)> cleaned RO-S (51.3 º).
17. Delete lines 414-416
Author Response
1. Line 76: correct citation style, Pororov et al. [16]
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have corrected the citation style as follow:
The previous study demonstrated that the use of the RO membranes as a primary stage of UPW production systems could result in produced UPW having the maximum electrical resistivity of 2 MΩ cm [16].
2. Line 103: why was the RO membrane operated with such a low recovery rate?
Author’s response: It is well-known that the RO membranes operated with a low recovery rate increases rejection of RO membrane. In UPW production system, the rejection of RO membrane is the most important factor. Therefore, the RO membrane is operated with a relatively low recovery rate for effective performance of the rejection.
3. Table 2: Specifications H+ OH-
Author’s response: The authors have corrected the specification.
4. Line 178: three-dimensional fluorescence excitation-emission matrix
Author’s response: The authors have corrected the specification.
5. Delete lines 189-200
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have deleted the lines 189-200.
6. Line 206, insignificant
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have deleted the lines 206.
7. Line 209, hydrophobic ?
Author’s response: We respectfully disagree with the review’s opinion. The increase of SUVA value means that the increase of NOM’s hydrophobicity. Therefore, hydrophilic DOM components was preferentially removed through the MFC filter
8. Line 228, fluorophores were only slightly lower
Author’s response: The authors strongly agreed the reviewer’s comment. the authors have corrected as follow:
the intensities of humic-like fluorophores were slightly lower after treatment with the MFC filter.
9. Please present Figures 2-6 in colour (for better understandability), showing the different intensities in EEM, using different line colours for chromatograms
Author’s response: The authors thank the reviewer for this comment. However, we don’t see that the figures presented in colour is necessary for better understandability. (No change made)
10. Line 272, membrane reversible(?) fouling
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have corrected as follow:
In general, membrane fouling can be categorized into two groups according to the adhesion strength of foulants to membrane surfaces: (i) irreversible fouling (i.e., desorbable using chemical cleaning) and (ii) reversible fouling (i.e., desorbable using hydraulically cleaning)
11. Line 273, (i) chemically-reversible fouling ?
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have corrected as follow:
In general, membrane fouling can be categorized into two groups according to the adhesion strength of foulants to membrane surfaces: (i) irreversible fouling (i.e., desorbable using chemical cleaning) and (ii) reversible fouling (i.e., desorbable using hydraulically cleaning)
12. Line 274, cleaning) and (ii) hydraulically-reversible fouling ?
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have corrected as follow:
In general, membrane fouling can be categorized into two groups according to the adhesion strength of foulants to membrane surfaces: (i) irreversible fouling (i.e., desorbable using chemical cleaning) and (ii) reversible fouling (i.e., desorbable using hydraulically cleaning)
13. Line 276, 10.9%; DOC fraction of RO-B = 30.0%; DOC fraction of RO-S = 42.5%)
Author’s response: We have deleted as the other reviewer’s comment.
14. Line 288, higher SUVA value of desorbed foulants from RO-B compared
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have corrected as follow:
This observation was strongly supported by the relatively higher SUVA value of desorbed foulants from RO-B compared to that of RO-S.
15. Lines 303 and 304: Emission < 380 nm corresponds to tryptophan-like (or protein-like) fluorescence and Emission > 380 nm for humic-like
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have corrected as follow:
Three different fluorophores were found for RO-A at Ex = 230/Em = 260 nm (RO-A I; protein-like fluorophore, the maximum intensity = 156 mV), Ex = 250/Em = 360 nm (RO-A II; humic-like fluorophore, the maximum intensity = 175 mV), and Ex = 290/Em = 360 nm (RO-A III; humic-like fluorophore, the maximum intensity = 156 mV)
16. Line 375, (51.7 º)> cleaned RO-S (51.3 º).
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have corrected as follow:
The hydrophobicity of the cleaned RO membrane surfaces decreased from cleaned RO-A (64.3 º) > cleaned RO-DI (56.7 º) > cleaned RO-B (51.7 º)> cleaned RO-S (51.3 º).
17. Delete lines 414-416.
Author’s response: The authors have deleted the lines 414-416.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The manuscript entitled “Reverse osmosis membranes as a pre-treatment of ion-exchange resin filters for ultrapure water production: An autopsy study” provides data concerning the fouling characteristics and the effect of different cleaning procedures on RO membranes used to produce ultrapure water. The manuscript is within the topics of Water journal and the work to the reviewer’s best knowledge is original. In fact, although the there are numerous papers concerning RO membrane fouling, only few of them focus on RO membranes used in ultrapure water production. Another strong point of the manuscript is the good quality of the English language and its logical structure which facilitates its reading. Finally, the manuscript provides significant quantity of new data, especially concerning analytical characterization of membrane fouling layers and desorbed foulants through RO chemical cleaning procedure; these data may be valuable for studying fouling mechanisms in RO systems for ultrapure water production.
Unfortunately, there are two significant issues that significantly weaken the manuscript value.
a. The desorption procedure with the different chemicals was performed only once. Given that there is significant spatial distribution in the fouling layer on RO membranes the obtained results, although interesting, can only be considered as representative of the specific coupons, not the whole membrane. The chemical cleaning procedure and the respective analysis should be repeated at least twice in order to minimize the possibilities for errors due to sampling shortcoming.
b. The interpretation of some results is considered weak, whereas some conclusions are only partially supported by the results.
Therefore, the authors are kindly advised either to focus mainly on the proper presentation of their significant experimental work (omitting most rather arbitrary conclusions) or to perform further experiments (including repetitions) to support their conclusions.
In more detail:
i. Lines 268-271: The conclusion is weak. The results denote that salt cleaning seems to remove hydrophilic organic fractions but the relating mechanism, e.g. “reduced intermolecular... gel layers” is not supported by the results themselves.
ii. Lines 274-280: It is not logical to assume that each cleaning procedure removes different fraction of the organic fouling layer, thus it is not correct to support that e.g. acid cleaning results in the desorption of 10.9% of DOC. Please remove these lines, as well as the same data from Table 4.
iii. Lines 341-356: The interpretation of the ATR-FTIR results is weak. How did the authors explain the different behavior of the same chemical bond (e.g. N-H amides) both increasing and decreasing due to the fouling layer? The conclusion from ATR-FTIR measurement (Lines 353-356) is too general. The authors are advised to either rewrite sub-chapter 3.3.4 with the help of an ATR-FTIR expert or provide the ATR-FTIR data as Supplementary Material.
iv. Lines 411-412: The recovery efficiency of the cleaning procedures is not exactly consistent with the physicochemical properties of the fouled and cleaned membranes.
v. Lines 427-429: Please see comment i.
vi. The Materials and Methods section should be also improved; when did the membrane samples were collected, after the 3-year operation? What does “…moderate stirring conditions.” mean, at what rpm value? The crossflow velocity does not have L/min units, please calculate the crossflow velocity based on the filtration cell’s geometry. Finally, please provide some more data on the analytical methods, e.g. on the HSPEC conditions.
vii. Line 206: Why did the authors think that conductivity and UVA reduction “…were significant”? The conductivity decrease is less than 2% and can be easily attributed to standard errors…
viii. Lines 189-200 and 414-416 are from the journals template, please remove them.
The authors are also kindly asked to address the following points of minor importance:
a. Line 19: Please replace “remarkable” with “high”;
b. Line 41: Please replace “4,000 million” with “4,0 billion”;
c. Line 45: Please remove “from liquids”;
d. Lines 78-80 and Line 82: Please provide the relating references;
e. Line 213: Please replace “not effectively” with “hardly”;
f. Line 375: Please correct “cleaned RO-A” to “cleaned RO-S”;
g. Line 394: Please replace “However” with “Moreover”;
h. Line 408: Please delete “Therefore”;
i. Line 423: Please delete “systematically”;
j. Line 437: It would be good to add a couple of future research directions that can help understanding of the fouling phenomena in RO membranes for ultrapure water production.
k. The title “An autopsy study of a fouled reverse osmosis membrane used for ultrapure water production” is more representative of the manuscript’s content.
Author Response
The manuscript entitled “Reverse osmosis membranes as a pre-treatment of ion-exchange resin filters for ultrapure water production: An autopsy study” provides data concerning the fouling characteristics and the effect of different cleaning procedures on RO membranes used to produce ultrapure water. The manuscript is within the topics of Water journal and the work to the reviewer’s best knowledge is original. In fact, although the there are numerous papers concerning RO membrane fouling, only few of them focus on RO membranes used in ultrapure water production. Another strong point of the manuscript is the good quality of the English language and its logical structure which facilitates its reading. Finally, the manuscript provides significant quantity of new data, especially concerning analytical characterization of membrane fouling layers and desorbed foulants through RO chemical cleaning procedure; these data may be valuable for studying fouling mechanisms in RO systems for ultrapure water production.
Unfortunately, there are two significant issues that significantly weaken the manuscript value.
a. The desorption procedure with the different chemicals was performed only once. Given that there is significant spatial distribution in the fouling layer on RO membranes the obtained results, although interesting, can only be considered as representative of the specific coupons, not the whole membrane. The chemical cleaning procedure and the respective analysis should be repeated at least twice in order to minimize the possibilities for errors due to sampling shortcoming.
Author’s response: The authors thank the reviewer for this comment. However, we selected 4 different coupons of the RO membrane at random as noticed in the materials and methods section. (Lines 136) Therefore, we see that the obtained results are well-representative the characteristics of organic foulants desorbed from whole membrane.
b. The interpretation of some results is considered weak, whereas some conclusions are only partially supported by the results.
Therefore, the authors are kindly advised either to focus mainly on the proper presentation of their significant experimental work (omitting most rather arbitrary conclusions) or to perform further experiments (including repetitions) to support their conclusions.
In more detail:
i. Lines 268-271: The conclusion is weak. The results denote that salt cleaning seems to remove hydrophilic organic fractions but the relating mechanism, e.g. “reduced intermolecular... gel layers” is not supported by the results themselves.
Author’s response: The authors thank the reviewer for this comment. We have provided additional information as follow:
This shows that the reduced the electrostatic interaction of charged membrane surface and NOM, leading to double layer compaction and charge screening due to structural variations in cross-linked fouling gel layers during exposure of the RO membrane surfaces to the salt solution may lead to the desorption of hydrophilic organic foulants [26]. Furthermore, some fraction of neutrals foulants can be easily desorbed by decreased hydrophobic/hydrophilic interaction with infiltrating chlorides ions into Stern layer [27-29].
ii. Lines 274-280: It is not logical to assume that each cleaning procedure removes different fraction of the organic fouling layer, thus it is not correct to support that e.g. acid cleaning results in the desorption of 10.9% of DOC. Please remove these lines, as well as the same data from Table 4.
Author’s response: The authors agree with the reviewer for this comment. We have deleted as per review’s comment.
iii. Lines 341-356: The interpretation of the ATR-FTIR results is weak. How did the authors explain the different behavior of the same chemical bond (e.g. N-H amides) both increasing and decreasing due to the fouling layer? The conclusion from ATR-FTIR measurement (Lines 353-356) is too general. The authors are advised to either rewrite sub-chapter 3.3.4 with the help of an ATR-FTIR expert or provide the ATR-FTIR data as Supplementary Material.
Author’s response: The authors agree with the reviewer for this comment. The ATR-FTIR data have been moved as Supplementary Material as per review’s comment.
iv. Lines 411-412: The recovery efficiency of the cleaning procedures is not exactly consistent with the physicochemical properties of the fouled and cleaned membranes.
Author’s response: The authors agree with the reviewer for this comment. We have corrected as per review’s comment as follow:
Therefore, recovery rates of the permeate flux through the cleaning using DI water, acid, base, and salt solutions were in the order of cleaned RO-DI (permeate flux = 3.6 L m-2 hr-1) < cleaned RO-A (permeate flux = 4.1 L m-2 hr-1) < cleaned RO-B (permeate flux = 4.8 L m-2 hr-1) < cleaned RO-S (permeate flux = 5.3 L m-2 hr-1), which is consistent with the recovery rates of the negative surface zeta potential and the contact angle with the exception of RO-A. This is because that acid cleaning is more effective for desorbing inorganic foulants than organic foulants. The recovery rates of surface features more influenced by the amount of desorbed organic foulants on membranes surface [21,23,25,30]. Therefore, permeate flux recovery rate of RO-A can be higher than that of RO-DI, even though the surface features recovery of RO-A is lower than that of RO-DI.
v. Lines 427-429: Please see comment i.
Author’s response: The authors thank the reviewer for this comment. We have provided additional information as follow:
This was attributed to the efficient desorption of hydrophilic organic foulants through the reduced the electrostatic interaction of charged membrane surface and NOM, leading to double layer compaction and charge screening due to structural variations in cross-linked fouling gel layers during exposure of the RO membrane surfaces to the salt solution and infiltrating of chlorides ions into Stern layer
vi. The Materials and Methods section should be also improved; when did the membrane samples were collected, after the 3-year operation? What does “…moderate stirring conditions.” mean, at what rpm value? The crossflow velocity does not have L/min units, please calculate the crossflow velocity based on the filtration cell’s geometry. Finally, please provide some more data on the analytical methods, e.g. on the HSPEC conditions.
Author’s response: The authors agree with the reviewer for this comment. We have provided additional information in Materials and method section as follow:
Foulants were extracted from the fouled RO membranes after 3-years operation using four different desorbing agents: (i) DI water (RO-DI), (ii) an acid (0.1 N HCl) solution (RO-A), (iii) a base (0.1 N NaOH) solution (RO-B), and (iv) a salt (0.1 N NaCl) solution (RO-S).
Four coupons of the fouled RO membranes (effective surface area of each coupon = 0.04 m2) collected from the first stage of the RO membrane modules were soaked in 500 mL of each desorbing solution for 10 hours under moderate stirring conditions (150 rpm).
The RO filtration unit was operated in re-circulation mode at 20 ± 1 ºC. During filtration tests, feed water pressure was maintained at 1000 kPa with a cross-flow velocity of 8.68 cm sec–1.
high-performance size-exclusion chromatography (HPSEC) equipped with a Protein-Pak 125 column (Waters, Milford, MA, USA), with fluorescence detection at excita-tion (Ex) wavelength 278 nm and emission (Em) wavelength 353 nm (RF-10AXL; Shimadzu, Kyoto, Japan) and UV detection at 254 nm (SPD-10AVP; Shimadzu, Kyoto, Japan), and phosphate buffer(96 mM NaCl + 2.4 mM NaH2PO4+ 1.6 mM Na2HPO4, ionic strength 0.1 M, pH 6.8), with a flow rate of 0.7 ml/min and a injection volume of 200 μl [18, 19].
vii. Line 206: Why did the authors think that conductivity and UVA reduction “…were significant”? The conductivity decrease is less than 2% and can be easily attributed to standard errors…
Author’s response: The authors agree with the reviewer for this comment. We have deleted as per review’s comment.
viii. Lines 189-200 and 414-416 are from the journals template, please remove them.
Author’s response: The authors agree with the reviewer for this comment. We have deleted as per review’s comment.
The authors are also kindly asked to address the following points of minor importance:
a. Line 19: Please replace “remarkable” with “high”;
Author’s response: The authors agree with the reviewer for this comment. We have corrected as per review’s comment.
b. Line 41: Please replace “4,000 million” with “4,0 billion”;
Author’s response: The authors agree with the reviewer for this comment. We have corrected as per review’s comment.
c. Line 45: Please remove “from liquids”;
Author’s response: The authors agree with the reviewer for this comment. We have corrected as per review’s comment.
d. Lines 78-80 and Line 82: Please provide the relating references;
Author’s response: The authors agree with the reviewer for this comment. We have provided as per review’s comment as follow:
However, very little is known about the RO membranes used for UPW production (i.e., removal of ionic materials and formation of membrane fouling) compared to the electrodeionization (EDI) processes [2, 3, 6].
Although a recent study has emphasized the importance of the RO membranes in UPW production systems with the latest trends occurring in the world water market, there has been no comprehensive research on fouling characteristics of the RO membranes used for UPW production [1, 3, 4, 6].
e. Line 213: Please replace “not effectively” with “hardly”;
Author’s response: The authors agree with the reviewer for this comment. We have corrected as per review’s comment.
f. Line 375: Please correct “cleaned RO-A” to “cleaned RO-S”;
Author’s response: The authors agree with the reviewer for this comment. We have corrected as per review’s comment.
g. Line 394: Please replace “However” with “Moreover”;
Author’s response: The authors agree with the reviewer for this comment. We have corrected as per review’s comment.
h. Line 408: Please delete “Therefore”;
Author’s response: The authors agree with the reviewer for this comment. We have corrected as per review’s comment.
i. Line 423: Please delete “systematically”;
Author’s response: The authors agree with the reviewer for this comment. We have corrected as per review’s comment.
j. Line 437: It would be good to add a couple of future research directions that can help understanding of the fouling phenomena in RO membranes for ultrapure water production.
Author’s response: The authors agree with the reviewer for this comment. We have provided additional information about future research direction as follow:
Furthermore, the in-depth mechanism of salt cleaning for fouled membrane and its implication for water treatment process based on membrane technology can be further extended.
k. The title “An autopsy study of a fouled reverse osmosis membrane used for ultrapure water production” is more representative of the manuscript’s content.
Author’s response: The authors agree with the reviewer for this comment. We have corrected as per review’s comment
Author Response File: Author Response.docx
Reviewer 3 Report
- Abstract should be improved; it is too specific and focusing on only the second part of the work (membrane cleaning). The first part (ultrapure water production) should be better introduced.
- Line 104: The mentioned nominal pore size (< 0.005 μm) fits to NF range not RO. Most of RO membranes have pore size < 2 nm!
- Table 2: Correct the charges: "H-" and "OH+" in the 2nd and 3rd column titles!
- Line 134: It is better to use the terms "cleaning agents" than "desorbing agents".
- Line 145: What are humectants?
- Lines 189-200 have to be deleted!
- Lines 206-207: According to Table 3, changes in conductivity and UVA254 are certainly not significant; this part is also contracting with lines 212-215. Please check.
- Table 3: Comparing the conductivity values for the feed and RO permeate will give a salt retention of 82 %, which is very low for RO membrane, Please comment.
- Table 3: How were the relative errors calculated? The relative error for TN, DOC, UVA, SUVA are too high, comparing with the average values; is some cases errors are > 50%. Please check and comment.
- Line 269:" ….. in cross-linked gel layers". Which gel layers are meant here? PA TFC RO fully-aromatic membranes does not exhibit surface gel layers.
- Please add a reliable interpretation regarding how the cleaning action of NaCl works and its impact on PA TFC membrane surface properties.
-Lines 274-278: Why is DOC of desorbed RO foulants using salt solution considered among irreversible fouling? It does not match with your assumed explanation (Lines 268-271).
Lines 282-284: This conclusion is not matching with values presnted for RO-S in Table 4.
-Table 4: What if these values (i.e., values of desorbed foulants) are compared with the actually adsorbed values, then the real / actual cleaning efficiency can be estimated. Please comment.
- Section 3.3.4. (ATR-FTIR): the interpretation of IR charts should be repeated; the current version does not make sense. First, the IR peaks assigned for different function groups existing in the virgin membrane should be separated / substracted, thereafter explaining the IR chart is made! IR is no quantitative tool and what is shown is an overlay of different chemistries belong to membrane and fouling layers!
- Lines 370-372: Please add a clear description regarding how the zeta potential for fouled membranes was measured? Were they cleaned before. How could you explain the action of cleaning agents on producing more negative surfaces?
- Table 5 : Considering the error range, no significant influence on contact angle values can be seen.
- Lines 396-398: "....due to the hindered back diffusion of salts ….". This argument is certainly not valid. The interaction between concentration polarization and formed cake layer is highly depending on the intrinsic properties of the formed cake layer (e.g., porosity, thickness). The formed cake layer can cause cake-enhanced concentration polarized or cake-reduced concentration polarization. According to literature, most of organic foulants (e.g. humic substances) cause cake-reduced concentration polarization. Please comment properly and support your argument.
- Lines 414-416 have to be deleted!
- Author contributions: Authors should use the Journal formula.
Author Response
- Abstract should be improved; it is too specific and focusing on only the second part of the work (membrane cleaning). The first part (ultrapure water production) should be better introduced.
Author’s response: The authors have changed abstract section as follow:
This study investigated the fouling and cleaning behaviors of reverse osmosis (RO) membranes in a lab-scale ultrapure water (UPW) production system via membrane autopsies and characterization of dissolved organic matter (DOM) and membrane foulants. Most of DOM were effectively removed by the MFC filter, with the exception of the peak at 150 Da. The RO membranes were effective in reducing conductivity, DOM, TN, and UVA254nm concentration; the polishing stage using IER filter resulted in ultra-trace levels of all these parameters required for semiconductor manufacturing (> 18.2 ΩM). The quantity of the desorbed RO membrane foulants, in terms of dissolved organic carbon (DOC), varied considerably depending on the type of desorbing agents: 0.1 N NaCl (65.12 mgC m-2) > 0.1 N NaOH (46.14 mgC m-2) > deionized water (25.39 mgC m-2) > 0.1 N HCl (15.95 mgC m-2). The high cleaning efficiency of the salt solution (0.1 N NaCl) was attributed to the efficient desorption of hydrophilic DOM foulants from the RO membrane surfaces. These results demonstrate that the salt cleaning may provide a promising option to recover the performance of the RO membranes fouled primarily by hydrophilic DOM fractions.
- Line 104: The mentioned nominal pore size (< 0.005 μm) fits to NF range not RO. Most of RO membranes have pore size < 2 nm!
Author’s response: The authors strongly agreed the reviewer’s comment. However, it was revealed that the pore size and MWCO of membrane can be changed depending on its operating condition (e.g., J/k ratio, velocity of crossflow, and permeate water flux) according to our previous research (Kang et al., Effects of hydrodynamic conditions (diffusion vs. convection) and solution chemistry on effective molecular weight cut-off of negatively charged nanofiltration membranes, Desalination 352 (2014) 136-141). Thus, it is not appropriate to classify the types of membrane based on only the pore size.
- Table 2: Correct the charges: "H-" and "OH+" in the 2nd and 3rd column titles!
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have corrected the specification.
- Line 134: It is better to use the terms "cleaning agents" than "desorbing agents".
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have changed the terms “cleaning agents” to “desorbing agents”.
- Line 145: What are humectants?
Author’s response: The authors have changed as follow:
The virgin RO membranes were rinsed several times using DI water and soaked in DI water overnight to remove membrane-coating materials (i.e. humectants) from membrane surfaces.
- Lines 189-200 have to be deleted!
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have deleted lines 189-200.
- Lines 206-207: According to Table 3, changes in conductivity and UVA254 are certainly not significant; this part is also contracting with lines 212-215. Please check.
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have deleted as reviewer’s comment.
- Table 3: Comparing the conductivity values for the feed and RO permeate will give a salt retention of 82 %, which is very low for RO membrane, Please comment.
Author’s response: The authors strongly agreed the reviewer’s comment. There is ion exchange resin filter process after RO membrane process in the UPW production system. Therefore, we operated RO membrane process with a relatively low rejection for long-term operation.
- Table 3: How were the relative errors calculated? The relative error for TN, DOC, UVA, SUVA are too high, comparing with the average values; is some cases errors are > 50%. Please check and comment.
Author’s response: The authors strongly agreed the reviewer’s comment. However, TOC/TN analyzer has its limitation (below 0.1 mg/L). Therefore, the concentration which is below the limitation can have relatively high errors.
- Line 269:" ….. in cross-linked gel layers". Which gel layers are meant here? PA TFC RO fully-aromatic membranes does not exhibit surface gel layers.
Author’s response: PA TFC RO fully-aromatic membranes does not exhibit surface gel layers on its structure. However, during the membrane is fouling, the fouling gel layer is formed on its surface (even PA TFC membrane). Therefore, the authors have corrected as follow:
This shows that the reduced the electrostatic interaction of charged membrane surface and NOM, leading to double layer compaction and charge screening due to structural variations in cross-linked fouling gel layers during exposure of the RO membrane surfaces to the salt solution may lead to the desorption of hydrophilic organic foulants [26].
- Please add a reliable interpretation regarding how the cleaning action of NaCl works and its impact on PA TFC membrane surface properties.
Author’s response: The authors provided the interpretation with a reference as follow:
Furthermore, some fraction of neutrals foulants can be easily desorbed by decreased hydrophobic/hydrophilic interaction with infiltrating sodium and chlorides ions into Stern layer [27-29].
-Lines 274-278: Why is DOC of desorbed RO foulants using salt solution considered among irreversible fouling? It does not match with your assumed explanation (Lines 268-271).
Author’s response: The authors have assumed the irreversible fouling is desorbable using chemical cleaning, and the reversible fouling is desorbable using hydraulically cleaning. Therefore, all desorbed foulants using chemical cleaning, except of using DI water, can be categorized into irreversible fouling. Because, the foulants desorbing using salt solution is also based on surface chemistry.
Lines 282-284: This conclusion is not matching with values presnted for RO-S in Table 4.
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have deleted as follow:
However, Al, Ca, Fe, and Mg were commonly observed to be dominant inorganic contents of the desorbed RO membrane foulants, likely due to their high concentrations in feed water.
-Table 4: What if these values (i.e., values of desorbed foulants) are compared with the actually adsorbed values, then the real / actual cleaning efficiency can be estimated. Please comment.
Author’s response: The authors strongly agreed the reviewer’s comment. However, it is impossible to accurately measure actual adsorbed value. Therefore, the analyzing desorbed foulants is commonly used for analyzing membrane foulants.
- Section 3.3.4. (ATR-FTIR): the interpretation of IR charts should be repeated; the current version does not make sense. First, the IR peaks assigned for different function groups existing in the virgin membrane should be separated / substracted, thereafter explaining the IR chart is made! IR is no quantitative tool and what is shown is an overlay of different chemistries belong to membrane and fouling layers!
Author’s response: The authors agreed with the reviewer’s comment. ATR-FTIR spectroscopy was used to analyze the functional groups composition of virgin and fouled membranes (not quantitative analyzing). However, we see that the IR intensity also can relatively quantify (not accurately quantify) at the same condition using ATR-FTIR. The reference which relatively quantified IR intensity using ATR-FTIR of virgin and fouled membrane are listed.
1. Jaeweon Cho, Gary Amy, John Pellegrino, Yeomin Yoon, "Characterization of clean and natural organic matter (NOM) fouled NF and UF membranes, and foulants characterization", Desalination, Vol. 118, 101-108, 1998.
2. Pei Xu, Jorg E. Drewes, Tae-Uk Kim, Christopher Bellon
3. Hyun-chul Kim, Jong-Hyun Hong, Seockheon Lee, Fouling of microfiltration membranes by natural organic matter coagulation treatment: A comparison of different initial mixing conditions, J. membr. Sci. 283, 266-272, 2006
4. Hojung Rho, Kangmin Chon, Jaeweon Cho (CA), Surface charge characterization of nanofiltration membranes by potentiometric titrations and electrophoresis: Functionality vs. zeta potential, Desalination 427, 19-26, 2018
- Lines 370-372: Please add a clear description regarding how the zeta potential for fouled membranes was measured? Were they cleaned before. How could you explain the action of cleaning agents on producing more negative surfaces?
Author’s response: The authors strongly agreed the reviewer’s comment. The authors have additionally provided in materials and method section as follow:
Prior to analyze surface feature, the virgin, fouled and cleaned RO membranes were soaked with DI water for overnight and were dried in a vacuum desiccator for 48 hours.
- Table 5 : Considering the error range, no significant influence on contact angle values can be seen.
Author’s response: It can bee seen that there is no significant influence on contact angle value depending on chemical desorbing. Therefore, the surface feature of membrane is commonly analyzed not only contact angle results but also zeta potential results. The zeta potential results show same trend with contact angle results depending on desorbing agents.
- Lines 396-398: "....due to the hindered back diffusion of salts ….". This argument is certainly not valid. The interaction between concentration polarization and formed cake layer is highly depending on the intrinsic properties of the formed cake layer (e.g., porosity, thickness). The formed cake layer can cause cake-enhanced concentration polarized or cake-reduced concentration polarization. According to literature, most of organic foulants (e.g. humic substances) cause cake-reduced concentration polarization. Please comment properly and support your argument.
Author’s response: The authors thank the reviewer for this comment. We have provided additional explanation as follow:
However, a significant decrease was observed for the salt rejection rate of the fouled RO membranes (88.2%) compared to the virgin RO membranes (97.5%). There may be two possible reasons; i) The interaction between concentration polarization and formed cake layer is highly depending on the intrinsic properties of the formed cake layer (e.g., porosity, thickness). The formed cake layer can cause cake-enhanced concentration polarized which cause increasement of salt passage through the RO membranes [32]. ii) the increased salt concentrations at the RO membrane surfaces (due to the hindered back diffusion of salts from the RO membrane surfaces by forming colloidal cake layers on those surfaces) can increase the passage of salts through the RO membranes [33].
- Lines 414-416 have to be deleted!
Author’s response: The authors have deleted lines 414-416.
- Author contributions: Authors should use the Journal formula.
Author’s response: The authors have corrected using Journal formula as follow:
H. R. analyzed the water sample and membrane. K. C. and J. C. designed the experiments and drew the conclusions.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
The authors have made appropriate changes in the manuscript
Reviewer 3 Report
Table 3: Comparing the conductivity values for the feed and RO permeate will give a salt retention of 82 %, which is very low for RO membrane, Please comment.
Author’s response: The authors strongly agreed the reviewer’s comment. There is ion exchange resin filter process after RO membrane process in the UPW production system. Therefore, we operated RO membrane process with a relatively low rejection for long-term operation.
Please add this clarification to the main text.
