Assembling covalent organic framework membranes with superior ion exchange capacity

Ionic covalent organic framework membranes (iCOFMs) hold great promise in ion conduction-relevant applications because the high content and monodispersed ionic groups could afford superior ion conduction. The key to push the upper limit of ion conductivity is to maximize the ion exchange capacity (IEC). Here, we explore iCOFMs with a superhigh ion exchange capacity of 4.6 mmol g−1, using a dual-activation interfacial polymerization strategy. Fukui function is employed as a descriptor of monomer reactivity. We use Brønsted acid to activate aldehyde monomers in organic phase and Brønsted base to activate ionic amine monomers in water phase. After the dual-activation, the reaction between aldehyde monomer and amine monomer at the water-organic interface is significantly accelerated, leading to iCOFMs with high crystallinity. The resultant iCOFMs display a prominent proton conductivity up to 0.66 S cm−1, holding great promise in ion transport and ionic separation applications.


Reviewer #1 (Remarks to the Author):
It is an interesting article which should be published after minor revision.

Reply:
Thanks the reviewer for the highly positive remarks. Fig. 1g cm3 Reply:

Correct in
Following this reviewer's valuable guidance, we have corrected "cm3" as "cm 3 " in Fig. 1g of the revised manuscript, as shown below.  Which thickness can be used without support? Reply: Thanks for the reviewer's valuable guidance. All iCOFMs in our manuscript were used and tested without support. Figure S16 is the digital photograph of the TpBD-(SO3H)2 membranes. When the membrane thickness exceeds 3 µm, the mechanical strength of the membrane is strong enough for testing without support. Only when the membrane thickness is as low as 0.5 µm, the mechanical strength of the membrane cannot be tested without support.

Reply:
Thanks for the reviewer's valuable guidance. We have added some speculations about possible applications in the Conclusions section of the revised manuscript. "…outperforming all the currently reported COF-based proton-conducting materials. We anticipate that our membranes can find diverse applications such as fuel cell, flow cell, lithium battery and nanofiltration. Moreover, our dual activation interface polymerization strategy can also inspire researchers in the field of chemical bond activation, and is expected to become a platform technology for organic framework membrane fabrication." 6. Literature citations need revision.

Reviewer #2 (Remarks to the Author):
In this manuscript, the authors reported the dual-activation interfacial polymerization strategy for the preparation of ionic covalent organic framework membranes (iCOFMs) with the highest ion exchange capacity ever reported. The iCOFMs exhibited higher crystallinity compared with the same type of COF powder. In addition, the membrane showed the highest conductivity, which was superior to other reported COF materials. However, some results are still missing in the manuscript to fully support the conclusions. I recommend to reconsider the manuscript after authors addressing the following points: 1. The defect-free iCOFMs exhibited an asymmetrical structure. The iCOFMs toward organic phase side showed a smooth and defect-free membrane surface while the iCOFMs toward water phase side showed loose membrane structure. Please discuss what causes this difference. Reply: Thanks for the reviewer's valuable guidance. When the interfacial polymerization started, the activated amine monomers diffused into the organic phase and reacted with the activated aldehyde monomers at the water-organic interface. After dual-activation, the higher reaction rate could generate COF particles instead of continuous COF membranes (Sci. Adv. 2020, 6, eabb1110). As the reaction time prolonged, the COF particles will stack into a loose layer, which reduced the further diffusion rate of amine monomers into the organic phase. This is similar to the self-sealing and self-terminating behavior in the interfacial polymerization process for polyamide membranes Therefore, the resultant iCOFMs exhibited an asymmetrical structure with a smooth and defect-free membrane surface toward organic phase side and a loose membrane structure toward water phase side.
The relevant discussion was added in the revised manuscript. "…When the interfacial polymerization started, the activated BD-(SO3H)2 monomers diffused into the organic phase and reacted with the activated Tp monomers at the water-organic interface. After dual-activation, the higher reaction rate could generate COF particles instead of continuous COF membranes 31 . As the reaction time prolonged, the COF particles will stack into a loose layer, which reduced the further diffusion rate of amine monomers into the organic phase. Accordingly, the interfacial polymerization was slowed down, leading to the formation of a compacted layer on top side. After 24 hours, defect-free iCOFMs finally formed, which exhibited an asymmetrical structure ( Figure   S1, Supplementary Information). The SEM and AFM images of the iCOFMs toward organic phase side showed a smooth and defect-free membrane surface with a roughness of 8.20 nm (Figure 1b)…" 2. The experimental PXRD pattern was highly consistent with the simulated reversed AA stacking PXRD patterns with the existence of the sharp diffraction peaks corresponding to the 100 planes.
Where are the sharp peaks located, and which crystal planes do the other peaks correspond to? Reply: Thanks for the reviewer's valuable guidance. The crystal plane corresponding to the diffraction peak has been inserted in Figure 2d in the revised manuscript, as show below. The diffraction peak at 3.3 o , 5.7 o and 26 were corresponding to 100, 110 and 002 crystal plane, respectively. Reply: Thanks for the reviewer's valuable guidance. A HTEM image with higher resolution has been presented in Figure 1f in the revised manuscript. We can see that the 002 crystal plane width was 0.34 nm, consistent with the crystal plane obtained by simulation. 4. For most state-of-the-art PEMs, increasing IEC value will impart higher water uptake, which in turn causes higher swelling ratio, originating from the flexible polymer networks nature (Adv. Mater. 2020, 2005565). But the rigid structure of iCOFMs is likely to reduce the swelling ratio. It is recommended to provide water uptake and swelling properties of the prepared membranes. Reply: Thanks for the reviewer's valuable guidance. We further tested the water uptake and swelling properties of the resulting iCOFMs ( Figure S18, Table S7). The water uptake of TpBD-(SO3H)2 membrane was 144 ± 5%, higher than most state-of-the-art PEMs. This was due to the high IEC value as well as the high specific surface area of the TpBD-(SO3H)2 membrane. Moreover, the TpBD-(SO3H)2 membrane exhibited the area and thickness swelling ratio of 18.3 ± 2.9% and 20.0 ± 2.4%, respectively, lower than most state-of-the-art PEMs. This was because the rigid framework structure of the TpBD-(SO3H)2 membrane was likely to considerably restrain the swelling.  Reply: Thanks for the reviewer's valuable guidance. We re-tested the iCOFM samples involved in Figure   3b, e for three times, and found only one spike in the re-tested XRD pattern. The peak at 2.6 degrees in the previous test may be caused by sample contamination. The XRD patterns in Figure 3b, e were updated in the revised manuscript as shown below. (e) Effect of the amine monomer activator addition amount on crystallinity of iCOFMs.
6. The iCOFMs had the high structural stability. In order to investigate the stability of the material under high humidity and high temperature conditions, it is necessary to run the membrane at 90 °C and 100% RH for a continuous period of time to prove the long-term thermal durability for further fuel cell applications.

Reply:
Following the reviewer's valuable guidance, we have carried out the long-term operation test of membrane proton conductivity at 90 °C and 100% RH. The results were added in Figure S27 in the revised manuscript, as shown below. During the test, and the proton conductivity of the TpBD-(SO3H)2 membrane only slightly decreased by about 7.6% within the first 5 hours, which was probably due to the instability of the amorphous regions of the membrane under high temperature and high humidity, and then remained stable within two weeks, verifying the membrane's excellent long-term thermal durability for further fuel cell applications.
The relevant discussion was added in the revised manuscript. "…Moreover, we have carried out the long-term operation test of membrane proton conductivity at 90 °C and 100% RH ( Figure S27).
During the test, and the proton conductivity of the TpBD-(SO3H)2 iCOFMs only slightly decreased by about 7.6% within the first 5 hours, which was probably due to the instability of the amorphous regions of the membrane under high temperature and high humidity, and then remained stable within two weeks, verifying the membrane's excellent long-term thermal durability for further fuel cell Reply: Thanks for the reviewer's valuable guidance. We have checked the written errors and unified the abbreviation for 2,4,6-triformylphloroglucinol as Tp in the revised manuscript.

Reviewer #3 (Remarks to the Author):
This work reports the synthesis of iCOF membranes by a dual-activation strategy and demonstrates high proton conductivity of membranes. The COF membrane formation mechanism is profoundly studied and discussed. However, the ion exchange capacity tests and discussions are not profound enough. The reported results are also similar to their previous works despite the high proton conductivity. Overall, the work may not satisfy the high quality of Nature Communication.
The detailed comments are listed below.
1. The increase of n-octanoic acid amount results in raised membrane thicknesses, leading to decreased proton conductivity. From Figure 3b, it can be seen that the membrane crystallinity is strongly influenced by the amount of n-octanoic acid. Thus, the crystallinity of these membranes may affect the proton conductivity as well.

Reply:
Thanks for the reviewer's valuable guidance. We have added some discussion about the effect of crystallinity on the proton conductivity of COF membranes in the Results and Discussion section.
"…As the n-octanoic acid addition amount increased from 5 to 15 mL, the 100 characteristic peak intensity was enhanced ( Figure 3b) and the proton conductivity of the membrane was increased from 0.33 S cm -1 to 0.66 S cm -1 . The enhanced proton conductivity was attributed to the ordered and continuous ionic channels formed in high crystallinity iCOFMs, which can effectively facilitate the proton transport 3 .…" 2. The synthesized TpPa-SO3H membranes show lower proton conductivity than TpBD-(SO3H)2.
Please explain this result.
The relevant discussion was added in the revised manuscript. "…The proton conductivity of the resultant TpPa-SO3H iCOFMs reached 0.2 S cm -1 at 90 o C, 100% RH ( Figure S26, Supplementary Information), lower than that of the TpBD-(SO3H)2 iCOFMs. This was because TpBD-(SO3H)2 membranes have a larger ion-exchange content (IEC) but a similar swelling ratio to TpPa-SO3H membranes (Table S7, Supplementary Information), implying a higher density of monodispersed sulfonic acid groups in the TpBD-(SO3H)2 membranes. Therefore, protons have a shorter hop distance via surface transport mechanism 44,45 . Moreover, owing to the higher IEC, the TpBD-(SO3H)2 membranes could absorb more water molecules ( Figure S7 and S25, Supplementary   Information), which play a critical role in proton transport based on Grotthuss and vehicle mechanisms 46 …"   Thanks for the reviewer's valuable guidance. To explore the selection criteria of activators, we further selected a variety of amine monomer activators (sodium hydroxide, sodium acetate and sodium benzoate) and aldehyde monomer activators (acetic acid, n-heptanoic acid, n-nonanoic acid and n-decanoic acid) for iCOFMs fabrication.
As shown in Figure S12, we cannot obtain membranes when using a stronger acid (acetic acid) as the aldehyde monomer activator. This was because the acetic acid may diffuse into water and the base cannot sufficiently activate the amine monomer in the water phase. Moreover, we cannot obtain membranes if the amine monomer activator was a strong base (sodium hydroxide), since the strong base would immediately neutralize the acid at the interface.
When using other amine monomer activators and aldehyde monomer activators, a series of iCOFMs can be obtained with different thickness and crystallinity. We found that the membrane thickness increased with the increase of the acidity of the aldehyde monomer activator ( Figure S13). This phenomenon can be explained by the membrane growth mechanism. As the acidity of the aldehyde monomer activator increased, aldehyde monomers were more easily to be activated and the initial interface polymerization rate became higher, leading to increased thickness of the loose layer of iCOFMs. Similarly, as the basicity of the amine monomer activator increased, the membrane thickness increased. In addition, the crystallinity of the iCOFMs can also be manipulated by varying the activators ( Figure S14). Therefore, we can obtain the selection criteria of the activators. The aldehyde monomer activator needs to be immiscible with water to form a stable interface and avoid diffusing into water phase.
The basicity of the amine monomer activator needs to be strong enough to extract protons from the ionic amine monomer, but should not be too strong to avoid neutralization with the acid in the organic phase. The addition amount as well as the acidity (basicity) of the activator can be adjusted according to the monomer reactivity as well as required membrane thickness and crystallinity.
The relevant discussion was added in the revised manuscript and Supplementary Information. "…According to the proposed dual activation interfacial polymerization mechanism, other similar Brønsted acids and bases also could be used for monomer activation. To explore the selection criteria of activators, we further selected a variety of amine monomer activators (sodium acetate and sodium benzoate) and aldehyde monomer activators (n-heptanoic acid, n-nonanoic acid and ndecanoic acid) for iCOFMs fabrication. As shown in Figure S13-14 (Supplementary Information), a series of iCOFMs can be obtained with different thickness and crystallinity. We found that the membrane thickness increased with the increase of the acidity of the aldehyde monomer activator.
This phenomenon can be explained by the membrane growth mechanism. As the acidity of the aldehyde monomer activator increased, aldehyde monomers were more easily to be activated and the initial interface polymerization rate became higher, leading to increased thickness of the loose layer of iCOFMs. Similarly, as the basicity of the amine monomer activator increased, the membrane thickness increased. In addition, the iCOFMs obtained by coupling sodium formate and n-octanoic acid exhibited highest crystallinity ( Figure S14 Supplementary Information). Therefore, sodium formate and n-octanoic acid were selected as model activators for further study…" "Discussion S3 As shown in Figure S12, we cannot obtain membranes when using a stronger acid (acetic acid) as the aldehyde monomer activator. This was because the acetic acid may diffuse into water and the base cannot sufficiently activate the amine monomer in the water phase. Moreover, we cannot obtain membranes if the amine monomer activator was a strong base (sodium hydroxide), since the strong base would immediately neutralize the acid at the interface. Therefore, we can obtain the selection criteria of the activators. The aldehyde monomer activator needs to be immiscible with water to form a stable interface and avoid diffusing into water phase. The basicity of the amine monomer activator needs to be strong enough to extract protons from the ionic amine monomer, but should not be too strong to avoid neutralization with the acid in the organic phase. The addition amount as well as the acidity (basicity) of the activator can be adjusted according to the monomer reactivity as well as required membrane thickness and crystallinity."   Thanks for the reviewer's valuable guidance. To analyze the reason for about 15% weight loss in 40-150 °C, we further carried out a vacuum water vapor sorption test on TpBD-(SO3H)2 and TpPa-SO3H membrane. It was found that at 30 °C and 50% relative humidity, the water vapor adsorption capacity was 229-247 mg g -1 ( Figure S7 and S25). The high water adsorption was due to the high specific area and abundant hydrophilic -SO3H groups within the membrane pores. As the membrane sample was kept under the conditions of 25 °C and 53 RH% and no special dehydration treatment was carried out before the TGA test, we derive that about 15% of the weight loss at 40-150 °C on the TGA test could be attributed to the evaporation of adsorbed water. 5. It is interesting to find that thus-prepared iCOF membranes hold higher crystallinity than their bulk powders synthesized under solvothermal conditions. It will be valuable to guide the synthesis of highly crystalline COF membranes if the underlying mechanism is discussed.  ., 2019, 141,   1807). On one hand, the solvothermal reaction system for COF powder synthesis was homogeneous.
We can only activate one type of monomer by adding either aldehyde monomer activator (acid) or amine monomer activator (base). On the other hand, during the homogeneous solvothermal synthesis for COF powder, the two monomers directly contact and polymerize rapidly at high temperatures. During the interfacial polymerization process for the fabrication of iCOFMs, the growth rate was limited by the diffusion rate of the monomers across the interface, which can maintain efficient self-correction for higher crystallinity.
The relevant discussion was added in the revised Supplementary Information. "Discussion S1 Previous studies have proved that suitable reaction rate and reversibility and low growth rate are reliable conditions for the synthesis of high crystallinity COFs 1-3 . On one hand, the solvothermal reaction system for COF powder synthesis was homogeneous. We can only activate one type of monomer by adding either aldehyde monomer activator (acid) or amine monomer activator (base).
On the other hand, during the homogeneous solvothermal synthesis for COF powder, the two monomers directly contact and polymerize rapidly at high temperatures. During the interfacial polymerization process for the fabrication of iCOFMs, the growth rate was limited by the diffusion rate of the monomers across the interface, which can maintain efficient self-correction for higher crystallinity." 6. Given the asymmetric structure of membranes, i.e. top dense films and loosely stacked particles, the membrane may contain a crystallinity asymmetry. The crystallinity of the film and particles should be presented separately.

Reply:
Thanks for the reviewer's valuable guidance. We used grazing incidence XRD (GIXRD) and GIWAXS to characterize the crystallinity of the top dense side and loosely stacked side of the COF membranes ( Figure S5-6). It was found that the crystallinity of the top dense side of the membrane was slightly higher than that of the loosely stacked side. This is because the crystallinity of COFs was significantly affected by the growth rate (Science, 2018, 361, 48). In this study, when the interfacial polymerization started, the two-phase monomers reacted rapidly, and the faster growth rate could result in COF particles with relatively low crystallinity (Sci. Adv. 2020, 6, eabb1110). As the reaction time prolonged, the COF particles will stack into a loose layer, which reduced the further diffusion rate of amine monomers into the organic phase. This is similar to the self-sealing and self-terminating behavior in the interfacial polymerization process for polyamide membranes (J. Mater. Chem. A, 2019, 7, 25641; J. Mater. A, 2020, 8, 23930). Accordingly, the interfacial polymerization was slowed down, leading to the formation of a compact layer on top side. The slower growth rate can maintain efficient selfcorrection of defects, leading to higher crystallinity of the top dense layer (J. Am. Chem. Soc. 2018,   140, 5145, Sci. Adv. 2020, 6, eabb1110).

Chem.
The relevant discussion was added in the revised Manuscript and Supplementary Information. "…Meanwhile, we conducted the grazing incidence wide-angle X-ray scattering (GIWAXS) and grazing incidence XRD (GIXRD) measurement to analyze the crystallinity of the membrane surface.
It was found that the crystallinity of the top dense side of the membrane was slightly higher than that of the loosely stacked side (Figure 1e and Figure S5-6, Supplementary Information)…" "Discussion S2 It was found that the crystallinity of the top dense side of the membrane was slightly higher than that of the loosely stacked side. This is because the crystallinity of COFs was significantly affected by the growth rate 1 . In this study, when the interfacial polymerization started, the two-phase monomers reacted rapidly, and the faster growth rate could result in COF particles with relatively low crystallinity 4 . As the reaction time prolonged, the COF particles will stack into a loose layer, which reduced the further diffusion rate of amine monomers into the organic phase. This is similar to the self-sealing and self-terminating behavior in the interfacial polymerization process for polyamide membranes 5,6 . Accordingly, the interfacial polymerization was slowed down, leading to the formation of a compact layer on top side. The slower growth rate can maintain efficient selfcorrection of defects, leading to higher crystallinity of the top dense layer 4,7 ."  Reply: Thanks for the reviewer's valuable guidance. We have carefully checked and corrected the grammatical errors in the revised manuscript. Some typical corrections are shown below.
"Benefited from the simplicity and scalability, interfacial polymerization (IP) has evolved as a platform technology for COF membrane fabrication by confining the polymerization reactions between monomers in two immiscible phases at the interface 19 , during which the monomer reactivity directly governs the membrane structure formation 24,31 ." was revised as "Benefited from the simplicity and scalability, interfacial polymerization (IP) has evolved as a platform technology for COF membrane fabrication by confining the polymerization reactions between monomers in two immiscible phases at the interface 19 . During IP process, the monomer reactivity directly governs the membrane structure formation 24,31 ." "However, the IP technology for fabricating nonioinic COF membranes, can not directly transplant into iCOFMs fabrication." was revised as "However, the IP technology for fabricating non-ionic COF membranes can hardly be transplanted into iCOFMs fabrication directly." "…indicating that the reactivity of amine monomer was decreased as the number of ionic groups increases." was revised as "indicating that the reactivity of amine monomer was decreased as the number of ionic groups increased." "…and COF nanosheets formed in the water phase (Figure 2b)." was revised as "…and COF nanosheets were formed in the water phase (Figure 2b)." "Under the dual-activation condition (Figure 2d), an iCOFMs formed with a thickness of 85 μm was formed,…" was revised as "Under the dual-activation condition (Figure 2d), an iCOFMs with a thickness of 85 μm was formed,…" "…the initial interface polymerization rate became faster, lead to the formation of more noncontinuous COF particles…" was revised as "…the initial interface polymerization rate became higher, leading to the formation of more non-continuous COF particles…" 10. The iCOF membranes give higher proton conductivities than others and commercial materials.
How about the stability of proton conductivity? Reply: Following the reviewer's valuable guidance, we have carried out the long-term operation test of membrane proton conductivity at 90 °C and 100% RH. The results were added in Figure S27 in the revised manuscript, as shown below. During the test, and the proton conductivity of the TpBD-(SO3H)2 membrane only slightly decreased by about 7.6% within the first 5 hours, which was probably due to the instability of the amorphous regions of the membrane under high temperature and high humidity, and then remained stable until two weeks, verifying the membrane's excellent long-term thermal durability for further fuel cell applications.
The relevant discussion was added in the revised manuscript. "…Moreover, we have carried out the long-term operation test of membrane proton conductivity at 90 °C and 100% RH ( Figure S27).
During the test, and the proton conductivity of the TpBD-(SO3H)2 membrane only slightly decreased by about 7.6% within the first 5 hours, which was probably due to the instability of the amorphous regions of the membrane under high temperature and high humidity, and then remained stable until two weeks, verifying the membrane's excellent long-term thermal durability for further fuel cell applications…" 5. The authors have re-tested the iCOFMs for three times, and found only one spike in the re-tested XRD pattern. The peak at 2.6 degree in the previous test may be caused by sample contamination.
But there are four samples in the original manuscript that all show double peaks, all of which were all caused by sample contamination? Reply: Thanks for the reviewer's valuable guidance. We re-prepared four kinds of iCOFMs in Figure 3b and 3e. The digital photos were shown in Figure R1. Three copies of each kind of membrane were prepared and performed for XRD tests. As shown in Figure R2, all four kinds of membranes show only one spike at 3.4 degree and no peak at 2.6 degree. Therefore, we can guarantee the validity and reproducibility of the XRD data in Figure 3b and 3e of the revised manuscript.
In addition, we tried to find the reason of the appearance of the peaks at 2.6 degree in the original manuscript. As all samples were rinsed with DMF and ethyl alcohol, and were carefully kept before tests, we suspect that the peaks at 2.6 degree shown in Figure 3b and 3e in the original manuscript may be caused by contamination during the XRD test process. As shown in Figure R3, during XRD test, the membrane sample was directly placed on a sample stage, which is a silicon wafer specially made by Rigaku Corporation and has no diffraction peak in the range of 2-120 degrees. Under normal wide-angle XRD test conditions, the X-ray detection depth could exceed 100 µm for lowdensity organic and porous materials (J. Pharm. Sci. 2010, 99, 3807-3814). If some contaminants existed on the sample stage, characteristic peaks of the contaminants could appear in the XRD pattern.
Moreover, all four samples, which exhibited diffraction peaks at 2.6 degree in the original manuscript, were tested sequentially using the same sample stage at one time. Therefore, we suspect that the sample stage we used may have been contaminated, probably due to the residuals from previous tests. Figure R1. Digital photo of the four kinds of TpBD-(SO3H)2 iCOFMs (Fabrication condition: a. 20 mL n-octanoic acid for aldehyde monomer activation and 2.0 eq sodium formate for amine monomer activation; b. 15 mL n-octanoic acid for aldehyde monomer activation and 2.5 eq sodium formate for amine monomer activation; c. 15 mL n-octanoic acid for aldehyde monomer activation and 2.0 eq sodium formate for amine monomer activation; d. 15 mL n-octanoic acid for aldehyde monomer activation and 1.5 eq sodium formate for amine monomer activation). 20 mL n-octanoic acid for aldehyde monomer activation and 2.0 eq sodium formate for amine monomer activation; b. 15 mL n-octanoic acid for aldehyde monomer activation and 2.5 eq sodium formate for amine monomer activation; c. 15 mL n-octanoic acid for aldehyde monomer activation and 2.0 eq sodium formate for amine monomer activation; d. 15 mL n-octanoic acid for aldehyde monomer activation and 1.5 eq sodium formate for amine monomer activation; Three copies of each kind of membrane were prepared and performed for XRD tests). Figure R3. XRD sample stage before and after putting membrane sample.
6. The authors have carried out the long-term operation test of membrane proton conductivity at 90 °C and 100% RH. The proton conductivity of the iCOFMs decreased by about 7.6% within the first 5 hours, and then remained stable within two weeks. Authors should discuss these two values separately in the manuscript, while not only mention the highest value.

Reply:
Following the reviewer's valuable guidance, we have added some discussions about long-term operation proton conductivity values in the Results and discussion section as well as in the Conclusion section of the revised manuscript.
"…During the test, and the proton conductivity of the TpBD-(SO3H)2 iCOFMs only slightly decreased by about 6% to 0.62 S cm -1 within the first 6 hours, which was probably due to the instability of the amorphous regions of the membrane under high temperature and high humidity.
The proton conductivity of TpBD-(SO3H)2 iCOFMs remained stable around 0.62 S cm -1 within two weeks, verifying the membrane's excellent long-term thermal durability for further fuel cell applications..." "…outperforming all the currently reported COF-based proton-conducting materials. Moreover, TpBD-(SO3H)2 iCOFMs also show excellent proton conductivity around 0.62 S cm -1 during the two-week long-term proton conductivity test (90 o C, 100% relative humidity). We anticipate that our membranes can find diverse applications such as fuel cell, flow cell, lithium battery and nanofiltration..." 7. Please check carefully the writing of Tp, e.g. in Figure 2e.

Reply:
Thanks for the reviewer's valuable guidance. We have checked the written errors and unified the abbreviation for 2,4,6-triformylphloroglucinol as Tp in the revised manuscript, as shown below.  "…The amine monomer solution was poured into the bottom of 100 mL beaker, and the Tp solution was added by drops on the top layer …" "…Typically, 0.2 mmol of 2,4,6-triformylphloroglucinol (Tp), 0.3 mmol (56.4 mg) of 2,5diaminobenzenesulfonic acid (Pa-SO3H), 3 mL mesitylene, 1 mL dioxane and acetic acid (6 mol L -