Journal of Membrane Science 604 (2020) 118087

Available online 25 March 20200376-7388/© 2020 Elsevier B.V. All rights reserved.

Influence of membrane hydrophilicity on water permeability: An experimental study bridging simulations

Aoze Miao, Mingjie Wei **, Fang Xu, Yong Wang *

State Key Laboratory of Materials-Oriented Chemical Engineering, and College of Chemical Engineering, Nanjing Tech University, Nanjing, 211816, Jiangsu, PR China

A R T I C L E I N F O

Keywords: Membrane Hydrophilicity Atomic layer deposition (ALD) Permeance Wetting states

A B S T R A C T

There has been an argument on the influence of membrane hydrophilicity on permeability between experimental and simulation works. Theoretical work finds that the membranes have a characteristic called a threshold pressure drop (ΔPT). Strongly hydrophobic membranes exhibit extremely high ΔPT, making membranes less wettable and consequently less permeable. In order to experimentally investigate the influence of wetting states of membranes, determined by membrane hydrophilicity, on permeability, we prepared a series of membranes with progressively changing hydrophilicities by atomic layer deposition (ALD) of TiO2 on polytetrafluoro-ethylene membranes while keeping the pore size unchanged. In addition, adding ethanol into the water and raising the operating pressure drops are also selected to improve the wetting degrees. The experimental results verify the presence of ΔPT for all membranes and the hydrophobic membranes having higher ΔPT. Since the hydrophilic membranes have better wetting degrees but higher transport resistance, there should be an opti-mized hydrophilicity, which is found to be determined by the feed solution and the operating ΔP. The optimized hydrophilic modification is the one when the membrane can be just adequately wetted by the feed solution at the target operating pressure drops. Highest water permeability is expected to be obtained at a moderate hydro-philicity, neither at the extreme hydrophilicity nor at the extreme hydrophobicity.

1. Introduction

Membrane separation plays a vital role in providing drinking water for daily life and industrial applications [1]. Since the efficiency of the membrane process is dictated by the permeability of membrane mate-rials, how to improve the permeability of membrane materials has drawn extensive attentions. A large number of studies have shown that membrane flux is closely related to the hydrophilicity of the membrane [2–4]. Generally, experimental studies have suggested that stronger hydrophilicity results in higher water flux [5–7]. However, molecular dynamics (MD) simulations show that hydrophobic pores are more conducive to the flux improvement, which is contrary to experimental results [8–12].

Very recently, we reveal the root cause for this controversy [13]. Hydrophobic membranes can be easily wetted in MD simulations, but are hardly wetted in experimental conditions because of the extreme difference in operating pressure drops (ΔPs). A hydrophobic membrane exhibits a high threshold pressure drops (ΔPT). Only when the operating ΔP exceeds ΔPT, the membrane can be wetted and thus produce water

permeance. The ΔPT of a hydrophobic membrane is often in the scale of several hundred megapascals, which cannot be achieved in experi-mental conditions. Typically, organic membranes, used in water treat-ment, have a certain degree of hydrophobicity [14–16] as a fully hydrophilic membrane material has poor stability in an aqueous envi-ronment and cannot achieve the purpose of long-term use. Therefore, wetting of organic membranes has become a ubiquitous issue in the applications of water treatment [17–19]. That is the reason why a huge number of studies are focused on the hydrophilic modification of organic membranes [20–22].

In order to verify these findings, we plan to select a series of mem-branes with progressively changing hydrophilic properties and study their permeability under different ΔPs. However, it is extremely chal-lenging to design a series of membranes that differ only in hydrophilicity while maintaining other parameters such as pore size, porosity, and membrane thickness unchanged. Usual modification methods include additive blending [23], adsorbed coatings [24], chemically induced grafting [25], etc. Although these methods have improved the mem-brane hydrophilicity to a certain degree, they cannot keep unchanged

* Corresponding author. ** Corresponding author.

E-mail addresses: mj.wei@njtech.edu.cn (M. Wei), yongwang@njtech.edu.cn (Y. Wang).

Contents lists available at ScienceDirect

Journal of Membrane Science

journal homepage: http://www.elsevier.com/locate/memsci

https://doi.org/10.1016/j.memsci.2020.118087 Received 21 February 2020; Received in revised form 16 March 2020; Accepted 22 March 2020

Journal of Membrane Science 604 (2020) 118087

2

other parameters of membranes that may influence the permeability. Fortunately, the development of atomic layer deposition (ALD) on

modifying membranes provides a very possible way to satisfy the above requirement. ALD is a technique in which gas phase precursors are alternately pulsed into a reactor and chemically adsorbed and then react on a substrate to form a conformal deposition layer [26–28]. As ALD has the characteristics of low reaction temperature [29,30], precise and controllable deposition thickness [31,32], good shape retention and uniformity of the substrate [33], it has recently been applied in the modification and functionalization of membrane [34–36] and remark-ably enhanced surface hydrophilicity has been achieved on a number of initially hydrophobic membranes by ALD of metal oxides [37,38]. Therefore, we expect that ALD is a desirable way to realize the goal of tuning the membrane hydrophilicity without changing other membrane characteristics, which is crucial for this work to separately verify the influence of hydrophilicity on membrane permeance. By adjusting the number of ALD cycles, the membrane surface and pores can be covered with hydrophilic metal oxides to different degrees, thereby achieving the purpose of progressively adjusting the membrane hydrophilicity.

In addition, we select another strategy of changing the wetting sates of membranes, that is, adjusting the composition of the liquid to be penetrate through the membrane [39,40]. For a hydrophobic mem-brane, it is difficult to be wetted by pure water. However, adding a polar solvent such as ethanol into water will greatly improve the wetting. Therefore, proper adjusting the composition of the liquid can investigate the influence of solution wettability on the permeability in a more convenient way.

In this work, we use ALD to deposit TiO2 on the highly hydrophobic polytetrafluoroethylene (PTFE) membrane. As the number of deposition cycles increases, the hydrophilicity of the membrane gradually in-creases. The dependence of wetting degrees on the membrane hydro-philicity are investigated by measuring the contact angles. Moreover, we choose another strategy of tuning the wetting degrees of membranes, that is adding a polar solvent into water, which is ethanol in this work. By adjusting the ethanol concentration (EC) of ethanol/water solution, the wetting degrees of membranes will be tuned as expected. The interference of these two strategies of adjusting membrane wetting de-grees is also investigated in details. Moreover, theoretical work [13] indicates that raising ΔPs will also improve the wetting degree of membranes. The dependence of wetting degrees on the operating ΔPs is focused on as well. The findings in this work verify the presence of ΔPT in hydrophobic membrane as predicted by previous theoretical work [13], and reveal that there is an optimized point for these three methods of adjusting wetting degrees, at which the highest permeance is expected to be observed.

2. Experimental section

2.1. Materials

PTFE microfiltration membranes with the pore size of 220 nm were purchased from Sartorius Stedim Biotech Co. Ltd. The diameter of membrane plates is 25 mm. Titanium tetrachloride (TiCl4, 99.99%) was obtained from Nanjing University. Deionized water (conductivity: 8–20 μS cm-1) was purchased from Wahaha. TiCl4 and deionized water were used as precursors in the TiO2 ALD process. Ethanol (99.98%) was purchased from Aladdin Chemicals and mixed with deionized water to obtain ethanol/water solutions in different proportions. All other

solvents were purchased from local suppliers and used as is.

2.2. ALD on PTFE

The PTFE membrane was deposited by TiO2. The ALD reaction was preheated to 120 �C, then the membrane was put into the reaction chamber as a substrate, and vacuum (~133 Pa) was evacuated before deposition for 1 h. The purge gas uses nitrogen (99.99%) and the carrier gas uses ultra-high purity nitrogen (99.999%). The carrier gas was used to pulse the precursor into the reaction chamber. TiCl4 and deionized water were alternately pulsed into the reaction chamber, and up to 300 deposition cycles were used in the ALD process.

Due to the hydrophobic characteristic of PTFE, the fiber surface of PTFE membranes is lack of reactive groups for ALD reaction. With reference to our previous research work [41], we have determined the conditions for the deposition of TiO2 ALD, as shown in Table 1.

2.3. Characterization

The membrane surface morphologies were characterized by a field emission scanning electron microscope (S-4800, Hitachi) working at 5 KV. A platinum layer is sputtered on the membrane to enhance the conductivity, and the sputtering time is controlled within 10–20 s. The contact angles of pure water ethanol/water solutions were measured by a contact angle tester (Dropmeter A-100, Maist) at ambient temperature to characterize the surface wettability of different samples. Each sample is measured at least 5 times at different locations and the average value is recorded. The thickness of TiO2 on the silicon wafer was measured by a spectroscopic ellipsometer (Complete EASE M � 2000U, J. A. Wool-lam) with the incidence angle of 70�.

2.4. Flux measurements

The flux of all membranes were measured using a dead-end stirred cell module (Amicon 8003, Millipore) at room temperature. The flux (L m-2 h-1) was calculated by the following equation:

Flux ¼V

S� t(1)

where V (L), S (m2), and t (h) are the volume of permeate, effective filtration area and test duration, respectively.

When measuring the flux of membranes by the ethanol/water solu-tions, we added a certain amount of deionized water to the collection- side bottles in advance. This method evidently reduced the EC in the ethanol/water solution in the collection-side bottles, thereby dramati-cally declined the volatilization rate of ethanol. Hence, the measured fluxes were reliable.

3. Results and discussion

3.1. ALD of TiO2 on PTFE membranes

Since the PTFE fibers are inert to the precursors of ALD reactions, the pulse time of precursors is increased as much as possible, so that enough precursors could be adsorbed on the surface of PTFE fibers. In addition, to make sure the reaction takes place predominantly inside membrane pores, we appropriately extend the exposure time to 10 s after several attempts. If the exposure time is too long, the reaction efficiency is reduced due to the self-restriction nature of atomic layer deposition, and some of the adsorbed precursors will be desorbed. If the exposure time is too short, the precursors cannot fully access the interior of membrane pores. An exposure time of 10 s is finally selected, so that TiO2 can be adequately deposited inside membrane pores.

To measure the deposition thickness of TiO2, 50, 100, 150, 200, and 250 cycles of depositions are performed on six single crystal silicon

Table 1 ALD operation parameters.

TiCl4 H2O

Pulse time (s) 0.015 0.03 Exposure time (s) 10 10 Purge time (s) 20 20

A. Miao et al.

Journal of Membrane Science 604 (2020) 118087

3

wafers, respectively. The deposition thickness is then measured by ellipsometry, and is shown in Fig. 1.

It is obvious that the growth rate of TiO2 is relatively stable. The deposition thickness increases linearly with the number of deposition cycles, and the growth rate is calculated to be ~ 0.64 Å cycle-1, which is close to the one in published work [41]. The linear increase also in-dicates that the growth of TiO2 is consistent with the self-limiting re-action between the two precursors. Furthermore, the variation of thickness at different places on the silicon wafer is smaller than 1 nm, which proves that the entire ALD growth is very uniform and the roughness is low. Such results indicate that ALD actually satisfies the requirement of keeping the membrane structure unchanged.

It is worth to mention that there are rich hydroxyl groups on silicon wafers, with which precursors have more opportunity to react. For the hydrophobic membranes, the growth rate is much smaller than that on the silicon wafer [41,42]. Hence, after 300 cycles of deposition, the thickness of TiO2 is much smaller than 21 nm. Compared with the pore diameter of 220 nm, the change in pore size caused by the deposition can

be ignored.

3.2. Surface morphology of PTFE membrane after TiO2 deposition

After depositing TiO2 on PTFE membranes, it is crucial to make sure the micro-structure of modified membranes is not significantly changed. As shown in Fig. 2, the morphologies of membranes deposited for 100, 200, and 300 cycles are similar to the pristine membrane. Each PTFE fiber is slightly thickened due to the deposition of TiO2. Thanks to the conformal deposition by ALD, the shape of PTFE fibers still maintains. Since the TiO2 trends to deposit at the defect location of PTFE fibers, large cycle number of TiO2 deposition will result in better coverage of TiO2 on the fibers. Hence, the large cycle number will lead to more hydrophilic membranes. As discussed above, the pore diameter is only slightly reduced after 300 depositions, hence the pore diameter has little effect on the overall permeability. In addition to the pore size, the roughness of the pore wall will also be influenced by ALD. As shown in Fig. 2, although the roughness is slighted increased when ALD cycle number rises, the deposition of TiO2 does not result in any observed large TiO2 particles due to the conformal deposition by ALD. Hence, we believe the change of roughness remains negligible when we measure the permeance of membranes.

In order to figure out the influence of pore size and roughness changes on permeability, we use an ethanol/water solution of 99.98% EC, which can sufficiently wet all membranes, to measure the per-meances of membranes with different deposition cycle numbers. As shown in Fig. 3, the permeance of the modified membranes change very little compared with the pristine membrane. This finding guarantees the important premise of subsequent experiments, that is, the membranes subjected to different hydrophilic modifications maintain the similar pore shape and structure. The coming measurement will be focused on the hydrophilicity only.

3.3. Contact angles of membranes with various hydrophilicities

As discussed in Introduction section, adding ethanol into water will promote the wetting degrees of a hydrophobic membrane. To verify this, we perform the contact angle measurement of pristine and modified

Fig. 1. The thickness of TiO2 layers deposited on silicon wafers as a function of various ALD cycles.

Fig. 2. The SEM images of the membrane surface with different ALD cycles: (a) pristine membrane; (b) 100 cycles; (c) 200 cycles; (d) 300 cycles. All images have the same magnification and the scale bar corresponding to 200 nm is given in (d).

A. Miao et al.

Journal of Membrane Science 604 (2020) 118087

4

membranes (shown in Fig. 4). It is evident that the contact angle drops gradually with the increase of EC (volume fraction), indicating that the solution is more likely to wet the membrane with the increase of EC. Furthermore, the pure water contact angle (EC being 0%) represents the hydrophilicity of membranes. It is obvious that the membrane with large deposition cycle number are more hydrophilic. For the case of a certain solution, the membrane contact angle gradually drops as the number of deposition cycles increases. indicating that more hydrophilic mem-branes tend to be easily wetted.

3.4. Influence of wetting states on the permeance

As shown Fig. 4a, the wetting degrees are quite different when the EC is low, but the wetting degrees are only slightly different when EC ¼99.98%. These findings indicate that membranes with various hydro-philicities will exhibit distinct wetting behaviors if the EC is different. Under the condition of keeping the ΔP constant (0.4 bar), the permeance of membranes with various hydrophilicities is firstly investigated as a function of EC.

It should be mentioned that the viscosity difference between the solutions is unneglectable (the specific viscosity value is in Table S1). Viscosity will have a certain effect on the fluxes of different concentra-tions of ethanol/water solutions, making their fluxes cannot be compared. According to the Hagen-Poiseuille equation:

Q¼π � r4 � ΔP

8� μ� l(2)

Qe¼Qμ (3)

where r (m), ΔP (bar), μ (Pa⋅s), l (m) are the pore radius, operating ΔP, liquid viscosity and channel length. There is an inverse relationship between the flux and the viscosity of the solution. Therefore, we can convert the flux of the solution into the comparable flux by excluding the viscosity. This method ensures that the only factor that influence the comparable flux in this work is the wetting degree. The flux, as well as the permeance based on it, are the effective ones, which is calculated through Eq. (3).

It is evident from Fig. 5 that there is no permeance for all membrane when EC is lower than 40%, because no membrane can be wetted by these ethanol/water solutions. When EC is 99.98%, both the pristine membrane and the modified membrane can be completely wetted by the solution. In this case the membrane hydrophobicity has little effect on the permeance, and the permeances of all membranes are almost the same. On the contrary, when the EC is below 40%, no membrane can be wetted and consequently no permeance is observed.

The permeance varies when the EC is between 40% and 80% as the wetting degrees are quite different for the membranes with various hydrophilicities. When the EC is 60%, the pristine membrane, which is the most hydrophobic one, cannot be wetted, thus no permeance is obtained. When the number of ALD cycles rises, indicating the increase of membrane hydrophilicity, the wetting degree is improved and thus

Fig. 3. Permeance of 99.98% EC through membranes subjected to ALD of TiO2 with different deposition times in contrast to the pristine membrane.

Fig. 4. Contact angle of membranes (a) with different deposition times as a function of EC; (b) Photographs of a pure water droplet on the surface of membranes without or with the ALD modification of various cycles. Each photograph is taken instantly when the water droplets contact the membrane.

Fig. 5. Permeances of membranes with various hydrophilicities as a function of ECs at 0.4 bar.

A. Miao et al.

Journal of Membrane Science 604 (2020) 118087

5

the hydrophilic membranes start to exhibit permeable. However, it is obvious that these permeances are significantly lower the permeance of 99.98% EC, which indicates that the wetting is not adequate. As pre-dicted in the theoretical work [13], such partially wetting degree actu-ally locates at the “running-stop” region. The “running-stop” implies a discontinuous water flow inside membranes. The running-stop transport is a metastable state with a wetting (running) and a nonwetting (stop) state repeating alternately rather than a stable continuous flow. The observed permeance is the time average of such discontinuous flow. If wetting state become dominant, which indicates the time of “running” much longer than the one of “stop”, the observed permeance is promoted.

For the cases of EC being 80%, the wetting degrees for all membranes are evidently improved. The state of “running” becomes dominant during the flux measuring. These findings could explain the observations of our previous ALD works [37,38]. The ALD of metal oxides on organic membranes will dramatically improve the permeability of organic membranes, because the hydrophilic modification turn membranes into more “running” and less “stop” and consequently improve the perme-ability of modified membranes.

Moreover, as theoretically predicted [13], the most hydrophobic membranes should have highest permeance if the wetting degree is high enough. However, for the cases of 99.98% EC, the advantage of hy-drophobic membrane cannot be emphasized because the interactions of ethanol with membranes of various hydrophilicities are all weak. The weak interaction originates from the ethanol molecule structure that carries both hydrophobic (methyl) and hydrophilic (hydroxyl) groups. The strong interaction of hydrophilic membrane with hydroxyl groups leads to the adsorption of ethanol molecules onto membrane pore wall, with methyl groups pointing out. Hence, the hydrophilic pore wall be-comes hydrophobic as they are uniformly covered by the methyl groups. Finally, the difference of hydrophilic or hydrophobic membranes become faint when they interact with ethanol molecules.

3.5. Dependence of membrane wettability on operating ΔPs

From the above analysis, it is evident that the wetting degree is dependent on the membrane hydrophilicity and the solution composi-tion. Theoretical work also demonstrates that the wetting degree can be also improved when the ΔP increases. To verify this finding, we subse-quently investigate the influence of ΔPs on the membrane wettability. We select the cases of 20% and 40% ECs and increase the ΔPs from 0.2 to 4 bar. As shown in Fig. 6, the starting ΔP (the minimum ΔP when the flux detected, denoted as ΔPm) of the flux measured by the hydrophilic membrane was lower than that of the hydrophobic membrane for both cases of ECs. It indicates that the hydrophobic membrane needs larger

ΔP to improve their wetting degree, so as to produce the permeance. It is worth to note that the starting ΔP is not ΔPT, but it is dependent

on ΔPT. According to the theoretical findings, higher ΔPT will lead to higher ΔPm. Since the adequate wetting occurs only when EC is 99.98%, the membranes are inadequately wetted for all rest cases. Hence, the cases of this section are still in the “running-stop” region. If the ΔP rises from 0.2 to 4 bar, the wetting degrees of membranes are continuously improved. However, the wetting degree of hydrophobic membranes is not adequate although ΔP reaches 4 bar. Hence, the observed fluxes for hydrophobic membranes are lower than those for hydrophilic mem-branes. As the membranes cannot endure higher operating ΔP, we did not raise it further.

By comparing the results of Fig. 6a and b, the starting ΔPs in the cases of 40% EC are obviously lower than those of 20% EC because the wetting state is better for the cases of 40% EC as discussed in the previous sec-tion. Table 2 lists ΔPm range of the membranes with various hydrophi-licities as well as the corresponding flux under these ΔPs.

In order to better exhibit the influence of ΔP, we replot the data of Fig. 6 by cases of various membrane hydrophilicities in detail (shown in Fig. 7). It is obvious that a higher EC has a lower starting ΔP, as dis-cussed above. When the ΔP further rises, the fluxes of 20% and 40% EC are getting close because the wetting degrees for 20% or 40% EC cases are both well improved at high ΔPs. It is interesting that the flux for 20% EC is equal to or even higher than those for 40% EC when the ΔP is larger than 3.5 bar. Since the ΔPm of 20% cases is higher, it is concluded that the 20% flux increases more rapidly.

Moreover, as 20% EC solution has higher contact angles than 40% EC solution does, the interaction of 20% EC solution with a certain mem-brane is weaker. The weaker interaction can be considered as “hydro-phobic” condition, thus the case of 20% EC can be considered as “hydrophobic” case compared to the 40% one. These findings are in well accordance with the theoretical prediction [13] that the hydrophobic membrane flux improves more significantly when the wetting degree increases. If ΔP is continuously increased (greater than 4 bar), the flux of the “hydrophobic membranes”, which are denoted for the case of 20% EC because of the weaker molecular interaction, is expected to surpass

Table 2 Starting pressure drop (ΔPm) measured at different ethanol concentrations (ECs).

ΔPm (bar)

ALD cycles 20% EC 40% EC 0 <2.00 <1.40 100 <1.60 <1.00 200 <1.00 <0.20 300 <0.60 <0.20

Fig. 6. Fluxes of membranes with various hydrophilicities as a function of ΔP: (a) 20% EC; (b) 40% EC.

A. Miao et al.

Journal of Membrane Science 604 (2020) 118087

6

that of the hydrophilic membrane as predicted in that theoretical work. By summarizing the influence of three methods of adjusting wetting

degrees, it is found that enhancing membrane hydrophilicity, increasing EC of ethanol/water solutions and raising ΔP will all promote the wet-ting degree of membrane. However, further enhancing membrane hy-drophilicity when membranes are all wetted will reduce the permeance due to the strong interaction between solution and membrane pore walls. Hence, there should be an optimized membrane hydrophilicity. When determining it, the other two parameters, liquid composition and ΔP should be taken into consideration. Since the liquid composition and the operating ΔP is determined by the feed solution and the limitation of experimental conditions, respectively, it is impossible to adjust the feed solution composition or to raise ΔP. Therefore, the optimized membrane hydrophilicity is expected to be the one that the hydrophilic modifica-tion just makes the membrane fully wetted at the given feed composition and operating ΔP, and no further hydrophilic modification is needed.

4. Conclusions

To investigate the influence of wetting state on membrane per-meance, we prepared a series of membranes with various hydrophilic-ities by applying ALD of TiO2 on a PTFE membrane with changing cycle numbers. In addition, ethanol/water solutions with ECs were selected to adjust wetting states of prepared membranes, so that the influence of wetting degrees on permeance will be investigated more sufficiently. First of all, we optimized the deposition conditions and obtained membranes with progressively increased hydrophilicity but similar pore sizes. The contact angles of prepared membranes are then measured by various ethanol/water solutions, which indicates better wetting degrees for membranes with stronger hydrophilicities. By comparing the cases of various EC, higher EC is proven to result in better wetting states for all membranes based on the measurement of contact angles. It is also found that the solution permeance will be observed for the membranes until

their wetting degrees are improved to a certain degree by increasing EC. Generally, hydrophobic membranes need higher EC to improve their wetting degrees. Moreover, membranes with various hydrophilicities are proven to have different ΔPT’s and hydrophobic membranes have higher ΔPT’s. The ΔPT for various membranes is also determined by the solution concentration, which is proven a crucial parameter for the wetting states. The experimental verification in this work demonstrates the importance of wetting states for membranes when they are involved with various liquid-related applications. The optimized hydrophilicity for membranes is found to be determined by the feed solution and operating ΔP, which will help to maximize the permeance of membranes.

Author statement

Aoze Miao: Investigation, Methodology, Experiment, Validation, Data curation, Writing - original draft, Writing - review & editing. Mingjie Wei: Conceptualization, Methodology, Writing - review & editing, Visualization, Funding acquisition. Fang Xu: Conceptualiza-tion, Formal analysis. Yong Wang: Writing - review & editing, Super-vision, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

Financial supports from the National Key Research and Development Program of China (2017YFC0403902), and the Jiangsu Natural Science Foundations (BK20190085), and the National Science Fund for

Fig. 7. Comparison of fluxes of 20% and 40% ECs: (a) 0 cycle; (b) 100 cycles; (c) 200 cycles; (d) 300 cycles.

A. Miao et al.

Journal of Membrane Science 604 (2020) 118087

7

Distinguished Young Scholars (21825803). We are also grateful to the Project of Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), and the Program of Excellent Innovation Teams of Jiangsu Higher Education Institutions.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.memsci.2020.118087.

References

[1] J.R. Werber, C.O. Osuji, M. Elimelech, Materials for next-generation desalination and water purification membranes, Nat. Rev. Mater. 1 (2016) 16018.

[2] J.K. Holt, H.G. Park, Y. Wang, M. Stadermann, A.B. Artyukhin, C.P. Grigoropoulos, A. Noy, O. Bakajin, Fast mass transport through sub-2-nanometer carbon nanotubes, Science 312 (2006) 1034–1037.

[3] M. Melillo, F. Zhu, M.A. Snyder, J. Mittal, Water transport through nanotubes with varying interaction strength between tube wall and water, J. Phys. Chem. Lett. 2 (2011) 2978–2983.

[4] N. Yan, Z. Wang, Y. Wang, Highly permeable membranes enabled by film formation of block copolymers on water surface, J. Membr. Sci. 568 (2018) 40–46.

[5] Q. Xu, Y. Yang, X. Wang, Z. Wang, W. Jin, J. Huang, Y. Wang, Atomic layer deposition of alumina on porous polytetrafluoroethylene membranes for enhanced hydrophilicity and separation performances, J. Membr. Sci. 415–416 (2012) 435–443.

[6] F. Yan, H. Chen, Y. Lü, Z. Lü, S. Yu, M. Liu, C. Gao, Improving the water permeability and antifouling property of thin-film composite polyamide nanofiltration membrane by modifying the active layer with triethanolamine, J. Membr. Sci. 513 (2016) 108–116.

[7] Y. Zhao, Z. Zhang, L. Dai, H. Mao, S. Zhang, Enhanced both water flux and salt rejection of reverse osmosis membrane through combining isophthaloyl dichloride with biphenyl tetraacyl chloride as organic phase monomer for seawater desalination, J. Membr. Sci. 522 (2017) 175–182.

[8] J. Goldsmith, C.C. Martens, Molecular dynamics simulation of salt rejection in model surface-modified nanopores, J. Phys. Chem. Lett. 1 (2010) 528–535.

[9] S. Joseph, N.R. Aluru, Why are carbon nanotubes fast transporters of water? Nano Lett. 8 (2008) 452–458.

[10] L. Wang, R.S. Dumont, J.M. Dickson, Nonequilibrium molecular dynamics simulation of pressure-driven water transport through modified CNT membranes, J. Phys. Chem. 138 (2013) 124701.

[11] T.D. Li, J. Gao, R. Szoszkiewicz, U. Landman, E. Riedo, Structured and viscous water in subnanometer gaps, Phys. Rev. B 75 (2007).

[12] F. Xu, Y. Song, M.J. Wei, Y. Wang, Water flow through interlayer channels of two- dimensional materials with various hydrophilicities, J. Phys. Chem. C 122 (2018) 15772–15779.

[13] F. Xu, M. Wei, X. Zhang, Y. Song, W. Zhou, Y. Wang, How pore hydrophilicity influences water permeability? Research 2019 (2019) 2581241.

[14] J. Zhao, L. Shi, C.H. Loh, R. Wang, Preparation of PVDF/PTFE hollow fiber membranes for direct contact membrane distillation via thermally induced phase separation method, Desalination 430 (2018) 86–97.

[15] S. Zhang, Q. Li, Y. Liu, C. Wu, W. Guo, Polytetrafluoroethylene (PTFE)/carbon black (CB) microporous membranes produced from PTFE/CB composite particles prepared by heterocoagulation process, High Perform. Polym. 29 (2017) 104–112.

[16] J. Pan, C. Xiao, Q. Huang, H. Liu, J. Hu, ECTFE porous membranes with conveniently controlled microstructures for vacuum membrane distillation, J. Mater. Chem. A 3 (2015) 23549–23559.

[17] A. Figoli, C. Ursino, F. Galiano, E. Di Nicol�o, P. Campanelli, M. Carnevale, A. Criscuoli, Innovative hydrophobic coating of perfluoropolyether (PFPE) on commercial hydrophilic membranes for DCMD application, J. Membr. Sci. 522 (2017) 192–201.

[18] L. Li, J.-W. Wang, H. Zhong, L.-Y. Hao, H. Abadikhah, X. Xu, C.-S. Chen, S. Agathopoulos, Novel α-Si3N4 planar nanowire superhydrophobic membrane prepared through in-situ nitridation of silicon for membrane distillation, J. Membr. Sci. 543 (2017) 98–105.

[19] K.-J. Lu, J. Zuo, T.-S. Chung, Tri-bore PVDF hollow fibers with a super- hydrophobic coating for membrane distillation, J. Membr. Sci. 514 (2016) 165–175.

[20] W. Sun, J. Liu, H. Chu, B. Dong, Pretreatment and membrane hydrophilic modification to reduce membrane fouling, Membranes 3 (2013) 226–241.

[21] B. Van der Bruggen, Chemical modification of polyethersulfone nanofiltration membranes: a review, J. Appl. Polym. Sci. 114 (2009) 630–642.

[22] S. Kalia, S. Boufi, A. Celli, S. Kango, Nanofibrillated cellulose: surface modification and potential applications, Colloid Polym. Sci. 292 (2014) 5–31.

[23] J. Hester, P. Banerjee, Y.-Y. Won, A. Akthakul, a.M. Acar, A. Mayes, ATRP of amphiphilic graft copolymers based on PVDF and their use as membrane additives, Macromolecules 35 (2002) 7652–7661.

[24] A. Fane, C. Fell, K. Kim, The effect of surfactant pretreatment on the ultrafiltration of proteins, Desalination 53 (1985) 37–55.

[25] E.M. Van Wagner, A.C. Sagle, M.M. Sharma, Y.-H. La, B.D. Freeman, Surface modification of commercial polyamide desalination membranes using poly (ethylene glycol) diglycidyl ether to enhance membrane fouling resistance, J. Membr. Sci. 367 (2011) 273–287.

[26] S. Wu, Z. Wang, S. Xiong, Y. Wang, Tailoring TiO2 membranes for nanofiltration and tight ultrafiltration by leveraging molecular layer deposition and crystallization, J. Membr. Sci. 578 (2019) 149–155.

[27] H. Van Bui, F. Grillo, J. Van Ommen, Atomic and molecular layer deposition: off the beaten track, Chem. Commun. (J. Chem. Soc. Sect. D) 53 (2017) 45–71.

[28] H.C. Guo, E. Ye, Z. Li, M.-Y. Han, X.J. Loh, Recent progress of atomic layer deposition on polymeric materials, Mater. Sci. Eng. C 70 (2017) 1182–1191.

[29] L. Guo, Z. Zhong, Y. Wang, Atomic layer deposition on block copolymer membranes with gyroidal nanopores toward periodically nanostructured vapor sensors: nanotubes versus nanorods, Adv. Mater. Interfaces 3 (2016) 1600017.

[30] S. Xiong, Z. Zhong, Y. Wang, Direct silanization of polyurethane foams for efficient selective absorption of oil from water, AIChE J. 63 (2017) 2232–2240.

[31] X. Jia, Z. Low, H. Chen, S. Xiong, Y. Wang, Atomic layer deposition of Al2O3 on porous polypropylene hollow fibers for enhanced membrane performances, Chin. J. Chem. Eng. 26 (2018) 695–700.

[32] H. Chen, S. Wu, X. Jia, S. Xiong, Y. Wang, Atomic layer deposition fabricating of ceramic nanofiltration membranes for efficient separation of dyes from water, AIChE J. 64 (2018) 2670–2678.

[33] H. Wang, M. Wei, Z. Zhong, Y. Wang, Atomic-layer-deposition-enabled thin-film composite membranes of polyimide supported on nanoporous anodized alumina, J. Membr. Sci. 535 (2017) 56–62.

[34] F. Li, Y. Yang, Y. Fan, W. Xing, Y. Wang, Modification of ceramic membranes for pore structure tailoring: the atomic layer deposition route, J. Membr. Sci. 397 (2012) 17–23.

[35] F. Li, X. Yao, Z. Wang, W. Xing, W. Jin, J. Huang, Y. Wang, Highly porous metal oxide networks of interconnected nanotubes by atomic layer deposition, Nano Lett. 12 (2012) 5033–5038.

[36] H.-C. Yang, R.Z. Waldman, Z. Chen, S.B. Darling, Atomic layer deposition for membrane interface engineering, Nanoscale 10 (2018) 20505–20513.

[37] S. Xiong, L. Kong, J. Huang, X. Chen, Y. Wang, Atomic-layer-deposition-enabled nonwoven membranes with hierarchical ZnO nanostructures for switchable water/ oil separations, J. Membr. Sci. 493 (2015) 478–485.

[38] H. Chen, Q. Lin, Q. Xu, Y. Yang, Z. Shao, Y. Wang, Plasma activation and atomic layer deposition of TiO2 on polypropylene membranes for improved performances of lithium-ion batteries, J. Membr. Sci. 458 (2014) 217–224.

[39] J. Feng, S. Xiong, Y. Wang, Atomic layer deposition of TiO2 on carbon-nanotube membranes for enhanced capacitive deionization, Separ. Purif. Technol. 213 (2019) 70–77.

[40] J. Feng, S. Xiong, Y. Wang, Atomic layer deposition of hybrid metal oxides on carbon nanotube membranes for photodegradation of dyes, Compos. Comm. 12 (2019) 39–46.

[41] Q. Xu, Y. Yang, J. Yang, X. Wang, Z. Wang, Y. Wang, Plasma activation of porous polytetrafluoroethylene membranes for superior hydrophilicity and separation performances via atomic layer deposition of TiO2, J. Membr. Sci. 443 (2013) 62–68.

[42] Q. Wang, X. Wang, Z. Wang, J. Huang, Y. Wang, PVDF membranes with simultaneously enhanced permeability and selectivity by breaking the tradeoff effect via atomic layer deposition of TiO2, J. Membr. Sci. 442 (2013) 57–64.

A. Miao et al.