Abstract
Controlled transport of water molecules through membranes and capillaries is important in areas as diverse as water purification and healthcare technologies1,2,3,4,5,6,7. Previous attempts to control water permeation through membranes (mainly polymeric ones) have concentrated on modulating the structure of the membrane and the physicochemical properties of its surface by varying the pH, temperature or ionic strength3,8. Electrical control over water transport is an attractive alternative; however, theory and simulations9,10,11,12,13,14 have often yielded conflicting results, from freezing of water molecules to melting of ice14,15,16 under an applied electric field. Here we report electrically controlled water permeation through micrometre-thick graphene oxide membranes17,18,19,20,21. Such membranes have previously been shown to exhibit ultrafast permeation of water17,22 and molecular sieving properties18,21, with the potential for industrial-scale production. To achieve electrical control over water permeation, we create conductive filaments in the graphene oxide membranes via controllable electrical breakdown. The electric field that concentrates around these current-carrying filaments ionizes water molecules inside graphene capillaries within the graphene oxide membranes, which impedes water transport. We thus demonstrate precise control of water permeation, from ultrafast permeation to complete blocking. Our work opens up an avenue for developing smart membrane technologies for artificial biological systems, tissue engineering and filtration.
This is a preview of subscription content, access via your institution
Access options
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$32.99 / 30 days
cancel any time
Subscribe to this journal
Receive 51 print issues and online access
$199.00 per year
only $3.90 per issue
Buy this article
- Purchase on SpringerLink
- Instant access to full article PDF
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
References
Karnik, R. et al. Electrostatic control of ions and molecules in nanofluidic transistors. Nano Lett. 5, 943–948 (2005).
Gravelle, S. et al. Optimizing water permeability through the hourglass shape of aquaporins. Proc. Natl Acad. Sci. USA 110, 16367–16372 (2013).
Liu, Z., Wang, W., Xie, R., Ju, X.-J. & Chu, L.-Y. Stimuli-responsive smart gating membranes. Chem. Soc. Rev. 45, 460–475 (2016).
Xiao, K. et al. Electrostatic-charge- and electric-field-induced smart gating for water transportation. ACS Nano 10, 9703–9709 (2016).
Wang, Z. et al. Polarity-dependent electrochemically controlled transport of water through carbon nanotube membranes. Nano Lett. 7, 697–702 (2007).
Hetherington, A. M. & Woodward, F. I. The role of stomata in sensing and driving environmental change. Nature 424, 901–908 (2003).
Borgnia, M. J., Nielsen, S., Engel, A. & Agre, P. Cellular and molecular biology of the aquaporin water channels. Annu. Rev. Biochem. 68, 425–458 (1999).
Zhao, C., Nie, S., Tang, M. & Sun, S. Polymeric pH-sensitive membranes—a review. Prog. Polym. Sci. 36, 1499–1520 (2011).
Kou, J. et al. Electromanipulating water flow in nanochannels. Angew. Chem. Int. Ed. 54, 2351–2355 (2015).
Li, J. et al. Electrostatic gating of a nanometer water channel. Proc. Natl Acad. Sci. USA 104, 3687–3692 (2007).
Gong, X. et al. A charge-driven molecular water pump. Nat. Nanotechnol. 2, 709–712 (2007).
Vaitheeswaran, S., Rasaiah, J. C. & Hummer, G. Electric field and temperature effects on water in the narrow nonpolar pores of carbon nanotubes. J. Chem. Phys. 121, 7955–7965 (2004).
Saitta, A. M., Saija, F. & Giaquinta, P. V. Ab initio molecular dynamics study of dissociation of water under an electric field. Phys. Rev. Lett. 108, 207801 (2012).
Qiu, H. & Guo, W. Electromelting of confined monolayer ice. Phys. Rev. Lett. 110, 195701 (2013).
Choi, E.-M., Yoon, Y.-H., Lee, S. & Kang, H. Freezing transition of interfacial water at room temperature under electric fields. Phys. Rev. Lett. 95, 085701 (2005).
Diallo, S. O., Mamontov, E., Nobuo, W., Inagaki, S. & Fukushima, Y. Enhanced translational diffusion of confined water under electric field. Phys. Rev. E 86, 021506 (2012).
Nair, R. R., Wu, H. A., Jayaram, P. N., Grigorieva, I. V. & Geim, A. K. Unimpeded permeation of water through helium-leak–tight graphene-based membranes. Science 335, 442–444 (2012).
Joshi, R. K. et al. Precise and ultrafast molecular sieving through graphene oxide membranes. Science 343, 752–754 (2014).
Sun, P., Wang, K. & Zhu, H. Recent developments in graphene-based membranes: structure, mass-transport mechanism and potential applications. Adv. Mater. 28, 2287–2310 (2016).
Liu, G., Jin, W. & Xu, N. Graphene-based membranes. Chem. Soc. Rev. 44, 5016–5030 (2015).
Abraham, J. et al. Tunable sieving of ions using graphene oxide membranes. Nat. Nanotechnol. 12, 546–550 (2017).
Radha, B. et al. Molecular transport through capillaries made with atomic-scale precision. Nature 538, 222–225 (2016).
Kao, K.-C. Dielectric Phenomena in Solids: With Emphasis on Physical Concepts of Electronic Processes Ch. 8 (Academic Press, Amsterdam, 2004).
Acik, M. et al. The role of oxygen during thermal reduction of graphene oxide studied by infrared absorption spectroscopy. J. Phys. Chem. C 115, 19761–19781 (2011).
Hontoria-Lucas, C., López-Peinado, A. J., López-González, J. D., Rojas-Cervantes, M. L. & Martín-Aranda, R. M. Study of oxygen-containing groups in a series of graphite oxides: physical and chemical characterization. Carbon 33, 1585–1592 (1995).
Konkena, B. & Vasudevan, S. Understanding aqueous dispersibility of graphene oxide and reduced graphene oxide through pKa measurements. J. Phys. Chem. Lett. 3, 867–872 (2012).
Jackson, J. D. Surface charges on circuit wires and resistors play three roles. Am. J. Phys. 64, 855–870 (1996).
Marcus, A. The electric field associated with a steady current in long cylindrical conductor. Am. J. Phys. 9, 225–226 (1941).
Chen, L. et al. Ion sieving in graphene oxide membranes via cationic control of interlayer spacing. Nature 550, 380–383 (2017).
Tielrooij, K. J., Garcia-Araez, N., Bonn, M. & Bakker, H. J. Cooperativity in ion hydration. Science 328, 1006–1009 (2010).
Siegel, J., Lyutakov, O., Rybka, V., Kolská, Z. & Svorčík, V. Properties of gold nanostructures sputtered on glass. Nanoscale Res. Lett. 6, 96 (2011).
O’Dwyer, J. J. Dielectric breakdown in solids. Adv. Phys. 7, 349–394 (1958).
Kim, S. K. et al. Conductive graphitic channel in graphene oxide-based memristive devices. Adv. Funct. Mater. 26, 7406–7414 (2016).
Eda, G. et al. Graphene oxide gate dielectric for graphene-based monolithic field effect transistors. Appl. Phys. Lett. 102, 133108 (2013).
Standley, B., Mendez, A., Schmidgall, E. & Bockrath, M. Graphene-graphite oxide field-effect transistors. Nano Lett. 12, 1165–1169 (2012).
Lee, J. S., Lee, S. & Noh, T. W. Resistive switching phenomena: a review of statistical physics approaches. Appl. Phys. Rev. 2, 031303 (2015).
Qin, S. et al. A physics/circuit-based switching model for carbon-based resistive memory with sp2/sp3 cluster conversion. Nanoscale 4, 6658–6663 (2012).
Chen, C. et al. Annealing a graphene oxide film to produce a free standing high conductive graphene film. Carbon 50, 659–667 (2012).
Borini, S. et al. Ultrafast graphene oxide humidity sensors. ACS Nano 7, 11166–11173 (2013).
Pei, S. & Cheng, H. The reduction of graphene oxide. Carbon 50, 3210–3228 (2012).
Park, S. et al. Colloidal suspensions of highly reduced graphene oxide in a wide variety of organic solvents. Nano Lett. 9, 1593–1597 (2009).
Ganguly, A., Sharma, S., Papakonstantinou, P. & Hamilton, J. Probing the thermal deoxygenation of graphene oxide using high-resolution in situ X-ray-based spectroscopies. J. Phys. Chem. C 115, 17009–17019 (2011).
Müller, R. A semiquantitative treatment of surface charges in DC circuits. Am. J. Phys. 80, 782–788 (2012).
Jackson, J. D. Classical Electrodynamics 3rd edn, Ch. 1, 12–14 (John Wiley & Sons, New York, 1999).
Geissler, P. L., Dellago, C., Chandler, D., Hutter, J. & Parrinello, M. Autoionization in liquid water. Science 291, 2121–2124 (2001).
Mafé, S., Ramírez, P. & Alcaraz, A. Electric field-assisted proton transfer and water dissociation at the junction of a fixed-charge bipolar membrane. Chem. Phys. Lett. 294, 406–412 (1998).
Pinkerton, T. D. et al. Electric field effects in ionization of water–ice layers on platinum. Langmuir 15, 851–856 (1999).
Wilson, N. R. et al. Graphene oxide: structural analysis and application as a highly transparent support for electron microscopy. ACS Nano 3, 2547–2556 (2009).
Loh, K. P., Bao, Q., Eda, G. & Chhowalla, M. Graphene oxide as a chemically tunable platform for optical applications. Nat. Chem. 2, 1015–1024 (2010).
Plimpton, S. Fast parallel algorithms for short-range molecular dynamics. J. Comput. Phys. 117, 1–19 (1995).
Vácha, R., Buch, V., Milet, A., Devlin, J. P. & Jungwirth, P. Autoionization at the surface of neat water: is the top layer pH neutral, basic, or acidic? Phys. Chem. Chem. Phys. 9, 4736–4747 (2007).
Vácha, R., Horinek, D., Berkowitz, M. L. & Jungwirth, P. Hydronium and hydroxide at the interface between water and hydrophobic media. Phys. Chem. Chem. Phys. 10, 4975–4980 (2008).
Mills, R. Self-diffusion in normal and heavy water in the range 1-45.deg. J. Phys. Chem. 77, 685–688 (1973).
Meyer, B. et al. Partial dissociation of water leads to stable superstructures on the surface of zinc oxide. Angew. Chem. Int. Ed. 43, 6641–6645 (2004).
Brodskaya, E., Alexander, P. L. & Aatto, L. Investigation of water clusters containing OH- and H3O+ ions in atmospheric conditions. A molecular dynamics simulation study. J. Phys. Chem. B 106, 6479–6487 (2002).
Huang, H. et al. Salt concentration, pH and pressure controlled separation of small molecules through lamellar graphene oxide membranes. Chem. Commun. 49, 5963–5965 (2013).
Acknowledgements
This work was supported by the Royal Society, Engineering and Physical Sciences Research Council, UK (EP/K016946/1, EP/N013670/1 and EP/P00119X/1), British Council (award reference number 279336045), European Research Council (contract 679689) and Lloyd’s Register Foundation. We thank J. Waters for assisting with X-ray measurements and G. Yu for electrical measurements.
Reviewer information
Nature thanks H. Fang, N. Koratkar, B. Mi and H. B. Park for their contribution to the peer review of this work.
Author information
Authors and Affiliations
Contributions
R.R.N. initiated and supervised the project. K.-G.Z. performed the experiment and analysed the data with help from K.S.V. and R.R.N.; K.S.V. carried out the PF TUNA and Raman characterization and analysis. C.T.C. carried out the mass spectroscopy. K.H., J.A. and Y.S. helped in sample preparation, characterization and data analysis. K.S.V., M.N.-A., H.G.-K., K.S.N. and F.M.P. performed the theoretical modelling and simulations. J.C.Z. and A.P. performed the XPS characterizations. O.P.M., V.G.K. and A.N.G. performed the infrared characterizations. A.K.G. contributed to theoretical discussions. R.R.N., K.S.V., K.-G.Z. and K.S.N. co-wrote the paper. All authors contributed to discussions.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
Extended Data Fig. 1 Metal–GO–metal sandwiched membranes.
a, Fabrication procedure for the metal–GO–metal sandwich membrane. b, Photograph of one of our metal–GO–metal sandwich membranes attached to the PET sheet (step 2 in a). Scale bar, 6 mm. This was further attached onto another plastic disk to seal the metal container for gravimetric testing. c, SEM image showing the discontinuities and voids in a 10-nm gold thin film on a GO membrane. Scale bar, 150 nm. d, Water permeation rate of metal–GO–metal sandwiched membranes as a function of gold electrode thickness. The dotted line is a guide to the eye. Water permeation rates of a bare porous silver (Ag) support (red filled circle) and a commercial polyamide nanofiltration membrane (green filled circle) are provided for comparison. Inset, SEM image of a 50-nm-thick gold thin film on a GO membrane. Scale bar, 150 nm.
Extended Data Fig. 2 Conducting filament formation in a GO membrane and its electrical characterization.
a, I–V characteristics during the first voltage sweep show a sudden increase in the current for membranes exposed to humid conditions, suggesting partial electrical breakdown of the GO membrane and conducting filament formation. b, In-plane and out-of-plane I–V characteristics of the GO membrane after filament formation. Inset, out-of-plane I–V characteristics of the GO membrane at 100% RH and vacuum.
Extended Data Fig. 3 Raman and AFM characterization of conducting filaments in GO membranes.
a, Topographical SEM image of a GO membrane after the formation of conducting filaments. b–d, Raman intensity ratio (ID/IG) mapping of D and G bands for a pristine GO membrane (b) and a GO membrane after conducting filaments have formed (c, d). c, Raman imaging from the membrane surface close to the positive electrode (about 200 nm away). d, Raman imaging from the membrane surface close to the negative electrode (about 100 nm away). e, f, The Raman spectra from the dark blue and green regions in c, respectively. g–j, Topography and the corresponding TUNA current image of pristine GO (g and h, respectively) and conducting GO (i and j) membranes (filament formed at 100% RH) exfoliated on a gold-thin-film-coated Si substrate. The conducting filaments are marked by red circles. Scale bars, 1 μm. k, Out-of-plane I–V characteristics of a conducting GO membrane with a diameter of about 7 mm before (parent membrane) and after dividing into four equal pieces. Inset, schematic of the structure of conducting carbon filaments in the GO membrane.
Extended Data Fig. 4 Influence of intercalated water and oxygen content on conducting filament formation in GO membranes.
a, b, Topography and the corresponding TUNA current images of GO membranes after filament formation at 40% RH (a and b, respectively) and inside liquid water (c and d). e, I–V characteristics of pristine and partially reduced (‘pr’) GO membranes during the first voltage sweep at 100% RH show partial breakdown. f, TUNA current image of a partially reduced GO membrane after filament formation. The conducting filaments are marked by red circles. Scale bars, 2 μm.
Extended Data Fig. 5 XPS characterization of GO membranes.
a, C 1s spectrum from a pristine GO membrane. b, c, C 1s spectra from GO membranes used for the electrically controlled permeation experiments after filament formation, from a freshly cleaved membrane surface close to the inner middle region (b) and close to the positive electrode (c). Black lines, raw data; red lines, the fitting envelope; blue lines, deconvolved peaks attributed to the chemical environments indicated.
Extended Data Fig. 6 Mass spectrometry to probe electrically controlled water permeation.
a, Schematic of the experimental set-up for mass spectrometry measurements. A throttle valve (TV) controls the gas inlet with a capacitance gauge (CG) used to measure the upstream pressure. An isolation valve (IV) isolates upstream and downstream sides of the membrane. A rotary pump (RP1) evacuates the feed and the permeate side to 1 mbar. The quadrupole mass spectrometer (QMS) measures the downstream partial pressure. A turbomolecular pump (TP) backed by a rotary pump (RP2) evacuates the high-vacuum chamber of the mass spectrometer. An active ion gauge (IG) measures the pressure down to 1 × 10−9 torr in the high-vacuum side. b, The partial pressure of He, H2, O2 and H2O at the permeate side as a function of time at different currents through the membrane. No detectable change is observed in the partial pressures of He, H2 and O2 under different currents through the membrane. c, The partial pressure of H2O as a function of the current across the GO membrane and the corresponding I–V characteristics (colour-coded axes).The dotted lines are guides to the eye.
Extended Data Fig. 7 Electrically controlled liquid water permeation in GO membranes.
a, Schematic of the experimental set-up. b, I–V characteristics of a Au/GO/Ag membrane during the first voltage sweep while it is immersed in liquid water in the experimental set-up. c, Liquid water permeation rate as a function of current across the membrane after filament formation and the corresponding I–V characteristics (colour-coded axes). Sample-to-sample variation in the permeation is less than 30% (three samples measured).
Extended Data Fig. 8 In situ membrane temperature and water absorption measurements.
a, Measured membrane temperature as a function of the current flowing across the membrane during the electrically controlled water permeation experiment. Error bars, standard deviation from 10 different measurements across the sample. b, The weight intake of a Au/GO/Ag membrane (1-μm-thick GO) at different humidity and electric current values. Weight intake is calculated with respect to the weight of the membrane at 0% RH. The shaded areas show the time during humidity sweeps.
Extended Data Fig. 9 Electric field around a current-carrying conductor.
a, Schematic of the application of a voltage V across an electrically conducting wire with radius a and length L; b is the point at which the potential decays to zero; r represents any point between a and b where the electric field E is calculated. b, Magnitude of E and its spatial distribution as a function of r and z around a conductive filament with 1-V potential difference across the ends and with 1-nA current flow.
Extended Data Fig. 10 Molecular dynamics simulations.
a, Side view of our molecular dynamics simulation set-up used to study the flow of water mixed with H3O+ and OH− ions in the graphene capillary. The model contains two boxes connected by a graphene capillary. At the beginning of the simulation, water was mixed with H3O+ and OH− ions (red and white dots). By moving the left wall (subjected to external pressure) of the box towards the capillary, the water flow is created and the right box is gradually filled. The arrow indicates the direction of the external pressure applied on the left wall of the box. b, c, Number of water molecules in the capillary (b) and number of water molecules in the right box (c) for pure water and water with ions once pressure is applied to the left box (colour-coded labels). d, Water flow rate as a function of the concentration of ions inside the capillary.
Rights and permissions
About this article
Cite this article
Zhou, KG., Vasu, K.S., Cherian, C.T. et al. Electrically controlled water permeation through graphene oxide membranes. Nature 559, 236–240 (2018). https://doi.org/10.1038/s41586-018-0292-y
Received:
Accepted:
Published:
Issue date:
DOI: https://doi.org/10.1038/s41586-018-0292-y
This article is cited by
-
Interfaces govern the structure of angstrom-scale confined water solutions
Nature Communications (2025)
-
Theoretical framework for confined ion transport in two-dimensional nanochannels
Nature Communications (2025)
-
Smart and solvent-switchable graphene-based membrane for graded molecular sieving
Nature Communications (2025)
-
Control of water for high-yield and low-cost sustainable electrochemical synthesis of uniform monolayer graphene oxide
Nature Communications (2025)
-
Bio-inspired design of next-generation ultrapermeable membrane systems
npj Clean Water (2024)
Alex Talyzin
I would like to ask two questions.
1) The most trivial explanation of presented permeation data would be electrolysis of water and formation of gas bubbles which block permeation. However, the paper claims that no electrolysis is observed even when 2V cycles are applied. Fig.3 shows data recorded at even higher voltages up to 5V. Smaller swelling shown in this figure is compatible with water electrolysis. There is no discussion about any possible reasons why there is no electrolysis in GO membranes which are water saturated at 100% humidity. Why there is no electrolysis in the membranes which support continuous flow of water and absorb 40% of water by weight? Did authors tried to estimate voltage between two neighboring "filaments" using values of electric field which they provided in simulations (10 000 000 V m−1) and average distance between the filaments?
2) It is well known that GO can be reduced electrochemically and exhibit quite complex electrochemistry due to variety of functional groups, see for example papers by M.Pumera. The electrochemistry of GO is not discussed in the paper. However, the XPS data provide evidence for formation of reduced graphene oxide close to electrode (Extended Data Fig. 5 C). Irreversible chemical modification of GO membranes could be expected under continuous 2V cycling. How authors explain absence of electrochemical degradation of their membranes and did they verify this by experiments with more than 5 cycles shown in Fig. 1 and Fig 3 ?
Thank you in advance for answers to these questions.
Alexander Talyzin
Alex Talyzin
It is very interesting paper and I would like to make several more comments. Here is the first one: why wet He leaked through GO membranes in 2012 and does not leak in 2018?
The only evidence for absence of electrolysis in GO membranes exposed to 100% humidity at 2-5V is the mass spectrometry data (Extended Data Fig. 6). The figure shows that no hydrogen and no oxygen are detected. However, the experiment shows also absence of He permeation.
Citing the paper:
"No detectable change is observed in the partial pressures of He, H2 and O2 under different currents through the membrane". The text provides citation to the ref 17 , the Science paper from 2012 with the title:
"Unimpeded Permeation of Water Through Helium-Leak–Tight Graphene-Based Membranes"
However, reading of ref 17 shows that the membrane is not permeated by He only at DRY conditions. When He is wet there is rather detectable permeation, see Fig. S4, supporting info for 2012 paper. Citation:
" Figure S4 shows the observed He-leak rate as a function of time for a 1-μm-thick GO membrane. One can see that, in high humidity, helium indeed starts permeating through GO. Initially its speed increases with time as the membrane gets hydrated and, after a few hours, the leak saturates to practically a constant value." By the way, note very slow rate of vapor saturation, it takes 4 hours to achieve maximal swelling.
Can authors explain why the He permeation is absent in new experiments where the He is clearly wet and there is swelling as evidenced by XRD? If the system is not sufficiently sensitive to detect He, it is easy to understand why oxygen and hydrogen were not detected as well.
Or may be the membranes in new study are different from old ones?
In which ways then and why this difference is not discussed?
Alex Talyzin
Next question is about XPS data. How authors calculated C/O ratio of their GO material?
No oxygen peak is shown in the paper, while the spectrum of reduced GO (recorded close to electrode) is mistakenly assigned to GO with C/O =3.6. What we see in the extended data Figure 5 c is clearly from reduced graphene oxide.
https://uploads.disquscdn.c...
Here I put together XPS from the paper published by the same group last year in Nature Materials (supplementary Figure 2) and Figure from 2018. Both suppose to show change of composition from C/O=3.2 to C/O=3.6 but look rather differently.
According to ref 42 (Ganguly et al) the spectrum recoded close to electrode is reduced GO with C/O over 5. Very similar rGO in the paper by Ganguly was produced by thermal reduction at 400 degrees C, Figure 2C.
Note that interpretation of XPS spectra dramatically changed between 2017 and 2018 (check C=O peaks) without any comments or explanation.
Electrochemical reduction (basically degradation) of GO is the trivial explanation for formation of rGO close to electrode.
The question remains: how the C/O ratios were calculated in the paper?
Following the statement of "full data availability" provided in the paper I sent request to authors asking for carbon and oxygen parts of XPS spectra. So far no reply received.
Alex Talyzin Replied to Alex Talyzin
Now it is clear that all C/O ratio numbers publisihed in this paper and in 2017 Nature Materials paper have no meaning. They did not recorded oxygen part of XPS spectra. Here is how Prof. R.Nair described their procedure to calculate C/O ratio:
" Similarly, we calculated the C/O ratio by using C 1s peak fitting rather than using both carbon and oxygen peaks. We prefer our procedures
because it avoids counting extra oxygen present on the sample surface (e.g., from water)."
This explanation was received from Prof. R.Nair through Nature Materials editor. The C/O number in this case depends on exact fitting procedure which are different in Nature Materials paper and new Nature paper. Both procedures can not not provide true C/O value. Editors of Nature Materials do not think that this information is of any interest for readers of their paper.
Alex Talyzin
Now two questions about modelling. I address these questions first of all to reviewers of this paper (authors are long aware of my opinion about non existence of graphene capillaries).
1) The GO membrane is not conductive (see text of paper and figure 1) and hydrophilic (see text of paper and Figure 8 in extended data, up to 48% of water by mass is adsorbed). It is electrically insulating along the membrane and hydrophilic even after formation of "filaments".
How the permeation across the dielectric and hydrophilic membrane can be modeled using single graphene capillary which is hydrophobic, electrically conductive and oriented along the membrane? How similar are thes etwo systems?
The paper actually provides evidence for absence of "graphene capillaries" in GO. The same methods which helped authors to reveal few 10 nm diameter "filaments" failed to reveal interconnected system of "graphene capillaries". The list of arguments against existence of these capillaries is rather long, see for example our papers from 2014. (DOI: 10.1039/C3NR04631A ) or DOI: 10.1021/nl5013689. No evidence which support existence of "graphene capillaries" in GO was presented by authors after they postulated it in 2012.
2) How the "Electrically controlled water permeation" can be modeled without electric field?
Extended data Figure 10 shows H3O+ and OH− ions in liquid water forming some kind of solution which diffuse into capillary. The ions are NOT in any way affected by electric field which suppose to create them. Electric field somehow help to dissociate water and at the same time does not cause any effect on ions. Should not it create electric current? What will happen with these ions at electrodes at 2-5 V?
It is a small detail after two previous points but why the size of graphene capillary is taken as 1 nm if the XRD of membranes shows 8.5Å? This suppose to give 8.5Å-3.5Å= 5Å diameter of "channels" that is how this group usually calculates them.
Final question: how close is this model to real membranes? Can electric field simultaneously exist to dissociate water and not exist in order to avoid electrical current?
R. R. N
At first glance, the comments by Talyzin may sound as written in good faith, unless a reader knows the history behind or is from another research group on the receiving end of his commentary.
Talyzin’s animosity stems from our Science 2012 paper that suggested that water permeates through graphene oxide (GO) laminates in a meander-like manner. It rapidly flows through ‘graphene capillaries’ along GO planes and, after reaching edges of individual sheets, diffuses to the next interlayer level. Talyzin promotes an alternative explanation (refs in his comment) which suggests diffusion through numerous holes in GO sheets. Our model has been consensually accepted by the community because it explains practically all the experimental evidence. Talyzin’s model is universally ignored for many glaring inconsistencies. To be fair, the meander model is not 100% proven yet. One of its problematic assumptions is the segregation into clean and functionalized regions to create ‘graphene capillaries’, which is needed to be consistent with a relatively high C/O ratios in GO laminates.
While we would be happy to accept any better explanation to explain experiments, we could not possibly accept Talyzin’s model. Over the last 5 years, my colleagues and I spent a lot of time and effort on explaining our position to Talyzin, invited him to Manchester and provided samples and data. Nothing marginally useful came out but repetitive baseless criticism. Then a barrage of malicious comments started after each paper we published. Sometimes they were anonymous (PubPeer) and, whenever possible, went to editors. As Nature now allows online comments, Talyzin could not overlook this venue, too.
All his commentary follows the same pattern. First, Talyzin offers his own explanation of our experiments (electrolysis of water in the present case) presuming that all the authors and all 4 expert reviewers are so stupid they did not consider trivialities (possibility of electrolysis in the paper having both electricity and water in the title!). Then Talyzin goes on to criticize the meander model, always on the basis of C/O ratios as mentioned above. Next, he discovers all sorts of seeming contradictions between the paper in question and previous papers, always focusing on his favorites, XRD and XPS measurements. Leaving aside that those measurements are fringe results used only for routine characterization, he completely ignores the fact that GO laminates in different papers are made differently to address specific problems and phenomena. The laminates have different thickness, flake sizes, morphology, etc. We avoid further details about our XRD and XPS measurements because they are provided in our Reply to yet another Talyzin’s comment on our earlier paper (with editors of Nature Materials).
Over the time, we have found it increasingly difficult and time consuming to deal with Talyzin’s attacks and stopped communicating. We reply only because this is a new venue for his trolling. If Talyzin wants to prove his diffusion-only model, we suggest him to publish a proper report on its merits in explaining the experimental evidence gained by many groups around the world. Let the community be a judge of it.
Needless to say, we welcome any genuine interest in our paper. We are ready to share our expertise and samples and would welcome visitors to Manchester who are interested in reproducing our results.
Alex Talyzin Replied to R. R. N
Prof.R.Nair have not answered my questions and attempts to replace discussion of the paper with personal comments and insults.
I suggest this discussion to focus on the current paper which is more interesting to readers than my persona.
Here the main questions again.
How can electrically controlled permeation be modeled without electrical current and electric field?
How can the solution of dissociated water in water (?)(with up to 10 % concentration) be used for modeling in principle? The model assumes that solution is stable for hours like NaCl. What will happen with these ions if electric field which produce them is actually added to the model?
The walls of graphene capullary are conductive while experiment proves that GO is not conductive. GO is hydrophylic as shown in experiments but capullary is hydrophobic. Why no electrolysis is observed at 2-5 V and 100% humidity?
There is also a set of very strange references which I will comment later.
Alex Talyzin
"Talyzin promotes an alternative explanation (refs in his comment) which suggests diffusion through numerous holes in GO sheets."
Not only through GO sheets but also through holes between GO flakes. See slide 6 from my talk made in 2014 at Graphene Week (link below) for quite large audience. Prof. R.Nair was chairmen on that session.
The Science paper from 2012 suggested unrealistic model with square-shaped and close packed GO flakes. That is how the "super-fast" flow was introduced: by assuming that each square GO 1 micron sized flake is surrounded by 0.7 nm permeation channels which account for 0.1% of total area of GO layer. The model also had some obvious errors .
After 5 years of fierce denial of very existence of holes across the flakes and around the flakes Prof. R.Nair ... published it as his own idea.
Citation from his Nature Materials paper, 2017:
"Pinholes in GO membranes originate from random stacking of
individual GO flakes and can also involve nanometre-size holes (ref 2)
within flakes."
Next step is to admit that thanks to holes there is no need for "superfast flow" and "graphene capillaries".
Citation from 2012 Science paper: "We SPECULATE that these empty spaces form a network of pristine-graphene capillaries within GO laminates."
Authors made no effort to provide any evidence for this speculation in 2012 or next papers until 2018. Please correct me with citation if I am wrong. This speculation is proved to be wrong by data presented in several Prof. R.Nair papers including the one we discuss here,
About three month ago I submitted correspondence letter to Nature Materials and now the comment is under review. It was first time ever when I contacted editors in connection of Prof. R.Nair publications.
Earlier I published my criticism in 3 papers on GO membranes, presented it at several conferences, seminars around EU and lectures for graduate students.
I enclose here link to file with my presentation made at Graphene Week in 2014 (Göteborg).
https://www.researchgate.ne...
Alex Talyzin
“he (Talyzin) completely ignores the fact that GO laminates in different papers are made differently to address specific problems and phenomena” R.R.Nair, see below.
Very good point. I would like to address question to authors and reviewers of this paper: how different are these membranes from previous reports and how were they prepared?
The paper is not reproducible as it does not provide basic characterization for precursor graphite oxide, for membranes prepared form this graphite oxide and even for the “filamented” membranes prepared by “controlled” breakdown.
1. I pointed out in my papers and presentations that even GRAPHITE oxides are different e.g. by degree and type of oxidation. It is very well known that graphite oxides are rather different depending on details of preparation procedure and on precursor graphite. The 2018 paper is first one where commercial material from Manchester based company BGT Materials is used. The company ignores my attempts to order their product. They advised me to make it by myself.
All previous studies were performed with homemade graphite oxide. Are these graphite oxides different in 2018 and back in 2012-2017? It is impossible to know because precursor graphite oxides were NEVER analyzed in any of membrane papers from this group 2012-2018. In my group we make characterization of every batch of graphite oxide using XRD, FTIR, XPS, TGA. None of these methods was ever used by R.Nair and his co-authors to characterize their graphite oxides. The membranes were prepared using unknown materials.
2. Next step is preparation of membranes. The only characterization of GO membrane available in 2018 paper is XPS spectrum for C1S. There is no XRD, FTIR, XPS, TGA for pristine GO membranes. Not at ambient conditions, not at any other humidity levels. Unknown material.
How different are these membranes compared to membranes prepared in 2012 Science paper and cited as reference in pervaporation experiments?
In 2012 the GO membranes were prepared by drop casting on Cu foil, foil was etched away by strong acid (never verified if the acid modified surface of GO membrane), in 2018 membranes were prepared by vacuum filtration. There is no discussion of possible differences. No XRD, FTIR, XPS, TGA were reported in in 2012 for membranes either. The same is for Science paper 2014- the study was performed using membranes with unknown structure and composition. The only available data in 2014 paper is small piece of XRD pattern. No XPS or any other data which could report oxidation degree or FTIR which could show something about functionalization. Paradoxically, the only data which report XPS and C/O (=2.8) of membranes prepared in Manchester at that period of time is our paper published in Nanoscale thanks to sample provided to me in 2012.
More characterization of GO membranes is provided in 2017 Nature Chemistry paper but only for very thin (few nm) and supported membranes. These membranes were of two different types: “common” and “highly laminated.” Conditions of sonication for GO dispersions are not provided in 2018 paper. No way to understand if these membranes are “common” , “highly laminated” or anything else.
3. Finally, in 2018 paper R.Nair and his co-authors report preparation of new material- GO membranes with conductive filaments. The material is major focus of the paper. How this material is characterized?
There is no reference XRD, FTIR or TGA data presented for ambient humidity or zero humidity. Readers are not given a possibility to see if there are any changes in FTIR or XRD due to the “controlled breakdown” due to absence of reference data from untreated membranes and even the data from “filamented” membranes. Any reduction, any graphitization, change in functionalization? No data.
The only data available for “filamented” membranes are small fragments of XRD and small piece of FTIR recorded at permeation conditions, supposedly with 100% humidity at least on one side of the membrane. Without reference point at ambient humidity these data are inconclusive. There is XPS spectrum from inner part and outer part of membrane with confusing C/O values, no oxygen part of spectra, there is no word about impurities (sulfur is very common for Hummers GO).
4. What will happen if somebody will be not able to reproduce results reported in this 2018 paper? We will possibly hear again that “all membranes are different”.
The issues with reproducibility were well addressed in this paper:
https://pubs.acs.org/doi/ab...
It provides also good way to deal with reproducibility issues: list of characterization methods which I support and advise to all GO membrane researchers.
Alex Talyzin
The paper states that conductive “filaments” produce 10 000 000 v/m electric
field but do not produce any Joule heating. Of course, if the temperature of this filament significantly exceeds boiling point of water and degradation point of GO it is an issue. Remarkably, the paper provides only estimation of electric field but no attempt to make simple calculation of temperature around the “filament”. The Joule heating is ruled out using simple temperature measurement recorded from the whole sample surface using IR thermometer with unspecified accuracy and area of spot.
“The influence of the current on the water transport can be due to Joule heating…. We measured the membrane temperature … and found no substantial variations (Extended Data Fig. 8a).”
Dear moderators who deleted this post previously, what is written below is not adertisement for IR thermometer. It is about improper use of cheap thermometer for research purpose which resulted in inconclusive and directly erroneous data. These data does not allow to "rule out" most trivial factor affecting permeation of membranes.
The N92FX, Maplin is cheap (20 pounds) hand held IR thermometer never intended for research.
The technical parameters of this device can be found directly on the package and correspond to the price. The accuracy of T measurement is 2.5 degrees. The Extended figure 8a of the Nature paper shows minimum 0.2 degrees error bar which is more than 10 times smaller relative to device accuracy.
The emissivity of N92FX is fixed to 0.97. That is good value for carbon but really bad for gold which covers the GO film. For polished gold emissivity is 0.025 and for not polished 0.47. For the gold film deposited over GO membrane it can be anything in between. As a result the temperature measurement can be as far as tens of degrees away from real.
Distance to spot ratio of this instrument is 8:1. Which mean that for the sample with diameter of 7-10 mm the measurements had to be done from the distance smaller than 56-80 mm, otherwise it would include not only whole sample but also everything around.
The measurement is inconclusive about joule heating of "conductive filaments" and space around them.
The model provided in the paper assumes 1 micrometer long wire (10 nm in diameter). Current is 1 nA, applied potential is 1V. It is standard problem to calculate dissipated power, then to make estimation for temperature of wire using e.g. heat capacity of graphite or other type of carbon materials.
On the next step one can estimate temperatures of water away from filament, like it is done for electric field in extended data Figure 9.
As a reviewer I would certainly ask authors ot make this calculation and to present results in their paper.
Alex Talyzin
I am still puzzled by absence of electrolysis at 2-5V in water saturated membranes at 100% humidity (see my very first comment at the bottom). See below zoomed figure 6 (extended data). The blue line which shows H2 actually shows change exactly at the moments when current was switched on and off. I tried to find red points in this graph (O2) but could not find any, not a single one. I wonder if authors could show hydrogen and oxygen curves with appropriate scale which help to see details and noise levels.
https://uploads.disquscdn.c...