A Method for Separating Fluidic Water from Impure Fluids and a Filter therefore

Abstract

A method of separating fluidic water from impure fluids is disclosed. The impure fluids comprising fluidic water and one or more substances having a kinetic diameter similar to that of water molecules. The kinetic diameter of the one or more substances is at most 50% and preferably 33% greater than that of the water molecules. The method comprises applying to a first side of a carbon nanomembrane the impure fluid; and collecting from the opposite of the carbon nanomembrane the fluidic water. The method can be used in filter applications.

Claims

1. A method of separating fluidic water from impure fluids, the impure fluids comprising fluidic water and one or more substances having a kinetic diameter similar to that of water molecules, comprising applying to a first side of a carbon nanomembrane the impure fluid; and collecting from the opposite of the carbon nanomembrane the fluidic water.

2. The method of claim 1, wherein the kinetic diameter of the one or more substances is at most 50% greater than that of the water molecules, and preferably at most 33% greater than that of the water molecules.

3. The method of claim 1, wherein the one or more substances are non-polar.

4. The method of claim 1, wherein the one or more substances are one of helium, neon, carbon dioxide, argon, oxygen, nitrogen, acetonitrile, n-hexane, ethanol, and 2-propanol.

5. The method of claim 1, wherein the carbon nanomembrane comprising laterally cross-linked aromatic compounds.

6. The method of claim 5, wherein the aromatic compounds are selected from the group consisting of polyphenyl compounds.

7. The method of claim 5, wherein the aromatic compounds are at least one of a terphenyl or quaterphenyl.

8. The method of claim 1, wherein the carbon nanomembrane has a thickness of between 0.5 nm and 100 nm.

9. The method of claim 1, wherein the carbon nanomembrane has pores with diameters in the range of 0.3 nm to 1.5 nm.

10. A filter for separating fluidic water from impure fluids, the impure fluids comprising fluidic water and one or more substances having a kinetic diameter similar to that of water molecules, wherein the filter comprises: a first container comprising the impure fluids; a second container for collecting the fluidic water; and a carbon nanomembrane located between the first container and the second container and arranged in such a manner that one surface of the carbon nanomembrane is in fluidic contact with the impure fluids.

11. The filter of claim 10, wherein the kinetic diameter of the one or more substances is at most 50% greater than that of the water molecules, and preferably at most 33% greater than that of the water molecules.

12. The filter of claim 10, wherein the one or more substances are non-polar.

13. The filter of claim 10, wherein the one or more substances are one of helium, neon, carbon dioxide, argon, oxygen, nitrogen, acetonitrile, n-hexane, ethanol, and 2-propanol.

14. The filter of claim 10, wherein the carbon nanomembrane substantially consists of laterally cross-linked aromatic compounds.

15. The filter of claim 14, wherein the aromatic compounds are selected from the group consisting of polyphenyl compounds.

16. The filter of claim 14, wherein the aromatic compounds are at least one of a terphenyl or quaterphenyl.

17. The filter claim 10, wherein the carbon nanomembrane has a thickness of between 0.5 nm and 100 nm.

18. The filter claim 10, wherein the carbon nanomembrane has pores with diameters in the range of 0.3 nm to 1.5 nm.

19. The filter of claim 10, wherein the carbon nanomembrane is radiation resistant.

20. (canceled)

21. A method for the extraction of potable water from a humid atmosphere, the humid atmosphere comprising fluidic water and one or more substances having a kinetic diameter substantially similar to that of water molecules, the method comprising applying to a first side of a carbon nanomembrane a humid atmosphere; and collecting from the opposite of the carbon nanomembrane the potable water.

22. A method of separating fluidic water from impure fluids using a carbon nanomembrane comprising laterally cross linked terphenyl or quaterphenyl compounds, the impure fluids comprising fluidic water and one or more substances, the method comprising: applying to a first side of a carbon nanomembrane the impure fluid; and collecting from the opposite of the carbon nanomembrane the fluidic water.

23. A filter for separating fluidic water from impure fluids, the impure fluids comprising fluidic water and one or more substances, wherein the filter comprises: a first container comprising the impure fluids; a second container for collecting the fluidic water; and a carbon nanomembrane comprising laterally cross linked terphenyl or quaterphenyl compounds located between the first container and the second container and arranged in such a manner that one surface of the carbon nanomembrane is in fluidic contact with the impure fluids.

Description

DESCRIPTION OF THE FIGURES

[0017] FIG. 1 shows an example of the filter using the carbon nanomembrane described in this document.

[0018] FIG. 2 shows the experimental set up for measuring the water permeance of the carbon nanomembrane.

[0019] FIG. 3 shows the water permeance of TPT CNMs as a function of the relative humidity in the feed chamber measured in a vacuum apparatus. The permeance was detected by a quadrupole mass-spectrometer (QMS). The hollow square is the value measured by the mass loss method (example 1).

[0020] FIG. 4 shows a comparison of the measured water permeance to the permeance for helium.

[0021] FIG. 5 displays water vapor transmission rate of the carbon nanomembrane in comparison to conventional membranes.

[0022] FIGS. 6A-6E show the morphology of TPT SAM and CNM. FIG. 6A is an STM image of TPT SAM measured at room temperature in ultra-high vacuum (UHV) (U.sub.Bias=500 mV, I.sub.T=70 pA). FIG. 6B is an AFM image of TPT CNM measured at 93 K in UHV via AFM tapping mode of operation (amplitude set point A=8.9 nm, center frequency f.sub.0=274.9 kHz). FIG. 6C is shows extracted line profiles in FIG. 6A. The profiles 1-2 of TPT SAM show a center-to-center intermolecular distance of 0.8 nm. FIG. 6D shows extracted line profiles in FIG. 6B. The profiles 3 and 4 of TPT CNM indicate a pore diameter of 0.6 nm. FIG. 6E shows the estimated pore diameter distributions (0.70.1 nm, the error bar denotes standard deviation) extracted from AFM images. The STM and AFM images shown were drift corrected.

[0023] FIG. 7 shows a comparison of single-channel water permeation coefficients between different membranes. Molecular dynamics simulation was used to study the permeation coefficients of CNTs ((5,5)CNT (B. Corry, Journal of Physical Chemistry B 112, 1427 (2008).), (6,6)CNT (B. Corry, Journal of Physical Chemistry B 112, 1427 (2008))), and a stopped-flow apparatus was employed to characterize aquaporins (AQP1 (T. Walz et al., Journal of Biological Chemistry 269, 1583 (1994)), AqpZ (M. J. Borgnia et al., Journal of Molecular Biology 291, 1169 (1999)). The permeation coefficient of TPT CNM was calculated by dividing the measured permeance by the areal density of nanochannels estimated from the AFM images

DETAILED DESCRIPTION OF THE INVENTION

[0024] The invention will now be described in detail. Drawings and examples are provided for better illustration of the invention. It will be understood that the embodiments and aspects of the invention described herein are only examples and do not limit the scope of protection in any way. The invention is defined by the claims and their equivalents. It will be understood that features of one aspect or embodiment of the invention can be combined with the features of a different aspect or aspects and or embodiments of the invention.

[0025] FIG. 1 shows an example of a filter 10 using a carbon nanomembrane 20 as described in this document. A first container 30 has an impure fluid 35. The impure fluid 35 comprises water with a number of other substances, for example low molecular weight materials, including but not limited to helium, neon, carbon dioxide, argon, oxygen, nitrogen, acetonitrile, n-hexane, ethanol, and 2-propanol. The impure fluid could also be sea water or other brackish water. The second container 40 on the other side of the carbon nanomembrane 20 has substantially pure water 45. The impure fluid 35 includes substances which have molecules with a similar kinetic diameter as that of water molecules and which are difficult to filter from the impure fluid 35 by prior art filter. In order to explain this surprising result, it is speculated that water transport through the carbon nanomembranes 20 could occur by a nanofluidic flow enhancement process, as will be explained below.

[0026] As noted in the introduction, the kinetic diameter is defined as the sphere of influence of the molecule that can lead to a scattering event. In the case of a water molecule the kinetic size is 265 pm. Helium and hydrogen molecules have similar kinetic diameter (260 pm and 289 pm) and thus these are particularly difficult species to remove from impure fluids. Other examples of the sizes of the kinetic diameter are generally known, for example from http://en.wikipedia.org/wiki/Kinetic_diameter (downloaded on 14 May 2018).

[0027] The carbon nanomembrane 20 used in the filter is produced by preparing a molecular thin layer of precursor compounds on a metallic or semi conductive substrate and crosslinking the molecular thin layer by electron beam or photon irradiation. The substrate may be selected from the group consisting of gold, silver, titanium, zirconium, vanadium, chromium, manganese, cobalt, tungsten, molybdenum, platinum, aluminum, iron, steel, copper, nickel, silicon, germanium, indium phosphide, gallium arsenide and oxides, nitrides or alloys or mixtures thereof, indium-tin oxide, sapphire, silicate or borate glasses, and aluminum coated polymer foils.

[0028] The carbon nanomembrane 20 is separated from the substrate and transferred to form free standing membranes or membranes supported by other surfaces or grids, see A. Turchanin, and A. Glzhuser, Carbon Nanomembranes, Adv. Mater. 28, 6075-6103 (2016); Turchanin et al., One Nanometer Thin Carbon Nanosheets with Tunable Conductivity and Stiffness, Adv. Mater. 21, 1233-1237 (2009), and P. Angelova et al., A Universal Scheme to Convert Aromatic Molecular Monolayers into Functional Carbon Nanomembranes, ACS Nano 7, 6489-6497 (2013). Alternatively, the carbon nanomembrane 20 can remain on the substrate and openings can be etched through the substrate to produce a filter 10 comprising the carbon nanomembrane 20 on a mechanically stable and permeable support.

[0029] Permeance and selectivity of the carbon nanomembrane 20 depend on a multitude of properties, such as but not limited to thickness of the carbon nanomembrane 20, diameter of pores through the carbon nanomembrane 20, density of the pores, and other properties of the material from which the carbon nanomembrane 20 is manufactured. The selection of the precursor molecules for manufacturing plays a role, since the length of the precursor molecules determines the thickness of the carbon nanomembrane 20 and/or the length of the pores through the carbon nanomembrane 20. It has been found that carbon nanomembranes 20 made from biphenyl, terphenyl and quaterphenyl compounds are suitable, but the invention is not limited thereto.

[0030] The pore diameter can be influenced by the shape of the precursor molecules e.g., linear precursor molecules, condensed precursor molecules, or bulky precursor molecules, see ACS Nano 7, 6489-6497 (2013). The degree of cross linking may influence the structure of the pores in the carbon nanomembrane 20. The carbon nanomembrane 20 used in the filter 10 has a degree of cross-linking of the molecules between 50%-100%, which is adjusted by varying the dose density of the radiation, and it is thought that a degree of crosslinking close to 100% is suitable. This cross-linking is for example achieved for cross-linking of biphenythiol layers on a gold substrate using electron flood-gun in a high vacuum (<510.sup.7 mbar) employing 100 eV electrons and a dose density of 50 mC/cm.sup.2.

[0031] The inventors have estimated that the carbon nanomembranes 20 should have the following properties. The carbon nanomembrane 20 substantially consists of laterally cross-linked aromatic compounds. The aromatic compounds are selected, for example, from the group consisting of polyphenyl compounds, such as but not limited to a terphenyl or quaterphenyl. The carbon nanomembrane 20 has a thickness of between 0.5 and 100 nm. It is thought that the carbon nanomembrane 20 should be between 1 nm and 5 nm, or up to 20 nm thickness to work optimally. The carbon nanomembrane 20 has pores with diameters in the range of 0.3 nm to 1.5 nm (measured with low-temperature AFM in ultrahigh vacuum).

EXAMPLES

Example 1

[0032] The carbon nanomembrane 20 used in the filter 10 can be manufactured as follows.

[0033] Preparation and Transfer of TPT-CNM

[0034] Cleaning of Glassware

[0035] Clean flask with piranha solution (a mixture of 95% H.sub.2SO.sub.4 and 30% H.sub.2O.sub.2 (v:v=7:3)). Rinse flask with Millipore water and let it dry in oven at 120 C.

[0036] Cleaning of Au/Mica Substrate

[0037] Cut Au/mica substrates (300 nm thermally evaporated gold on mica, Georg Albert PVD-Coatings) into small pieces and clean the surface with nitrogen. Place the substrates into UV-Ozone chamber and clean for 3 min. When finished, put the substrate into ethanol for at least 20 min and then rinse the surface of the substrate with ethanol and blow the substrate dry with nitrogen.

[0038] SAM Preparation

[0039] Connect the cleaned flask with a Schlenk line (vacuum/nitrogen manifold) and degas the flask by exchanging the content alternatively with vacuum and nitrogen (for at least three times). Fill the flask at the end with nitrogen. Put the cleaned Au/mica substrate into the flask, carry out degassing procedures a few times until the pressure reaches 10.sup.2 mbar. If necessary, heat the flask as well to get rid of any water vapor. Add 5-10 ml of dry dimethylformamide (DMF) to the flask (do the addition under a nitrogen atmosphere) and degas the solvent several times until no bubbles are seen. Add a very small amount of 1,1,4,1-Terphenyl-4-thiol (TPT) molecules (Sigma-Aldrich) to the flask, degas the system again until no bubbles are seen. Keep the flask under nitrogen and heat the solution to 70 C. After 24 h, take the sample out, rinse the sample first with DMF and then ethanol, and blow the sample dry with nitrogen. Store the sample under argon gas.

[0040] Electron Irradiation

[0041] Crosslinking of TPT-SAMs into CNMs is achieved using an electron flood-gun in a high vacuum (<510.sup.7 mbar) employing 100 eV electrons and a dose density of 50 mC/cm.sup.2.

[0042] Transfer of CNMs onto Silicon Nitride Membranes/Silicon Wafers

[0043] A 4% butyl acetate/ethyl lactate solution of polymethyl methacrylate (PMMA) 50K (ALLRESIST GmbH) is spin-coated on to the CNM/Au/mica surface at 4000 rpm for 40 s, then cured on a hot plate at 90 C. for 5 min. Subsequently, a 4% butyl acetate/ethyl lactate solution of PMMA 950K (ALLRESIST GmbH) is spin-coated at 4000 rpm for 40 s, then cured on a hot plate at 90 C. for 5 min. Transfer the sample to an I.sub.2/KI/H.sub.2O (w:w:w=1:4:40) etching bath for 3-5 min. Detach the mica layer from the PMMA-CNM-Au structure and then transfer the PMMA-CNM-Au structure back to the 12/KI/H.sub.2O solution for 10 min to dissolve the Au. After etching, clean the PMMA-CNM structure first with water, then with KI/H.sub.2O (w:w=1:10) solution for 2 min, and then clean with water 3 times. Transfer the PMMA-CNM structure onto a silicon nitride membrane/silicon wafer with a single hole (membrane size: 0.1 mm0.1 mm, membrane thickness: 500 nm, hole size: 5-50 m, Silson Ltd), let the PMMA-CNM structure dry overnight. Dissolve PMMA with acetone. The immersion time for dissolution of the PMMA layer is 1 h.

[0044] The carbon nanomembrane is then ready.

[0045] Evaluation of Water Permeation

[0046] To evaluate the water permeation through the carbon nanomembrane 20, an upright cup method is employed, as shown schematically in FIG. 2. The carbon nanomembrane is transferred onto a silicon nitride membrane 22 supported by a Si frame 23 where the silicon nitride membrane 22 has a regular hole 24 to form a test sample 28 (as described before). Then the test sample 28 is glued onto a metal container 31 which is filled with a specified amount of water 36. The metal container 31 with the test sample 28 is then placed into an enclosed oven 41 with a constant temperature (300.1 C.). The water vapor 46 inside the oven is controlled to a relative humidity (RH) of 15%2% by a saturated LiCl solution 43. The water vapor 37 above the water 36 inside the metal container 31 will reach a relative humidity of 100% since the metal container 31 contains pure water inside. Due to the differential water vapor pressure inside and outside the metal container 31 the water 37 will be transported through the carbon nanomembrane 20. The weight loss of water 36 inside the metal container 31 is measured after several days by using a balance 50. The water permeance of the carbon nanomembrane 20 can be calculated by the following equations:

[00001]$\mathrm{Permeance}=\frac{\mathrm{weigh}.Math..Math.\mathrm{loss}.Math..Math.\mathrm{rate}\left(\frac{.Math..Math.w}{t}\right)}{\mathrm{membrane}.Math..Math.\mathrm{area}\left(A\right)\mathrm{pressuredifference}\left(.Math..Math.P\right)}$$.Math..Math.P=\mathrm{satured}.Math..Math.\mathrm{vapor}.Math..Math.\mathrm{pressure}\left(1-\mathrm{RH}\right)$

TABLE-US-00001 TABLE 1 measured permeance for terphenyl (TPT) and for quaterphenyl (QPT) based membranes. They are both (1.2 0.2) 10.sup.4 mol m.sup.2 s.sup.1 Pa.sup.1. Water Permeance (mol m.sup.2 s.sup.1 Pa.sup.1) Samples TPT-CNM QPT-CNM 1 1.08E04 1.39E04 2 8.97E05 1.19E04 3 1.13E04 1.19E04 4 1.27E04 5 1.27E04 6 1.13E04 Average value 1.13E04 1.26E04 Standard deviation 1.37858E05 1.17453E05

[0047] It was found that, for other (polar and non-polar) liquids like acetonitrile (kinetic diameter of about 0.34 nm), n-hexane (kinetic diameter of about 0.43 nm), ethanol (kinetic diameter of about 0.43 nm) and 2-propanol (kinetic diameter of 0.47 nm), no weight loss was detected, indicating that CNMs have a high selectivity of water against other liquids with small kinetic diameter.

[0048] The carbon nanomembranes described in this document are produced by crosslinking with electron beam or photon irradiation. Subsequent irradiation therefore does not significantly change their properties. This feature makes them suitable for use in locations in which they experience significant radiation. Examples include, but are not limited to, spacecraft or power stations. The carbon nanomembranes are likely to suffer less damage from the radiation compared to other materials.

[0049] The high water permeance of CNMs was independently confirmed by vapor transport measurements in vacuum. One side of the CNMs was exposed to water vapor under controlled relative humidity (RH) and the flow of permeating molecules was detected by a quadrupole mass spectrometer placed behind the other side of the CNMs. Within the level of experimental accuracy, the water permeance at saturation conditions (100% RH) agrees well with the gravimetric results (FIG. 3). At lower levels of humidity, the permeance dropped, indicating that the higher permeance at saturation pressure was caused by water condensation. Unlike graphene oxide membranes of the prior art, the permeance of CNMs did not vanish with decreasing humidity but remained at 2.010.sup.5 mol.Math.m.sup.2.Math.s.sup.1.Math.Pa.sup.1 at RH below 20%. This is likely related to a transition between different transport mechanisms. Interestingly, the permeance of helium (4.510.sup.8 mol.Math.m.sup.2.Math.s.sup.1.Math.Pa.sup.1) is 2,500 times lower than that of water although they have similar kinetic diameters (0.265 nm for water and 0.26 nm for helium). No noticeable permeation was detected for other gas molecules with kinetic diameters larger than 0.275 nm (Ne, CO.sub.2, Ar, O.sub.2, N.sub.2).

[0050] FIG. 4 shows the measured permeance for water in comparison to the one for helium. It can be seen from the figure that the measured permeances of water is higher than that of helium by more than three orders of magnitude even if the kinetic diameter of water is larger than that of helium. The method of separating the fluidic water from the impure fluids of one or more substances having a similar kinetic diameter as water, like helium, nitrogen, and oxygen and the filter for such a separation thus shows a very high selectivity. Examples of the one or more substances with a similar kinetic diameter are given in the following table:

TABLE-US-00002 TABLE 2 examples for substances with a similar kinetic diameter of water. Values are from http://en. wikipedia.org/wiki/Kinetic_diameter (downloaded on 14 May 2018) with exception oft hose for methanol, ethanol, n-hexane, acetonitrile (all from supporting information to S. Van der Perre et al., Langmuir 30, 8416 (2014)), and 2-propanol (S. Wannapaiboon, Journal of Materials Chemistry A3, 23385 (2015)) Kinetic Molecule Molecular diameter Name Formula weight (pm) Hydrogen H.sub.2 2 289 Helium He 4 260 Methane CH.sub.4 16 380 Ammonia NH.sub.3 17 260 Water H.sub.2O 18 265 Neon Ne 20 275 Acetylene C.sub.2H.sub.2 26 330 Nitrogen N.sub.2 28 364 Carbon monoxide CO 28 376 Ethylene C.sub.2H.sub.4 28 390 Nitric oxide NO 30 317 Oxygen O.sub.2 32 346 Methanol CH.sub.4O 32 380 Hydrogen sulfide H2S 34 360 Hydrogen chloride HCl 36 320 Argon Ar 40 340 Acetonitrile C.sub.2H.sub.3N 41 340 Propylene C.sub.3H.sub.6 42 450 Carbon dioxide CO.sub.2 44 330 Nitrous oxide N.sub.2O 44 330 Propane C.sub.3H.sub.8 44 430 Ethanol C.sub.2H.sub.6O 46 430 2-Propanol C.sub.3H.sub.8O 60 470 Sulfur dioxide SO.sub.2 64 360 Chlorine Cl.sub.2 70 320 Benzene C.sub.6H.sub.6 78 585 Hydrogen bromide HBr 81 350 Krypton Kr 84 360 n. Hexane C.sub.6H.sub.14 86 430 Xenon Xe 131 396 Sulfur hexafluoride SF.sub.6 146 550 Carbon tetrachloride CCl.sub.4 154 590 Bromine Br.sub.2 160 350

[0051] FIG. 5 shows the water vapor transmission rate of the terphenyl based carbon nanomembrane based on the measured permeance in comparison to conventional membranes. It will be seen that the rate is orders of magnitude higher.

[0052] The carbon nanomembranes have a nanofluidic flow enhancement. The transport rate for water does not depend significantly on the thickness of the carbon nanomembrane, or on the length of the precursor molecules, see table 1. The selectivity to non-polar small molecules can be expected to increase with the thickness, or with the length of the precursor molecules as shown by the following calculation.

[0053] Assuming the transport of water and non-aqueous air molecules through the carbon nanomembrane with a thickness x in the time t with diffusion constant D can be modelled by the diffusion function for the concentration c behind the membrane (see for example the disclosure in http://demonstrations.wolfram.com/DiffusionInOneDimension/ downloaded on 14 May 2017)

c(exit)=c.sub.0/2sqrt(Dt)exp(x.sup.2/(4Dt))

[0054] The ratio g of the water concentration c.sub.1 to the concentration of non-aqueous air components c.sub.2 will be

g=c.sub.1/c.sub.2=c.sub.01/c.sub.02sqrt(D.sub.2/D.sub.1)exp(x.sup.2/(4t)(1/D.sub.11/D.sub.2))

[0055] The permeability P across a membrane is proportional to the diffusion constant D (see exemplarily http://www.tiem.utk.edu/gross/bioed/webmodules/permeability.htmdownloaded on 14 May 2017). If D1 is expressed as

D.sub.1=h*D.sub.2

with the values in FIG. 3, h is about 10.sup.3 to 10.sup.4. Thus

g=c.sub.01/c.sub.02sqrt(1/h)exp(x.sup.2/(4tD.sub.2)(1/h1))

[0056] Neglecting 1/h in the exponent gives

g=c.sub.01/c.sub.02sqrt(1/h)exp(x.sup.2/(4tD.sub.2))

[0057] Comparing a quaterphenyl based carbon nanomembrane to the terphenyl based one, the ratio of the selectivities of the quaterphenyl based carbon nanomembrane to the terphenyl based one g.sub.q/g.sub.t becomes

g.sub.q/g.sub.t=exp(1/(4tD.sub.2)(x.sub.q.sup.2x.sub.t.sup.2)

[0058] Assuming the thicknesses of the two membranes follow x.sub.q=4/3 x.sub.t we get

g.sub.q/g.sub.t=exp(x.sub.t.sup.2/(4tD.sub.2)((4/3).sup.21)=exp(7/9x.sub.t.sup.2/(4tD.sub.2)).

[0059] Since the diffusion length for non-aqueous air components 2 sqrt (D2 t) is very small compared to the thickness x.sub.t of the carbon nanomembrane, this ratio is high and a significant improvement of the selectivity of the quaterphenyl membrane over the terphenyl one can be expected. A corresponding reasoning applies to a comparison of a terphenyl based membrane with a biphenyl based one. Since the mechanical stability of membrane will also increase with the thickness, a terphenyl based membrane is preferred compared to a biphenyl based one, and quaterphenyl based or those made from even longer precursor molecules like polyphenyl compounds are even more preferred.

Example 2

[0060] To explore the morphology of TPT SAMs and CNMs (prepared as in Example 1), scanning tunneling microscopy (STM) and atomic force microscopy (AFM) was employed. (FIGS. 6A-6E). The STM image of TPT SAM (FIG. 6A) was obtained by using a multi-chamber UHV system (Omicron) with a base pressure of 510.sup.11 mbar. The measurement was operated at room temperature.

[0061] The tunneling tip was prepared by electrochemical etching (3 mol.Math.l-1 NaOH solution) of a tungsten wire and further processed in situ by sputtering with Ar.sup.+-ions (pAr=310.sup.10 mbar, E=1 keV, t=1-2 min). The AFM images of TPT SAM and CNM (FIG. 6B) were acquired using an RHK UHV 7500 system (510.sup.11 mbar) with R9 controller.

[0062] The measurements of TPT SAM and TPT CNM were conducted in the non-contact operation mode and the amplitude-modulated tapping operation mode respectively at 93 K using a liquid nitrogen flow cryostat. Before measurements, the TPT SAM samples were first annealed in UHV for 1 h at 323 K and later for 1 h at 333 K for removal of residual adsorbates. The TPT CNM samples were annealed in UHV at 348 K for 30 min. The AFM tips were sputtered with Ar.sup.+-ions at 680 eV for 90 s. For the AFM images, Tap300Al-G force sensors (40 N/m, 280 kHz, Q10000, Budget Sensors) were used. Analysis and post-processing (including corrections for thermal drift and polynomial background subtraction) of the STM and AFM data occurred in the open-source software package Gwyddion (34) (http://gwyddion.net/).

[0063] These pore diameters d.sub.pore were estimated manually by measuring the area of the pores (A.sub.pore) shown in AFM images using the mask drawing tool in Gwyddion. The pore diameter was calculated by assuming that all pores are circular.

[00002]$d_{\mathrm{pore}}=\sqrt{\frac{4.Math.A_{\mathrm{pore}}}{}}$

[0064] The topography of a TPT SAM showed that the TPT molecules were grown in different but highly oriented and densely packed monolayer domains on Au(111) surface (FIGS. 6A, 6C). X-ray photoelectron spectroscopy (XPS) measurements revealed that the TPT molecules were arranged in a densely packed 1.2-nm-thin monolayer. Surprisingly, after electron irradiation, this monolayer structure is completely reorganized; and the resulting CNM contains a dense channel network with an average channel diameter of 0.70.1 nm and an areal density of 10.sup.18 m.sup.2 (FIGS. 6B, 6D, 6E).

[0065] It will be noted that the AFM measurement was conducted in the tapping-operation mode, which is essentially dominated by short-range repulsive interaction forces. See, Garcia et al., Physical Review B 60, 4961 (1999). Since the water permeation in the channel is also affected by (attractive) long-range forces, the effective pore diameter is assumed to be considerably smaller than the 0.7 nm measured in AFM.

[0066] Assuming that all sub-nanometer channels in the CNMs are active in mass transport, the permeation coefficient of TPT CNM was calculated by dividing the measured permeance by the areal density of nanochannels estimated from the AFM images. The single-channel permeation coefficient is approximately 66 water molecules.Math.s.sup.1.Math.Pa.sup.1. This value compares well with the values obtained for carbon nanotubes and aquaporin proteins (FIG. 7). It is known that water molecules confined in sub-nanometer channels form water chains attributed to the strong and short time hydrogen-bonding character between neighboring molecules (see J. Kofinger et al., Physical Chemistry chemical Physics 13, 15403 (2011)), which allows water to rapidly rush through as a single file. This cooperative effect can well explain the unexpectedly high water permeance through CNMs.

[0067] There was no reason for the skilled person knowing the art to expect that the water transport through the pores of a nanometer thin carbon nanomembrane disclosed in this document would resemble more the transport through channels like CNT or aquaporins than the ballistic transport through pores in thin sieves.

Example 3

[0068] The CNMs were prepared from biphenyl-based precursor molecules on aluminized polymer films according to the methods disclosed in international Patent Application No WO2017/072272. Analogue to example 1, these were transferred to a silicon nitride membrane 22 supported by a Si frame 23 where the silicon nitride membrane 22 has a regular hole 24 to form a test sample 28 and characterized accordingly for their water permeation. An average value for the water permeance of 6.510.sup.5 mol m.sup.2 s.sup.1 Pa.sup.1 was measured, which is about half of that of the TPT- and QPT-CNMs in example 1.

Applications

[0069] In addition to the application for the use in a radiation environment mentioned above, the carbon nanomembrane could also be used in clothing, for dehumidification of gas, as well as for dehydration of materials, such as organic materials. It would also be possible to use the carbon nanomembrane for desalination, for example from sea water.

[0070] One application could be for recovery of potable water from a humid atmosphere or from foul water. It would be possible to use the carbon nanomembrane of this document to obtain water from the enclosed atmosphere of a spacecraft. This is useful in space due to the radiation resistance of the carbon nanomembrane. In this latter case, the atmosphere would be the impure fluid 35 and the potable water would be the fluidic water 45 shown in FIG. 1.