PROCESS FOR PRODUCING CARBON FIBER MATERIALS FOR SEPARATING CO2 OR NH3 FROM GAS MIXTURES, CARBON FIBER MATERIALS AND ITS USE

Abstract

A method for producing a carbon fiber material is disclosed, the method comprising the steps of a) Preparation of a solution of polyacrylonitrile in a suitable organic solvent b) Electrospinning of the solution obtained in a) and drying of the obtained fiber material c) Crosslinking of the obtained fiber material by heating to 150 to 350° C. in an air or oxygen atmosphere for 1 to 30 h d) Carbonization of the obtained fiber material in an inert gas atmosphere at a temperature in the range of 500 to 2,500° C., characterized in that no silicon, sulfur, metal compounds, intermetallic compounds, silicon compounds and/or sulfur compounds are added to the polyacrylonitrile solution in step a) and that neither stabilization nor surface modification steps are carried out with the fiber material by treatment with chemical reagents and/or exposure to tensile stress.

Also disclosed is a carbon fiber material obtainable by the above process, as well as a carbon fiber material, characterized in that it has a proportion of ultramicropores V.sub.<0.4 nm of 0.01 to 10, more preferably 0.02 to 5, even more preferably 0.025 to 0.1, most preferably 0.03 to 0.06 cm.sup.3/g, determined by CO.sub.2 adsorption tests and evaluation with DFT and GC-MC simulation.

Claims

1. A method for producing a carbon fiber material comprising the steps of a) Preparing a solution of polyacrylonitrile in a suitable organic solvent, the solution consisting only of solvent and polyacrylonitrile, b) Electrospinning of the solution obtained in a) and drying of the obtained fiber material, c) Crosslinking of the obtained fiber material by heating to 150 to 350° C. in an air or oxygen atmosphere for 1 to 30 h d) Carbonization of the obtained fiber material in an inert gas atmosphere at a temperature in the range of 600 to 900° C. without applying tensile stress.

2. The method of claim 1, wherein the organic solvent is selected from DMF, DMSO, DMAC, acetone, methyl ethyl ketone, and ethanol and mixtures thereof.

3. The method of claim 1, wherein the solvent is DMF.

4. (canceled)

5. The method of claim 1, wherein the polyacrylonitrile is a polyacrylnitrile homopolymer or a polyacrylonitrile copolymer having a comonomer content of from 5 to 10 wt %.

6. The method of claim 1, wherein the polyacrylonitrile has a molecular weight in g/mol of from 50,000 to 200,000.

7. (canceled)

8. (canceled)

9. A use of the carbon fiber material obtained by the process of claim 1 for separating NH.sub.3 from a gas mixture.

10. The use according to claim 9, wherein the gas mixture is exhaust gas from combustion engines, natural gas, biogas or air.

Description

DESCRIPTION OF THE FIGURES

[0049] The relationship between carbonization temperature and adsorption properties for Ar and CO.sub.2 is shown in FIG. 1.

[0050] FIG. 2 and FIG. 4 show the adsorbed volume of Ar and CO.sub.2, respectively, per g of carbon fiber material at different relative pressures as a function of carbonization temperature.

[0051] FIG. 3 shows the dependence of the quotient of adsorbed CO.sub.2 volume and surface area of the material as a function of carbonization temperature. It can be clearly seen how the quotient drops steeply at about 850° C., and hardly changes at all from about 980° C. onwards.

[0052] FIG. 5 shows the cumulative pore volume as a function of pore size for different carbonization temperatures.

[0053] FIG. 6 shows the adsorbed or desorbed volume Ar at different relative pressures for carbonization temperatures of 600° C. and 700° C., respectively.

[0054] FIG. 7 shows SEM images (30,000 magnification, 20 kV) of carbon fibers according to the invention.

[0055] FIG. 8 to FIG. 10 show NH.sub.3 isotherms measured at 240 K, 273 K and 298 K, respectively, on samples of the invention carbonized at different temperatures.

[0056] FIG. 11 shows a comparison of the isotherms of NH.sub.3 and CO.sub.2, each measured at 273 K on samples according to the invention carbonized at different temperatures. The left part of the figure shows the adsorption of NH.sub.3, the right part the adsorption of CO.sub.2.

[0057] FIG. 12 shows the NH.sub.3/N.sub.2 selectivity (IAST) of a gas composition with a molar ratio of 0.001:99.999 (NH.sub.3/N.sub.2) with samples according to the invention carbonized at different temperatures.

[0058] FIG. 13 shows the NH.sub.3/N.sub.2 selectivity (IAST) of a gas composition with a molar ratio of 10:90 with a sample according to the invention carbonized at 1000° C.

[0059] FIG. 14 shows the isosteric adsorption enthalpy of NH.sub.3 measured with samples according to the invention carbonized at different temperatures. The enthalpy of adsorption was determined by the isosteric method. The data used for this purpose were obtained from the isotherms by linear interpolation between two isotherm points.

[0060] FIG. 15 shows the CO.sub.2/N.sub.2 selectivity at a CO.sub.2:N.sub.2 gas volume ratio of 5:95, measured by dynamic gas adsorption for samples carbonized at different temperatures according to the invention.

[0061] FIG. 16 shows CO.sub.2/N.sub.2 selectivity at a CO.sub.2:N.sub.2 gas volume ratio of 10:90, measured by dynamic gas adsorption for samples according to the invention carbonized at different temperatures.

[0062] FIG. 17 shows IAST calculations for CO.sub.2/N.sub.2 selectivity.

DETAILED DESCRIPTION

[0063] The carbon fiber material obtained by the method according to the invention has mainly ultramicropores in the fibers, whereas hardly any meso- or macropores. The carbon fibers are porous with pores of various sizes. According to the invention, the ultramicropores represent the largest proportion of the pores present in the fibers, while fewer meso- or macropores are present. As a result, a large amount of CO.sub.2 or NH.sub.3 can be reversibly adsorbed by the carbon fiber material according to the invention.

[0064] After carbonization in step c), the carbon fiber material is obtained as a nonwoven (nanofiber mat), which can be used directly as adsorption material in gas separation processes without further processing steps.

[0065] The thickness of the nonwoven obtained in step c) is typically 100 to 2000 μm, preferably 400 to 1000 μm, more preferably 600 to 900 μm, particularly preferably 800 μm, depending on the spinning time, when produced on a laboratory scale. However, the size and thickness of the nanofiber mat can be adjusted to much larger values after adjusting the electrospinning equipment.

[0066] According to the invention, the polyacrylonitrile (PAN) that can be used is a polyacrylonitrile homopolymer or a polyacrylonitrile copolymer containing up to about 5 to 10% by weight of one or more comonomers. Suitable comonomers include acrylic acid, methacrylic acid, itaconic acid and/or acrylamide. The polyacrylonitrile is either commercially available (e.g. from sigma-aldrich or BOC Sciences) or can be prepared in the usual way. The molecular weight of the PAN polymer in g/mol is preferably 50,000 to 200,000, preferably 80,000 to 180,000, particularly preferably 100,000 or 150,000.

[0067] Any organic solvent that can dissolve PAN is suitable as a solvent. In particular, a polar solvent such as dimethylformamide (DMF), dimethyl sulfoxide (DMSO), dimethylacetamide (DMAC), acetone, methyl ethyl ketone, ethanol or mixtures thereof are suitable. DMF is particularly suitable.

[0068] According to the invention, the PAN solution in step a) has a concentration of 5 to 20 wt %, preferably 10 to 15 wt %, particularly preferably about 10 wt %. It is prepared by adding the PAN to the solvent and stirring, optionally with heating, until the PAN is completely dissolved (optical control). It is stirred, for example, at room temperature for 2 days.

[0069] Electrospinning is a fiber manufacturing process that uses electrical force to draw charged filaments from polymer solutions or polymer melts up to a fiber diameter of several hundred nanometers. The process does not require the use of coagulation chemistry or high temperatures to produce solid filaments from solution. This makes the process particularly suitable for producing fibers from large and complex molecules. Standard laboratory apparatus for electrospinning consists of a spinneret (spinning die, typically one or more injection needle(s)), for example 4 or 16 needles, connected to a high-voltage DC power supply (e.g., 5 to 100 kV), a syringe pump, and a collector, which may be grounded or used at negative voltages of several kV. A polymer solution, sol-gel, particulate suspension, or melt is loaded into the syringe and this liquid is extruded from the needle tip at a constant rate by a syringe pump. Alternatively, the droplet at the tip of the spinneret can be replenished by feeding from a header tank at constant feed pressure. This constant pressure feed works better for low viscosity feedstocks. When a sufficiently high voltage is applied to a liquid droplet, the liquid body becomes charged and electrostatic repulsion counteracts the surface tension and the droplet is stretched; at a critical point, a liquid stream erupts from the surface. This breakout point is known as the Taylor cone. If the molecular cohesion of the liquid is sufficiently high, no stall occurs (if so, the droplets are electrosprayed) and a charged liquid jet is formed. As the jet dries in flight, the mode of current flow changes from ohmic to convective as the charge migrates to the surface of the fiber. The jet is then extended by an impact process initiated by electrostatic repulsion at small bends in the fiber until it is finally deposited on the grounded collector. The elongation and thinning of the fiber resulting from this bending instability leads to the formation of uniform fibers with nanometer diameters. The collector rotates on one side and the needle is also reciprocated laterally, resulting in a uniform thickness of the fiber mat. Drying of the fiber during spinning results in chaotic movement of the fiber and thus disorderly separation of the spun fiber and formation of a fiber mat. The collector can be, for example, a plate or a drum or also (in large-scale plants) a moving mat.

[0070] In the process according to the invention, electrospinning is carried out according to a standard procedure. Usually, an electrospinner, e.g. from IME Technologies, is used at 10 to 60%, preferably 15 to 35%, particularly preferably 20 or 30% relative humidity and 20 to 30° C., preferably 25° C. However, it has been shown that these environmental parameters have no significant effect on the properties of the carbon fiber material. For example, the flow rate is 20 to 60, preferably 30 to 50, and particularly preferably 40 μl/min per needle used, and the volume is 1.5 to 20 ml, preferably 2 to 15 ml. The spinning time depends on the volume and is, for example, 1 h or 2 h for 2.2 ml and 6 h for 14.6 ml. The voltage is e.g. 21 kV at the anode and −4 kV at the cathode. The collector rotates e.g. at 1000 to 2000 rpm, preferably 1500 rpm. The needle-collector distance is e.g. 100 to 200 mm, preferably 130 to 190 mm, especially 140 or 180 mm. The inner needle diameter is e.g. 0.5 to 1 mm, preferably 0.7 to 0.9 mm, in particular 0.8 mm. The lateral movement of the needle occurs over a total distance of 80 to 150 mm, preferably 90 to 130 mm, particularly preferably 100 or 120 mm. The speed of the lateral needle movement is e.g. 20 mm/s, and the reversal delay is e.g. 500 ms. In this way, PAN nanofiber mats are obtained.

[0071] After electrospinning in step b), the obtained nanofiber mats are preferably first dried, e.g. at 100 to 200° C., preferably about 150° C., for 1 to 24 h, preferably 1 to 2 h, particularly preferably for about 1 h in air atmosphere.

[0072] Subsequently, they are crosslinked in an air or oxygen atmosphere, e.g. at 150 to 300° C., preferably 200 to 300° C., particularly preferably 250° C. for 8 to 30 h, e.g. overnight.

[0073] According to the invention, the decisive step for adjusting the desired properties of the carbon fiber material is the carbonization of the nanofiber mats thus obtained in step c), and in particular the temperature. The temperature is an easily adjustable and controllable parameter, but its influence on the pore structure and gas adsorption properties has not yet been investigated.

[0074] Carbonization is a pyrolysis process. In this process, PAN fibers are transformed by cyclization, dehydrogenation and N.sub.2 elimination, as shown schematically below:

##STR00001##

[0075] The cyclization (step a)) in this case takes place partly already in the crosslinking step b), as explained above.

[0076] Accordingly, depending on the progress of the reactions, nitrogen atoms may be present on the fiber surface in different hybridization states. These can be determined, for example, by XPS (X-ray photoelectron spectroscopy).

[0077] According to the invention, carbonization is carried out, for example, in a furnace at 500 to 2500° C., preferably 600 to 900° C., even more preferably 800 to 875° C., i.e. preferably at comparatively low temperatures of 600 to 900° C., particularly preferably at about 800° C. to 850° C.

[0078] Alternatively, carbonization can also be carried out with laser or IR heating as well as with microwave treatment.

[0079] According to the invention, it is not the type of heating that is important, but rather maintaining the specified temperatures during carbonization, in order to achieve the claimed properties of the fiber material.

[0080] The pore size in the carbon fiber material obtained after carbonization decreases continuously from low carbonization temperatures (about 600° C.) to higher carbonization temperatures (about 1000° C.), as the inventors have found. Above 850° C., the pore size becomes so small that CO.sub.2 or NH.sub.3 can no longer diffuse into the pores, so that a sudden change in the adsorption properties of the material with respect to CO.sub.2 and NH.sub.3 can be observed above this temperature. A temperature of 800 to 850° C. is particularly suitable, since CO.sub.2 is still adsorbed well here, but the pores for N.sub.2 and Ar are already too small.

[0081] The conductivity of the carbon fiber material obtained also depends on the carbonization temperature, as the inventors have found. When carbonized in the temperature range of 800 to 850° C., the electrical conductivity is higher than, for example, when carbonized at 600° C. or 700° C., which makes the material more suitable for use as an electrode.

[0082] According to the invention, the carbonization treatment is carried out under a protective gas (inert gas). Suitable gases include nitrogen, argon or helium, most preferably argon. The inert gas can be introduced into the furnace, for example, and a gas flow can be maintained during the reaction period. Prior to the carbonization treatment, the furnace is preferably evacuated and flooded with the inert gas, e.g. argon, to prevent the introduction of air. The atmosphere in the furnace is as free as possible from O.sub.2 and CO.sub.2 due to an inert gas atmosphere. According to the invention, carbonization is carried out, for example, at atmospheric pressure under a protective gas atmosphere, typically under argon at a flow rate of about 200 l/h, for example. It is expected that the process can be carried out at other flows or pressures without significant effects on the product being observed. The duration of the carbonization treatment is typically 1.5 to 5 h, preferably 2 to 4 h, more preferably 3 h. The heating rate is typically 200 to 400° C./h, preferably 250 to 350° C./h, most preferably 300° C./h. The cooling rate is typically 150 to 250° C./h, preferably 200° C./h. In each case, the heating and cooling times are not included in the specified carbonization time, but are additional.

[0083] The carbon fiber material obtained by the process according to the invention shows flat surfaces of the fiber and, due to the lateral movement of the needle during electrospinning, no preferred orientation (i.e., the fiber is not coiled in an orderly manner or arranged in parallel loops). The smooth surface of the fiber was confirmed by SEM (FIG. 7). The fiber diameter is uniform and decreases with increasing carbonization temperature from about 250 nm to 220 nm. The fiber diameter is estimated by scanning electron microscopy (SEM) and ImageJ® software and is based on average values of 20-30 measurements.

[0084] The pore size of the carbon fiber material is referred to herein as defined in the IUPAC Technical Report (https://doi.org/10.1515/pac-2014-1117; part 2, General definitions and terminology). It is for [0085] ultramicropores <0.7 nm [0086] micropores <2 nm [0087] mesopores 2-50 nm [0088] macropores >50 nm.

[0089] The “kinetic diameter” of gases is understood here as the smallest diameter that a molecule can present to its environment. It differs from the atomic diameter, which indicates the atomic size as the size of the electron shell and which is generally much smaller than the kinetic diameter. There are several ways of determining the kinetic diameter according to the CRS Handbook of Chemistry and Physics, namely from viscosity data, the Van der Waals equation and thermal conductivity. A calculation from bond angles, bond lengths, and the Van der Waals semimeters is also possible. In most cases, the kinetic diameters calculated by the various methods differ by only 2-3%.

[0090] In the context of the present application, the following kinetic diameters for gases reported in the literature are assumed: CH.sub.4 0.38 nm.sup.[1], Ar 0.340 nm.sup.[2], N.sub.2 0.364 nm.sup.[2], CO.sub.2 0.330 nm.sup.[2], H.sub.2 0.29 nm.sup.[1], NH.sub.3 0.260 nm.sup.[1] or, better matching the experimental results, 0.326 nm.sup.[3] ([1] B. Scholes, C. A. et al., Recent Patents on Chemical Engineering, 2008, 1, 52-66; [2] D. W. Beck, Zeolite Molecular Sieves: Structure, Chemistry and Use, Wiley, New York 1974, 593-724, [3] M. E. van Leeuwen, Fluid Phase Equilib., 1994, 99, 1-18).

[0091] According to the invention, the pore size in carbon fiber material can be determined by single gas adsorption measurements of several gases with different kinetic diameters, e.g., Ar, N.sub.2, CO.sub.2, and O.sub.2. The basic principles for the measurement methods are given in the IUPAC Technical Report mentioned above. Accordingly, measurement with argon is suitable for the determination of larger pores, but not for the measurement of ultramicropores. For this purpose, measurement with CO.sub.2 is required.

[0092] For example, the pore size is determined by adsorption of Ar, as shown in FIG. 2. Samples carbonized at 600° C. and 700° C. show very slow adsorption kinetics and pseudoirreversibility, which prevents reasonable measurement (cf. FIG. 6). This indicates pores with very small diameters or strong tortuosity of the pores. For samples carbonized at 800° C. and higher, the adsorption of argon is fully reversible without hysteresis.

[0093] The BET surface area is determined by the standard method of adsorption of argon gas at 87K followed by evaluation of the range of monolayer adsorption according to the BET model. The instrument software ASiQwin® (Quantachrome) is used for this purpose.

[0094] BET surface area measurements with argon are shown in Table 1 below:

TABLE-US-00001 TABLE 1 BET surface area of samples from Ar adsorption experiments. Sample S.sub.BET (m.sup.2/g) C800 13.4 C825 12.4 C850 13.1 C875 13.2 C900 14.6 C925 15.0 C950 14.3 C975 16.0 C1000 14.8 C1100 20.0 C800 denotes a carbon fiber sample carbonized at 800° C.; the other designations are analogous.

[0095] Thus, the BET surface area S.sub.BET of the carbon fiber material of the invention, as determined by argon adsorption, is relatively low and ranges from 12 to 20 m.sup.2/g with a large variation that increases slightly from samples carbonized at 800° C. to samples carbonized at 1100° C. due to the reduction in fiber diameter (thus larger geometric surface area).

[0096] FIG. 4 shows the CO.sub.2 adsorption isotherms of the fibers according to the invention. Unlike the Ar measurements, CO.sub.2 adsorption could be easily measured for the samples carbonized at 600° C. and 700° C., implying a kinetic inhibition of Ar adsorption, since the adsorption of CO.sub.2 at 273 K is much faster than that of Ar at 87 K. The isotherms show a similar shape from C600 to C850, namely a steep slope that flattens at higher pressures. The initial slope is slightly higher for C600 and C700 than for C800 and C850, indicating a higher adsorption energy due to a higher number of narrower pores or a change in surface chemistry. However, the difference is not very large and the total pore volume at 1 bar is almost identical. C900 shows a smaller initial slope but only a slightly reduced amount of adsorbed CO.sub.2 at 1 bar, while the initial slope and total adsorbed volume decrease significantly at carbonization temperatures of 950° C. and above. C1100 achieves a slightly larger adsorbed volume than C1000, which may be attributed to the slightly increased surface area without significant change in porosity or surface chemistry.

[0097] FIG. 3 shows that adsorbed CO.sub.2 volume divided by BET surface area is a measure of surface affinity for CO.sub.2. The value is practically constantly 5 cm.sup.3/m.sup.2 for C800 to C875, and drops steeply to less than 1 cm.sup.3/m.sup.2 at higher carbonization temperatures >950° C. The 5 cm.sup.3/m.sup.2 value is exceptionally high compared to other carbon materials, which typically require a surface area of several hundred m.sup.2 to achieve a similarly high CO.sub.2 adsorption capacity. This phenomenon exhibited by the carbon fiber material of the invention could be attributed to a high chemical affinity for CO.sub.2 or a high ultramicroporous volume, where the ultramicropores are not accessible to Ar and therefore contribute nothing to the BET surface area.

[0098] The textural properties of the carbon fiber materials of the invention derived from CO.sub.2 adsorption tests compared to three known commercially available carbons (carbon blacks) are summarized in Table 2 below:

TABLE-US-00002 TABLE 2 Texture properties derived from CO.sub.2 adsorption experiments Carbonisation S.sub.micro V.sub.<0.4 nm V.sub.>0.4 nm V.sub.ges. temperature (MC) (MC) (MC) (MC) ° C. m.sup.2/g cm.sup.3/g cm.sup.3/g cm.sup.3/g 600 626 0.053 0.119 0.172 700 618 0.051 0.114 0.165 800 615 0.039 0.141 0.180 825 599 0.039 0.133 0.172 850 610 0.034 0.145 0.179 875 656 0.035 0.156 0.191 900 551 0.011 0.179 0.190 925 318 0.004 0.132 0.136 950 274 0 0.118 0.118 975 160 0 0.071 0.071 1000 151 0 0.069 0.069 1100 165 0 0.070 0.070 SuperP 299 0 0.134 0.134 GP300 321 0.004 0.126 0.130 BP2000 1173 0.007 0.470 0.477 S.sub.micro is the surface area of the micropores V.sub.<0.4 nm is the volume of pores with a pore size < 0.4 nm V.sub.>0.4 nm is the volume of pores with a pore size > 0.4 nm V.sub.ges. is the total pore volume E.sub.ads is the asorption energy MC means determined according to a Monte Carlo simulation (Grand Canonical Monte Carlo Simulation (GC-MC)) SuperP is Carbon black SuperP ® of the company Alfaesar GP300 is Graphene platelets from Sigma-Aldrich with a specified BET surface area of 300 m.sup.2/g BP2000 is Black Pearls ® 2000 Carbon Black from Cabot

[0099] The pore volume was calculated from GC-MC and shown divided into pores larger and smaller than 0.4 nm. The pore volume <0.4 nm decreases with increasing carbonization temperature, while the total pore volume increases slightly from C600 to C900, as does the micropore surface area. C1000 shows no pores below 0.4 nm and a sharp decrease in total pore volume and surface area.

[0100] The commercial carbon materials show no pores <0.4 nm at all, which explains their low CO.sub.2 uptake at low pressures. In contrast, the carbon fiber material according to the invention shows a high percentage of ultramicropores V.sub.<0.4 nm of 0.01 to 10, more preferably 0.02 to 5, even more preferably 0.025 to 0.1, most preferably 0.03 to 0.06 cm.sup.3/g, determined by CO.sub.2 adsorption and evaluation with DFT and GC-MC simulation.

[0101] FIGS. 8 to 10 show adsorption isotherms of ammonia measured on different samples according to the invention at different temperatures.

[0102] Surprisingly, NH.sub.3 and CO.sub.2 were found to show the same tendency with the materials according to the invention: very high adsorption was obtained at C600 to C800. C900 proved to be kinetically hindered. At C1000, only a low adsorption of the respective gas was observed. NH.sub.3 as a Lewis base and CO.sub.2 as a Lewis acid require different surface chemistry for effective adsorption. It is very unlikely that chemically different molecules will exhibit the same adsorption behavior due to changes in the surface chemistry of the adsorbent. Therefore, in the materials according to the invention, the surface structure or chemistry is not the decisive factor for adsorption, but the accessibility of the ultramicropores.

[0103] Comparison of the adsorption of NH.sub.3 and CO.sub.2 (FIG. 11) shows that at C600 to C800 NH.sub.3 is adsorbed in greater amounts than CO.sub.2. However, at C1000, CO.sub.2 is adsorbed more than NH.sub.3. This could be due to the fact that NH.sub.3 has access to a larger number of pores, the pores can be packed more densely with NH.sub.3, i.e., more molecules of NH.sub.3 fit into the same pore than molecules of CO.sub.2, or that there is a different surface chemistry in the pores.

[0104] It is known from the literature (M. Goncalves et al., Environ. Sc. Technol. 2011, 45, 10605-10610) that the NH.sub.3 adsorption capacity increases linearly with the amount of oxygen surface groups. However, no correlation between surface chemistry (oxygen or nitrogen) and NH.sub.3 adsorption capacity was observed in the samples according to the invention, as shown in the following tables.

[0105] The elemental composition of the samples was determined by elemental analysis and XPS and is summarized in the following tables:

TABLE-US-00003 TABLE 3 CHNO analysis Sample C (wt.-%) N (wt.-%) O (wt.-% H (wt.-%) C600 63.5 23.0 11.8 2.1 C700 67.9 19.9 11.1 1.8 C800 72.6 16.2 10.6 1.4 C900 77.6 11.6 9.5 1.2 C1000 91.5 7.1 3.1 0.4 C1100 95.8 3.7 0.8 0.5

TABLE-US-00004 TABLE 4 XPS analysis Sample C (at.-%) N (at.-%) O (at.-%) C600 76.5 19.7 3.8 C700 80.4 16.8 2.8 C800 84.1 13.7 2.2 C900 86.6 11.0 2.4 C1000 92.6 4.3 3.1 C1100 95.1 3.1 1.8

[0106] This represents a further indication that, according to the invention, the pore size is the determining factor for the adsorption capacity, and not the surface chemistry.

[0107] For the fiber materials of the invention, the adsorption data correlate strongly with the kinetic diameter of the adsorbate.

[0108] The correlation between kinetic diameter and adsorption capacity of the samples according to the invention is shown in the following table:

TABLE-US-00005 TABLE 5 Kinetic diameter and adsorption capacity.   Gas N.sub.2 Ar CO.sub.2 NH.sub.3 Kinetic diameter (pm) 364 340 330 260/326 Samples with high adsorption capacity 0600*-0700*  0600-0900*  0600-0900* [00002] *kinetically hindered

[0109] The arrow on the right means that the adsorption capacity increases from top to bottom.

[0110] The smaller the molecules or atoms, the smaller the pores into which they can penetrate. The pore size decreases with increasing carbonization temperature. Therefore, the adsorption capacity changes drastically from one sample to another if molecules can no longer penetrate the pores. Therefore, the kinetic diameter and the pore size effect consistently explain the adsorption behavior of the nanofiber mats according to the invention.

[0111] 0.26 nm is a value often reported in the literature as the kinetic diameter for NH.sub.3. However, it has been shown that this value cannot adequately explain experimental data, especially for pore diffusion. In contrast, the value determined by van Leeuwen (0.326 nm) is in good agreement with the findings of Kanezashi et al. and is therefore used here. (M. Kanezashi, A. Yamamoto, T. Yoshioka, T. Tsuru, A. I. Ch. E., 2010, 56, 1204-1212, M.E. van Leeuwen, Fluid Phase Equilib., 1994, 99, 1-18).

[0112] CO.sub.2 and NH.sub.3 show similar behavior because they have comparable molecular size. CO.sub.2 and NH.sub.3 have very slow adsorption kinetics on C900 (very small pores). However, surface chemistry has only a minor effect on adsorption.

[0113] NH.sub.3 is adsorbed more strongly than CO.sub.2 (C600 to C800) on samples with a larger number of accessible pores. CO.sub.2, in contrast, is slightly favored when pores are not accessible (C1000). Surface chemistry favors CO.sub.2 adsorption somewhat, at least for C1000.

[0114] FIG. 14 shows the isosteric adsorption enthalpies of NH.sub.3 on the samples. Higher values indicate stronger adsorption of NH.sub.3 on the sample.

[0115] IAST (ideal adsorbed solution theory) calculations predict excellent selectivity of the materials of the invention for NH.sub.3 over N.sub.2. The results are shown in FIGS. 12 and 13.

[0116] However, the calculations should only be considered as approximations, since strongly different adsorbing components can lead to deviations in IAST calculations. The IAST results can be well explained by a molecular sieve effect. C600 to C800 have pores that are accessible to NH.sub.3 but not to N.sub.2. This results in high values for selectivity. At C1000, i.e. for a material without accessible pores, the selectivity for NH.sub.3 decreases strongly.

[0117] FIGS. 15 and 16 show measured values for CO.sub.2/N.sub.2 selectivity that are in agreement with IAST predictions (about 25-55 for C600-C800). To the knowledge of the inventors, these values are among the highest measured to date for CO.sub.2/N.sub.2 on carbon-based samples.

[0118] For comparison, IAST calculations on CO.sub.2/N.sub.2 selectivity are shown in FIG. 17. These are slightly higher than the measured values (high double-digit to low triple-digit range).

[0119] The carbon fiber material according to the invention is particularly applicable for the separation of CO.sub.2 or NH.sub.3 from gas mixtures. The gas mixtures can be, for example, air, combustion exhaust gases, natural gas, biogas or other gas mixtures.

[0120] H.sub.2 can also be separated from suitable gas mixtures (e.g. CH.sub.4/H.sub.2) by means of the carbon fiber materials according to the invention due to its small kinetic diameter.

[0121] The materials according to the invention can advantageously be used as adsorbents in the Pressure Swing, alternatively in the Temperature Swing as well as, with particular advantage due to the especially high selectivity at low pressures, in the Vacuum Swing adsorption process.

[0122] The carbon fibers of the invention, especially those carbonized at 1000° C. (C1000), can also adsorb large amounts of water. They could therefore be used advantageously for drying gases.

[0123] For the same reason, before separating CO.sub.2 or NH.sub.3 from gases, it is important to dry these gases, e.g. by cooling, since water vapor competes with CO.sub.2 or NH.sub.3, respectively, and would therefore reduce the separation efficiency of the membrane.

EXAMPLES

Synthesis of Carbon Fibers

Example 1

[0124] PAN from sigma-aldrich or BOC Sciences with a molecular weight of 150,000 g/mol was used.

[0125] Carbon fibers were prepared by electrospinning. First, solutions of 10% polyacrylonitrile (PAN) in dimethylformamide (DMF) were prepared. For this purpose, 4 g of PAN was dissolved in 36 g of DMF under stirring with a magnetic stirrer at room temperature for two days. The different samples were prepared from the same solution. The solution was spun directly into fiber mats in a controlled atmosphere by an electrospinner from IME Technologies. The parameters shown in Table 6 were used for this purpose.

TABLE-US-00006 TABLE 6 Standard parameters for elektrospinning Climate Temperature 25° C. Humidity 20% Material Flow rate 40 μL/min Volume 2.2 mL Spinning time 1 h/2 h PAN wt % 10 Lateral Movement Total distance 100 mm Start position −50 mm Distance 100 mm Speed 20 mm/s Reversal delay 500 ms Distances Needle-collector 140 mm Inner needle diameter 0.8 mm Collector movement Rotation 1500 Upm Voltage Positive electrode 21 kV Negative Electrode −4 kV

[0126] After the electrospinning process, the mats were dried in the oven in air at 150° C. overnight and then crosslinked in air atmosphere at 250° C. overnight. Finally, after reaching the target temperature, the crosslinked mats were carbonized for 3 hours at 1000° C. in the previously evacuated and argon-flooded oven and then cooled. The heating rate was 300° C./h, the cooling rate 200° C./h, and the argon flow rate 200 L/h.

Example 2

[0127] A 10 wt % solution of PAN (Mw 150,000 g/mol) in DMF was prepared by adding 8 g PAN to 72 g DMF and stirring at room temperature for two days. This solution was used in the electrospinning process to prepare PAN nanofiber mats. The process was carried out in a controlled atmosphere with the parameters given in Table 7:

TABLE-US-00007 TABLE 7 Parameters for the electrospinning process using a 10 wt % solution of PAN in DMF Climate Temperature 25° C. Humidity 30% Material Flow rate 40 μL/min Volume 14.6 mL Spinning time 6 h PAN wt % 10 Lateral movement Total distance 120 mm Starting position −60 mm Distance 120 mm Speed 20 mm/s Reversal delay 500 ms Distances Needle-collector 180 mm Inner needle diamater 0.8 mm Collector diameter 60 mm Collector movement Rotation 1500 rpm Voltage Positive electrode 21 kV Negative electrode −4 kV

[0128] Subsequently, the obtained PAN nanofiber mats were dried at 150° C. for 1 h. Crosslinking was carried out in an air atmosphere at 250° C. overnight (18 h). For carbonization, the samples were heated at a heating rate of 300° C./h under a stream of argon (200 L/h) until the desired carbonization temperature was reached, which was maintained for 3 h. The cooling rate was 200° C./h. The carbonization temperature was maintained until the desired carbonization temperature was reached. The cooling rate was 200° C./h. Samples for the 600 to 1000° C. temperature series (C600 to C1000) were prepared from the same solution.

Analytics

REM

[0129] SEM images were acquired using an ETD detector on an FEI Quanta® FEG 650 with a magnification of 30,000×. The accelerating voltage was 20 kV. During sample preparation, the samples were glued to the sample holder and grounded with a copper strip. The diameters of the fibers were estimated using the ImageJ® program. The average diameter of a sample is based on a measurement of 20 to 30 fiber diameters.

Gas Adsorption

[0130] Gas adsorption experiments were performed on a Quantachrome Autosorb® iQ2 instrument equipped with a cryocooler. For preparation, samples were cut into small strips approximately 1 mm wide, of which 30-100 mg were transferred to the sample tube. The samples were then baked at 300° C. for 8 hours. To determine the sample mass, both the empty sample tube and the tube filled with sample were weighed three times each after bake-out. The weight of the sample was calculated from the difference in the average masses. The gas adsorption measurements with the gases argon, nitrogen and oxygen were carried out in different sequences on the same sample. Several measurements were taken at different temperatures for each gas and sample. The order of the measurements was not determined systematically and was based on metrological considerations at the time of the measurement.

Dynamic Gas Adsorption

[0131] Dynamic gas adsorption experiments were performed with a mixSorb® SHP from 3P Instruments. Control and evaluation were performed using mixSorb Manager® software (version 1.2.3.0) and 3P Sim® (version 1.0.7).

[0132] The composition of the injected gases was adjusted by means of several mass flow controllers (MFC). Measurement of the gas composition downstream of the sample column was performed using a Pfeiffer mass spectrometer (MS) (Pfeiffer Omnistar® GSD 320). The sample column was temperature controlled using either a water/ethylene glycol bath (−10 to 70° C.) or a heating jacket (90 to 400° C.). The stainless steel sample column is filled with sample to a height of 8.5 cm, has an inner diameter of 0.45 cm and a sample volume of about 1.0 cm.sup.3. A temperature sensor is inserted 2.5 cm into the bulk from the top.

[0133] Before a measurement, 110-160 mg of the corresponding sample was first cut into strips about 1 cm long and 1 mm wide and introduced into the sample column using a funnel. Voids were filled by careful stuffing with a rod. The empty weight and the weight of the sample column with sample were each determined three times.

[0134] The first heating of a sample was performed at 300° C. for 2 h at an applied vacuum and a flow of the carrier gas (helium) of 7 mL/min to remove adsorbed water and other gases. Further bakeout after a measurement sequence was performed for 1 h at 70° C., with vacuum applied and a flow of carrier gas of 7 mL/min.

[0135] To start a measurement, the specified pressure was built up with the carrier gas helium. When temperature as well as pressure were stable, the experiment was started by adding the adsorptives CO.sub.2 and N.sub.2 to the sample column. The corresponding parameters can be taken from Table 8. After completion of a breakthrough experiment, the next breakthrough experiment of the sequence was started by increasing the adsorptive fraction to the specified value of the next measurement point. After the end of a sequence, bakeout was performed as described and the next sequence was performed analogously.

TABLE-US-00008 TABLE 8 Parameters of the measurement sequences. Helium was used as carrier gas, CO.sub.2 and N.sub.2 as adsorptive gas. Total pressure Temperature 1. measuring 2. measuring [bar] [° C.] point point 1. Sequence 5.0 0 CO.sub.2: 0.5% CO.sub.2: 1.0% N.sub.2: 9.5% N.sub.2: 19.0% He: 90.0% He: 80.0% 2. Sequence 5.0 0 CO.sub.2: 1.0% CO.sub.2: 2.0% N.sub.2: 9.0% N.sub.2: 18.0% He: 90.0% He: 80.0%

[0136] For each sequence, a reference measurement was also taken using glass beads (0.4-0.6 mm diameter) as filler material. The adsorbed amount of CO.sub.2 and N.sub.2 was obtained from the difference between the sample and reference measurements.

[0137] The selectivity was determined with the following formula:

[00001]$S=\frac{n_{{\mathrm{CO}}_2}}{n_{N_2}}.Math.\frac{y_{N_3}}{y_{{\mathrm{CO}}_2}}$ [0138] n.sub.i—adsorbed amount of substance i in the adsorbed phase [0139] y.sub.i—Volume amount of i in the gas phase

BET

[0140] The quantitative evaluation of the measured gas adsorption isotherms according to the BET method was performed using the function implemented in the ASiQwin® (Quantachrome) instrument software. The range of values used for the evaluation was determined by the “Micropore BET assistant” available in the software in order to achieve a reproducible point selection according to the recommended criteria. Since the parallel measurement of the saturation vapor pressure (actually recommended.sup.2 for highest accuracy especially for the determination of mesopores) with the use of the cryocooler means giving up half of the measurement capacity and is not possible for vapor pressures >1 bar/760 Torr anyway, the saturation vapor pressures for all measurements were taken from the NIST database. They are summarized in Table 9 below. For adsorption methods, Torr is the usual unit, so that this unit and not the corresponding SI unit is used here.

TABLE-US-00009 TABLE 9 Saturation vapor pressures of the gases argon, nitrogen, oxygen and carbon dioxide as a function of temperature. Temperature [K] P.sub.0(Ar) [Torr] P.sub.0(N.sub.2) [Torr] P.sub.0(O.sub.2) [Torr] 77.35 212.3*  759.55 155.61 87 771.94** 2072.1 537.44 273 CO.sub.2: 26 141 *own measurement, because below the triple point (only literature values available for solid or gaseous phase) **value for 87.45 K

[0141] Further, the following values for the area of a gas molecule were used to calculate BET surface areas:

TABLE-US-00010 TABLE 10 used values for the cross-sectional area of argon, nitrogen and oxygen on a surface according to the BET model. Argon @ Nitrogen @ Oxygen @ Gas 87 K 77 K 77 K area of a gas molecule [A.sup.2] 14.2 16.2 14.1

DFT and GC-MC

[0142] The evaluation of the measured isotherms according to the DFT method (Density Functional Theory) and the GC-MC method (Grand Canonical Monte Carlo) was performed by functions implemented in the Quantachrome ASiQwin® software (version 5.0). The given model for carbon dioxide at 273 K was used for slit pores. For Ar at 87 K, the QS-DFT model for slit pores was used. The carbon reference stored in the software was selected as the material.

CHNO

[0143] A Varion EL cube elemental analyzer (Elementar, Germany) was used for elemental analysis. 2 mg sample of each fiber material was combusted in CHN mode, and 10 mg sample was combusted in O mode. In CHN mode, the samples were burned, and the combustion products were separated and detected. In O mode, the pores were treated in reducing atmosphere, in which O-containing fragments were converted to CO and detected as such. This procedure was repeated three times for each sample in each mode. Polyethylene was added to the samples carbonized at 900° C. for better combustion.

XPS-Analysis

[0144] XPS analyses were used with a Phi5000 VersaProbe II (ULVAC-Phi Inc., US). Monochromatic 1.486 keV Al.sub.Kα radiation was used for each measurement. Peak analysis was performed with using CasaXPS with Shirley background and instrument-specific corrections. The spectra were calibrated to 284.4 eV using the C1s signal.