Selective Permeation of Hydrogen Gas Using Cellulose Nanofibril Film

Note pubs.acs.org/Biomac

Selective Permeation of Hydrogen Gas Using Cellulose Nanofibril Film Hayaka Fukuzumi, Shuji Fujisawa, Tsuguyuki Saito, and Akira Isogai* Department of Biomaterials Sciences, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Tokyo 113-8657, Japan ABSTRACT: Biobased membranes that can selectively permeate hydrogen gas have been developed from aqueous dispersions of 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO)-oxidized cellulose nanofibrils (TOCN) prepared from wood cellulose: TOCN-coated plastic films and self-standing TOCN films. Compared with TOCNs with sodium, lithium, potassium, and cesium carboxylate groups, TOCN with free carboxyl groups (TOCN-COOH) had much high and selective H2 gas permeation performance. Because permeabilities of H2, N2, O2, and CO2 gases through the membranes primarily depended on their kinetic diameters, the gas permeation behavior of the various TOCNs can be explained in terms of a diffusion mechanism. Thus, the selective H2 gas permeability for TOCN-COOH was probably due to a larger average size in free volume holes present between nanofibrils in the layer and film than those of other TOCNs with metal carboxylate groups. The obtained results indicate that TOCN-COOH membranes are applicable as biobased H2 gas separation membranes in fuel cell electric power generation systems.

■

INTRODUCTION Gas barrier films have been used for packaging materials, electronic components, and back sheets for solar cells.1−4 Control of gas permeability using polymer and polymer composite membranes is also significant for practical application in gas separation devices.5−8 The carbon dioxide, nitrogen, oxygen, and other gas separation behavior of carbon nanotube/polymer composites and their corresponding gas separation mechanisms have recently been studied.9,10 Gas separation using membranes can be regarded as an energyefficient technique applicable to traditional gas production processes, and some polymer composite membranes are currently used for purification of nitrogen from air, removal of carbon dioxide from natural gas, and recovery of hydrogen from gas mixtures.9 In particular, efficient H2 gas separation using membranes has become increasingly required in the past few years as electric power generation systems based on fuel cells using the chemical reaction between H2 and O2 have seen remarkable development.11,12 Recently, self-standing and layered films consisting of nanosized cellulose fibrils have gained a great deal of interest as oxygen barrier materials.13 In particular, those consisting of surface-charged nanofibrillated celluloses have shown quite low oxygen permeabilities under dry conditions,14,15 superior to those of ethylene-vinyl alcohol copolymers used as commercial oxygen-barrier films. In our previous studies, 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO)-mediated oxidation of wood cellulose was used for position selective oxidation of C6−OH groups exposed on crystalline cellulose microfibril surfaces to C6-carboxylates to prepare TEMPO-oxidized cellulose nanofibrils (TOCNs).16−18 The TOCNs prepared from wood © XXXX American Chemical Society

cellulose are 3−4 nm in width and generally >500 nm in length. The average density of C6-carboxylates present on the nanofibril surfaces is approximately 1.7 groups nm−2.18 TOCN/water dispersions have been used to prepare selfstanding TOCN films and TOCN-coated plastic films, which have extremely high oxygen barrier properties under dry conditions.15,19,20 A recent study using positron annihilation lifetime spectroscopy revealed that self-standing wood TOCN films have quite small free volumes uniformly present from the film surface to inside, with an average hole diameter of 0.47 nm.21 Such small hole sizes and the highly crystalline structures of TOCN films having sodium C6-carboxylate groups (TOCNCOONa films) bring about their extremely high oxygen-barrier properties under dry conditions. Thus, biobased cellulose nanofibril films have potential application as environmentally friendly alternatives to conventional petroleum-based polymer gas barrier membranes. In this study, we demonstrate the gas permeabilities of selfstanding TOCN and TOCN-coated plastic films for not only oxygen but also nitrogen, carbon dioxide, hydrogen, and their mixtures, to investigate the gas permeation selectivity of TOCNs having either metal carboxylates or free carboxyl groups. It was found that TOCN films having free carboxyl groups (TOCN-COOH) had excellent H2 gas permeation selectivity than other gases. Received: March 12, 2013 Revised: April 16, 2013

A

dx.doi.org/10.1021/bm400377e | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

■

Note

spectra of the self-standing TOCN films were recorded using a Jasco FT/IR-6100 spectrometer in attenuated total reflection (ATR) mode from 400 to 4000 cm−1 at 4 cm−1 resolution. Tensile tests of selfstanding TOCN films ∼13 μm in thickness were carried out at 23 °C and 50% relative humidity using a Shimadzu EZ-TEST tensile tester equipped with a 500 N load cell. Specimens 20 mm and 2 mm in length and width, respectively, were measured for a 10 mm span length at a speed of 1.0 mm min−1. At least five specimens were measured for each sample.

MATERIALS AND METHODS

Preparation of TOCN/Water Dispersions. Aqueous dispersions of TEMPO-oxidized cellulose nanofibrils with sodium carboxylate groups (TOCN-COONa) and free carboxyl groups (TOCN-COOH) were prepared according to a previously reported method.19 Briefly, TEMPO-mediated oxidation was applied to a never-dried softwood bleached kraft pulp containing ∼90% α-cellulose using a TEMPO/ NaBr/NaClO system in water at pH 10 with 10 mmol g−1-pulp NaClO. A 0.1% (w/v) slurry of TEMPO-oxidized cellulose in water (250 mL) was first disintegrated with a double cylinder-type homogenizer (Physcotron, Microtec Nition Co. Ltd., Japan) at 7500 rpm for 30 s and then with an ultrasonic homogenizer (probe tip diameter: 26 mm; US-300T, Nissei) at a frequency of 19.5 kHz and an output power of 300 W for 8 min. The TOCN-COONa/water dispersion thus obtained was centrifuged with an Avanti HP-26 XP Centrifuge (Beckman Coulter, Inc.) at 10 400g for 10 min to remove the small amounts of unfibrillated and partly fibrillated fractions. TOCN-COOH was prepared as follows: 0.1 M HCl was slowly added to the 0.1% TOCN-COONa/water dispersion under magnetic stirring to set the pH to ∼2.0. After stirring for 30 min, the TOCNCOOH gel particles thus formed were sequentially washed with 0.01 M HCl (once) and distilled water (three times) by repeated centrifugation until the supernatant became neutral pH. The washed gel particles were then suspended in water at 0.1% (w/v) and sonicated for 1 min under the same conditions described above, to obtain a transparent TOCN-COOH/water dispersion at pH of 4−5. The aqueous TOCN-COOLi dispersion was prepared from the TOCN-COOH/water dispersion by slowly adding 0.05 M LiOH until the pH reached 8.0. The aqueous TOCN-COOK and TOCN-COOCs dispersions were prepared as follows: 100 mM KCl or CsCl was added to a 0.1% (w/v) slurry of TEMPO-oxidized cellulose in water (250 mL) and stirred for 2 h. The slurry was then washed with distilled water three times using a centrifuge to remove the excess salt. The TEMPO-oxidized celluloses with potassium and cesium carboxylates were suspended in water (250 mL) at 0.1% (w/v) solid content and sonicated for 8 min to prepare their aqueous nanodispersions at pH of 6−7. Film Preparation. Commercial poly(ethylene terephthalate) films (PET, thickness 50 and 12 μm, area 9 cm × 9 cm) were hydrophilized by plasma sputtering at 5 mA for 5 min in vacuum using a soft-etching device (SEDA-PFA, Meiwafosis). The 0.1 (w/v) % TOCN/water dispersions (5 mL) were uniformly cast onto the hydrophilized surface of the PET film and allowed to dry at room temperature for 2 days. The 50 μm thick-PET films were used for coating TOCN-COONa and TOCN-COOH dispersions, and the 12 μm-thick PET films were used for coating TOCN dispersions with other metal carboxylate groups. Self-standing TOCN-COONa films were prepared by casting the dispersion onto Petri dishes and drying at 40 °C for 1 day. Some TOCN-COONa films were soaked in 0.01 M HCl for 0.5−30 min, washed thoroughly with distilled water, and dried at room temperature to obtain self-standing TOCN-COOH films. A TOCN-COONa film soaked in distilled water was prepared in the same procedure and used as a reference sample. Analyses. Gas permeation tests were performed using a differential-pressure method (GTR-31A, GTR Tec Co., Japan) based on JISK7126A and ASTMD1434 standards. Air was used as a N2/O2 mixed gas, and H2, N2, or CO2 was used as a single gas in the permeation test. Each film was degassed in the chamber of the test system for 0.5 h before starting the measurement. Each gas passed through the films was determined using a gas chromatography attached with a thermal conductivity detector (G2700, Yanaco, Japan), which was equipped with the GTR-31A. The light transmittance of the self-standing TOCN films and the light reflectance of the TOCNcoated PET films were measured using a Shimadzu UV−vis−NIR spectrophotometer (Jasco V-670, Japan). Thicknesses of the selfstanding TOCN films and TOCN-coated PET films were calculated from the UV−vis interference spectra using an attached analytical program. Here, refractive indices of 1.55 and 1.65 were used for TOCN and PET, respectively.19 Fourier transform infrared (FTIR)

■

RESULTS AND DISCUSSION TOCN/water dispersions having sodium carboxylate groups (TOCN-COONa) and free carboxyl groups (TOCN-COOH) were prepared, the latter of which was prepared by acidifying the TOCN-COONa/water dispersion and redispersing the resulting gels in water by sonication. Because the pKa value of free C6-carboxyl groups are ∼3.6, the TOCN-COOH nanofibrils were dispersible in pure water of pH ∼4.5 through the dissociation of almost all of the C6-carboxyl groups of the TOCN-COOH under the conditions used.19 First, TOCN-coated films were prepared by coating the aqueous 0.1 (w/v) % TOCN-COONa and TOCN-COOH dispersions on PET films and then drying. The N2, O2, CO2, and H2 transmission rates of the TOCN-coated PET films were then measured by the differential pressure method. Gas permeability coefficients of TOCN-COONa and TOCNCOOH layers coated on PET films were calculated using the following equation: T (TOCN − PET film) P(TOCN − PET film) T (PET film) T (TOCN film) = + P(PET film) P(TOCN film)

(1)

where T and P are the thickness and gas permeability coefficient, respectively, of the PET film, TOCN layer, or TOCN-coated PET film. Figure 1a shows various gas permeabilities through TOCNCOONa and TOCN-COOH layers coated on PET films. In the case of N2, O2, and CO2, the difference in gas permeability between TOCN-COONa and TOCN-COOH layers was small, and each of the values was close to the detection limit of the test apparatus used. However, the H2 permeability of the TOCN-COOH layer was 1 order of magnitude greater than that of the TOCN-COONa layer. Figure 1b depicts the gas permeabilities plotted against the kinetic diameters of the permeant gases. The size of the permeant gas correlated well with the corresponding gas permeability; the smaller the kinetic diameter of the gas molecule, the higher the gas permeability of both TOCN layers. In general, the gas permeability behavior of nonporous polymer films is explained by the solubility and diffusivity of the permeant gas in the films.22 The good correlations between the kinetic gas diameter and gas permeability in Figure 1 indicate that the gas permeation behavior through the TOCN layers is primarily explained by the diffusion mechanism rather than the gas solubility in the layers.23 Because the CO2 permeability for the TOCN-COONa layer was relatively higher than the value on the correlation line, the solubility factor of CO2 in the TOCN-COONa layer may not be negligible. The relative permeability ratio between gases A2 and B2, calculated by P(A2)/P(B2), of the TOCN-COONa and TOCN-COOH layers, and cellophane, PET, and poly(ethylene) (PE) films are summarized in Table 1. The H2/ B

dx.doi.org/10.1021/bm400377e | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Note

Figure 2. H2 permeability and relative H2/N2 permeability ratio for TOCN-COOH, TOCN-COOLi, TOCN-COONa, TOCN-COOK, and TOCN-COOCs layers coated on PET films.

Thus, the significantly high H2 permeability for the TOCNCOOH layer was not due to the size of counterions of carboxylate groups but probably ascribed to packing structures of TOCN elements in the layers. The gas molecules are likely to diffuse through the free volume holes present between nanofibrils in the TOCN layers under dry conditions.21 Because TOCNs with metal carboxylate groups formed close packing structures in the layers due to electrostatic repulsions efficiently working between TOCN elements not only in dispersion states but also during drying process of the dispersions, thus resulting in smaller free volume holes and also lower H2 gas permeabilities for these TOCN layers. In contrast, the average size of free volume holes present between nanofibrils in the TOCN-COOH layer may be somewhat larger, because interfibrillar hydrogen bonds prevailing against interfibrillar electrostatic repulsions are preferably formed in the dried layer. Thus, the H2 molecules having the smallest kinetic diameter in the gases tested here are permeable more easily through the paths between nanofibrils in the TOCN-COOH layer. Similar results were obtained for self-standing TOCNCOONa and TOCN-COOH films. The as-prepared selfstanding TOCN-COONa film soaked in 0.01 M HCl for 0.5− 30 min to convert the sodium carboxylate groups to free carboxyl groups, washed thoroughly with water, and then dried at room temperature (Figure 3a). Figure 3b shows the FTIR spectra of the treated self-standing TOCN films. The absorption peak intensity at 1600 cm−1, due to the CO stretching vibration of carboxylate groups in the TOCNCOONa film, gradually decreased with increasing soaking time. This peak completely disappeared after 5 min soaking treatment. The band around 1630 cm−1 was due to the bending vibration of water molecules present in a small quantity in the films. The peak intensity at 1720 cm−1 due to the CO stretching vibration of free carboxyl groups increased correspondingly. Moreover, the absorption around 1405 cm−1 due to the C−O stretching of dissociated carboxyl groups in the TOCN-COONa film disappeared after 5 min soaking treatment.19 Thus, the sodium carboxylate groups in the TOCNCOONa film were completely converted to free carboxyl groups to form TOCN-COOH by the simple soaking treatment in 0.01 M HCl for 5 min or more. Light transmittance of the self-standing TOCN-COONa film was slightly decreased by soaking in water or 0.01 M HCl, probably because of a slight increase of free volume size in the

Figure 1. (a) Gas permeabilities of PET films coated with TOCNCOONa or TOCN-COOH. (b) Relation between kinetic diameters of the permeant gases and their permeabilities through the TOCNcoated PET films.

Table 1. Relative Gas Permeability Ratios of TOCN and Commercial Films TOCN-COONa TOCN-COOH cellophane PET PE

H2/N2

H2/CO2

H2/O2

O2/N2

CO2/N2

350 2200 220 190 8.5

12 24 14 5.3 0.58

49 290 39 30 3.0

7.2 7.4 5.6 6.2 2.8

29 92 16 35 15

N2 permeability ratio of the TOCN-COOH layer was 2200, which is much higher than that of the TOCN-COONa layer or the other films, and is also the highest value for any polymer film reported so far.24 Thus, the TOCN-COOH layer had sufficiently high hydrogen gas permeation capability in comparison with N2 gas. Moreover, the TOCN-COOH layer had relatively high relative permeability ratios for H2/O2 and CO2/N2 compared with those of the TOCN-COONa layer or the other films in Table 1. The relative gas permeability ratios for the TOCN-COONa layer were similar to those of cellophane film for all gases examined. Figure 2 shows the H2 permeability and relative H2/N2 permeability ratio for TOCN layers, the counterions of which had H+, L+, Na+, K+, and Cs+. FTIR analysis of the corresponding self-standing films showed that all carboxyl groups in each TOCN had the corresponding metal ions or protons when the procedures described in the experimental section were used. Except for the TOCN-COOH layer, both H2 permeability and H2/N2 permeability ratio were similarly low for the TOCNs with carboxylate monovalent metal ions. C

dx.doi.org/10.1021/bm400377e | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Note

Figure 3. (a) Scheme to showing the conversion of TOCN-COONa to TOCN-COOH self-standing film. (b) FTIR spectra of TOCN-COONa films soaked in 0.01 M HCl for 0.5−30 min. (c) Light transmittance of TOCN-COONa and TOCN-COOH films. (d) Elastic modulus (black bar) and tensile strength (gray bar) of TOCN-COONa and TOCN-COOH films. (e) H2 permeability of TOCN-COONa and TOCN-COOH selfstanding films.

■

treated films due to swelling. Nevertheless, all TOCN films had high light transmittance; more than 85% at 600 nm wavelength (Figure 3c). Elastic modulus and tensile strength of the TOCNCOONa film slightly decreased after soaking treatment in water or 0.01 M HCl, again indicating the occurrence of partial cleavage of interfibrillar hydrogen bonds originally present in the TOCN-COONa film by swelling. Figure 3e shows the H2 permeability of the self-standing TOCN-COONa and TOCN-COOH films. The H2 permeability was increased by 1 order of magnitude by the 0.1 M HCl soaking treatment. This result was quite similar to that observed for TOCN-COONa and TOCN-COOH coated on PET films (Figure 1). Therefore, the different H2 permeabilities between TOCN-COONa and TOCN-COOH layer-coated PET films were also evaluated for self-standing films. The gas transmission rates of other gases (N2, O2, and CO2) could not be measured for the thick self-standing films, because the values were too low to be detected by the test apparatus. Thus, only H2 gas is permeable by TOCN-COOH films or TOCNCOOH-coated polymer films using their relatively high H2 transmission rates.

CONCLUSIONS

New biobased materials that can selectively permeate H2 gas compared with N2, O2, and CO2 gases have been developed in two forms: TOCN-COOH coated on PET film and selfstanding TOCN-COOH film. The obtained results indicate that the TOCN-COOH layer and film have gas permeability behavior primarily due to the gas molecule diffusion mechanism. The difference in H2 permeability between TOCN-COOH and other TOCNs with metal carboxylate groups is likely to be due to the different sizes of free volume holes present between nanofibrils in the TOCN layers and films under dry conditions. As selective gas separation membranes, both high separation selectivity and high permeability are required for practical application.22 Because these two factors trade-off on one another, it is generally difficult at present for separation membrane preparation to satisfy these two demands simultaneously. The present TOCN-COOH layers formed on polymer substrate films or self-standing TOCN-COOH films are, therefore, possible candidates for biobased and efficient H2 gas separation membranes. D

dx.doi.org/10.1021/bm400377e | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

■

Note

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was supported by Grants-in-Aid for Scientific Research (Grants 21228007 and 21-6112) from the Japan Society for the Promotion of Science (JSPS). We would like to thank Dr. Yoshiaki Kumamoto and Mr. Kenta Mukai (Kao Corporation, Japan) for providing commercial PET and PE films.

■

REFERENCES

(1) Alexandre, M.; Dubois, P. Mater. Sci. Eng. R 2000, 28, 1−63. (2) Lange, J.; Wyser, Y. Packag. Tehnol. Sci. 2003, 16, 149−158. (3) Usuki, A.; Hasegawa, N.; Kato, M.; Kobayashi, S. Inorganic Polymeric Nanocomposites and Membranes; Advances in Polymer Science Series; Springer-Verlag: Berlin Heidelberg, 2005; Vol. 179, pp 135−195. (4) Jørgensen, M.; Norrman, K.; Krebs, F. C. Sol. Energy Mater. Sol. Cells 2008, 92, 686−714. (5) Allen, S. M.; Fujii, M.; Stannett, V.; Hopfenberg, H. B.; Williams, J. A. J. Membr. Sci. 1977, 2, 153−163. (6) Exama, A.; Arul, J.; Lenckl, R. W.; Lee, L. Z.; Toupin, C. J. Food Sci. 1993, 58, 1365−1370. (7) Guilbert, S.; Cuq, B.; Gontard, N. Food Addit. Contam. 1997, 14, 741−751. (8) Nobel, R. D.; Koval, C. A. Review of Facilitated Transport Membranes. In Materials Science of Membranes for Gas and Vapor Separation; Yampolskii, Y., Pinnau, I., Freeman, B. D., Eds.; John Wiley & Sons: Chichester, U.K., 2006; pp 411−435. (9) Cong, H. L.; Radosz, M.; Towler, B. F.; Shen, Y. Q. Sep. Purif. Technol. 2007, 55, 281−291. (10) Ismail, A. F.; Rahim, N. H.; Mustafa, A.; Matsuura, T.; Ng, B. C.; Abdullah, S.; Hashemifard, S. A. Sep. Purif. Technol. 2011, 80, 20−31. (11) Yoshida, K.; Hirano, Y.; Fujii, H.; Tsuru, T.; Asaeda, M. J. Chem. Eng. Jpn. 2001, 34, 523−530. (12) Baker, R. W. Ind. Eng. Chem. Res. 2002, 41, 1393−1411. (13) Lavoine, N.; Desloges, I.; Dufresne, A.; Bras, J. Carbohydr. Polym. 2012, 90, 735−764. (14) Aulin, C.; Gällstedt, M.; Lindström, T. Cellulose 2010, 17, 559− 574. (15) Fukuzumi, H.; Saito, T.; Iwata, T.; Kumamoto, Y.; Isogai, A. Biomacromolecules 2009, 10, 162−165. (16) Saito, T.; Kimura, S.; Nishiyama, Y.; Isogai, A. Biomacromolecules 2007, 8, 2485−2491. (17) Saito, T.; Nishiyama, Y.; Putaux, J.-L.; Vignon, M.; Isogai, A. Biomacromolecules 2006, 7, 1687−1691. (18) Isogai, A.; Saito, T.; Fukuzumi, H. Nanoscale 2011, 3, 71−85. (19) Fujisawa, S.; Okita, Y.; Fukuzumi, H.; Saito, T.; Isogai, A. Carbohydr. Polym. 2011, 84, 579−582. (20) Fukuzumi, H.; Saito, T.; Isogai, A. Influence of TEMPOoxidized cellulose nanofibril length on film properties. Carbohydr. Polym. 2013, 93, 172−177. (21) Fukuzumi, H.; Saito, T.; Iwamoto, S.; Kumamoto, Y.; Ohdaira, T.; Suzuki, R.; Isogai, A. Biomacromolecules 2012, 12, 4057−4062. (22) Ghosal, K.; Freeman, B. D. Polym. Adv. Technol. 1994, 5, 673− 697. (23) Anderson, M. R.; Mattes, B. R.; Reiss, H.; Kaner, R. B. Science 1991, 252, 1412−1415. (24) Robeson, L. M. J. Membr. Sci. 2008, 320, 390−400.

E

dx.doi.org/10.1021/bm400377e | Biomacromolecules XXXX, XXX, XXX−XXX