Subscriber access provided by University of Sussex Library
Article
High Temperature Proton Conduction in Nanocellulose Membranes: Paper Fuel Cells Thomas Bayer, Benjamin V. Cunning, Roman Selyanchyn, Masamichi Nishihara, Shigenori Fujikawa, Kazunari Sasaki, and Stephen M. Lyth Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/acs.chemmater.6b01990 • Publication Date (Web): 23 Jun 2016 Downloaded from http://pubs.acs.org on June 27, 2016
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Chemistry of Materials is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 13
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Chemistry of Materials
High Temperature Proton Conduction in Nanocellulose Membranes: Paper Fuel Cells Thomas Bayera,b, Benjamin V. Cunninga, Roman Selyanchyna, Masamichi Nishiharaa, Shigenori Fujikawaa, Kazunari Sasakia-d, Stephen M. Lytha,e* a
International Institute for Carbon-Neutral Energy Research (WPI-I2CNER);
b
Department of Mechanical Engineering, Faculty of Engineering;
c
Next-Generation Fuel Cell Research Center (NEXT-FC);
d
International Research Center for Hydrogen Energy; Kyushu University, 744 Motooka, Nishi-ku, Fukuoka 8190395, Japan
e
Energy2050, Department of Mechanical Engineering, University of Sheffield, S10 2TN, UK
ABSTRACT: Polymer electrolyte membrane fuel cells (PEFCs) are an efficient and clean alternative power source, but the high cost impedes widespread commercialization. The fuel cell membrane, e.g. Nafion, contributes significantly to this cost, and therefore novel alternatives are required. Temperature is also an important factor; high temperature operation leads to faster reaction kinetics, lower electrocatalyst loading, and improved water management, thereby further reducing cost. However, higher temperature puts greater demands on the membrane. Conductivity is related strongly to humidification and therefore this generally decreases above 100°C. Nanocellulose membranes for fuel cells, in which the proton conductivity increases up to 120°C, are here reported for the first time. The hydrogen barrier properties are far superior to conventional ionomer membranes. Fuel cells with nanocellulose membranes are successfully operated at 80°C. Additionally, these membranes are environmentally friendly and biodegradable.
INTRODUCTION Cellulose is the most abundant polymer on earth. It is the major constituent of the cell walls of plants, providing structural integrity to trees. Wood pulp contains around 40-50% cellulose, cotton around 90 wt%, and it can even be produced by bacteria.1 Cellulose is a simple biopolymer made up of carbon, oxygen and hydrogen (Figure 1).2–4 The polymer chains group together during biosynthesis to form microfibrils via intermolecular hydrogen bonds and van der Waals forces. These microfibrils have diameters up to several tens of nanometers, with lengths of several microns.2,4–6 Due to the strong intra- and intermolecular hydrogen-bonding network, they exhibit high mechanical strength (up to 1080 MPa) and are highly stable.4,5 In the presence of pectin or lignin these microfibrils cluster together to form thicker macrofibrils with a diameter of up to 50 nm. This structure is what gives the cell wall in plants high stiffness and strength.4,5 Cellulose is a renewable, biodegradable and non-toxic polymer that has been used by mankind for thousands of years. Papyrus is a cellulose-based paper and was used to write on as early as 3000 BC. With the discovery of papermaking from woodbark and cloth in China in 105 AD, cellulose became the most important material for writing documents.6 Pure cellulose was first isolated and named in 1837 by Anselme Payen.7
Figure 1. Chemical structure of cellulose. a) Naturally occurring cellulose. b) Cellulose with carboxyl groups formed during the pulping process. c) Cellulose with sulfate ester 8 groups formed during acid hydrolysis.
ACS Paragon Plus Environment
Chemistry of Materials
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 13
Despite the microstructure of cellulose fibers being known, it was not possible to isolate the individual microfibrils until the late seventies. In 1977 Turbak et al.,9,10 ran wood pulp fibers through a high pressure milk homogenizer and observed cellulose nanofibers for the first time. This ignited a strong scientific and economic interest in using this material for alternative applications. Cellulose nanofibers (CNFs) are generally produced from purified cellulose fiber pulps by mechanical treatments such as high pressure homogenization, grinding, ultrasonic treatment, or cryocrushing.2,5,11 CNFs contain both amorphous and crystalline regions of cellulose and have a high aspect ratio.5,12 Another form of cellulose can be obtained via hydrolysis of CNFs in e.g. hydrochloric or sulfuric acid, to form cellulose nanocrystals (CNCs).3,12,13 In this process the amorphous regions of the CNFs are dissolved, leaving only shorter, more crystalline regions. The CNC dimensions can vary depending on the source of cellulose and the hydrolysis conditions (ranging from e.g. 3 to 5 nm width and 50 to 500 nm length).3,5 Due to their negatively charged surface (sulfuric acid reacts with the hydroxyl groups forming sulfate esters), CNCs can be dispersed readily in water.3,12,14 Just as conventional wood pulp is used to make paper (Figure 2), nanocellulose can be processed to form “nanocellulose paper”. In 2008 Henriksson reported nanocellulose paper produced from wood pulp via an enzymatic hydrolysis pretreatment, combined with mechanical beating and mechanical homogenization. The resulting paper had very high toughness, a Young’s Modulus of 13.2 GPa, and high tensile strength (214 MPa). The excellent mechanical properties were related to the nanofiber network.15 In 2009, Nogi et al. fabricated nanocellulose paper with high transparency (71.6% at λ = 600 nm), and the optical properties were maintained even after heating up to 150°C.16,17 They showed that the nanocellulose paper could be written on, and folded just like conventional paper. Nanocellulose paper has good oxygen barrier properties, which can be tailored by thermal or chemical treatment.18,19 Fang et al. used nanocellulose paper as a substrate for solar cells due to their ultrahigh optical transparency and low optical haze.20 Koga et al. reported the fabrication of highly transparent conductive networks of silver nanowires and carbon nanotubes on cellulose nanofiber paper, offering new possibilities for future paper electronics.21 Recently Choi et al. reported the use of nanocellulose as a separator membrane for flexible lithium ion batteries, enabling new possibilities for nextgeneration energy storage in field of wearable consumer electronics.22 Other application such as bio-composites, bio-medicine and gas barrier films for e.g. food packaging are also being investigated.19,23,24
Figure 2. Photographs of: (a) cellulose nanofiber (CNF) slurry; b) cellulose nanocrystal (CNC) slurry; c) (left to right) conventional cellulose-based paper, CNF paper, and CNC paper (all similar thickness, ~90 µm). Optical micrographs of (d) CNF and (e) CNC paper. AFM images of (f) CNF and (g) CNC on silicon substrates. FESEM images of (h) CNF and (i) CNC paper.
Fuel cells are electrochemical energy conversion devices, which directly convert the chemical energy of a fuel (such as hydrogen) into electrical energy, for as long as reactants are supplied and the products are removed. The basic structure comprises a proton conducting electrolyte membrane, sandwiched by anode and cathode electrocatalyst layers (usually platinum nanoparticles on a carbon black support). These components are collectively known as a membrane electrode assembly (MEA). Oxygen is supplied to the cathode via a carbon paper gas diffusion layer (GDL), and hydrogen is similarly supplied to the anode. Fuel cells have potential to revolutionize the way we utilize energy and help society move away from fossil fuels. They can run on carbon-neutral fuels such as hy-
ACS Paragon Plus Environment
Page 3 of 13
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Chemistry of Materials
drogen or bio-methanol, and have extremely high efficiency, not limited by the Carnot cycle. Polymer electrolyte fuel cells (PEFCs) are so-called low-temperature fuel cells, utilizing a solid membrane as electrolyte and generally operating below 80°C. These are used primarily in residential fuel cell systems (such as Ene-Farm in Japan); in fuel cell vehicles (FCVs) such as the Toyota MIRAI; and in portable power applications such as the Upp from Intelligent Energy. PEFCs are popular due to their high power density, fast start up, and dynamic response to changes in load. However, hydrogen storage is still an important issue that remains largely unsolved.25,26 Direct methanol fuel cells (DMFCs) allow for the usage of liquid fuel, making storage and transport easy, enabling instantaneous refill, and giving DMFC systems an overall very high energy density.25,26 For both PEFCs and DMFCs, the most commonly used ionomer membrane is Nafion®, which has a high proton conductivity (> 100 mS cm-1), good mechanical strength and chemical stability.27 However, Nafion is expensive; it is a fluoropolymer which is difficult to recycle at end-of-life; it has proton conductivity which depends strongly on humidification; its mechanical stability is compromised at > 110°C;27and it has relatively high methanol cross-over in DMFCs.26,27 Nanocellulose paper could be an alternative ionomer membrane for fuel cells due to the low cost, good gas barrier properties, mechanical toughness and acidic oxygen functional groups. For example, cellulose has been investigated in composites with Nafion. Jiang et al. blended Nafion with bacterial cellulose resulting in proton conductivity of 71 mS cm-1 at 100% RH and 30°C.28 Their membranes exhibited improved power density of 106 mW/cm2 when used in a PEFC.28 Recently, Gadim et al. continued on Jiang’s work and mixed Nafion with bacterial cellulose, obtaining slightly lower conductivity (40 mS cm-1 at 40°C and 98% RH) and much lower fuel cell power density.29 Hasani-Sadrabadi et al. incorporated acidhydrolyzed cellulose nanocrystals into Nafion, suppressing methanol cross-over, increasing the proton conductivity, and maintaining the conductivity at higher temperature due to improved water uptake.30 Jiang et al. immersed bacterial cellulose biofilms into H3PO4 or phytic acid, achieving proton conductivities at 20°C of 80 and 50 mS cm-1, respectively, and PEFC power densities of 17.9 and 23 mW cm-2, respectively, at 25°C.31 Lin et al. modified bacterial cellulose membranes with 2-acrylamido-2methyl-1-propanesulfonic acid (AMPS) using ultraviolet light-induced grafting polymerization, resulting in membranes with about half the methanol permeability of Nafion 115 and DMFC power density of 16 mW cm-2.32 Kasai et al. showed that cross-linked cellulose sulfate membranes have half the methanol permeation compared to Nafion 112, whilst exhibiting proton conductivity of 81 mS cm-1 at room temperature.33 Ladhar et al. mixed nanocllulose with natural rubber and reported a maximum conductivity of 6 x 10-3 mS cm-1 for a 15 wt% CNCcontaining composite.34 However, the conductivity of pure nanocellulose was not reported. Finally, Smolarkiewicz et al. doped cellulose with imidazole, achieving a
proton conductivity of approximately 2 x 10-3 mS cm-1 at 160°C under anhydrous conditions.35 In all the aforementioned studies, dopants were used to increase the proton conductivity, or cellulose was used as a nanofiller to improve the properties of Nafion. Only Smolarkiewicz et al. measured the proton conductivity of a pure cellulose reference sample. The microcrystalline cellulose pellets, compressed at room temperature under a pressure of 10 MPa, had a maximum conductivity of 2 x 10-6 mS cm-1 at 70°C (humidity not defined).35 To the best of our knowledge the proton conductivity of nanocellulose membranes has not yet been investigated, especially at higher temperature, over a wide range of relative humidities, and with activation energies to give insight into the conduction mechanism. Pure, unadulterated cellulose or nanocellulose membranes have not been investigated for application in hydrogen fuel cells. Here the hydrogen gas permeability of nanocellulose paper, specifically cellulose nanofibers (CNF) and cellulose nanocrystals (CNC) is measured. The proton conductivity is investigated over a wide temperature and humidity range, giving insights into the conduction mechanism. The application of nanocellulose paper as an ionomer membrane for PEFC-type hydrogen fuel cells is reported.
EXPERIMENTAL SECTION Nanocellulose Paper Preparation. Nanocellulose materials (CNF and CNC slurry, 3.0 wt% and 11.8 wt% solids, respectively) were purchased from University of Maine. Nanocellulose slurry (CNF = 100.2 g, CNC = 52.9 g) was mixed with 500 ml of deionized water and stirred for 24 hours at 500 rpm. The dispersions were then vacuumfiltered onto Millipore filters (hydrophilic PTFE, 0.1 µm pore size). After filtration the resulting CNF membranes were placed between two Teflon sheets, hot-pressed for 20 min (at 110°C and 1.1 MPa), and then carefully peeled off the filter. CNC membranes were simply peeled of the filter. The resulting CNF papers are free-standing and flexible, whilst the CNC membranes are more brittle and break more easily. The thicknesses of the resulting CNF and CNC electrolyte membranes were 32 and 30 μm, respectively. The preparation process is shown schematically in the supplementary information (Supplementary Figure 1). Morphology, Pore Size Distribution, Chemical Composition and Crystallinity. The microstructure of the nanocellulose fibers was investigated by atomic force microscopy (Seiko Instrument scanning probe microscope unit SPA300HV with SPI 3800 N probe station and cantilever type SN-AF01, SiN, 100µm height). AFM samples were prepared by dropping 100 µl of nanocellulose dispersion onto 1 cm2 silicon substrates, which were then dried in air. The assessed area was 4 µm x 4 µm. The average roughness was calculated by using 10 horizontal and 10 line roughness profiles obtained by AFM. The surface of nanocellulose membranes was investigated by laser microscopy (Keyence One-shot 3D Measurement Macroscope VR-3000). The morphology of the membrane surface was also analyzed using field emission scanning elec-
ACS Paragon Plus Environment
Chemistry of Materials
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
tron microscopy (FESEM Hitachi S-5200, 5 kV. To prevent charging, samples were coated with a thin platinum layer before observation, deposited by using an ion sputter (Hitachi E-1030). Cross-sections of the membrane electrode assemblies were investigated by SEM (Hitachi S5200, 1.5 kV). MEAs were cut using a scalpel. Specific surface area and pore size distribution were determined by N2 physisorption at 77K using a Belsorp mini-II automated system (Bel Japan Inc.) The CNC and CNF membranes were cut into small pieces and first degassed in the Belprep II device (Bel Japan Inc.) at 100°C for 20 h prior to the analysis by N2 adsorption at −196°C. Pore size distribution was determined from N2 adsorption at a relative vapor pressure of 0.01−0.99 following a BJH model.36 Chemical composition was determined by x-ray photoelectron spectroscopy (XPS) (PHI 5000 Versa Probe II with Al Kα radiation, 15 kV, 25 W). Before analysis the data were charge-corrected to the C-C bond of the C 1s signal, to account for the electronically insulating nature of the samples.[53] Infrared spectra were collected in attenuated total reflection (ATR) mode on a Fourier transform infrared spectrometer (Thermo Scientific, Nicolet iN10 MX) equipped with a germanium ATR tip. Data was averaged from 128 acquisitions in the region of 675-4000 cm-1, with a spectral resolution of 4 cm-1. For crystallinity determination, X-ray diffraction (XRD) was measured using a Smartlab X-ray Diffractometer with Cu Kα radiation (γ = 1.5418 Å, 0.01°/step, data collection from 5 to 40°). The crystallinity of both materials was calculated by using Equation 1, where Itotal is the intensity of the crystalline and amorphous parts (peak intensity ~22.3°) and Iamorph the intensity of the amorphous part (peak intensity ~18.5°) of the nanocellulose.
IC =
I total − I amorph I total
⋅ 100% [1]
Gas Permeability Measurements. For gas permeance measurements the membrane area was masked with Kapton® and alumina tape to provide a circle of desired diameter (d = 1 cm) and area (S = 0.785 cm2). Additionally, to prevent deformation during measurements, membranes were placed on a porous polycarbonate support filter (1.2 µm pore size). The dry hydrogen permeation rate through the nanocellulose CNF and CNC papers (as well as a Nafion 112 membrane) was measured between room temperature and 80°C using a GTR-11A/31A gas barrier testing system (GTR Tec Corp., Japan). This uses a differentialpressure method where gas permeation is induced by vacuum on the permeate side and extra pressure applied at the feed side, as described in more detail in our previous studies.37 The total pressure difference was 200 kPa. The sample collection time after vacuuming the sweep side of the membrane was 30 minutes. The collected gas was transferred to a gas chromatograph and the volume was measured. The gas chromatograph is combined with a thermal conductivity detector (TCD) (Yanaco G3700T, Japan).
Page 4 of 13
Ionic Conductivity. Ionic conductivity was investigated for a range of temperatures and relative humidities using a membrane testing device (MTS-740, Scribner) coupled with an impedance analyzer (Solatron SI1260).38 CNF and CNC membranes (with 21 µm and 33 µm thickness, respectively) were measured with an AC amplitude of 10 mV in a frequency range from 30 MHz to 10 Hz. The impedance was measured at 30°C with isothermal changes in relative humidity (RH) from 100 to 0% RH. Before measurements, the sample was pretreated at 100% RH for 4 hours, followed by 1 hour pretreatment at each humidity. Then the impedance was measured from 40 to 120°C at 100% RH to show the influence of temperature on proton conductivity. At temperatures above 100°C the measurement chamber was pressurized at 230 kPa total pressure. Impedance plots are fitted to an equivalent circuit using ZPlot (Scribner), from which the membrane resistance, R (Ω), was determined from the high frequency intercept. Conductivity σ was calculated using Equation 2, with membrane thickness L (cm), membrane resistance R (Ω) and cross-sectional area A (0.5 cm2) respectively.39
σ=
L [2] R⋅ A
Fuel Cell Performance. A membrane electrode assembly (MEA) was prepared using nanocellulose membranes with a thickness of ~30 µm. Catalyst ink was prepared by mixing Pt/C electrocatalyst (Tanaka Kikinzoku Kogyo K.K., 46.2 wt % Pt) with 5 wt% Nafion polymer solution (Aldrich), ethanol (Chameleon), and deionized water. The catalyst ink was stirred overnight, then sonicated before use for 30 min (SMT Ultra Sonic Homogenizer UH-600). The catalyst ink was sprayed onto the nanocellulose membranes (Nordson K.K. Spraying Device, C-3J) using a mask to create an electrode size of 0.5 cm2 with a catalyst loading of 0.3 mgPt/cm2 for both electrodes. Hydrophobic carbon paper (EC-TP1-060T) gas diffusion layers (GDLs) were hot-pressed onto the electrocatalyst layers (0.2 kN for 190s at 132°C). The MEA was installed into a NEDO single cell holder (1 cm2) and its performance was investigated at 80°C. The cell was preconditioned for two hours at 80°C and a nitrogen gas flux of 100 ml/min (95% RH). The gas flux was changed to hydrogen and air (100 ml/min, 95% RH) and after ten minutes a performance test was conducted. Polarization curve and power density were investigated using a potentiostat (VersaSTAT 4, Amtek).40,41
RESULTS AND DISCUSSION Morphology of Nanocellulose. Photographs of the nanocellulose slurries are shown in Figure 2a-b. Photographs of the CNF and CNC membrane are compared with conventional cellulose-based paper (Fuji Xerox, G70) in Figure 2c. Conventional cellulose paper is completely
ACS Paragon Plus Environment
Page 5 of 13
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Chemistry of Materials
opaque and white. The CNF paper is slightly transparent, and the CNC paper is more transparent. Both CNF and CNC membranes look very similar under an optical microscope (Figure 2d-e). AFM images (Figure 2f-g) show much more clearly the difference in microstructure between CNF and CNC. CNF consists of a network of randomly interwoven fibers with large diameters ranging from ~50 to 180 nm and lengths extending beyond the imaging area. The thicker fibers may comprise bundles of smaller fibers. The surface roughness is 36.60 ± 11.25 nm. In contrast, CNC comprises more highly aligned fibers with much smaller diameter (
