- Email: [email protected]
Hydrogen characteristics and ordered structure of mono-mesogen type liquid-crystalline epoxy polymer
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Hydrogen characteristics and ordered structure of mono-mesogen type liquid-crystalline epoxy polymer Shuji Kawamoto a, Hirotada Fujiwara b, Shin Nishimura c,* a
Department of Hydrogen Energy Systems, Graduate School of Engineering, Kyushu University, Japan Research Center for Hydrogen Industrial Use and Storage, Kyushu University, Japan c Department of Mechanical Engineering, Faculty of Engineering, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka 819-0395, Japan b
article info
abstract
Article history:
Fuel cell systems, such as those used in electric vehicles require lightweight hydrogen fuel
Received 12 October 2015
storage. Polymer composites are candidate materials for storage vessels requiring high
Received in revised form
fracture resistance for high pressure cycling, as well as high thermal conductivity. Here, we
7 March 2016
investigate 4,40 -diglycidyloxybiphenyl (DGOBP) epoxy for such applications, and clarified
Accepted 20 March 2016
the relationship between the hydrogen penetration properties and the ordered structure of
Available online 10 April 2016
the epoxy. We controlled the ordered structure of the DGOBP polymer by adjusting the curing temperature of a 4,40 -diaminodiphenyl methane (DDM) curing system. We obtained
Keywords:
a series of cured DGOBP with DDM samples with crystallinity values of 19%, 27%, 32%, and
Epoxy
36%, corresponding to curing temperatures of 130, 120, 110, and 100 C, respectively. The
Polymer
thermal conductivities of these samples were 0.26e0.31 Wm1 K1. Cured DGOBP showed
Crystallization
hydrogen contents of 403e1271 ppm (25e75% smaller than that of conventional epoxies).
Liquid crystal
Liquid crystalline epoxy polymers can be a candidate material for hydrogen storage sys-
Thermal conductivity
tems requiring well-controlled hydrogen penetration properties.
Hydrogen properties
Copyright © 2016, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Introduction Hydrogen energy systems are being investigated for nextgeneration clean energy systems. However, safety is a primary concern for commercial application and must be considered in the design of hydrogen energy systems. For hydrogen energy to be feasible and widely implemented in transport and stationary technologies, it needs to be safely and efficiently stored [1]. Compressed gas is currently the
preferred solution for storing hydrogen in fuel cell vehicles (FCV). To satisfy the desired driving range and convenience of FCVs, hydrogen should be stored at 70 MPa and filled at a rapid rate, reaching its storage pressure within 3 min. To avoid thermal damage to the storage vessel, the hydrogen gas needs to be pre-cooled to 40 C at the hydrogen filling station. Because the compression can greatly elevate the temperature of the gas [2], vessel materials must effectively radiate heat. Moreover, hydrogen storage vessels installed in FCVs need to be lightweight [3] in order to achieve high fuel efficiency.
* Corresponding author. Tel.: þ81 92 802 3248. E-mail address: [email protected] (S. Nishimura). http://dx.doi.org/10.1016/j.ijhydene.2016.03.124 0360-3199/Copyright © 2016, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
The current FCV hydrogen vessel is a “type IV vessel”, which consists of a fully-wrapped carbon fibre reinforced plastic, such as an epoxy polymer with a plastic liner [4]. A disadvantage of polymeric materials for this application is that they generally have low thermal conductivity. In contrast, liquid-crystalline epoxy polymers, such as twinmesogen type liquid-crystalline epoxy polymers, have been reported as strong thermal conductors [5]. Therefore, we consider that liquid-crystal polymers are suitable liner materials or matrices for carbon fibre reinforced plastics in type IV vessels. The high heat dissipation of liquid-crystal polymers should allow a higher pre-cooling temperature during filling, reducing the difficulty of sealing hydrogen gas at low temperature. In addition, it was thought that the highly dense structure of the liquid-crystal polymer should provide an effective barrier to hydrogen permeation [6,7]. In this study, we design a liquid-crystal epoxy polymer suitable as a gas-barrier material for type IV hydrogen storage vessels. A number of studies investigating the hydrogen properties of rubber as a sealing material have been reported [8e14]. It was shown that, when exposed to high-pressure hydrogen gas, the polymer materials were penetrated by hydrogen molecules. The amount of hydrogen penetration into the materials and the volume change ratio have been suggested as criteria for standardising polymer materials for high-pressure hydrogen equipment. Low penetration and volume change ratio are preferred for hydrogen barrier materials from the viewpoint of avoiding fractures and gas leaks. In addition, the thermal conductivity of the hydrogen barrier materials should be much lower than the current state-of-theart materials considering safety and degradation of the polymers. The hydrogen penetration and thermal conductivity are expected to correlate with the “ordered structure”, such as the crystal domain or liquid-crystal structure of the material. Understanding the effect of this ordered structure is important for designing a reliable structural material for type IV vessels to store high-pressure hydrogen gas. Here, we investigate the influence of the structure of an epoxy polymer on the hydrogen penetration properties and thermal conductivity. We selected a mono-mesogen type liquid-crystal polymer, which has a controllable ordered structure, as a model material having a similar structure, molecular size and reactive groups as conventional epoxy polymers such as bis(4diglycidyloxyphenyl)propane. The selected material for this study was 4,40 -diglycidyloxybiphenyl (DGOBP) which has one biphenyl group as mesogen. A number of studies of DGOBP have been reported, using curing systems such as 2,7-diaminofluorene [15], sulfanilamide [16e20], 4,40 -diaminodiphenylsulfone [21], 4,40 -diaminobiphenyl and 4-methyl-m-phenylenediamine [22,23], and 4,40 -diaminodiphenyl methane (DDM) [5,24e27]. However, the relationship between the ordered structure and the hydrogen properties, in particular for DGOBP cured by DDM, have not been reported. In this study we fabricated a series of DGOBP samples cured with different DDM concentrations to optimize the ordered structure and investigate the effects on the hydrogen penetration and thermal conductivity. Such an analysis is expected to provide useful information for proposing guidelines for the design of polymeric materials for hydrogen systems at the molecular level.
7501
Materials and methods Materials 4,40 -biphenol and DDM were purchased from Tokyo Kasei Kogyo Co., Ltd. Chloromethyloxirane, tetrabuthylammonium bromide (TBAB), sodium hydroxide and sulfanilamide were purchased from Wako Pure Chemical Industries Co., Ltd. Epikote828, also known as Bisphenol A (BisA), was purchased from Mitsubishi Chemical Co., Ltd. as conventional epoxy polymer. The main ingredient of Epikote828 is 2,2-bis(4diglycidyloxyphenyl)propane. All compounds were used as received without further purification. The mono-mesogen type liquid-crystal epoxy monomer DGOBP was synthesised as previously described [16]. The synthesis route for the DGOBP monomer and the chemical structures of the DGOBP monomer and curing agent are shown in Scheme 1. The product was purified by reprecipitation from water and methanol. To observe the ordered structures of the DGOBP samples cured with DDM, the monomer and curing agent were combined at a molar ratio of 2:1. This stoichiometric mixture of DGOBP and DDM was melted and mixed by stirring at 150 C on a hot plate for 1 min. Afterwards the curing reaction was quenched by liquid nitrogen. To prepare samples for evaluating the conversion of the epoxy group in DGOBP to the amino group in DDM, this same process was employed. However, the curing reaction was undertaken for 2, 3, 5, or 7 min. Disk-shaped specimens of DGOBP cured with DDM (diameter: 40 mm, thickness: 2 mm) were prepared from a stoichiometric mixture of the polymer and curing agent (molar ratio 2:1). The DGOBPeDDM mixture was prepared as described above, stirred for 3 min, and then cured in a mould for 4 h. The curing temperature was varied: 100, 110, 120, 130, 140, 150 or 180 C. A stoichiometric mixture of BisA and DDM
Scheme 1 e (a) Synthesis route for the DGOBP monomer. (b) Chemical structure of the BisA monomer. (c) Chemical structure of the DDM curing agent.
7502
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
(molar ratio 2:1) was cured at 100 C for 4 h. The cured polymers were cut into test pieces for subsequent analyses.
measurements at room temperature were undertaken with a scanning speed of 2 min1, resolution of 0.02 , and a 2q range of 10e50 .
Characterization Dynamic mechanical analysis (DMA) 1
H and
13
C nuclear magnetic resonance (NMR)
NMR spectra were obtained using an AvanceII 500-MHz spectrometer (Bruker BioSpin Co., Ltd.) operated at room temperature (1H NMR at 500 MHz and 13C NMR at 125 MHz). The chemical shifts were measured relative to the peaks of tetramethylsilane as a standard (d in 0.00 ppm).
Liquid chromatography Liquid chromatography was performed using an HPLC system (HLC-8320GPC, Tosoh Co., Ltd.) equipped with two TSKgel SuperHZ-M columns and one TSKgel SuperHZ2000 column. Measurements were acquired in size exclusion chromatography mode with a reflective index detector, tetrahydrofuran elution solvent, flow rate of 0.35 mL min1, and injection volume of 0.1 mL for a 1% w/v sample.
Fourier transform infrared spectroscopy (FT-IR) FT-IR spectra of the DGOBP monomer, pre-cured samples, and DGOBP cured with DDM were measured using a FT-IR Spectrum100s (PerkinElmer Co., Ltd.) by the KBr disc method in the wavelength range 450e4000 cm1 with a resolution of 4 cm1. Eight scans were performed.
Differential scanning calorimetry (DSC) DSC measurements of stoichiometric DGOBP and DDM mixtures (5 mg powder samples) were carried out using a DSC 204 HP instrument (Netzsch GmbH) in a nitrogen atmosphere from room temperature to 300 C. The heating rate was varied: 5.0, 7.5, or 10.0 C min1. Isothermal DSC measurements were performed on a 5 mg molten DGOBPeDDM sample. The heating rate was 50 C min1 from room temperature to the test temperature of 130.0, 132.5, 135.0, 137.5, 140.0, 150.0, 160.0, or 170.0 C, where the sample was held for 20 min in a nitrogen atmosphere. The DGOBP cured with DDM at different curing temperatures and the BisA cured with DDM (10 mg solid samples) were measured at a heating rate of 10 C min1 from room temperature to 300 C in a nitrogen atmosphere. The glass transition temperature was determined by the temperature at the half height of the heat capacity change in the DSC curve.
Polarized optical microscopy The phase transition behaviour of the molten DGOBPeDDM samples was observed using optical microscopy (Eclipse LV150, Nikon Co., Ltd., 10 eyepiece, 10 or 20 objective lenses) equipped with a crossed polarizer and a hot stage (CSS450WCX, Japan High Tech Co., Ltd.). The temperature of the hot stage was calibrated by measuring the melting points of DDM (90 C), TBAB (104 C), and sulphanilamide (165 C).
Wide-angle X-ray scattering (WAXS) Samples of DGOBP cured with DDM were cut into disc-shaped specimens (diameter: 13 mm, thickness: 2 mm) for the WAXS measurements. A MultiFlex X-ray diffraction unit (Rigaku Co., Ltd.) with Cu-Ka radiation was used. Continuous
DMA was performed using a DVA-200s instrument (IT Keisokuseigyo Co., Ltd.) from 30 to 300 C at a heating rate of 10 C min1 in tensile mode, with a strain amplitude of 0.08% and a frequency of 10 Hz. Rectangular specimens cut from cured samples with a cross-sectional area of (1.5 1.5) mm2 and a length of 15 mm were used. The samples were fixed at a distance of 10 mm between the grips of the DMA machine and measured a 2 C intervals. The molecular weight between crosslink points of the cured samples was calculated from the storage modulus (tan d) at 280 C (where tan d converged for every cured polymer). Here, d is the stressestrain phase lag.
Thermal conductivity measurements Disc-shaped specimens (diameter of 10 mm; thickness of 0.4 mm) were cut from DGOBP cured with DDM, spray-coated with thin gold and graphite electrodes. The thermal diffusivity of the cured samples was measured by the half-time method using a xenon flash analyser (LFA 467 Hyper Flash, NETZSCH GmbH) at 25 C. The specific heat capacity was measured by DSC with Al2O3 as the reference material, as described in ISO11357-4. The density of DGOBP samples cured with DDM was measured using a Sartorius LA230S balance equipped with the density determination kit YDK01 (Sartorius Co., Ltd.), using ethanol at 25 C. The thermal conductivity of the cured samples was calculated from the thermal diffusivity, heat capacity, and density. The measurement error of the thermal conductivity was 8%.
Hydrogen content, diffusion coefficient, and volume change after hydrogen exposure Disc-shaped specimens with a diameter of 13 mm and a thickness of 2 mm were cut from DGOBP cured with DDM and exposed to 90 MPa hydrogen gas at 30 C for 24 h in a highpressure vessel. The hydrogen release profile after decompression was measured using a thermal desorption analyser (JSH-201, J-Science Lab Co., Ltd.). The amount of hydrogen released from the sample was determined by gas chromatography at 30 C in a flow of argon every 5 min. The diffusion properties of hydrogen were estimated by fitting the data. The hydrogen content immediately after decompression (CH 0 ; t ¼ 0) and the diffusion coefficient D were estimated by fitting Eq. (1) to the remaining hydrogen content of the specimen using the least-squares method [8].
CHR ðtÞ
n . o# " 2 ∞ exp ð2n þ 1Þ p2 Dt l2 0 32 H X ¼ 2 $C0 $ p ð2n þ 1Þ2 n¼0 " # ∞ X exp Db2n t r20 $ b2n n¼1
(1)
where CHR ðtÞ (wt. ppm) is the remaining hydrogen content at time t (s), the time elapsed since decompression. CH0 (wt. ppm) is the hydrogen content at t ¼ 0, D (m2 s1) is the diffusion coefficient, bn is the root of the zero-order Bessel function, and l (m) and r (m) are the thickness and radius of the specimen, respectively. The measurement errors for the hydrogen
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
7503
content and diffusion coefficient were estimated from the accuracy of the TDA chromatography and the measurement error of specimens from the propagation of errors from Equation (1). The measurement error for the hydrogen content and diffusion coefficient were 5% and 14%, respectively. The volume change of the disc-shaped specimens after hydrogen exposure was observed at 30 C by a laser scan micrometre (TM-3000, Keyence Co., Ltd.) with a TM-065 sensor head and an InGaN green LED light source. The disc area was measured at 5 min intervals. The volume was calculated by cubing the square root of the area and was expressed as a proportion of the initial sample dimensions (i.e. the volume change ratio).
Results and discussion Synthesis of DGOBP 4,4’-dihydroxybiphenyl and chloromethyloxirane reacted stoichiometrically with NaOH (aq.) and TBAB (see Scheme 1). The product was analysed by liquid chromatography using a differential refractive index detector where the peak area was calculated from the chromatogram. The reaction products of Scheme 1 comprised 95.1% DGOBP monomer, 3.7% DGOBP dimer, and 1.2% DGOBP trimer. The chemical structure of the synthesized DGOBP was confirmed by the 1H and 13C peaks identified in NMR and IR spectra. The data for the 1H peak from NMR (500 MHz, CDCl3, d) are as follows: 7.49 ppm (d, J ¼ 9 Hz, 4H, Ar H), 7.00 ppm (d, J ¼ 9 Hz, 4H, Ar H), 4.27 ppm (dd, J ¼ 11 and 3 Hz, 2H, glycidyl CH2), 4.03 ppm (dd, J ¼ 11 and 6 Hz, 2H, glycidyl CH2), 3.40 ppm (m, 2H, glycidyl CH2), 2.94 ppm (dd, J ¼ 5 and 4 Hz, 2H, epoxy CH2), and 2.80 ppm (dd, J ¼ 5 and 3 Hz, 2H, epoxy CH2). The data for the 13C peak determined by NMR (125 MHz, CDCl3, d) are as follows: 157.8 ppm (OCCHCHCAr), 134.0 ppm (OCCHCHCAr), 127.8 ppm (OCCHCHCAr), 115.0 ppm (OCCHCHCAr), 69.0 ppm (CH2OCHCH2O), 50.1 ppm (CH2OCHCH2O), and 44.7 ppm (CH2OCHCH2O). The IR (KBr) data are as follows: n ¼ 3063 cm1 (w; nas (Ar CH)), 3041 cm1 (w; ns (Ar CH)), 3008 cm1 (w; ns (epoxy CH)), 2929 cm1 (w; nas (CH)), 2877 cm1 (w; ns (CH)), 1606 cm1 (s; n (Ar ring)), 1501 cm1 (s; d (Ar ring)), 1244 cm1 (s; nas (COC)), 1180 cm1 (m; d (Ar CH)), 1134 cm1 (m; d (Ar CH)), 1037 cm1 (s; ns (COC)), 911 cm1 (m; nas (epoxy COC)), 814 cm1 (s; g (Ar CH)), and 762 cm1 (m; g (Ar CH)), melting point ¼ 160 C.
Curing behaviour The curing behaviour was determined by DSC as described in Section Materials and methods. The curing temperature was determined from the exothermic peak of the DSC traces of the stoichiometric DGOBPeDDM mixture [25]. At each heating rate (5.0, 7.5, and 10.0 C min1), the exotherm of the curing reaction appears at temperatures above 130 C, higher than the endotherm related to the melting of the DDM (around 85 C for this purity of monomer). Fig. 1 shows the isothermal DSC traces of the melting process of the DGOBPeDDM mixtures cured at different temperatures. The isothermal DSC traces show a broad peak related to the curing reaction after the samples melted. The peak width decreased with increasing isothermal temperature, indicating an increase in the reaction rate at
Fig. 1 e Isothermal DSC curves of the DGOBP monomer cured with DDM at different temperatures. The arrows indicate the onset of exothermic peaks.
higher isothermal temperatures. The DSC traces of the mixtures cured below 135 C showed an additional exothermic peak, which appeared 3e4 min after reaching the target temperature. This peak was absent in mixtures cured above 137.5 C, suggesting that it originates from the ordered structure in the samples (as will be discussed later). The DGOBPeDDM mixtures used in the isothermal curing study were prepared for observation using polarised optical microscopy by mixing at 150 C for 1 min, followed by rapid cooling in liquid nitrogen [26]. Note that although DGOBP was dissolved in the molten DDM at 150 C, its melting point is actually 160 C. Fig. 2 shows polarised optical micrographs of the DGOBPeDDM mixtures obtained at 130 C and 135 C. After 3e4 min, a birefringent domain began to emerge in the sample cured at 130 C and a birefringent texture began to emerge in the sample cured at 135 C. The ordered structure of such birefringent domains in the sample cured at 130 C was confirmed by wide angle X-ray patterns reported by Mititelu et al. [15]. A magnified view of the birefringent domain of the sample cured at 130 C is shown in Fig. 3. The nucleation and growth (to the order of 10 mm) of a spherulite-like structure is clearly visible in these micrographs. The polarised optical micrographs of samples cured at 135 C showed no definite crystallite structure; the spherulite-like structure was absent in these samples. Li et al. attributed the extra exothermic peak
7504
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
Fig. 2 e Polarised optical microscopy images (100 £ magnification) of the DGOBP monomer mixed with DDM at a curing temperature of 130 C and 135 C. (a) 1 min, (b) 4 min, (c) 5 min, and (d) 10 min.
Fig. 3 e Polarised optical microscopy images (200 £ magnification) of the DGOBP monomer mixed with DDM at a curing temperature of 130 C. (a) 3 min 30 s, (b) 3 min 40 s, (c) 3 min 50 s, and (d) 4 min.
in the isothermal DSC traces to the smectic phase [16]. Hence, we infer that the birefringence observed in the polarised optical microscopy images of samples cured at high temperature (above 135 C) reflects the nematic phase. This suggests that the exothermic peaks in the isothermal DSC traces originate from the crystal phase. Fig. 4 shows the timeetemperatureephase transformation (TTT) diagram indicating the regions of the isotropic, liquidcrystal, and crystal phases of the DGOBP isothermally cured with DDM [28e33]. The solid curves represent the onset of the
liquid-crystal phase and the crystal phase transition from the isotropic phase. These onsets were determined from the polarised optical micrographs of the ordered phase observed during isothermal curing. The broken line marks the boundary between the liquid-crystal and crystal phases. Therefore, the ordered structure of DGOBP cured with DDM can be controlled by the curing temperature. Thus, to control the amount of ordered structure in the cured DGOBP with DDM, cured polymer samples were prepared at 100, 110, 120, 130, 140, 150, and 180 C for 4 h.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
7505
Fig. 4 e Timeetemperatureephase transformation diagram of the isothermal curing of DGOBP mixed with DDM.
Conversion of the epoxy group The conversion of the epoxy group was determined from the FT-IR and DSC measurements. The conversion of the aminebased curing system has been studied by many authors [34e37]. In terms of the methodology of these studies, the calibration curve for the conversion of the epoxy group by the consumption of the DGOBP monomer and DDM was defined in this study. The area under the exothermic peak in the DSC traces of the pre-cured samples was assumed as the residual heat of the curing reaction. The total heat of reaction was assumed to be the residual heat of the curing reaction of the uncured DGOBPeDDM mixtures. The overall conversion was calculated from dividing the residual heats by the total heats of the curing reaction. In the FT-IR spectra of the pre-cured samples, we computed the ratio of the peaks attributed to the CO stretching vibration of the epoxy group (at 910 cm1) and the in-plane vibration of the phenyl ring (at 1610 cm1). The decrease in the height of the 910 cm1 peak was attributed to the consumption of monomer. The peak at 1610 cm1 was adopted as the internal standard for the spectra. Using the calibration curve method, the conversion of the epoxy group can be evaluated regardless of the overlap of the peaks originating from other chemical species. By analysing the changes in the 910 cm1/1610 cm1 peak ratios, we derived the epoxide conversion in the DGOBP cured with DDM. Fig. 5 shows the FT-IR spectra of the DGOBP cured with DDM at different temperatures. The KBr disc method used to acquire the data allowed us to evaluate the conversion both on the surface and in the bulk of the samples. The cured samples show reduced absorption intensity of the epoxide group at 910 cm1. The percentage conversion values of the epoxide groups at each curing temperature, obtained by the peak ratio of 910 cm1/1610 cm1, are summarised in Table 1. Above 140 C, the epoxide group was completely consumed. Even at lower curing temperatures, more than 94% of the epoxide groups were consumed during the curing process. From these results, it can be concluded that the curing of DGOPB with DDM was successful at all temperatures investigated here.
Fig. 5 e FT-IR spectra of DGOBP cured with DDM at different temperatures.
Observation of the ordered structure Fig. 6 shows the WAXS patterns of DGOBP cured with DDM at different temperatures. Samples cured below 130 C exhibited three diffraction peaks, at 2q ¼ 19 , 22 , and 27 , corresponding to d-spacings of 0.46, 0.40, and 0.32 nm, respectively. At curing temperatures below 130 C, crystalline phases should be present in the cured polymer. The d-spacing of the sample cured above 140 C was attributed to the stacking of mesogen groups, implying that the liquid crystal structure was frozen in the cured polymer [7]. We evaluated the degree of crystallinity by integrating the intensity of the crystal peak and broad peak (halo patterns) from the WAXS patterns [38]. The crystallinity values of our DGOBPeDDM samples were 19%, 27%, 32%, and 36%, where the degree of crystallinity increased as the curing temperature decreased. This crystallization gives rise to the exothermic peaks observed in the isothermal DSC traces (see Fig. 1). Douglas et al. and Cho et al. have reported a relationship between the phase transition and the gel point [31,33]. Smaller spherulite particles were formed in DGOBP cured at 130 C compared to those in samples cured at 100 C. This was attributed to a decline in the mobility of the molecular chains as a result of gelation, preventing an increase in the spherulite size. Fig. 7 shows the storage modulus and tan d from DMA measurements (Fig. 7 a, b, respectively) of DDM-cured BisA
7506
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
Table 1 e Epoxide conversion and thermal, physical, and structural properties of DGOPB cured with DDM. Curing temperature ( C) 100 110 120 130 140 150 180 BisA (100) a b c d e f g h i j k
Conversion E0 at r.ta (%) (Pa)
Tgb ( C)
E0 at Tgc (Pa)
Tcd ( C)
E0 at Tce (Pa)
109 109 109 109 109 109 109 109
127 151 170 179 192 203 234 140
1.42 1.25 9.25 7.44 5.00 4.28 3.00 5.63
109 109 108 108 108 108 108 107
213 216 214 208 e e e e
2.77 2.82 4.72 4.65
94 95 97 98 100 100 100 e
1.94 1.99 2.02 2.07 2.21 2.21 2.30 2.68
e e e e
108 108 108 108
E0 at Tmf (Pa) 1.98 2.01 2.18 2.18 1.66 1.13 8.63 4.90
108 108 108 108 108 108 107 107
rMcg (g mol1)
Tgh ( C)
Dc i (%)
86 86 71 66 72 105 162 317
144 149 159 163 173 186 237 132
35.8 31.8 27.4 18.7 0.0 0.0 0.0 0.0
dj rk (nm) (g cm3) 0.46 0.45 0.46 0.46 0.47 0.45 0.45 0.50
1.270 1.268 1.267 1.255 1.244 1.246 1.242 1.191
The storage modulus measured at 30 C. The glass transition temperature as measured by the peak of tan d. The storage modulus measured at the glass transition temperature. The melting temperature of the crystal structure. The storage modulus measured at Tc. The storage modulus measured at 300 C (before melting). The molecular weight between crosslink points. The glass transition temperature from DSC measurement. The degree of crystallinity calculated from WAXS measurement. The d-spacing as measured by peak top of halo pattern from WAXS measurement. The density.
Fig. 6 e WAXS patterns of DGOBP cured with DDM at different temperatures. and DGOBP samples and their corresponding DSC traces (Fig. 7 c). The peak height of tan d of DGOBP was smaller than that of BisA. The storage modulus of DGOBP cured with DDM widely varies from 150 to 250 C; the resulting peak in the tan
d curves is thought to indicate the glass transition temperature. The storage moduli of the samples at room temperature increased slightly with increasing curing temperature in the case of the DGOBP sample, which means that the amount of liquid crystal or crystal structure does not affect the storage modulus of the sample at room temperature. The decreasing storage modulus in the rubbery region differs between the crystalline samples and the liquid crystalline samples. The molecular weight between crosslink points of DGOBP cured with DDM at 130 C showed the lowest value of all samples tested. It was thought that the lower curing temperature samples formed low chemical crosslink points, attributed to the inhibition of the curing reaction as a result of the increased crystallinity. The temperature dispersion of tan d of DGOBP cured with DDM below 130 C (see Fig. 7 b) showed two peaks, one at approximately 210 C corresponding to melting of the crystal phase and the other the glass transition of the amorphous phase (which depended on the curing temperature). The peak height of tan d (indicating the glass transition) of the amorphous phase was smaller than that of the crystal phase at temperatures 120 C. The tan d curves of DGOBP cured with DDM above 140 C exhibited a single peak, indicating that these samples lacked the crystal phase. The tan d curve of BisA cured with DDM exhibited a single peak, however, the DSC trace (Fig. 7 c) showed an endothermic peak before a change in the heat capacity. This peak can be assigned to the enthalpy relaxation, according to a previous study of a similar cured BisA system [39,40]. The glass transition temperature of DGOBP cured with DDM derived from DSC measurements depended on the curing temperature. This phenomenon has also been observed in conventional epoxy polymer systems [41]. The relationship between curing temperature and glass transition temperature is in agreement with the DMA measurements. Neither of the DSC traces for cured BisA or DGOBP showed an exothermic peak of melting. Both the DMA and DSC measurements
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
7507
Fig. 7 e Thermal measurements of BisA and DGOBP cured with DDM at different curing temperatures (a) Temperature dependence of the storage modulus. (b) Temperature dependence of the dynamic loss tangent (tan d) of BisA and DGOBP cured with DDM at different curing temperatures. (c) DSC traces of BisA and DGOBP cured with DDM at different curing temperatures. The arrows shows the glass transition temperature.
demonstrated that we can obtain DGOBP cured with DDM with a crystal structure, which is supported by the WAXS results. Along with the epoxide conversions, Table 1 lists the storage moduli at room temperature, the glass transition temperatures, and molecular weight between crosslink points estimated from the storage moduli in the rubbery state. The glass transition temperatures estimated by the DSC measurements and the degree of crystallinity calculated from the WAXS measurements are also listed in Table 1. The degree of crystallinity can be calculated from the density of crystalline and amorphous phases; although the precise density of the crystal structure has not been clarified for DGOBP cured with DDM. However, the sample densities follow the trend of the WAXS measurements. Both sets of results support a crystalline phase in DGOBP cured with DDM, which diminishes as the curing temperature increases. Reducing the curing temperature increased the density of DGOBP cured with DDM, which is synonymous with increasing the degree of crystallinity. To clarify this point, BisA cured with DDM (also presented in Table 1) is less dense than DGOBP cured with DDM and exhibited no evidence of a crystal phase in the WAXS measurements. From the WAXS, DMA, and density measurements, we infer that DGOBP cured with DDM at low temperature comprises both crystalline and amorphous component. Thus, we can control the degree of crystallinity of DGOBP cured with DDM by adjusting the curing temperature.
Thermal conductivity Fig. 8 shows the thermal conductivity of DGOBP cured with DDM at different curing temperatures. The mean value of the thermal conductivity from 100 C to 130 C was 0.26e0.28 Wm1 K1 which is slightly lower than that of samples cured from 140 C to 180 C, 0.27e0.31 Wm1 K1. A
Fig. 8 e Thermal conductivity of DGOBP cured with DDM at different curing temperatures. The broken line is the conductivity of BisA cured with DDM. The error bars show the measurement error.
7508
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
Fig. 9 e Time dependence of the remaining hydrogen content in DGOBP cured with DDM at different curing temperatures.
decline in the thermal conductivity with increasing crystallinity for polyethylene terephthalate has been reported by Choy [42]. This decline in the thermal conductivity is attributed to the static scattering of phonons from the crystal interface. However, the thermal conductivity of the DGOBPDDM system always exceeded that of BisA-DMM (within experimental uncertainty) for the conditions of these experiments. The improvement in thermal conductivity for different monomers is attributed to the densely packed atoms in the liquid crystal structures and crystal structures.
Hydrogen properties Fig. 9 presents the hydrogen release profiles obtained from TDA measurements of the DGOBP cured with DDM by showing the remaining hydrogen content in the specimens after decompression from exposure to high-pressure hydrogen gas. The slope of the hydrogen release profiles moderately decreases with increasing degree of crystallinity, indicating a small diffusion coefficient of hydrogen in the polymer. The contents and diffusion coefficients of hydrogen in DGOBP cured with DDM obtained by least-squares fitting, are listed in Table 2.
Fig. 10 shows the hydrogen content and diffusion coefficient of hydrogen as functions of the curing temperature. Both of these properties decrease with increasing crystallinity of the DGOBP sample, and both are lower in DGOBP cured with DDM than in BisA cured with DDM. The differences in the hydrogen content values are larger than the estimated error in this study, and are hence significant. The hydrogen content of DGOBP cured with DDM are 25e75% smaller than that of conventional epoxy polymer. The volume change ratios of DGOBP cured with DDM at each curing temperature depended on the hydrogen content, and in the case of samples cured below 130 C, the volume change ratio decreased due to the high crystallinity. The hydrogen properties were not significantly correlated with the molecular weight between crosslink points. Fig. 11 shows the hydrogen properties as a function of the degree of crystallinity of DGOBP cured with DDM. The hydrogen penetration properties are strongly inversely correlated with the amount of crystal structure in the polymer, despite the low correlation with the molecular weight of crosslinking points. The inverse relationship between crystallinity and hydrogen content likely reflects the difficulty of penetrating the crystalline phase. The relationship between crystallinity and gas properties has been reported for polyethylene by Michaels et al. [43,44]. It is thought that hydrogen also penetrates the amorphous region or the liquid crystal region of the cured epoxy polymer, indicating that the hydrogen penetration properties could be estimated from the amount of hydrogen entering these regions. In other words, the hydrogen penetration property of DGOBP cured with DDM can be determined from the degree of crystallinity.
Conclusions In this study, we investigated the relationship between the hydrogen penetration properties and the ordered structure of a mono-mesogen type liquid-crystalline epoxy polymer, a candidate material for hydrogen storage vessels. We selected 4,40 -diglycidyloxybiphenyl (DGOBP) as a model epoxy polymer and prepared a series of samples with controlled ordered structures. The ordered structure of DGOBP was controlled by
Table 2 e Structural properties, thermal conductivity, and hydrogen penetration properties of DGOPB cured with DDM. Curing temperature ( C) 100 110 120 130 140 150 180 BisA (100) a b c d e
rMca (g mol1) 86 86 71 66 72 105 162 317
Dcb (%) 35.8 31.8 27.4 18.7 0.0 0.0 0.0 0.0
The molecular weight between crosslink points. The degree of crystallinity calculated from WAXS measurement. The thermal conductivity. The diffusion coefficient of hydrogen molecular. The hydrogen content.
Kc (Wm1 K1) 0.26 0.28 0.27 0.27 0.31 0.29 0.29 0.21
DHd (m2 s1) 3.77 3.89 4.58 5.17 9.90 1.07 1.11 4.35
12
10 1012 1012 1012 1012 1011 1011 1011
CHe (wt.ppm) 403 458 511 629 1271 1183 1206 1682
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
7509
range of 0.26e0.31 Wm1 K1, regardless of their degrees of crystallinity. Cured DGOBP showed hydrogen contents in the range of 403e1271 ppm, which is 25e75% smaller than that of conventional epoxy polymer. The hydrogen content and diffusion coefficient of DGOBP cured with DDM reduced with increasing crystallinity, indicating that both factors can be controlled by the degree of crystallinity. Because the ordered structure governs the diffusivity of hydrogen in liquid-crystal epoxy polymer, we conclude that the degree of crystallinity of the polymer strongly correlates with the hydrogen penetration. According to these results, DGOBP polymer with liquid-crystal structure can be a candidate polymer material for components in hydrogen systems, which require well-controlled hydrogen penetration properties. This suggests that we can design a suitable epoxy polymer with low hydrogen penetration properties by understanding and carefully controlling the development of the liquid-crystalline phase.
Acknowledgements This paper is based on results obtained from a project commissioned by the New Energy and Industrial Technology Development Organization (NEDO). The authors thank Hitoshi Koizumi and Mayuko Shido (NETZSCH Japan) for performing the thermal diffusivity measurements. Fig. 10 e Hydrogen content (a) and diffusion coefficient (b) of DGOBP cured with DDM at different curing temperatures. The broken line is the value of BisA cured with DDM. The error bars show the measurement error.
Fig. 11 e Relationship between hydrogen content, diffusion coefficient, and degree of crystallinity in DGOBP cured with DDM.
adjusting the curing temperature of the 4,40 -diaminodiphenyl methane (DDM) curing system. Crystal structures were observed in DGOBP cured with DDM below 130 C, and the degree of crystallinity decreased with increasing curing temperatures. We obtained a series of samples having 19%, 27%, 32%, and 36% crystallinity, corresponding to curing temperatures of 130, 120, 110, and 100, respectively. The thermal conductivity values of DGOBP cured with DDM were in the
references
[1] Strubel V. Final activity report. STORHY; 2008 Oct 15. Project No.502667. [cited 2015 Dec 18] Available form: http://www. storhy.net/pdf/StorHy_FinalPublActivityReport_FV.pdf/. [2] Monde M, Mitsutake Y, Woodfield PL, Maruyama S. Characteristics of heat transfer and temperature rise of hydrogen during rapid hydrogen filling at high pressure. Heat Transf Asian Res 2007;36(1):13e27. [3] Takeichi N, Senoh H, Yokota T, Tsuruta H, Hamada K, Takeshita HT, et al. “Hybrid hydrogen storage vessel”, a novel high-pressure hydrogen storage vessel combined with hydrogen storage material. Int J Hydrogen Energy 2003;28(10):1121e9. [4] Sakaguchi Y, Yamamoto Y, Nishiwaki H. Aluminum lined high pressure hydrogen cylinder. J Hydro Energy Syst Soc Jpn 2009;34(4):39e45. [5] Akatsuka M, Takezawa Y. Study of high thermal conductive epoxy resins containing controlled high-order structures. J Appl Polym Sci 2003;89(9):2464e7. € . Barrier properties of liquid crystalline polymers [6] Flodberg Go and their blends with PE and PETP. Sweden: KTH; 2002, ISBN 91-72-83-273-8. p. 26e34. [7] Weinkauf DH, Paul DR. Gas transport properties of thermotropic liquid-crystalline copolyesters. I. The effects of orientation and annealing. J Polym Sci Part B Polym Phys 1992;30(8):817e35. [8] Yamabe J, Nishimura S. Influence of fillers on hydrogen penetration properties and blister fracture of rubber composites for O-ring exposed to high-pressure hydrogen gas. Int J Hydrogen Energy 2009;34(4):1977e89. [9] Yamabe J, Matsumoto T, Nishimura S. Application of acoustic emission method to detection of internal fracture of
7510
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18] [19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
sealing rubber material by high-pressure hydrogen decompression. Polym Test 2011;30(1):76e85. Yamabe J, Nishimura S. Nanoscale fracture analysis by atomic force microscopy of EPDM rubber due to high-pressure hydrogen decompression. J Mater Sci 2011;46(7):2300e7. Yamabe J, Nishimura S. Influence of carbon black on decompression failure and hydrogen permeation properties of filled ethylene-propyleneedieneemethylene rubbers exposed to high-pressure hydrogen gas. J Appl Polym Sci 2011;122(5):3172e87. Yamabe J, Fujiwara H, Nishimura S. Fracture analysis of rubber sealing material for high pressure hydrogen vessel. J Environ Eng 2011;6(1):53e68. Koga A, Uchida K, Yamabe J, Nishimura S. Evaluation on high-pressure hydrogen decompression failure of rubber Oring using design of experiments. Int J Automot Eng 2011;2(4):123e9. Fujiwara H, Ono H, Nishimura S. Degradation behavior of acrylonitrile butadiene rubber after cyclic high-pressure hydrogen exposure. Int J Hydrogen Energ 2015;40(4):2025e34. Mititelu A, Cascaval CN. Liquid crystalline epoxy thermoset obtained from biphenyl mesogen. Polym Plast Technol Eng 2005;44(1):151e62. Li Y, Badrinarayanan P, Kessler MR. Liquid crystalline epoxy resin based on biphenyl mesogen: thermal characterization. Polymer 2013;54(12):3017e25. Li Y, Kessler MR. Liquid crystalline epoxy resin based on biphenyl mesogen: effect of magnetic field orientation during cure. Polymer 2013;54(21):5741e6. Li Y, Kessler MR. Creep-resistant behavior of self-reinforcing liquid crystalline epoxy resins. Polymer 2014;55(8):2021e7. Li Y, Kessler M. Cure kinetics of liquid crystalline epoxy resins based on biphenyl mesogen. J Therm Anal Calorim 2014;117(1):481e8. Su W-FA, Chen KC, Tseng SY. Effects of chemical structure changes on thermal, mechanical, and crystalline properties of rigid rod epoxy resins. J Appl Polym Sci 2000;78(2):446e51. Su W-FA. Thermoplastic and thermoset main chain liquid crystal polymers prepared from biphenyl mesogen. J Polym Sci Part A Polym Chem 1993;31(13):3251e6. € cher M. “Liquid crystalline“ thermosets Mormann W, Bro from 4,4’-bis(2,3-epoxypropoxy)biphenyl and aromatic diamines. Macromol Chem Phys 1996;197(6):1841e51. Carfagna C, Amendola E, Giamberini M, Filippov AG, Bauer RS. Curing kinetics of liquid-crystalline epoxy resins. Liq Cryst 1993;13(4):571e84. Lee J, Song Y, Kim S, Lee H. Dielectric and molecular dynamic studies of sub-Tg by thermally stimulated current (TSC) analysis for liquid crystalline epoxy thermosets. Mater Chem Phys 2003;77(2):455e60. Lu MG, Shim MJ, Kim SW. Curing reaction and phase change in a liquid crystalline monomer. Macromol Chem Phys 2001;202(2):223e30. Lin Q, Yee AF, Earls JD, Hefner Jr RE, Sue H-J. Phase transformations of a liquid crystalline epoxy during curing. Polymer 1994;35(12):2679e82. Shiraishi T, Motobe H, Ochi M, Nakanishi Y, Konishi I. Effect of network structure on thermal and mechanical
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42] [43] [44]
properties of cured epoxide resins. Polymer 1992;33(14):2975e80. Ortiz C, Kim R, Rodighiero E, Ober CK, Kramer EJ. Deformation of a polydomain, liquid crystalline epoxybased thermoset. Macromolecules 1998;31(13):4074e88. Gavrin AJ, Douglas EP. Isothermal curing of acetylene functionalized liquid crystalline thermoset monomers. Macromolecules 2001;34(17):5876e84. Liu G, Gao J, Song L, Hou W, Zhang L. Synthesis and curing of liquid-crystalline epoxy resins containing a biphenyl mesogen. Macromol Chem Phys 2006;207(23):2222e31. Douglas EP. Structure-property relationships in liquid crystalline thermosets. J Macromol Sci Part C 2006;46(2):127e41. Harada M, Ando J, Yamaki M, Ochi M. Synthesis, characterization, and mechanical properties of a novel terphenyl liquid crystalline epoxy resin. J Appl Polym Sci 2015;132(1):41296. Cho S, Douglas EP, Lee JY. Transition diagrams for a liquid crystalline thermoset containing a rigid-rod epoxy. Polym Eng Sci 2006;46(5):623e9. € cher M. ‘Liquid crystalline’ thermosets Mormann W, Bro (LCT) from diaromatic diepoxides and aromatic diamines. Polymer 1998;39(20):4905e10. Barton J. The application of differential scanning calorimetry (DSC) to the study of epoxy resin curing reactions. In: Dusek K, editor. Epoxy resins and composites I. Berlin Heidelberg: Springer-Verlag; 1985. p. 111e54. Min B-G, Stachurski Z, Hodgkin J. Cure kinetics of elementary reactions of a DGEBA/DDS epoxy resin: 1. Glass transition temperature versus conversion. Polymer 1993;34(23):4908e12. Achilias DS, Karabela MM, Varkopoulou EA, Sideridou ID. Cure kinetics study of two epoxy systems with fourier transform infrared spectroscopy (FTIR) and differential scanning calorimetry (DSC). J Macromol Sci Part A 2012;49(8):630e8. Vonk CG. Computerization of Ruland's X-ray method for determination of the crystallinity in polymers. J Appl Crystallogr 1973;6(2):148e52. Neuser S, Michaud V. Effect of aging on the performance of solvent-based self-healing materials. Polym Chem 2013;4:4993e9. Bogaerts K, Lavrenova A, Spoelstra AB, Boyard N, Goderis B. Curing kinetics and morphology of a nanovesicular epoxy/ stearyl-block-poly(ethylene oxide) surfactant system. Soft Matter 2015;31(11):6212e22. Carbas RJC, Marques EAS, da Silva LFM, Lopes AM. Effect of cure temperature on the glass transition temperature and mechanical properties of epoxy adhesives. J Adhes 2014;90(1):104e19. Choy C. Thermal conductivity of polymers. Polymer 1977;18(10):984e1004. Michaels AS, Parker RB. Sorption and flow of gases in polyethylene. J Polym Sci 1959;41(138):53e71. Michaels AS, Bixler HJ. Flow of gases through polyethylene. J Polym Sci 1961;50(154):413e39.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Hydrogen characteristics and ordered structure of mono-mesogen type liquid-crystalline epoxy polymer Shuji Kawamoto a, Hirotada Fujiwara b, Shin Nishimura c,* a
Department of Hydrogen Energy Systems, Graduate School of Engineering, Kyushu University, Japan Research Center for Hydrogen Industrial Use and Storage, Kyushu University, Japan c Department of Mechanical Engineering, Faculty of Engineering, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka 819-0395, Japan b
article info
abstract
Article history:
Fuel cell systems, such as those used in electric vehicles require lightweight hydrogen fuel
Received 12 October 2015
storage. Polymer composites are candidate materials for storage vessels requiring high
Received in revised form
fracture resistance for high pressure cycling, as well as high thermal conductivity. Here, we
7 March 2016
investigate 4,40 -diglycidyloxybiphenyl (DGOBP) epoxy for such applications, and clarified
Accepted 20 March 2016
the relationship between the hydrogen penetration properties and the ordered structure of
Available online 10 April 2016
the epoxy. We controlled the ordered structure of the DGOBP polymer by adjusting the curing temperature of a 4,40 -diaminodiphenyl methane (DDM) curing system. We obtained
Keywords:
a series of cured DGOBP with DDM samples with crystallinity values of 19%, 27%, 32%, and
Epoxy
36%, corresponding to curing temperatures of 130, 120, 110, and 100 C, respectively. The
Polymer
thermal conductivities of these samples were 0.26e0.31 Wm1 K1. Cured DGOBP showed
Crystallization
hydrogen contents of 403e1271 ppm (25e75% smaller than that of conventional epoxies).
Liquid crystal
Liquid crystalline epoxy polymers can be a candidate material for hydrogen storage sys-
Thermal conductivity
tems requiring well-controlled hydrogen penetration properties.
Hydrogen properties
Copyright © 2016, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Introduction Hydrogen energy systems are being investigated for nextgeneration clean energy systems. However, safety is a primary concern for commercial application and must be considered in the design of hydrogen energy systems. For hydrogen energy to be feasible and widely implemented in transport and stationary technologies, it needs to be safely and efficiently stored [1]. Compressed gas is currently the
preferred solution for storing hydrogen in fuel cell vehicles (FCV). To satisfy the desired driving range and convenience of FCVs, hydrogen should be stored at 70 MPa and filled at a rapid rate, reaching its storage pressure within 3 min. To avoid thermal damage to the storage vessel, the hydrogen gas needs to be pre-cooled to 40 C at the hydrogen filling station. Because the compression can greatly elevate the temperature of the gas [2], vessel materials must effectively radiate heat. Moreover, hydrogen storage vessels installed in FCVs need to be lightweight [3] in order to achieve high fuel efficiency.
* Corresponding author. Tel.: þ81 92 802 3248. E-mail address: [email protected] (S. Nishimura). http://dx.doi.org/10.1016/j.ijhydene.2016.03.124 0360-3199/Copyright © 2016, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
The current FCV hydrogen vessel is a “type IV vessel”, which consists of a fully-wrapped carbon fibre reinforced plastic, such as an epoxy polymer with a plastic liner [4]. A disadvantage of polymeric materials for this application is that they generally have low thermal conductivity. In contrast, liquid-crystalline epoxy polymers, such as twinmesogen type liquid-crystalline epoxy polymers, have been reported as strong thermal conductors [5]. Therefore, we consider that liquid-crystal polymers are suitable liner materials or matrices for carbon fibre reinforced plastics in type IV vessels. The high heat dissipation of liquid-crystal polymers should allow a higher pre-cooling temperature during filling, reducing the difficulty of sealing hydrogen gas at low temperature. In addition, it was thought that the highly dense structure of the liquid-crystal polymer should provide an effective barrier to hydrogen permeation [6,7]. In this study, we design a liquid-crystal epoxy polymer suitable as a gas-barrier material for type IV hydrogen storage vessels. A number of studies investigating the hydrogen properties of rubber as a sealing material have been reported [8e14]. It was shown that, when exposed to high-pressure hydrogen gas, the polymer materials were penetrated by hydrogen molecules. The amount of hydrogen penetration into the materials and the volume change ratio have been suggested as criteria for standardising polymer materials for high-pressure hydrogen equipment. Low penetration and volume change ratio are preferred for hydrogen barrier materials from the viewpoint of avoiding fractures and gas leaks. In addition, the thermal conductivity of the hydrogen barrier materials should be much lower than the current state-of-theart materials considering safety and degradation of the polymers. The hydrogen penetration and thermal conductivity are expected to correlate with the “ordered structure”, such as the crystal domain or liquid-crystal structure of the material. Understanding the effect of this ordered structure is important for designing a reliable structural material for type IV vessels to store high-pressure hydrogen gas. Here, we investigate the influence of the structure of an epoxy polymer on the hydrogen penetration properties and thermal conductivity. We selected a mono-mesogen type liquid-crystal polymer, which has a controllable ordered structure, as a model material having a similar structure, molecular size and reactive groups as conventional epoxy polymers such as bis(4diglycidyloxyphenyl)propane. The selected material for this study was 4,40 -diglycidyloxybiphenyl (DGOBP) which has one biphenyl group as mesogen. A number of studies of DGOBP have been reported, using curing systems such as 2,7-diaminofluorene [15], sulfanilamide [16e20], 4,40 -diaminodiphenylsulfone [21], 4,40 -diaminobiphenyl and 4-methyl-m-phenylenediamine [22,23], and 4,40 -diaminodiphenyl methane (DDM) [5,24e27]. However, the relationship between the ordered structure and the hydrogen properties, in particular for DGOBP cured by DDM, have not been reported. In this study we fabricated a series of DGOBP samples cured with different DDM concentrations to optimize the ordered structure and investigate the effects on the hydrogen penetration and thermal conductivity. Such an analysis is expected to provide useful information for proposing guidelines for the design of polymeric materials for hydrogen systems at the molecular level.
7501
Materials and methods Materials 4,40 -biphenol and DDM were purchased from Tokyo Kasei Kogyo Co., Ltd. Chloromethyloxirane, tetrabuthylammonium bromide (TBAB), sodium hydroxide and sulfanilamide were purchased from Wako Pure Chemical Industries Co., Ltd. Epikote828, also known as Bisphenol A (BisA), was purchased from Mitsubishi Chemical Co., Ltd. as conventional epoxy polymer. The main ingredient of Epikote828 is 2,2-bis(4diglycidyloxyphenyl)propane. All compounds were used as received without further purification. The mono-mesogen type liquid-crystal epoxy monomer DGOBP was synthesised as previously described [16]. The synthesis route for the DGOBP monomer and the chemical structures of the DGOBP monomer and curing agent are shown in Scheme 1. The product was purified by reprecipitation from water and methanol. To observe the ordered structures of the DGOBP samples cured with DDM, the monomer and curing agent were combined at a molar ratio of 2:1. This stoichiometric mixture of DGOBP and DDM was melted and mixed by stirring at 150 C on a hot plate for 1 min. Afterwards the curing reaction was quenched by liquid nitrogen. To prepare samples for evaluating the conversion of the epoxy group in DGOBP to the amino group in DDM, this same process was employed. However, the curing reaction was undertaken for 2, 3, 5, or 7 min. Disk-shaped specimens of DGOBP cured with DDM (diameter: 40 mm, thickness: 2 mm) were prepared from a stoichiometric mixture of the polymer and curing agent (molar ratio 2:1). The DGOBPeDDM mixture was prepared as described above, stirred for 3 min, and then cured in a mould for 4 h. The curing temperature was varied: 100, 110, 120, 130, 140, 150 or 180 C. A stoichiometric mixture of BisA and DDM
Scheme 1 e (a) Synthesis route for the DGOBP monomer. (b) Chemical structure of the BisA monomer. (c) Chemical structure of the DDM curing agent.
7502
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
(molar ratio 2:1) was cured at 100 C for 4 h. The cured polymers were cut into test pieces for subsequent analyses.
measurements at room temperature were undertaken with a scanning speed of 2 min1, resolution of 0.02 , and a 2q range of 10e50 .
Characterization Dynamic mechanical analysis (DMA) 1
H and
13
C nuclear magnetic resonance (NMR)
NMR spectra were obtained using an AvanceII 500-MHz spectrometer (Bruker BioSpin Co., Ltd.) operated at room temperature (1H NMR at 500 MHz and 13C NMR at 125 MHz). The chemical shifts were measured relative to the peaks of tetramethylsilane as a standard (d in 0.00 ppm).
Liquid chromatography Liquid chromatography was performed using an HPLC system (HLC-8320GPC, Tosoh Co., Ltd.) equipped with two TSKgel SuperHZ-M columns and one TSKgel SuperHZ2000 column. Measurements were acquired in size exclusion chromatography mode with a reflective index detector, tetrahydrofuran elution solvent, flow rate of 0.35 mL min1, and injection volume of 0.1 mL for a 1% w/v sample.
Fourier transform infrared spectroscopy (FT-IR) FT-IR spectra of the DGOBP monomer, pre-cured samples, and DGOBP cured with DDM were measured using a FT-IR Spectrum100s (PerkinElmer Co., Ltd.) by the KBr disc method in the wavelength range 450e4000 cm1 with a resolution of 4 cm1. Eight scans were performed.
Differential scanning calorimetry (DSC) DSC measurements of stoichiometric DGOBP and DDM mixtures (5 mg powder samples) were carried out using a DSC 204 HP instrument (Netzsch GmbH) in a nitrogen atmosphere from room temperature to 300 C. The heating rate was varied: 5.0, 7.5, or 10.0 C min1. Isothermal DSC measurements were performed on a 5 mg molten DGOBPeDDM sample. The heating rate was 50 C min1 from room temperature to the test temperature of 130.0, 132.5, 135.0, 137.5, 140.0, 150.0, 160.0, or 170.0 C, where the sample was held for 20 min in a nitrogen atmosphere. The DGOBP cured with DDM at different curing temperatures and the BisA cured with DDM (10 mg solid samples) were measured at a heating rate of 10 C min1 from room temperature to 300 C in a nitrogen atmosphere. The glass transition temperature was determined by the temperature at the half height of the heat capacity change in the DSC curve.
Polarized optical microscopy The phase transition behaviour of the molten DGOBPeDDM samples was observed using optical microscopy (Eclipse LV150, Nikon Co., Ltd., 10 eyepiece, 10 or 20 objective lenses) equipped with a crossed polarizer and a hot stage (CSS450WCX, Japan High Tech Co., Ltd.). The temperature of the hot stage was calibrated by measuring the melting points of DDM (90 C), TBAB (104 C), and sulphanilamide (165 C).
Wide-angle X-ray scattering (WAXS) Samples of DGOBP cured with DDM were cut into disc-shaped specimens (diameter: 13 mm, thickness: 2 mm) for the WAXS measurements. A MultiFlex X-ray diffraction unit (Rigaku Co., Ltd.) with Cu-Ka radiation was used. Continuous
DMA was performed using a DVA-200s instrument (IT Keisokuseigyo Co., Ltd.) from 30 to 300 C at a heating rate of 10 C min1 in tensile mode, with a strain amplitude of 0.08% and a frequency of 10 Hz. Rectangular specimens cut from cured samples with a cross-sectional area of (1.5 1.5) mm2 and a length of 15 mm were used. The samples were fixed at a distance of 10 mm between the grips of the DMA machine and measured a 2 C intervals. The molecular weight between crosslink points of the cured samples was calculated from the storage modulus (tan d) at 280 C (where tan d converged for every cured polymer). Here, d is the stressestrain phase lag.
Thermal conductivity measurements Disc-shaped specimens (diameter of 10 mm; thickness of 0.4 mm) were cut from DGOBP cured with DDM, spray-coated with thin gold and graphite electrodes. The thermal diffusivity of the cured samples was measured by the half-time method using a xenon flash analyser (LFA 467 Hyper Flash, NETZSCH GmbH) at 25 C. The specific heat capacity was measured by DSC with Al2O3 as the reference material, as described in ISO11357-4. The density of DGOBP samples cured with DDM was measured using a Sartorius LA230S balance equipped with the density determination kit YDK01 (Sartorius Co., Ltd.), using ethanol at 25 C. The thermal conductivity of the cured samples was calculated from the thermal diffusivity, heat capacity, and density. The measurement error of the thermal conductivity was 8%.
Hydrogen content, diffusion coefficient, and volume change after hydrogen exposure Disc-shaped specimens with a diameter of 13 mm and a thickness of 2 mm were cut from DGOBP cured with DDM and exposed to 90 MPa hydrogen gas at 30 C for 24 h in a highpressure vessel. The hydrogen release profile after decompression was measured using a thermal desorption analyser (JSH-201, J-Science Lab Co., Ltd.). The amount of hydrogen released from the sample was determined by gas chromatography at 30 C in a flow of argon every 5 min. The diffusion properties of hydrogen were estimated by fitting the data. The hydrogen content immediately after decompression (CH 0 ; t ¼ 0) and the diffusion coefficient D were estimated by fitting Eq. (1) to the remaining hydrogen content of the specimen using the least-squares method [8].
CHR ðtÞ
n . o# " 2 ∞ exp ð2n þ 1Þ p2 Dt l2 0 32 H X ¼ 2 $C0 $ p ð2n þ 1Þ2 n¼0 " # ∞ X exp Db2n t r20 $ b2n n¼1
(1)
where CHR ðtÞ (wt. ppm) is the remaining hydrogen content at time t (s), the time elapsed since decompression. CH0 (wt. ppm) is the hydrogen content at t ¼ 0, D (m2 s1) is the diffusion coefficient, bn is the root of the zero-order Bessel function, and l (m) and r (m) are the thickness and radius of the specimen, respectively. The measurement errors for the hydrogen
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
7503
content and diffusion coefficient were estimated from the accuracy of the TDA chromatography and the measurement error of specimens from the propagation of errors from Equation (1). The measurement error for the hydrogen content and diffusion coefficient were 5% and 14%, respectively. The volume change of the disc-shaped specimens after hydrogen exposure was observed at 30 C by a laser scan micrometre (TM-3000, Keyence Co., Ltd.) with a TM-065 sensor head and an InGaN green LED light source. The disc area was measured at 5 min intervals. The volume was calculated by cubing the square root of the area and was expressed as a proportion of the initial sample dimensions (i.e. the volume change ratio).
Results and discussion Synthesis of DGOBP 4,4’-dihydroxybiphenyl and chloromethyloxirane reacted stoichiometrically with NaOH (aq.) and TBAB (see Scheme 1). The product was analysed by liquid chromatography using a differential refractive index detector where the peak area was calculated from the chromatogram. The reaction products of Scheme 1 comprised 95.1% DGOBP monomer, 3.7% DGOBP dimer, and 1.2% DGOBP trimer. The chemical structure of the synthesized DGOBP was confirmed by the 1H and 13C peaks identified in NMR and IR spectra. The data for the 1H peak from NMR (500 MHz, CDCl3, d) are as follows: 7.49 ppm (d, J ¼ 9 Hz, 4H, Ar H), 7.00 ppm (d, J ¼ 9 Hz, 4H, Ar H), 4.27 ppm (dd, J ¼ 11 and 3 Hz, 2H, glycidyl CH2), 4.03 ppm (dd, J ¼ 11 and 6 Hz, 2H, glycidyl CH2), 3.40 ppm (m, 2H, glycidyl CH2), 2.94 ppm (dd, J ¼ 5 and 4 Hz, 2H, epoxy CH2), and 2.80 ppm (dd, J ¼ 5 and 3 Hz, 2H, epoxy CH2). The data for the 13C peak determined by NMR (125 MHz, CDCl3, d) are as follows: 157.8 ppm (OCCHCHCAr), 134.0 ppm (OCCHCHCAr), 127.8 ppm (OCCHCHCAr), 115.0 ppm (OCCHCHCAr), 69.0 ppm (CH2OCHCH2O), 50.1 ppm (CH2OCHCH2O), and 44.7 ppm (CH2OCHCH2O). The IR (KBr) data are as follows: n ¼ 3063 cm1 (w; nas (Ar CH)), 3041 cm1 (w; ns (Ar CH)), 3008 cm1 (w; ns (epoxy CH)), 2929 cm1 (w; nas (CH)), 2877 cm1 (w; ns (CH)), 1606 cm1 (s; n (Ar ring)), 1501 cm1 (s; d (Ar ring)), 1244 cm1 (s; nas (COC)), 1180 cm1 (m; d (Ar CH)), 1134 cm1 (m; d (Ar CH)), 1037 cm1 (s; ns (COC)), 911 cm1 (m; nas (epoxy COC)), 814 cm1 (s; g (Ar CH)), and 762 cm1 (m; g (Ar CH)), melting point ¼ 160 C.
Curing behaviour The curing behaviour was determined by DSC as described in Section Materials and methods. The curing temperature was determined from the exothermic peak of the DSC traces of the stoichiometric DGOBPeDDM mixture [25]. At each heating rate (5.0, 7.5, and 10.0 C min1), the exotherm of the curing reaction appears at temperatures above 130 C, higher than the endotherm related to the melting of the DDM (around 85 C for this purity of monomer). Fig. 1 shows the isothermal DSC traces of the melting process of the DGOBPeDDM mixtures cured at different temperatures. The isothermal DSC traces show a broad peak related to the curing reaction after the samples melted. The peak width decreased with increasing isothermal temperature, indicating an increase in the reaction rate at
Fig. 1 e Isothermal DSC curves of the DGOBP monomer cured with DDM at different temperatures. The arrows indicate the onset of exothermic peaks.
higher isothermal temperatures. The DSC traces of the mixtures cured below 135 C showed an additional exothermic peak, which appeared 3e4 min after reaching the target temperature. This peak was absent in mixtures cured above 137.5 C, suggesting that it originates from the ordered structure in the samples (as will be discussed later). The DGOBPeDDM mixtures used in the isothermal curing study were prepared for observation using polarised optical microscopy by mixing at 150 C for 1 min, followed by rapid cooling in liquid nitrogen [26]. Note that although DGOBP was dissolved in the molten DDM at 150 C, its melting point is actually 160 C. Fig. 2 shows polarised optical micrographs of the DGOBPeDDM mixtures obtained at 130 C and 135 C. After 3e4 min, a birefringent domain began to emerge in the sample cured at 130 C and a birefringent texture began to emerge in the sample cured at 135 C. The ordered structure of such birefringent domains in the sample cured at 130 C was confirmed by wide angle X-ray patterns reported by Mititelu et al. [15]. A magnified view of the birefringent domain of the sample cured at 130 C is shown in Fig. 3. The nucleation and growth (to the order of 10 mm) of a spherulite-like structure is clearly visible in these micrographs. The polarised optical micrographs of samples cured at 135 C showed no definite crystallite structure; the spherulite-like structure was absent in these samples. Li et al. attributed the extra exothermic peak
7504
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
Fig. 2 e Polarised optical microscopy images (100 £ magnification) of the DGOBP monomer mixed with DDM at a curing temperature of 130 C and 135 C. (a) 1 min, (b) 4 min, (c) 5 min, and (d) 10 min.
Fig. 3 e Polarised optical microscopy images (200 £ magnification) of the DGOBP monomer mixed with DDM at a curing temperature of 130 C. (a) 3 min 30 s, (b) 3 min 40 s, (c) 3 min 50 s, and (d) 4 min.
in the isothermal DSC traces to the smectic phase [16]. Hence, we infer that the birefringence observed in the polarised optical microscopy images of samples cured at high temperature (above 135 C) reflects the nematic phase. This suggests that the exothermic peaks in the isothermal DSC traces originate from the crystal phase. Fig. 4 shows the timeetemperatureephase transformation (TTT) diagram indicating the regions of the isotropic, liquidcrystal, and crystal phases of the DGOBP isothermally cured with DDM [28e33]. The solid curves represent the onset of the
liquid-crystal phase and the crystal phase transition from the isotropic phase. These onsets were determined from the polarised optical micrographs of the ordered phase observed during isothermal curing. The broken line marks the boundary between the liquid-crystal and crystal phases. Therefore, the ordered structure of DGOBP cured with DDM can be controlled by the curing temperature. Thus, to control the amount of ordered structure in the cured DGOBP with DDM, cured polymer samples were prepared at 100, 110, 120, 130, 140, 150, and 180 C for 4 h.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
7505
Fig. 4 e Timeetemperatureephase transformation diagram of the isothermal curing of DGOBP mixed with DDM.
Conversion of the epoxy group The conversion of the epoxy group was determined from the FT-IR and DSC measurements. The conversion of the aminebased curing system has been studied by many authors [34e37]. In terms of the methodology of these studies, the calibration curve for the conversion of the epoxy group by the consumption of the DGOBP monomer and DDM was defined in this study. The area under the exothermic peak in the DSC traces of the pre-cured samples was assumed as the residual heat of the curing reaction. The total heat of reaction was assumed to be the residual heat of the curing reaction of the uncured DGOBPeDDM mixtures. The overall conversion was calculated from dividing the residual heats by the total heats of the curing reaction. In the FT-IR spectra of the pre-cured samples, we computed the ratio of the peaks attributed to the CO stretching vibration of the epoxy group (at 910 cm1) and the in-plane vibration of the phenyl ring (at 1610 cm1). The decrease in the height of the 910 cm1 peak was attributed to the consumption of monomer. The peak at 1610 cm1 was adopted as the internal standard for the spectra. Using the calibration curve method, the conversion of the epoxy group can be evaluated regardless of the overlap of the peaks originating from other chemical species. By analysing the changes in the 910 cm1/1610 cm1 peak ratios, we derived the epoxide conversion in the DGOBP cured with DDM. Fig. 5 shows the FT-IR spectra of the DGOBP cured with DDM at different temperatures. The KBr disc method used to acquire the data allowed us to evaluate the conversion both on the surface and in the bulk of the samples. The cured samples show reduced absorption intensity of the epoxide group at 910 cm1. The percentage conversion values of the epoxide groups at each curing temperature, obtained by the peak ratio of 910 cm1/1610 cm1, are summarised in Table 1. Above 140 C, the epoxide group was completely consumed. Even at lower curing temperatures, more than 94% of the epoxide groups were consumed during the curing process. From these results, it can be concluded that the curing of DGOPB with DDM was successful at all temperatures investigated here.
Fig. 5 e FT-IR spectra of DGOBP cured with DDM at different temperatures.
Observation of the ordered structure Fig. 6 shows the WAXS patterns of DGOBP cured with DDM at different temperatures. Samples cured below 130 C exhibited three diffraction peaks, at 2q ¼ 19 , 22 , and 27 , corresponding to d-spacings of 0.46, 0.40, and 0.32 nm, respectively. At curing temperatures below 130 C, crystalline phases should be present in the cured polymer. The d-spacing of the sample cured above 140 C was attributed to the stacking of mesogen groups, implying that the liquid crystal structure was frozen in the cured polymer [7]. We evaluated the degree of crystallinity by integrating the intensity of the crystal peak and broad peak (halo patterns) from the WAXS patterns [38]. The crystallinity values of our DGOBPeDDM samples were 19%, 27%, 32%, and 36%, where the degree of crystallinity increased as the curing temperature decreased. This crystallization gives rise to the exothermic peaks observed in the isothermal DSC traces (see Fig. 1). Douglas et al. and Cho et al. have reported a relationship between the phase transition and the gel point [31,33]. Smaller spherulite particles were formed in DGOBP cured at 130 C compared to those in samples cured at 100 C. This was attributed to a decline in the mobility of the molecular chains as a result of gelation, preventing an increase in the spherulite size. Fig. 7 shows the storage modulus and tan d from DMA measurements (Fig. 7 a, b, respectively) of DDM-cured BisA
7506
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
Table 1 e Epoxide conversion and thermal, physical, and structural properties of DGOPB cured with DDM. Curing temperature ( C) 100 110 120 130 140 150 180 BisA (100) a b c d e f g h i j k
Conversion E0 at r.ta (%) (Pa)
Tgb ( C)
E0 at Tgc (Pa)
Tcd ( C)
E0 at Tce (Pa)
109 109 109 109 109 109 109 109
127 151 170 179 192 203 234 140
1.42 1.25 9.25 7.44 5.00 4.28 3.00 5.63
109 109 108 108 108 108 108 107
213 216 214 208 e e e e
2.77 2.82 4.72 4.65
94 95 97 98 100 100 100 e
1.94 1.99 2.02 2.07 2.21 2.21 2.30 2.68
e e e e
108 108 108 108
E0 at Tmf (Pa) 1.98 2.01 2.18 2.18 1.66 1.13 8.63 4.90
108 108 108 108 108 108 107 107
rMcg (g mol1)
Tgh ( C)
Dc i (%)
86 86 71 66 72 105 162 317
144 149 159 163 173 186 237 132
35.8 31.8 27.4 18.7 0.0 0.0 0.0 0.0
dj rk (nm) (g cm3) 0.46 0.45 0.46 0.46 0.47 0.45 0.45 0.50
1.270 1.268 1.267 1.255 1.244 1.246 1.242 1.191
The storage modulus measured at 30 C. The glass transition temperature as measured by the peak of tan d. The storage modulus measured at the glass transition temperature. The melting temperature of the crystal structure. The storage modulus measured at Tc. The storage modulus measured at 300 C (before melting). The molecular weight between crosslink points. The glass transition temperature from DSC measurement. The degree of crystallinity calculated from WAXS measurement. The d-spacing as measured by peak top of halo pattern from WAXS measurement. The density.
Fig. 6 e WAXS patterns of DGOBP cured with DDM at different temperatures. and DGOBP samples and their corresponding DSC traces (Fig. 7 c). The peak height of tan d of DGOBP was smaller than that of BisA. The storage modulus of DGOBP cured with DDM widely varies from 150 to 250 C; the resulting peak in the tan
d curves is thought to indicate the glass transition temperature. The storage moduli of the samples at room temperature increased slightly with increasing curing temperature in the case of the DGOBP sample, which means that the amount of liquid crystal or crystal structure does not affect the storage modulus of the sample at room temperature. The decreasing storage modulus in the rubbery region differs between the crystalline samples and the liquid crystalline samples. The molecular weight between crosslink points of DGOBP cured with DDM at 130 C showed the lowest value of all samples tested. It was thought that the lower curing temperature samples formed low chemical crosslink points, attributed to the inhibition of the curing reaction as a result of the increased crystallinity. The temperature dispersion of tan d of DGOBP cured with DDM below 130 C (see Fig. 7 b) showed two peaks, one at approximately 210 C corresponding to melting of the crystal phase and the other the glass transition of the amorphous phase (which depended on the curing temperature). The peak height of tan d (indicating the glass transition) of the amorphous phase was smaller than that of the crystal phase at temperatures 120 C. The tan d curves of DGOBP cured with DDM above 140 C exhibited a single peak, indicating that these samples lacked the crystal phase. The tan d curve of BisA cured with DDM exhibited a single peak, however, the DSC trace (Fig. 7 c) showed an endothermic peak before a change in the heat capacity. This peak can be assigned to the enthalpy relaxation, according to a previous study of a similar cured BisA system [39,40]. The glass transition temperature of DGOBP cured with DDM derived from DSC measurements depended on the curing temperature. This phenomenon has also been observed in conventional epoxy polymer systems [41]. The relationship between curing temperature and glass transition temperature is in agreement with the DMA measurements. Neither of the DSC traces for cured BisA or DGOBP showed an exothermic peak of melting. Both the DMA and DSC measurements
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
7507
Fig. 7 e Thermal measurements of BisA and DGOBP cured with DDM at different curing temperatures (a) Temperature dependence of the storage modulus. (b) Temperature dependence of the dynamic loss tangent (tan d) of BisA and DGOBP cured with DDM at different curing temperatures. (c) DSC traces of BisA and DGOBP cured with DDM at different curing temperatures. The arrows shows the glass transition temperature.
demonstrated that we can obtain DGOBP cured with DDM with a crystal structure, which is supported by the WAXS results. Along with the epoxide conversions, Table 1 lists the storage moduli at room temperature, the glass transition temperatures, and molecular weight between crosslink points estimated from the storage moduli in the rubbery state. The glass transition temperatures estimated by the DSC measurements and the degree of crystallinity calculated from the WAXS measurements are also listed in Table 1. The degree of crystallinity can be calculated from the density of crystalline and amorphous phases; although the precise density of the crystal structure has not been clarified for DGOBP cured with DDM. However, the sample densities follow the trend of the WAXS measurements. Both sets of results support a crystalline phase in DGOBP cured with DDM, which diminishes as the curing temperature increases. Reducing the curing temperature increased the density of DGOBP cured with DDM, which is synonymous with increasing the degree of crystallinity. To clarify this point, BisA cured with DDM (also presented in Table 1) is less dense than DGOBP cured with DDM and exhibited no evidence of a crystal phase in the WAXS measurements. From the WAXS, DMA, and density measurements, we infer that DGOBP cured with DDM at low temperature comprises both crystalline and amorphous component. Thus, we can control the degree of crystallinity of DGOBP cured with DDM by adjusting the curing temperature.
Thermal conductivity Fig. 8 shows the thermal conductivity of DGOBP cured with DDM at different curing temperatures. The mean value of the thermal conductivity from 100 C to 130 C was 0.26e0.28 Wm1 K1 which is slightly lower than that of samples cured from 140 C to 180 C, 0.27e0.31 Wm1 K1. A
Fig. 8 e Thermal conductivity of DGOBP cured with DDM at different curing temperatures. The broken line is the conductivity of BisA cured with DDM. The error bars show the measurement error.
7508
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
Fig. 9 e Time dependence of the remaining hydrogen content in DGOBP cured with DDM at different curing temperatures.
decline in the thermal conductivity with increasing crystallinity for polyethylene terephthalate has been reported by Choy [42]. This decline in the thermal conductivity is attributed to the static scattering of phonons from the crystal interface. However, the thermal conductivity of the DGOBPDDM system always exceeded that of BisA-DMM (within experimental uncertainty) for the conditions of these experiments. The improvement in thermal conductivity for different monomers is attributed to the densely packed atoms in the liquid crystal structures and crystal structures.
Hydrogen properties Fig. 9 presents the hydrogen release profiles obtained from TDA measurements of the DGOBP cured with DDM by showing the remaining hydrogen content in the specimens after decompression from exposure to high-pressure hydrogen gas. The slope of the hydrogen release profiles moderately decreases with increasing degree of crystallinity, indicating a small diffusion coefficient of hydrogen in the polymer. The contents and diffusion coefficients of hydrogen in DGOBP cured with DDM obtained by least-squares fitting, are listed in Table 2.
Fig. 10 shows the hydrogen content and diffusion coefficient of hydrogen as functions of the curing temperature. Both of these properties decrease with increasing crystallinity of the DGOBP sample, and both are lower in DGOBP cured with DDM than in BisA cured with DDM. The differences in the hydrogen content values are larger than the estimated error in this study, and are hence significant. The hydrogen content of DGOBP cured with DDM are 25e75% smaller than that of conventional epoxy polymer. The volume change ratios of DGOBP cured with DDM at each curing temperature depended on the hydrogen content, and in the case of samples cured below 130 C, the volume change ratio decreased due to the high crystallinity. The hydrogen properties were not significantly correlated with the molecular weight between crosslink points. Fig. 11 shows the hydrogen properties as a function of the degree of crystallinity of DGOBP cured with DDM. The hydrogen penetration properties are strongly inversely correlated with the amount of crystal structure in the polymer, despite the low correlation with the molecular weight of crosslinking points. The inverse relationship between crystallinity and hydrogen content likely reflects the difficulty of penetrating the crystalline phase. The relationship between crystallinity and gas properties has been reported for polyethylene by Michaels et al. [43,44]. It is thought that hydrogen also penetrates the amorphous region or the liquid crystal region of the cured epoxy polymer, indicating that the hydrogen penetration properties could be estimated from the amount of hydrogen entering these regions. In other words, the hydrogen penetration property of DGOBP cured with DDM can be determined from the degree of crystallinity.
Conclusions In this study, we investigated the relationship between the hydrogen penetration properties and the ordered structure of a mono-mesogen type liquid-crystalline epoxy polymer, a candidate material for hydrogen storage vessels. We selected 4,40 -diglycidyloxybiphenyl (DGOBP) as a model epoxy polymer and prepared a series of samples with controlled ordered structures. The ordered structure of DGOBP was controlled by
Table 2 e Structural properties, thermal conductivity, and hydrogen penetration properties of DGOPB cured with DDM. Curing temperature ( C) 100 110 120 130 140 150 180 BisA (100) a b c d e
rMca (g mol1) 86 86 71 66 72 105 162 317
Dcb (%) 35.8 31.8 27.4 18.7 0.0 0.0 0.0 0.0
The molecular weight between crosslink points. The degree of crystallinity calculated from WAXS measurement. The thermal conductivity. The diffusion coefficient of hydrogen molecular. The hydrogen content.
Kc (Wm1 K1) 0.26 0.28 0.27 0.27 0.31 0.29 0.29 0.21
DHd (m2 s1) 3.77 3.89 4.58 5.17 9.90 1.07 1.11 4.35
12
10 1012 1012 1012 1012 1011 1011 1011
CHe (wt.ppm) 403 458 511 629 1271 1183 1206 1682
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
7509
range of 0.26e0.31 Wm1 K1, regardless of their degrees of crystallinity. Cured DGOBP showed hydrogen contents in the range of 403e1271 ppm, which is 25e75% smaller than that of conventional epoxy polymer. The hydrogen content and diffusion coefficient of DGOBP cured with DDM reduced with increasing crystallinity, indicating that both factors can be controlled by the degree of crystallinity. Because the ordered structure governs the diffusivity of hydrogen in liquid-crystal epoxy polymer, we conclude that the degree of crystallinity of the polymer strongly correlates with the hydrogen penetration. According to these results, DGOBP polymer with liquid-crystal structure can be a candidate polymer material for components in hydrogen systems, which require well-controlled hydrogen penetration properties. This suggests that we can design a suitable epoxy polymer with low hydrogen penetration properties by understanding and carefully controlling the development of the liquid-crystalline phase.
Acknowledgements This paper is based on results obtained from a project commissioned by the New Energy and Industrial Technology Development Organization (NEDO). The authors thank Hitoshi Koizumi and Mayuko Shido (NETZSCH Japan) for performing the thermal diffusivity measurements. Fig. 10 e Hydrogen content (a) and diffusion coefficient (b) of DGOBP cured with DDM at different curing temperatures. The broken line is the value of BisA cured with DDM. The error bars show the measurement error.
Fig. 11 e Relationship between hydrogen content, diffusion coefficient, and degree of crystallinity in DGOBP cured with DDM.
adjusting the curing temperature of the 4,40 -diaminodiphenyl methane (DDM) curing system. Crystal structures were observed in DGOBP cured with DDM below 130 C, and the degree of crystallinity decreased with increasing curing temperatures. We obtained a series of samples having 19%, 27%, 32%, and 36% crystallinity, corresponding to curing temperatures of 130, 120, 110, and 100, respectively. The thermal conductivity values of DGOBP cured with DDM were in the
references
[1] Strubel V. Final activity report. STORHY; 2008 Oct 15. Project No.502667. [cited 2015 Dec 18] Available form: http://www. storhy.net/pdf/StorHy_FinalPublActivityReport_FV.pdf/. [2] Monde M, Mitsutake Y, Woodfield PL, Maruyama S. Characteristics of heat transfer and temperature rise of hydrogen during rapid hydrogen filling at high pressure. Heat Transf Asian Res 2007;36(1):13e27. [3] Takeichi N, Senoh H, Yokota T, Tsuruta H, Hamada K, Takeshita HT, et al. “Hybrid hydrogen storage vessel”, a novel high-pressure hydrogen storage vessel combined with hydrogen storage material. Int J Hydrogen Energy 2003;28(10):1121e9. [4] Sakaguchi Y, Yamamoto Y, Nishiwaki H. Aluminum lined high pressure hydrogen cylinder. J Hydro Energy Syst Soc Jpn 2009;34(4):39e45. [5] Akatsuka M, Takezawa Y. Study of high thermal conductive epoxy resins containing controlled high-order structures. J Appl Polym Sci 2003;89(9):2464e7. € . Barrier properties of liquid crystalline polymers [6] Flodberg Go and their blends with PE and PETP. Sweden: KTH; 2002, ISBN 91-72-83-273-8. p. 26e34. [7] Weinkauf DH, Paul DR. Gas transport properties of thermotropic liquid-crystalline copolyesters. I. The effects of orientation and annealing. J Polym Sci Part B Polym Phys 1992;30(8):817e35. [8] Yamabe J, Nishimura S. Influence of fillers on hydrogen penetration properties and blister fracture of rubber composites for O-ring exposed to high-pressure hydrogen gas. Int J Hydrogen Energy 2009;34(4):1977e89. [9] Yamabe J, Matsumoto T, Nishimura S. Application of acoustic emission method to detection of internal fracture of
7510
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18] [19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 7 5 0 0 e7 5 1 0
sealing rubber material by high-pressure hydrogen decompression. Polym Test 2011;30(1):76e85. Yamabe J, Nishimura S. Nanoscale fracture analysis by atomic force microscopy of EPDM rubber due to high-pressure hydrogen decompression. J Mater Sci 2011;46(7):2300e7. Yamabe J, Nishimura S. Influence of carbon black on decompression failure and hydrogen permeation properties of filled ethylene-propyleneedieneemethylene rubbers exposed to high-pressure hydrogen gas. J Appl Polym Sci 2011;122(5):3172e87. Yamabe J, Fujiwara H, Nishimura S. Fracture analysis of rubber sealing material for high pressure hydrogen vessel. J Environ Eng 2011;6(1):53e68. Koga A, Uchida K, Yamabe J, Nishimura S. Evaluation on high-pressure hydrogen decompression failure of rubber Oring using design of experiments. Int J Automot Eng 2011;2(4):123e9. Fujiwara H, Ono H, Nishimura S. Degradation behavior of acrylonitrile butadiene rubber after cyclic high-pressure hydrogen exposure. Int J Hydrogen Energ 2015;40(4):2025e34. Mititelu A, Cascaval CN. Liquid crystalline epoxy thermoset obtained from biphenyl mesogen. Polym Plast Technol Eng 2005;44(1):151e62. Li Y, Badrinarayanan P, Kessler MR. Liquid crystalline epoxy resin based on biphenyl mesogen: thermal characterization. Polymer 2013;54(12):3017e25. Li Y, Kessler MR. Liquid crystalline epoxy resin based on biphenyl mesogen: effect of magnetic field orientation during cure. Polymer 2013;54(21):5741e6. Li Y, Kessler MR. Creep-resistant behavior of self-reinforcing liquid crystalline epoxy resins. Polymer 2014;55(8):2021e7. Li Y, Kessler M. Cure kinetics of liquid crystalline epoxy resins based on biphenyl mesogen. J Therm Anal Calorim 2014;117(1):481e8. Su W-FA, Chen KC, Tseng SY. Effects of chemical structure changes on thermal, mechanical, and crystalline properties of rigid rod epoxy resins. J Appl Polym Sci 2000;78(2):446e51. Su W-FA. Thermoplastic and thermoset main chain liquid crystal polymers prepared from biphenyl mesogen. J Polym Sci Part A Polym Chem 1993;31(13):3251e6. € cher M. “Liquid crystalline“ thermosets Mormann W, Bro from 4,4’-bis(2,3-epoxypropoxy)biphenyl and aromatic diamines. Macromol Chem Phys 1996;197(6):1841e51. Carfagna C, Amendola E, Giamberini M, Filippov AG, Bauer RS. Curing kinetics of liquid-crystalline epoxy resins. Liq Cryst 1993;13(4):571e84. Lee J, Song Y, Kim S, Lee H. Dielectric and molecular dynamic studies of sub-Tg by thermally stimulated current (TSC) analysis for liquid crystalline epoxy thermosets. Mater Chem Phys 2003;77(2):455e60. Lu MG, Shim MJ, Kim SW. Curing reaction and phase change in a liquid crystalline monomer. Macromol Chem Phys 2001;202(2):223e30. Lin Q, Yee AF, Earls JD, Hefner Jr RE, Sue H-J. Phase transformations of a liquid crystalline epoxy during curing. Polymer 1994;35(12):2679e82. Shiraishi T, Motobe H, Ochi M, Nakanishi Y, Konishi I. Effect of network structure on thermal and mechanical
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42] [43] [44]
properties of cured epoxide resins. Polymer 1992;33(14):2975e80. Ortiz C, Kim R, Rodighiero E, Ober CK, Kramer EJ. Deformation of a polydomain, liquid crystalline epoxybased thermoset. Macromolecules 1998;31(13):4074e88. Gavrin AJ, Douglas EP. Isothermal curing of acetylene functionalized liquid crystalline thermoset monomers. Macromolecules 2001;34(17):5876e84. Liu G, Gao J, Song L, Hou W, Zhang L. Synthesis and curing of liquid-crystalline epoxy resins containing a biphenyl mesogen. Macromol Chem Phys 2006;207(23):2222e31. Douglas EP. Structure-property relationships in liquid crystalline thermosets. J Macromol Sci Part C 2006;46(2):127e41. Harada M, Ando J, Yamaki M, Ochi M. Synthesis, characterization, and mechanical properties of a novel terphenyl liquid crystalline epoxy resin. J Appl Polym Sci 2015;132(1):41296. Cho S, Douglas EP, Lee JY. Transition diagrams for a liquid crystalline thermoset containing a rigid-rod epoxy. Polym Eng Sci 2006;46(5):623e9. € cher M. ‘Liquid crystalline’ thermosets Mormann W, Bro (LCT) from diaromatic diepoxides and aromatic diamines. Polymer 1998;39(20):4905e10. Barton J. The application of differential scanning calorimetry (DSC) to the study of epoxy resin curing reactions. In: Dusek K, editor. Epoxy resins and composites I. Berlin Heidelberg: Springer-Verlag; 1985. p. 111e54. Min B-G, Stachurski Z, Hodgkin J. Cure kinetics of elementary reactions of a DGEBA/DDS epoxy resin: 1. Glass transition temperature versus conversion. Polymer 1993;34(23):4908e12. Achilias DS, Karabela MM, Varkopoulou EA, Sideridou ID. Cure kinetics study of two epoxy systems with fourier transform infrared spectroscopy (FTIR) and differential scanning calorimetry (DSC). J Macromol Sci Part A 2012;49(8):630e8. Vonk CG. Computerization of Ruland's X-ray method for determination of the crystallinity in polymers. J Appl Crystallogr 1973;6(2):148e52. Neuser S, Michaud V. Effect of aging on the performance of solvent-based self-healing materials. Polym Chem 2013;4:4993e9. Bogaerts K, Lavrenova A, Spoelstra AB, Boyard N, Goderis B. Curing kinetics and morphology of a nanovesicular epoxy/ stearyl-block-poly(ethylene oxide) surfactant system. Soft Matter 2015;31(11):6212e22. Carbas RJC, Marques EAS, da Silva LFM, Lopes AM. Effect of cure temperature on the glass transition temperature and mechanical properties of epoxy adhesives. J Adhes 2014;90(1):104e19. Choy C. Thermal conductivity of polymers. Polymer 1977;18(10):984e1004. Michaels AS, Parker RB. Sorption and flow of gases in polyethylene. J Polym Sci 1959;41(138):53e71. Michaels AS, Bixler HJ. Flow of gases through polyethylene. J Polym Sci 1961;50(154):413e39.