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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/he

High-pressure gaseous hydrogen permeation test

method -property of polymeric materials for high-
pressure hydrogen devices (1)-

Hirotada Fujiwara a,*, Hiroaki Ono a, Kiyoaki Onoue a, Shin Nishimura a,b
a
Research Center for Hydrogen Industrial Use and Storage Kyushu University (HYDROGENIUS), HY-10, 744
Motooka, Nishi-ku, Fukuoka 819-0395, Japan
b
Department of Mechanical Engineering in Faculty of Engineering, Kyushu University, Japan

highlights graphical abstract

High pressure hydrogen gas

permeation test device was
developed.

Evaluation of polymer permeation
in equilibrium state up to 100 MPa
is possible.

Advantage of the method was
compared with the non-
equilibrium state method.

Hydrogen permeability de-
teriorates with the increase of gas
pressure.

Hydrostatic compression effect
can affect diffusion deterioration
under higher pressure.

article info abstract

Article history: Polymeric materials are widely used in hydrogen energy system such as FCEV and
Received 13 June 2020 hydrogen refueling stations under high-pressure condition. The permeation property (co-
Received in revised form efficients of permeation, diffusion and solubility) of polymers under high-pressure
19 July 2020 hydrogen condition should be discussed as parameters to develop those devices. Also
Accepted 23 July 2020 the property should be determined to understand influence of the compression by the
Available online 21 August 2020 pressure on polymer materials. A device which can measure gas permeation property of
polymer materials accurately in equilibrium state under high-pressure environment is
Keywords: developed, and the reliability of the measurements is ensured. High-pressure hydrogen gas
High-pressure gas seal permeability characteristics up to 100 MPa are measured for high-density polyethylene. An
Gas permeation advantage of the method is discussed comparing with the non-equilibrium state method,

* Corresponding author.
E-mail addresses: [email protected] (H. Fujiwara), [email protected] (H. Ono), onoue.kiyoaki.
[email protected] (K. Onoue), [email protected] (S. Nishimura).
https://doi.org/10.1016/j.ijhydene.2020.07.215
0360-3199/© 2020 The Author(s). Published by Elsevier Ltd on behalf of Hydrogen Energy Publications LLC. This is an open access article under the CC BY-NC-
ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4 29083

Gas diffusion focusing on the hydrostatic pressure effect. Deterioration of hydrogen permeability is
Gas solubility observed along with the decrease of diffusion coefficient, which is supposedly affected by
Storage tanks hydrostatic compression effect with the increase of environment pressure.
Dispensing hoses © 2020 The Author(s). Published by Elsevier Ltd on behalf of Hydrogen Energy Publications
LLC. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).

HRS is limited to the study using practical parts such as tanks

Introduction or hoses [8].
In this study, hydrogen permeability measurement system
Recently, hydrogen energy system has been paid much for high-pressure hydrogen up to 90 MPa has been developed.
attention as a solution to fossil fuel depletion and environ- Our research group are continuously working on the
mental issues [1], since hydrogen does not deteriorate analysis of the polymeric materials for the sealing or barrier
regardless of long term storage, and is able to generate a large materials for high-pressure hydrogen devices. In our previous
power, it has been put into practical use as an energy carrier study, it has been already reported that hydrogen absorbed in
medium [2]. Nevertheless, fossil fuel such as gasoline is still rubber by high-pressure hydrogen exposure initiates bubbles
dominant energy especially for transportation sector. In order during decompression, resulting in fracture such as blister,
to convert the fuel of the vehicles, electric vehicles, such as overflow and buckling due to the inflation of the rubber ma-
battery electric vehicles (BEV) and fuel cell electric vehicles terials [9e12]. Furthermore, the inflated polymeric materials
(FCEV) are much attracted for the decades. Since 2014, FCEVs by the high-pressure hydrogen exposure shrink after the
have been on the market and have been gradually dissemi- hydrogen elimination by decompression. The polymeric ma-
nated in US, EU countries, Korea and Japan. In these countries, terials are inflated and contracted repeatedly by the change of
hydrogen refueling stations (HRS) are developed. Because of the pressure. Since the elastic modulus at the volume inflation
low energy density of gaseous hydrogen, the hydrogen is state is reduced and the reduction ratio corresponds to the
compressed up to 70 MPa and stored in the high-pressure change in the cross-sectional area, it is concluded that the
hydrogen vessels on the vehicle. In order to promote the reduction in the number of molecular chains responsible for
dissemination of the FCEV, mass-production and cost- the elastic modulus per unit area due to volume expansion is
reduction of the high-pressure hydrogen devices and storage not the result of hydrogen affecting the rubber molecular
tanks are required [3]. In these devices and storage tanks, motion [13]. It is also reported that, despite the chemical
polymeric materials are used for the sealing devices such as structure is not changed by the repeated exposure and
O-rings or gaskets, inner liners for the type 4 tanks on the decompression, the physical deterioration such as the
vehicle and the flexible hydrogen dispensing hoses of HRS. decrease of filler network and the boundary rubber layer
These polymeric materials are directly exposed to the high- progresses, resulting the lowering of the elastic modulus [14].
pressure hydrogen and are should to seal or barrier the We also reported a quantitative evaluation method using an
high-pressure hydrogen. For example, international standard attenuation of the amount of light as an index of the state of
for the type 4 tank in GTR 13 requires the hydrogen perme- the blister fracture [15].
ation to be less than 46 Ncm3 h1 dm3 at 70 MPa, 55  C [4]. Polymeric materials such as crystalline polymers or rubber
ISO19880-5 requires that the permeation amount of hydrogen materials have free volume between molecular chains caused
form the flexible dispensing hose to be less than by the segmental motion of the molecules. Therefore, the
500 Ncm3 h1 m1 at 87.5 MPa. HySafe proposed a basic hydrogen gas easily penetrates into and permeates from the
method to estimate an allowable upper limit for hydrogen material. Numbers of studies have been conducted on the gas
permeation in automotive applications, and a maximum permeability of polymeric materials. Salame reported “Per-
hydrogen permeation ratio from road vehicles into enclosed machor parameter” calculated from empirically derived fac-
structures [5]. To design the type 4 tank or hydrogen hoses, tors for each chemical group in the polymer repeat unit to
suitable materials shall be selected, based on the hydrogen derive gas-permeability coefficients based on free volume
permeability of the inner liner polymeric materials simulating fractions and aggregation density energies [16]. A calculation
the condition under high-pressure environment of practical method considering the cohesive energy, packing, and rota-
use. Hydrogen permeability, diffusivity, and penetrability are tional degrees of freedom of the polymer is derived by Bicer-
important parameters for designing or selecting the materials ano et al. [17]. Besides them, Paul et al. calculated the free
suitable for those usage [6]. For the purpose, development of volume fraction based on the occupied volume and density
direct measurement method of hydrogen permeability is calculations and reported that the result can be used for pre-
necessary. Although there is a report about hydrogen dicting the gas permeability as the earned free volume frac-
permeation coefficients of epoxy resin with glass lacquer, tion showed good correlations with the gas permeability [18].
aiming to enable long-term storage of hydrogen in tanks, the As typical early studies on gas permeation in pressurized
pressure applied in the test was as low as 1.5 bar (0.15 MPa) [7]. environments there are reports by Henly et al., Stern et al. and
But the evaluation of permeability under the pressure in the Li et al. on PEs and PPs [19e25]. Henry evaluated the
type 4 on-board tank or for the practical operation pressure of
29084 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4

permeability of inorganic and organic gases at pressurization pressurized instantaneously to determine the diffusion coef-
conditions of up to about 200 psi (1.4 MPa), and Li up to 1500 psi ficient. Another method is measuring by using electrical
(10 MPa). They proved that the permeation coefficients in properties such as piezoelectric electro sorption in the limit of
logarithmic notation show linearity with pressurization 10 MPa [31]. The reasons why this method can not be applied
pressure and that the relation of pressure to permeability for high-pressure hydrogen gas is that sealing property can
differs depending on the gas species. Li evaluated the disso- not be secured where the electric wiring is installed and that
lution amount and reported that the linear approximation of the wiring can be deteriorated by the hydrogen gas. In short,
the low pressure was manifested and the Henry’s law was there are currently no measuring instruments that satisfy
applicable below the critical pressure, but the considerable those necessary requirements.
deviation occurred above the critical pressure. Stern have We have also proposed a thermal desorption analysis
analyzed the permeation behavior of inorganic gas e.g. hydro (TDA) method for measuring the amount of gas penetration
olefins and Fluor olefin-based organic gases up to 54.4 atm and the diffusion coefficient in a non-equilibrium state by
(5.5 MPa) under the conditions of-18  Ce70  C and discussed measuring hydrogen eliminated from a polymeric material
the relation to free volume. It is explained that the cause of after decompression following high-pressure hydrogen
change in permeation coefficient with pressurization is the exposure by gas chromatograph [11,32,33]. The TDA method is
effect of correlation between the increased free volume due to a convenient method for determination of amount of pene-
the absorption of penetrant and the decrease of free volume trated gas and the diffusion coefficient of the sample. How-
due to hydrostatic compression. Naito et al. proposed a ever, the method is controversial about the influence of the
method to divide the pressure dependency of permeability environment in which the test specimen is placed after
into two factors related to hydrostatic pressure and concen- decompression, and it has not been judged whether or not the
trations by considering the effects of gas molecular sizes and extrapolated value is a true value.
solubility on diffusions [26,27]. We have developed a system capable of permeation test in
When high-pressure gas is applied to one side of polymer equilibrium state for polymeric materials with high-pressure
materials, there is a possibility of the increases in permeation gases up to 100 MPa, especially hydrogen gas which is easy
driving force, reduction in free volume, and progress of plas- to leak. Advantage of this system is that sheet-shaped speci-
ticization. Those phenomena can affect permeation property mens can be used, that is, easy to prepare with various ma-
of the materials interacting each other in a complicated terials, and easy to determine thickness, i.e. permeation
manner. It will be easily understood that actual measurement distance, and permeation area. Determination of thickness
of the permeation behavior under practical hydrogen pressure and area is essential to earn coefficients such as permeability,
above 80 MPa is indispensable. diffusion, and solubility for relative assessment of the poly-
Currently, there are equipments to measure the gas mer materials. The structure of the equipment enables to
permeability characteristic in a high-pressure gas equilibrium contribute to get accurate and reliable data, preventing
state up to 130 atm (about 13 MPa), but there is no reports of deformation of the samples. In this paper, the gas permeation
actual evaluation for polymer materials in such a high- behavior at high-pressure in equilibrium state and determi-
pressure environment [26]. This is because there is no appa- nation of the permeability and amount of absorbed hydrogen
ratus which can analyze materials in a high-pressure equi- are described. As a model material, high-density polyethylene
librium state. There are several potentially applicable was selected, which is not affected by humidity and molding
measurement methods. For example, an equipment consists conditions, and has been reported a lot, not practical mate-
of high-pressure compartment, and the compartment in rials which is difficult to be supplied and disclosed. The results
which a sample is held, divided by a valve to evaluate gas are compared with those by TDA method. The evaluation of
absorption and a diffusion coefficient. In this method, internal the reliability of the equipment is also described.
pressure is measured which decreases according with the gas
amount absorbed into the specimen, comparing with the
space volume measured after the gas is released [28,29]. Experimental
However, this method needs a pressure gauge which can
measure very wide and dynamic range of pressure accurately Basic theory of permeation tests
for the evaluation at high-pressure condition. For example, to
detect some hPa conversion under high-pressure condition as To evaluate the gas permeation property, a differential pres-
high as 100 MPa. The absence of those accurate pressure sure method has been reported by Barrer, Kammermeyer,
gauge makes it difficult to apply the method for the evaluation Amerongen [34e36]. The method has been standardized in
at high-pressure condition. ASTM1434 and ISO7229, and many evaluations applied the
The volume of the sample changes under hydrostatic high- method have been reported [8]. In the method, permeation
pressure, and true value of diffusion coefficient of the sample coefficient, diffusion coefficient, and solubility coefficient
cannot be calculated. Absorption behavior measurement showing the amount of absorbed gas were analyzed based on
method using a magnetic suspension balance is also consid- the permeation behavior of the gas penetrates through the
ered as a candidate [30]. In this case, it is necessary to provide specimen set between a high-pressure cell and a low-pressure
a measuring instrument with such high precision that a cell. The gas permeation amount per unit time is measured
change of several tens of ppm in weight ratio can be observed. until the permeated gas reaches steady state, and value
It is necessary to measure only the true specimen weight normalized by permeation area, test pressure, and test tem-
change without drift even when a material is highly perature is expressed as a gas permeability GP [mol/
 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4 29085

(m2
s
Pa)]. By normalizing gas permeability with the perme- specimen when high-pressure gas is applied (permeation
ation distance, the value converted into the gas permeation plugs are inserted therein), an auto gas sampler that collects
coefficient, and is expressed by the following equation. the carrier gas with permeated gas at regular intervals and
introduces it into gas chromatography, and a gas chroma-
P ¼ GP  d (1)
tography system that can measure the amount of permeated
2
where P is the permeation coefficient [mol
m/(m
s
Pa)], and gas in the carrier gas. The chiller circulates the heat medium
d is the specimen thickness [m], to adjust the temperature of the gas to a test temperature in
advance, and the gas is maintained at the test temperature
273:15  V during the measurement. The high-pressure gas shut-off
GP ¼ (2)
A  Dp  T  0:0227 valves A and B which can be opened and closed instanta-
where V is the Amount (volume) of gas permeated to the low- neously are provided on the gas piping. A safety valve is
pressure side in the steady state [cm3/sec], A is the permeation equipped between the gas mixture and gas chromatography
area [m2], Dp is the pressure difference (Pa), and T is temper- to provide against unexpected excessive pressure over 1 MPa.
ature (K). Pressure gauge: KJ16-6FH [Nagano-keiki CO. LTD.] with NR-
In this method, the permeation coefficient can be obtained TH08 and KFD2-STC4-Ex1 (accuracy ± 0.5%). Thermometer:
by quantifying the permeated gas. Since the permeation co- EXS-T [OKAZAKI MANUFACTURING COMPANY] with NR-
efficient is derived from the actual amount of permeated gas, TH08 (accuracy ± 0.4%).
it is more reliable than with the volume method described
above. High-pressure permeation cell
When the time dependency of the permeated gas amount
is drawn as an integral curve (the horizontal axis is time, the The high-pressure permeation cell is equipped with the
vertical axis is total amount of the permeated gas), and the Permeation plug consists of Upper bolt, Lower bolt and Frame
intersection point when the straight line section indicating shown in Fig. 1(b), and is designed to withstand deformation
the steady state is extrapolated to the time axis is “time-lag” q, when high-pressure gases are applied from the top of the
the diffusion coefficient can be expressed by the following drawing with the test specimen interposed there between
equation. Sintered metal filters (porous stainless steel disc [SMC Cor-
poration] [38]). In order to suppress deformation at the time of
l2 gas application, ring shaped anchors 0.5 mm in height are
D¼ (3)
6q provided on the low-pressure side, and a gas-sealing O-Ring is
where D denotes the diffusion coefficient (cm2/s). provided on the high-pressure side. The gas permeated to the
Further, the solubility coefficient in which the amount of low-pressure side, shown in the lower part of the figure needs
gas absorbed per unit volume of the test specimen normalized to be mixed instantaneously with the gas chromatography
by pressure is expressed by the following equation. carrier gas with high efficiency. Correspondingly, the perme-
ated gas is introduced as a mixed gas into the auto sampler
P
S¼ (4) through the outer piping while being mixed by the carrier gas
D
(50 cc/min) flowing from below through the inside of the
where S is the solubility coefficient (mol/m3
Pa). double tube structure in the sintered metal filter.
Although there have been reports about high-pressure
permeation tests using practical tanks [8] or metal plates [37] Specifications of the components of the system
to evaluate gas permeation property, there have been no re-
ports for permeation tests using polymer sheet materials. The High-pressure gas shut-off valves: [NOVA] NV1-60-6H-ATO
thickness and permeation area of the specimens should be and [Fujikin Co.Ltd.] M3R4.
specified to calculate each coefficient. Gas Chromatography: [JP-SCIENCE LAB Co.,Ltd] GC-7000-T,
Important conditions for the test under a high-pressure Column: Molecular Sieves, Analytical Period: 0.5 vol-ppm
environment include: 1: stability of the permeation area of (flow-rate 50 cc/min, 4 kpa) 4 min, Detector: Thermal Con-
the test specimen in high-pressure environment; 2: accurate ductivity Detector (TCD) Measurements, Auto sampler: [JP-
determination of the thickness of the sample, which is the SCIENCE LAB Co.,Ltd] 5 min intervals into a gas chromato-
permeation distance l, 3: instantaneous pressurization of the graph with a GSL-700A 1 mL measuring tube.
specimen to the target pressure to determine the accurate
time-lag; 4: avoidance of gas leakage in the condition of Test procedure
instantaneous pressurization and the damage of test spec-
imen; and 5: the stability of the temperature in the whole The temperatures in the buffer tank and permeation cell
system. showed in Fig. 1(a) were maintained at 30  C ± 0.2  C. Test gas
from the gas supply is pressurized with a compressor to create
System overview of the high-pressure permeation test a test pressure gas in the buffer tank with the Valve-A open
equipment and-B closed, the gas for the test cylinder pressure gas. It is
necessary to pressurize with an additional pressure predicted
As shown in Fig. 1(a), the high-pressure permeation tester to loose, which can be calculated from the volume of the
consists of a 100 MPa-class compressor, buffer tank (500 cc), opposite side of Valve-B. When the temperature of the gas in
pressure gage, and a permeation cell that holds the test the buffer tank reaches the test temperature, the Valve-A is
29086 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4

Fig. 1 e (a) Schematic of Permeation system. (b) Schematic of Permeation plug.

closed and-B is opened, and the gas is instantaneously applied initial hydrogen amount and diffusion coefficient of zero time
to the specimen. Conversion of pressure and temperature as unknowns from the diffusion equation shown in the for-
when Valve B opened is shown in appendix Figure A1. mula Eq. (5).
The applied gas permeates through the specimen, and the "   #
permeated gas is introduced into the gas chromatograph 1 at 32 X∞
exp ð2n þ 1Þ2 p2 Dt [2
CH ðtÞ ¼  CHð0Þ 
intervals of 5 min to be quantified. The gas permeation in the p2 n¼0 ð2n þ 1Þ2
"    # (5)
transverse direction is likewise introduced into the gas chro- X∞
exp Db2n t r2
matograph 2 and quantified. All seal materials (O-Rings and b2n
n¼1
buck-up rings) are replaced for each experiment for avoiding
influence of cyclical use. where CH(t) is the remaining hydrogen content (wt･ppm), D is
diffusion coefficient (mm2/s), bn is the root of the zero-order
Sample Bessel function, [ is the thickness of sample and r is the
radius of disk specimen.
The employed material was high density polyethylene (HDPE)
Novatec HB111R [Japan Polychem Corporation], The plate
specimen with the dimension of 150 mm  150 mm  4 mm Results
was prepared by the hot press method according to the JIS
K6922-2, Molding Temperature: 180  C, Average Cooling Rate: Reliability evaluation of test equipment
15  C/min, Molding Pressure: 5 MPa, Pressurization Time:
5min. After preparation of the specimen, the skin layer of the Examination of seal structure
plate specimen were removed by MC-working to form 1.1 mm, In order to prove the sealing performance of the high-pressure
1.5 mm, 2 mm, and 3.1 mm plates with surface roughness, permeation plug shown in Fig. 1(b), the effect of the presence
F27 mm disks were punched from the plate for the or absence of anchors was examined using a 1.5 mm thick
experiments. sheet specimens. In the absence of the anchor, excessive
The properties of the test specimens were shown Table 1. amount of lateral gas leakage was detected with test gas
pressurization. On the other hand, in the seal mechanism
Thermal Desorption Analyzer (TDA) using the anchors in combination, even when the high-
pressure gas is instantaneously applied, gas leakage in the
The measurement method previously reported is described
here again to confirm. A disk-shaped test piece of
f13 mm  2 mm thickness exposed to high-pressure
hydrogen is held in an electric oven held at 30  C by a Ther- Table 1 e The properties of the test specimens.
mal Desorption Analyzer JSH-201 [JP-SCIENCE LAB Co.,Ltd] Density 0.942 g/cm3
Argon gas of 50 cc per min 4 kPa as a GC carrier gas is circu- Molecular weight Mn:1.3  104 g/mol,
lated in the electric furnace, and hydrogen gas eliminated Mw:2.7  105 g/mol
Degree of crystallinitya 77e78%,
from the specimen is sampled at intervals of 5 min and
Glass transition temperatureb 115  C
introduced into the GC to measure the concentration of the
Tensile yield stresses 24.1 MPa
hydrogen gas. The absorbed gas in the specimen is completely Elongation at break 818%
eliminated by heating to 175  C. during the elapse of 24 h. The Modulus 645 MPa.
time conversion of the remaining hydrogen amount in the Surface roughness, Ra: 1.45e2.86 mm,
specimen with the time passage is obtained from the integral Rz: 10.45e17.50 mm
curve of the elimination amount. The penetrated hydrogen a
Determined by wide-angle X-ray measurement.
b
amount and diffusion coefficient are obtained by fitting the Determined by differential scanning calorimetry.
 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4 29087

Fig. 2 e Permeation curve on 90 MPa at 30  C Transmission side area 201 mm2. (a) Differential Permeation curve (b) Integral
Permeation curve.

lateral direction, which is caused by the breakage of sealing with the frame, to suppress the deformation of the specimen.
mechanism was not observed, and even when it is held for a As a consequence, the sealing mechanism is secured, and it
long time, gas leakage was not detected. Transmission curve becomes possible to seal the hydrogen gas efficiently.
was successfully obtained as shown in Fig. 2. After about 5 h,
permeation reached the steady state, after which no variation Fineness(Pore size) of the sintered metal filter
in the permeation amount is observed. 3D image of a spec- Stern et al. compared the permeability determined by using a
imen after high-pressure permeation test is shown in Fig. 3. 14  14 mesh steel screen and that earned by using porous
The deformation of the specimen of polymeric material was stainless steel disc as a sample-holding mechanism [39].
not observed in the sealing system using the anchors, but in There was a difference between the two results, and it was
the mechanism using only the O-Ring, the specimen concluded that the permeability is greater when steel screen
deformed. The photograph (b) depicts the deformation in the was used, and that the difference was greater when pressures
lower right part of the specimen of polymeric material along were higher. The reasons include (1) a decrease in perme-
with the trace of the outward movement of the O-Ring. In the ability of the porous metal support due to a decrease in
planar self-sealing system with the O-Ring, the O-Ring is effective membrane area, or (2) an increase in permeability of
pressed against the contacting surface of the specimen by a the screen support due to deformation of the membrane. Steel
force higher than the pressure of the applied gases. As a result, screen was not used in this test because of the large defor-
the specimen of polymeric material is deformed in the outer mation of polymeric materials and the possibility of breakage,
circumferential direction and becomes thin to make a gap, and the relation between the gas-permeability coefficient and
then the seal mechanism cannot be secured because of the the gas-diffusion coefficient values for the roughness of the
gap. Therefore, the mechanism in which the specimens de- mesh of sintered metal filters was evaluated. The results are
forms as in the case of the structures represented by ISO JIS- shown in Fig. 4. Although there is a difference between the
K7126it is not appropriate for the high-pressure permeation coefficient values depending on the nominal filtration accu-
test of samples such as polymeric material or rubber. In racy, which represents the fineness of the filter, the difference
contrast, our new system provides anchor and system to is small, and a specific tendency is not found. Stern assumes
restrain the movement of the specimen and reduce the gap that the true value is intermediate between the two values

Fig. 3 e 3D image of a specimen after high-pressure permeation test. (a) Permeation plug with anchors and an O-Ring. (b)
Permeation plug with only an O-Ring. (c) Schematic of O-Ring Self sealing mechanism. The O-Ring is oppressed by applied
hydrogen pressure, and compresses the specimen with the pressure exceeding the applied pressure. The resin is deformed
with the outward pressure by the O eRing.
29088 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4

earned by the two structures, but is not necessarily true. As area, and the diffusion coefficient calculated from the time-
the influence of the use of the sintered metal filter, there were lag. When sintered metal filters having the same permeation
concerns about reduction of the permeation amount due to area are used for both the low-pressure side and the high-
reduction of the effective area and suppression of the pressure side (black circles in the figure), the gas permeation
permeated gas diffusion, but both the permeation coefficient amount shows a good proportional relationship, and the
and the diffusion coefficient were almost unchanged in our extrapolated value shows zero. When a smaller filter of
method. It is considered that one possible cause of the sta- F5 mm (permeation area 19.6 mm2) is used, same value
bility is that even when the finest 1 mm filter is used, the should be shown as the same kind of test specimen is used,
permeation coefficient of the filter is several hundred times but different values are shown. When a O-Ring larger than the
larger than the permeation coefficient of the test specimen. filter in the low-pressure side was used (shown by marks in
Moreover, the permeated hydrogen is mixed and conveyed orange, green, blue), the permeation amount and permeation
with high efficiency by the carrier gas, and the influence of the coefficient was as theoretical, but the diffusion coefficient
transmission distance is negligibly small. The particle diam- showed different from the theoretical value. When the
eter used for filters with a Nominal filtration accuracy of 1 mm permeation area on the high-pressure side was smaller than
is 63e75 mm, the surface roughness is Ra: 7.5 mm, and the Rz: that on the low-pressure side, the difference in the perme-
53.7 mm is extremely larger than Rmax6.3, which ensures good ation amount was different, and the values of the permeation
sealing performance in the metallic gaskets system [40]. Since coefficient and the diffusion coefficient were also different
the size of the hydrogen molecules is overwhelmingly small (shown by a mark in red). This is considered to be because gas
with respect to the size of the gap formed between the parti- diffusion in the thickness direction is not isotropic. The above
cles of the filter and the specimen, it is considered that the results demonstrate that gas-sealing is performed by O-Ring,
reduction of the hydrogen permeation amount due to the and a mechanism in which the area on the low-pressure side
presence of the filter is negligible. Even if there is an influence, is smaller than that on the high-pressure side must not be
the permeation value extrapolated to zero is almost equiva- used. In addition, it was ensured F5 mm small plug having a
lent to the value when 1 mm of nominal filtration accuracy was small permeation area (19.6 mm2) can not be used. It was also
employed. Since the objective of our study is to evaluate the found that the structure having same area size in the low
relatively on the basis of this method, it is judged that there is pressure side and the high-pressure side is preferable.
no problem to use the filter of 1 mm to perform the experiment. Furthermore, it was found that the influence of the entrap-
ment of the gas from the outer periphery in the high-pressure
Examination of permeation area environment was negligible, since the extrapolated value
The effect of the sealing system with sintered metal filters and shows zero, and it was possible to accurately determine the
the combination of anchors and O-Ring was evaluated in gas permeation area from the inner diameter of the anchor.
terms of the permeation area. Various combination of the Therefore, the system with O-Ring F16 mm (permeation area
structure of the high-pressure side and the low-pressure side 201.1 mm2), which has the largest area of this equipment, and
was examined under a pressure condition of 90 MPa using a plugs P21 was selected for further experiments.
1.5 mm thick material as shown in Table 2. It is theoretically
correct that the permeation amount increases proportionally Thickness of the specimen
as the permeation area increases. It suggests that no Since the diffusion is proportional to the square of the thick-
destruction or deformation of the specimen occurs in the ness of the specimen, the relationship between the delay time
measurement area. Fig. 5 shows the amount of permeate gas (time-lag), which is a value for determining the diffusion co-
relative to the area of the permeate side (Low-pressure side), efficient, and the thickness of the specimen was evaluated.
the permeation coefficient calculated based on the permeate The results are shown in Fig. 6. The correlation between the

Fig. 4 e Influence of the fineness of the sintered filter. Transmission side area of high-low is 211 mm2, Applied pressure
90 MPa at 30  C
 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4 29089

Table 2 e Combination of plugs with different transmission areas on the low-pressure side and high-pressure side.
Test conditions High-Low pressure side Same transmission area High-Low pressure side Diffrent
transmission area
Marker

schematic of
permeation
plug

High O-Ring Inner 21 18 12 9 21 21 21 12

pressure side diameter (mm)

Sintered filter 201.1 132.7 50.3 19.6 50.3 50.3 201.1 50.3
area (mm2)

Low Sintered filter 201.1 132.7 50.3 19.6 50.3 201.1 50.3 201.1
pressure side area (mm2)

square of the thickness of the specimen and the time-lag was and under the condition of 90 MPa, the permeation coefficient
confirmed, and it was judged that the specimen up to about was approximately half of the value obtained at the time of
3 mm in thickness could be used in the present apparatus. The atmospheric pressure as an extrapolated value. As reported by
extrapolated value passes through zero, which is a theoretical Li, Stern and Naito, the logarithm of permeation coefficients
value. This also demonstrates that the effect of the fineness of vs. pressures was linear, with positive slopes for highly solu-
the filter does not affect the permeation mechanism. The ble gases such as organic vapors and CO2, and negative slopes
above results indicate that the existence of the anchor does for less soluble gases such as N2 and He [41]. The pressure
not affect the permeation distance, and it was judged our dependence of the permeability consists of a component
high-pressure permeation test apparatus satisfies the speci- related to the hydrostatic pressure and a component related to
fications sufficient to evaluate the gas permeability the concentration of the permeate gas. Li and Henley pro-
performance. posed plasticization of the polymer network caused by pene-
trant, which is the events gas penetrates into the polymers
Results of the high-pressure permeation test and expands the matrices, resulting in accelerated segmental
motion and thereby gas diffusions [19]. The results of our
Fig. 7 shows the results of the high-pressure permeation test approach indicate that the plasticization effects caused by
with respect to pressure (each value of the measurement is hydrogen gas does not occur even at high-pressure condi-
shown in appendix). The permeated hydrogen gas amount tions. In other words, the hydrostatic effect due to pressuri-
increased as the pressure increased, but the increase ratio zation is expected to be the dominant factor of this evaluation.
slowed as the pressure increased. The concerned phenomena The diffusion coefficient and solubility on coefficient also
of the increase of permeation driving force was not observe. show decreasing tendency as the pressure increases, and the
On the supposition that HDPE was used for the tank liner ratio is slightly larger in the diffusion coefficient. It is sup-
material, the increase ratio of the gas permeation is not as posed that the suppression of diffusion is a major factor of the
large as the increase of the pressure in the tank. The gas decrease of the permeation coefficient. Penetrant is thought to
permeation measured with this method exhibited safer value move around free volume, which is a void caused by thermal
than the one extrapolated from lower pressure value. The vibration of the polymer chain, and to diffuse. The shrinkage
permeation coefficient decreased with increasing pressure, of free volume under high-pressure environment is thought to

Fig. 5 e Effect of permeation area on gas permeation mechanism in 90 MPa hydrogen gas pressurization at 30  C (a)
Permeated hydrogen gas amount on stationary state. (b) Permeability coefficient (c) Diffusion coefficient.
29090 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4

quickly after decompression. It is necessary to perform mea-

surement in an equilibrium state as it shows plasticization.
We are currently conducting high-pressure hydrogen in
situ analysis of high-order structure using a high-pressure cell
with a diamond window and an optical technique. Free vol-
ume fraction is evaluated based on the volume change. The
details of the experiment and the results will be reported in
the near future.

Comparison with TDA method

TDA method measures the desorption behavior by keeping

the test specimen after the hydrogen exposure in an argon gas
flow environment. Fig. 8 shows the profile of -(a) hydrogen
elimination behavior, -(b)remaining hydrogen amount and
-(c)remaining hydrogen amount under Argon gas environ-
ment fitted with diffusion Eq. (5), all in the HDPE specimen
Fig. 6 e Effect of specimen thickness derived from the
exposed to hydrogen at 90 MPa. The amount of hydrogen
Relationship between specimen thickness and time-lag on
indicated by “Argon” in the figure represents the result the
gas permeation test. Transmission side area of high-low is
specimen was kept in the argon gas circulation 50 cm3 per min
211 mm2 and nominal filtration accuracy is 1.0 mm. Applied
4 kPa 30  C after 10 min elapse, which is the time required for
pressure 90 MPa at 30  C.
taking the specimen out of the high-pressure vessel to install
in TDA. “Helium” represents the value measured by TDA after
being kept under helium-flow condition (4 kPa, 50 cm3 per
be a cause of the suppression of diffusion. However, the pos- min.) at 30  C for 1 h “Vacuumed” value was obtained by
sibility of an increase of the crystal phase, which inhibits the measuring with TDA after being kept in vacuumed condition
diffusion of the gas should be considered. The change in the of about 10Pa (G) with a vacuum. The amount of remaining
fraction of free volume as an influence of the high-pressure hydrogen does not differ after 1 h regardless of the difference
hydrogen gas pressurization should also be considered. The of the environments, and it was proven that the elimination
change in crystallinity has been analyzed in detail by Ito et al. until then was considered to be approximately equal as in
under hydrostatic pressure conditions [41]. Polyethylene is vacuum.” This fact shows the hydrogen elimination behavior
compressed and densified by hydrostatic pressure while of HDPE is not conformity with elimination-controlling but
maintaining the orthorhombic structure, but it is restored with diffusion-controlling, and the high-pressure hydrogen

Fig. 7 e Results of high pressure permeation test at 30  C: influence of applied pressure on each value (a) Permeated
hydrogen gas amount, (b) Permeability, (c) Diffusion, (d) Solubility. Measurement condition: transmission side area of high-
low is 211 mm2, nominal filtration accuracy is 1.0 mm.
 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4 29091

Fig. 8 e Influence of environments condition on remaining hydrogen content in specimens measured by argon carrier TDA
after decompression following the hydrogen exposure to 90 MPa at 30  C for 24 h “Argon” specimen is kept under argon
environment 15 min after decompression, “Helium” and “Vacuumed” are kept under helium environments condition or
vacuum condition 15 min after decompression and under argon environment 60 min after decompression. (a) Hydrogen
elimination from the specimen (b) Profile of remaining hydrogen amount (c) Profile of remaining hydrogen amount under
argon gas, and fitting with diffusion equation.

elimination can be discussed with diffusion-controlling. The

remaining hydrogen content extrapolated to zero hour is
equivalent for the amount of penetrated gas under high-
pressure equilibrium condition.
Fig. 9 shows the amount of penetrated gas in HDPE, which
is the weight of the penetrated hydrogen per unit of HDPE
weight, calculated based on the solubility coefficient obtained
by the high-pressure permeation test method. As the test
pressure increased, the absorption amount increased, but the
ratio of increase slowed slightly as the pressure increased, and
it was found that the pressure deviated from Henry’s law.
Figs. 9(b) and 10(a) show correlations between amount of
penetrated hydrogen gas and diffusion coefficients obtained by
the TDA method. The amount of penetrated hydrogen gas
showed the same tendency as those obtained with TDA
measured at near non-equilibrium, but the diffusivity as a
whole showed larger values with TDA. The amount of pene-
trated hydrogen measured by TDA and high-pressure
Fig. 9 e Pressure dependency of amount of penetrated
hydrogen permeation test is considered to be the amount of
hydrogen gas in HDPE determined by high-pressure
hydrogen penetrates into a sample with shrinkage of free vol-
permeation test.
ume under pressure. As the specimens for both measurements
are in the same hydrogen saturated state, the amount of

Fig. 10 e High-pressure hydrogen gas properties in HDPE using different measurements of equilibrium and no equilibrium
States (a) Amount of penetrated hydrogen gas (b) Diffusion coefficient.
29092 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4

penetrated hydrogen shows the same value. In the high- behavior can not be analyzed, as it can be affected by the
pressure permeation test, the diffusion coefficient is compression of the free volume due to hydrostatic pres-
measured in the steady state with shrunk free volume as it is, sure. The high-pressure permeation testing method we
then the result reflects the behavior of the state with the report in this paper enables to provide very accurate and
decrease of the free volume. In contrast, the TDA method is a precise data of permeation and diffusion coefficient and
measurement in which the free volume recovers the state amount of penetrated hydrogen gas in high-pressure envi-
under atmospheric pressure, with the removal of external ronment, which have been difficult to obtain. Those data
pressure. Because the diffusion coefficient is small when the are useful not only for selecting materials in the practical
free volume is small and large when the free volume is large, use under high-pressure hydrogen environment, but also
the TDA method shows a larger value than the high-pressure for general theoretical discussions about penetration and
permeation test method. The higher the test condition pres- diffusion of penetrant in polymer materials under a high-
sure is, the greater the difference between the two results be- pressure environment. We will measure the size of
comes. Therefore, it is inferred that the difference between the shrunk free volume by utilizing the molecular size of
TDA method and the high-pressure permeation test method penetrant, or the interaction of polymer materials with
became smaller under the low-pressure condition, that is, the penetrant to discuss the permeation mechanism by using
condition in which the diffusion coefficient is smaller. hydrogen, nitrogen, helium etc in high-pressure conditions.
This method is expected to play an important role both in
industrially and scientifically.
Conclusion

An equipment capable of evaluating hydrogen permeation Declaration of competing interest

behavior of polymeric materials in an equilibrium state
under high-pressure hydrogen gas conditions has been The authors declare that they have no known competing
developed. The reliability of the equipment was ensured in financial interests or personal relationships that could have
terms of the accuracy in the measurement of the perme- appeared to influence the work reported in this paper.
ation area, permeation distance, avoiding the effect of fil-
ters, and keeping stability of pressure without gas leakage.
The gas permeation amount in the environment up to
Acknowledgment
90 MPa was successfully measured, and the permeation
coefficient was obtained. The diffusion coefficient, solubil-
A part of this research is based on the results obtained from a
ity coefficient, and gas penetration amount were calculated
project commissioned by the New Energy and Industrial
from the delay time. Precise and reliable values of those
Technology Development Organization (NEDO), Japan.
parameters are necessary to evaluate reliability of poly-
meric materials such as packing for hydrogen shut-off Appendix
valves, O-Rings for joint sealing, backup rings, lining for
hoses and tanks, multi layered materials, and the materials Influence of reading 7error in pressure gauge and ther-
in which blister destruction progresses to conduct mometer on the results is shown in Table A3. Discrepancy of
comparative analysis. The comparative analysis of the data results are all negligibly small, supposing in 90 MPa case. The
earned by this method with the previously reported TDA influence of ± 0.4% error in applied pressure on permeation
data clarified that the diffusion coefficient was smaller than coefficient is 4.91e4.96  1016 [mol･m/(m2 ・ s ・ Pa)]. The
the TDA data, though the amount of penetrated hydrogen influence of ± 0.4% error in thermometer on permeation co-
was almost same as in TDA data. Although TDA method efficient is 4.94e4.96  1016 [mol･m/(m2 ・ s ・ Pa)] and
has an advantage of simplicity and easiness for the mea- amount of penetrated hydrogen is 1118e1121 wt ・ ppm.
surement of amount of penetrated hydrogen gas, diffusion

Table A1 e High-pressure permeation test value.

Target pressure Permeability coefficient Diffusion coefficient Solubility coefficient Amount of penetrated
[MPa] [mol･m/(m2 ・ s ・ Pa)] [cm2/s] (x106) [mol/m3・Pa] (x106) [wt・ppm]
(x1016)
10 9.25 1.35 6.87 143
20 8.78 1.32 6.65 289
30 8.05 1.28 6.28 409
40 7.40 1.20 6.17 526
50 6.76 1.08 6.25 671
60 6.32 1.01 6.25 799
70 5.88 0.99 5.96 881
80 5.40 0.90 6.00 1016
90 5.08 0.91 5.57 1073
90 4.95 0.84 5.86 1120
 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4 29093

Table A2 e DATA measured by TDA mesod after high-pressure gas exposure

Target Diffusion coefficient Amount of
pressure [MPa] [cm2/s] (x106) penetrated [wt・ppm]
10 2.09 143
30 2.11 401
50 1.98 646
70 2.03 831
90 1.82 1050

Table A3 e Influence of reading error in pressure gauge and thermometer

Factor of error Pressure Temperature Permeability Diffusion Solubility Amount of
[MPa] [ C] coefficient coefficient coefficient penetrated
[molm/(m2 ・ s ・ Pa)] [cm2/s] (x106) [mol/m3・Pa] [wt・ppm]
(x1016) (x106)
90 MPa/30  C reference measurement 90.00 30.00 4.93 0.84 5.86 1120
results (actual measurement)
When the pressure is different 90.45 30.00 4.91 0.84 5.83 1120
by þ 0.45 MPa (assumption)
When the pressure is different 89.45 30.00 4.96 0.84 5.90 1120
by 0.45 MPa (assumption)
When the temperature is different 90.00 30.50 4.93 0.84 5.85 1118
by þ0.5  C (assumption)
When the temperature is different 90.00 29.50 4.94 0.84 5.87 1121
by 0.5  C (assumption)

Figure A1 e Conversion of pressure and temperature when right after measurement starts. Time zero corresponds to the
time when Valve B opened.

references [3] Mausa Steffen, Hapkeb Jobst, Ranongb Chakkrit Na,

chnera Erwin Wü, Friedlmeiera Gerardo, Wenger David.
Filling procedure for vehicles with compressed hydrogen
tanks. Int J Hydrogen Energy 2008;33(17):4612e21.
[1] Kiyoaki Onoue, Murakami Yukitaka, Petros Sofronis. Japan’s
[4] Global technical regulation on hydrogen and fuel cell
energy supply: mid-to-long-term scenario -A proposal for a
vehicles Contents. Global technical regulation No. 13. https://
new energy supply system in the aftermath of the March 11
www.unece.org/fileadmin/DAM/trans/main/wp29/wp29wgs/
earthquake. Int J Hydrogen Energy 2012;37(10):8123e32.
wp29gen/wp29registry/ECE-TRANS-180a13e.pdf.
[2] IEA Technology roadmap - hydrogen and fuel cells. http://
[5] Adams P, Bengaouer A, Cariteau B, Molkov V, Venetsanos
ieahydrogen.org/pdfs/
Allowable AG. Hydrogen permeation rate from road vehicles.
TechnologyRoadmapHydrogenandFuelCells-(1).aspx; 2015.
Int J Hydrogen Energy 2011;36(3):2742e9.
29094 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 5 ( 2 0 2 0 ) 2 9 0 8 2 e2 9 0 9 4

[6] Castagnet Sylvie, Grandidier Jean-Claude, Comyn Mathieu. membranes. II. Pure Ar, SF6, CF4, and C2H2F2 in polyethylene.
Guillaume Benoıt Hydrogen influence on the tensile J. Poly. Sci PART B Polym. Phys 1986;24(10):2149e66.
properties of mono and multi-layer polymers for gas [26] Naito Yasutoshi, Mizoguchi Keishin, Terada Katsuhiko,
distribution. Int J Hydrogen Energy 2010;35(14):7633e40. Kamiya Yoshinori. The effect of pressure on gas permeation
[7] Prewitz Marc, Gaber Martin, Müller Ralf, Marotztke Christian, through semicrystalline polymers above the glass transition
Holtappels Kai. Polymer coated glass capillaries and temperature. J Polym Sci B Polym Phys 1991;29(4):457e62.
structures for high-pressure hydrogen storage: permeability [27] Kamiya Yoshinori, Naito Yasutoshi, Bourbon Dominique.
and hydrogen tightness. Int J Hydrogen Energy Sorption and partial molar volumes of gases in
2018;43(11):5637e44. poly(ethylene-co-vinyl acetate). J Polym Sci Polym Phys Ed
[8] B. Acosta, P. Moretto, N. Frischauf, F. Harskamp. Gastef: the 1994;32(2):281e6.
JRC-IE compressed hydrogen gas tanks testing facility. [28] Sato Yoshiyuki, Yurugi Masashi, Fujiwara Koji,
https://core.ac.uk/download/pdf/34994655.pdf. Takishima Shigeki, Masuoka Hirokatsu. Solubilities of
[9] Koga Atsushi, Uchida Kenichi, Junichiro Yamabe, carbon dioxide and nitrogen in polystyrene under high
Shin Nishimura. Evaluation on high-pressure hydrogen temperature and pressure. Fluid Phase Equil
decompression failure of rubber O-ring using design of 1996;125(1e2):129e38.
experiments. Int J Automot Eng 2011;2(4):123e9. [29] Sato Yoshiyuki, Iketani Toru, Takishima Shigeki,
[10] Junichiro Yamabe, Koga Atsushi, Shin Nishimura. Failure Masuoka Hirokatsu. Solubility of hydrofluorocarbon (HFC-
behavior of rubber O-ring under cyclic exposure to high- 134a, HFC-152a) and hydrochlorofluorocarbon (HCFC-142b)
pressure hydrogen gas. Eng Fail Anal 2013;35:193e205. blowing agents in polystyrene. Fluid Phase Equil
[11] Junichiro Yamabe, Shin Nishimura. Influence of fillers on 1996;40(6):1369e75.
hydrogen penetration properties and blister fracture of [30] Sato Yoshiyuki, Takikawa Tadao, Sorakubo Atsushi,
rubber composites for O-ring exposed to high-pressure Takishima Shigeki, Masuoka Hirokatsu, Imaizumi Mitsuhiro.
hydrogen gas. Int J Hydrogen Energy 2009;34(4):1977e89. Solubility and diffusion coefficient of carbon dioxide in
[12] Koga Atsushi, Tadahisa Yamabe, Sato Hiroyuki, biodegradable polymers. Ind Eng Chem Res
Uchida Kenichi, Nakayama Junichi, Junichiro Yamabe, 2000;39(12):4813e9.
Shin Nishimura. A visualizing study of blister initiation [31] David C. Bonner Yang-leh Cheng A new method for
behavior by gas decompression. Tribol Online 2013;8(1):68e75. determination of equilibrium sorption of gases by polymers
[13] Fujiwara Hirotada. Analysis of acrylonitrile butadiene rubber at elevated temperatures and pressures. J Polym Sci Polym
(NBR) expanded with penetrated hydrogen due to high Lett Ed 1975;13(5):259e64.
pressure hydrogen exposure. Int Polym Sci Technol [32] Junichiro Yamabe, Nakao Masatoshi, Fujiwara Hirotada,
2017;44(3):T41e8. Shin Nishimura. Influence of fillers on hydrogen
[14] Fujiwara Hirotada, Ono Hiroaki, Shin Nishimura. penetration properties and blister fracture of EPDM
Degradation behavior of acrylonitrile butadiene rubber after comosites exposed to 10 MPa hydrogen gas. Nihon Kikai
cyclic high-pressure hydrogen exposure. Int J Hydrogen Gakkai Ronbunshu A Hen Trans Jpn Soc Mech Eng Part A
Energy 2015;40(4):2025e34. 2008;74(7):971e81.
[15] Ono Hiroaki, Fujiwara Hirotada, Kiyoaki Onoue, Ono Hiroaki, [33] Ono Hiroaki, Fujiwara Hirotada, Shin Nishimura. Penetrated
Shin Nishimura. Influence of repetitions of the high-pressure hydrogen content and volume inflation in unfilled NBR
hydrogen gas exposure on the internal damage quantity of exposed to high-pressure hydrogeneWhat are the
high-density polyethylene evaluated by transmitted light characteristics of unfilled-NBR dominating them? Int J
digital image. Int J Hydrogen Energy 2019;44(41):23303e19. Hydrogen Energy 2019;44(41):23303e19.
[16] Salame Morris. Prediction of gas barrier properties of high [34] Barrer RM. Permeation, diffusion and solution of gases in
polymers. Polym Eng Sci 1986;26(22):1543e6. organic polymers. Trans Faraday Soc 1939;35:628e43.
[17] Bicerano Jozef. Prediction of polymer properties. New York: [35] Brubaker David William, Kammermeyer Karl. Flow of Gases
Marcel Dekker; 1993. ISBN 0-203-91011-7. through Plastic Membranes. Indus Eng Chem
[18] Park JY, Paulr DR. Correlation and prediction of gas 1953;45(5):1148e52.
permeability in glassy polymer membrane materials via a [36] Van Amerongen GJ. The permeability of different rubbers to
modified free volume based group contribution method. J gases and its relation to diffusivity and solubility. J Appl Phys
Membr Sci 1997;125:23e39. 1946;17(11):972e85.
[19] Li NN, Henley EJ. Permeation of gases through polyethylene [37] Nagao Akihide, Takagi Shusaku. Gas-phase hydrogen
films at elevated pressures. AIChE J 1964;10(5):666e70. permeation test device and method of protecting gas-phase
[20] Reilly JW, Henley EJ, Staffin HK. Separation of gaseous hydrogen permeation test device United States patent
mixtures by permeation through polyethylene film. AIChE J patent. 2019; No.: US 10,295,453 B2.
1964;16(3):353e5. [38] SMC sintered metal element catalog. http://ca01.smcworld.
[21] Li NN, Long RB. Permeation through plastic films. AIChE J com/catalog/en/ifilter_metal/EB-ES-E/6-11-p0103-0118-
1969;15(1):73e80. shouket_en/data/6-11-p0103-0118-shouket_en.pdf.
[22] Stern SA, Mullhaupt JT, Gareis PJ. The effect: of pressure on [39] Stern SA, Fang S-M, Jobbins RM. Permeation of gases at high
the permeation of gases and vapors through polyethylene. pressures. J. Macromol. Sci Phys. B5 1971;1:41e69.
Usefulness of the corresponding states principle. AIChE J [40] Matsuzaki Yoshio, Kazamakic Tsuneji. Effect of
1969;15(1):64e73. surface roughness on compressive stress of static
[23] Stern SA, Fang S-M, Frisch HL. Effect of pressure on gas seales nippon kikai gakkai ronbunshu. Chimia
permeability coefficients. A new application of "free volume" 1987;53(489):1091e5.
Theory. J Poly Sci Part A-2 1972;(10):201e19. [41] Naito Yasutoshi, Bourbon Dominique, Terada Katsuhiko,
[24] CO2, CH4, C2H4, and C3H8 Stern A, Kulkarni SS, Frisch HL. Kamiya Yoshinori. Permeation of high-pressure gases in
Tests of a “free-volume” model of gas permeation through poly(ethylene-co-vinyl acetate). J. Poly. Sci PART B Polym.
polymer membranes. I. Pure CO2, CH4, C2H4, and C3H8 in Phys. 1993;31:693e7.
polyethylene. J Polym Sci Polym Phys Ed 1983;21(3):467e81. [41] Ito Taisuke, Marur Harumasa. Pressure-strain and pressure-
[25] SF6, CF4, and C2H2F2 Stern SA, Sampat SR, Kulkarni SS. Tests volume relationships in the crystal lattice of polyethylene at
of a “free-volume” model of gas permeation through polymer 293  K. Polym J 1971;2(6):768e82.