Y2O3-stabilized ZrO2 composite membranes under different atmospheres

Journal of Membrane Science 370 (2011) 149–157

Contents lists available at ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

Preparation, characterization and stability of vanadium/Y2 O3 -stabilized ZrO2 composite membranes under different atmospheres Jung Hoon Park ∗ , Edoardo Magnone, Sung Il Jeon, Il Hyun Baek Green House Gas Research Center, Korea Institute of Energy Research, 71-2, Jang-dong, Yuseong-gu, Daejeon 305-343, Republic of Korea

a r t i c l e

i n f o

Article history: Received 28 July 2010 Received in revised form 30 December 2010 Accepted 9 January 2011 Available online 19 January 2011 Keywords: Hydrogen permeation Composite membrane Vanadium Y2 O3 -stabilized ZrO2 ceramic phase Processing parameters

a b s t r a c t In this experimental work vanadium/Y2 O3 -stabilized ZrO2 composite membrane was successfully fabricated by a two-step sintering process. The formation of a well-defined ceramic metal composite membrane was confirmed under high vacuum sintering conditions. Their surface morphology, stability and structural properties were studied as a function of temperature up to 1100 ◦ C by thermogravimetry, differential thermal analysis, scanning electron microscopy, energy dispersive X-ray spectroscopy, and X-ray diffraction analysis. Significant differences were observed in the crystalline structure of the vanadium/Y2 O3 -stabilized ZrO2 composite membranes when they were treated at high temperature in different atmospheres (vacuum, air, hydrogen, nitrogen, and helium). The first hydrogen permeation investigation of vanadium/Y2 O3 -stabilized ZrO2 composite membrane is reported. Preliminary hydrogen permeation experiments have been confirmed that hydrogen flux was 1.08 ml min−1 cm2 for a dense vanadium/Y2 O3 -stabilized ZrO2 composite membrane (thick: 380 ␮m) at 300 ◦ C using 100% H2 as the feed gas. Both high temperature XRD and ex situ study showed deterioration of membranes and the formation of a small amount of vanadium nitride (VNx ) and vanadium hydride (VHx ) solid solutions under nitrogen and hydrogen gas conditions, respectively. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Hydrogen production is a key component of a fuel-cell hydrogen economy. Inorganic membranes for pure hydrogen production and purification are a key technology in the strategy to develop more environmentally friendly power sources [1]. As a result, considerable interest has arisen in the research area of high performance hydrogen membranes under the severe conditions typical of gasification of fuels reforming or hydrogen production. For the production of hydrogen, diverse methods and sources are available [2–5]. Large amounts of data on hydrogen permeation through inorganic membranes of varying composition have been reported by J. Sunarso et al. [6], Lu et al. [1], and Kikuchi [7]. In addition to its potential role in generating fuel-cell hydrogen and gas purification, the water–gas shift (WGS) reaction have current importance in several commercial operations [8]. The literature on palladium is extensive. Over the last decade, researchers have focused on thin palladium membranes supported by porous substrates and ceramic metal composite membranes

∗ Corresponding author at: Korea Institute of Energy Research, Climate Change Technology Research Division, 71-2, Jang-dong, Yuseong-gu, Daejeon 305-343, Republic of Korea. Tel.: +82 42 860 3766; fax: +82 42 860 3134. E-mail address: [email protected] (J.H. Park). 0376-7388/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2011.01.010

for high-pressure, high-temperature applications, as well as the development of other metal–palladium alloy systems [9]; nevertheless, recent efforts in the hydrogen purification and separation membrane have focused on the development and evaluation of non-palladium based membranes that offer a lower cost, high flux, and high durability [10]. A critical component within a membrane reactor unit is the membrane stability behavior against different atmospheres. Hydrogen extraction, in particular, requires high-quality membranes that should exhibit high thermal-chemical stability. Despite this, the effect of changing the atmospheric gas on non-palladium based membrane material properties has been scarcely studied. According to the recent work reported in the literature, the point is that for the middle temperature in the region of 300–600 ◦ C, the crystalline and amorphous palladium or palladium-alloy membranes have been researched extensively to get high flux and solve the hydrogen embitterment of palladium because it is common knowledge that the main disadvantages of this material are its low stability for feedstock poisons such as carbon monoxide, carbon dioxide, ammonia or hydrogen sulfide [11], and its high cost [10]. Exhaustive background reviews of the main limitations of Pd-based membranes for hydrogen separation were recently prepared by Lu et al. [1] and Armor [12]. Alternatively, metals from the Group V including vanadium are currently being evaluated by various investigators because

150

J.H. Park et al. / Journal of Membrane Science 370 (2011) 149–157

Table 1 Coefficient of thermal expansion (TEC) of selected hydrogen permeability metal to Y2 O3 -stabilized ZrO2 ceramic phase and some other selected solid oxide ionic conductors. Metal

Temperature (K)

˛ (×10−6 K−1 )

Ref.

Vanadium

600 700 1000 600 700 1000 298–1273 298–1273 300–1050

10.2 10.5 11.6 13.6 14.1 15.6 10.5 12.3 11.5

[26] [26] [26] [26] [26] [26] [27] [28] [29]

Palladium

ZrO2 stabilized with 8 wt% Y2 O3 Ce0.8 Sm0.2 O2−ı Ce0.8 Gd0.2 O2−ı

these alloys have shown some promise in comparison to palladium [13,14]. For example, Group V elements have over an order of magnitude more permeability for hydrogen than palladium [15–17]. As matters stand, the group V elements have a very high solubility for H2 resulting in a high H/Metal ratio during hydrogen permeation, which can lead to embitterment of the metal that may lead to subsequent cracking of the membrane [18]. Hydrogen permeation is a consequence of the dissolution and diffusion of hydrogen in the dense membranes, and it can be strongly influenced by interactions that occur at material surface and/or at interfaces between metal and another phase components on ceramic metal composite membranes [19–23]. Van Deventer et al. [19] and Namba et al. [21] reported that the surface impurity layers on the pure vanadium membrane give a significant influence on its hydrogen permeation properties and could be well described using an oxide barrier model in according to the study by Strehlow and Savage [23]. Additionally, the value of the thermal expansion coefficient (TEC) should match with those of the other membrane components. In fact, it is well known that the TECs of the different components must be as close to one another as possible in order to minimize thermal stresses which could lead to cracking and mechanical failure. The reason for this possible outcome is the fact that if the various metal and oxide components in a ceramic metal composite membrane are mismatched in thermal expansion characteristics, the membranes can crack during thermal cycling and render the ceramic metal composite membrane ineffective or at least less effective. In searching for a support which will be compatible with high-permeability vanadium-based membrane, it is deliverable to minimize stress at the metal vanadium interfaces which can lead to the formation of cracks. For the most part, stress at the interface can be minimized if the material components of ceramic metal composite membrane have similar TEC over the anticipated temperature range of composite membrane fabrication and use. As a matter of fact, we noted that high-permeability vanadium-based membrane (i) exhibited high hydrogen conductivity in a relatively low-temperature range around 300–500 ◦ C [24] and (ii) the TEC values consistent with 8 mol% Y2 O3 -stabilized ZrO2 ceramic phase. In fact, the TECs for both phases are nearly equal to about 10.5 × 10−6 K−1 [25,26], whereas especially remarkable is the fact that the TEC results on palladium (i.e., 14.1 × 10−6 K−1 at 700 K) [25] are quite different from those of the Y2 O3 -stabilized ZrO2 ceramic phase, or other selected ionic conducting oxide (Table 1). In addition, it is also important to note that TECs of Ce0.8 Sm0.2 O2−ı [27] and Ce0.8 Gd0.2 O2−ı [28] are higher than that of the vanadium phase. Briefly, we hypothesize that the ceramic metal composite membrane stability is governed not only by intrinsic mechanical resistance of single membrane components (i.e. metallic vanadium or Y2 O3 -stabilized ZrO2 ceramic phase) in different gas, but also by thermalstructural stability at the ceramic/metal interfaces.

The present paper has three objectives. First, we provide the first description of the novel dense vanadium/Y2 O3 -stabilized ZrO2 composite membranes for hydrogen production, separation and related processes. It is believed that it is a new idea about vanadium-based composite membrane and a novel and unique engineering approaches in response to current energy issues. Our hypothesis is supported by the fact that vanadium TEC is coincident with Y2 O3 -stabilized ZrO2 ceramic phase on the required temperature range. The second, to give an overview of the preparation and characterization of dense vanadium/Y2 O3 -stabilized ZrO2 composite membrane were carried out under different atmospheric conditions. Therefore, the purpose of this study was also to investigate the stability of metallic vanadium in a new palladium-free composite membrane under different atmospheric condition. The last is to examine some questions concerning the membrane performance. We also compare hydrogen flux measurements from dense vanadium/Y2 O3 -stabilized ZrO2 composite membrane with published measurements from other palladium-free ceramic metal composite membranes. 2. Experimental 2.1. Sample preparation A 60 vol% vanadium/Y2 O3 -stabilized ZrO2 composite membrane was prepared by mechanically mixing vanadium metal powder (325 mesh, 99.5% Alfa Aesar) with Y2 O3 -stabilized ZrO2 ceramic phase (i.e., Tosoh-zirconia) and conventional PVB (Polyvinyl butyral, Sigma) as organic binder. The purpose of PVB, a common binder used in electroceramic processing, is to impart sufficient strength to the vanadium/Y2 O3 -stabilized ZrO2 precursor mixture so that ceramic metal composite membrane will maintain its shape through pre-sintering processing procedures. The mixture was milled with acetone and ball for 1 h. The obtained powder after dry were compressed into disks of 20 mm in diameter and ca. 1.2 mm of thickness in a stainless steel mold under a hydraulic load of 8–20 ton on an area of 3.14 cm2 by unilateral press (model 25601 series, Specac Limited, U.K.). The vanadium/Y2 O3 -stabilized ZrO2 precursor mixtures were pre-sintered at 300 ◦ C under helium condition for 10 h and then sintered at 1600 ◦ C for 2 h in vacuum furnace (10−3 –10−5 Torr). The sintered disk was polished with 600 grit SiC on both surfaces by Grinder (Beta series, Buehler LTD., U.S.A.) in order to remove any potentially contaminating substances. Helium gas-tightness tests of as-prepared membranes were conducted at room temperature with unsteady state gas permeation setup. A detailed description of the helium gas tightness test is given elsewhere [29]. 2.2. Ceramic metal composite membrane characterization Thermogravimetric analysis (TG) and differential thermal analysis (DTA) of the metallic vanadium, Y2 O3 -stabilized ZrO2 ceramic phase, PVB, and vanadium/Y2 O3 -stabilized ZrO2 mixture powder was performed by Thermal Analyzer-SDT600 (TA instrument, U.S.A.) at a heating rate of 5 ◦ C min−1 in helium gas and air to determine the optimum regimen of sintering. The crystal structure of the vanadium/Y2 O3 -stabilized ZrO2 composite membrane were characterized with an X-ray diffractometer (XRD, Rigaku Co Model D/Max 2200-Ultimaplus, Japan) using CuK␣ radiation in the Bragg angle range 20◦ < 2 < 80◦ . In addition, in order to investigate the relation between reactivity and stability of ceramic metal composite membrane the vanadium metal was analyzed by high temperature XRD as a function of gas atmosphere. The morphology and composition of the dense vanadium/Y2 O3 -stabilized ZrO2 composite membrane was ana-

J.H. Park et al. / Journal of Membrane Science 370 (2011) 149–157

151

Fig. 1. The schematic diagrams of permeation test cell (1) bomb; (2) mass flow; (3) ceramic metal composite; (4) permeation controller membrane cell; (5) furnace; (6) pressure controller; (7) temperature controller; (8) back indicator pressure regulator; (9) mass flow; (10) vacuum pump; (11) gas chromatograph; (12) data acquisition system.

lyzed with a scanning electron microscope and energy dispersive X-ray spectroscopy (SEM-EDX, Model 1530, LEO Co., Germany). Micrographs and X-ray dot maps were taken in digital form. 2.3. Hydrogen permeation measurement Fig. 1 shows schematically the process of detecting hydrogen permeation through a vanadium/Y2 O3 -stabilized ZrO2 composite membrane. Polished ceramic metal composite membrane was affixed to stainless steel ring using brazing method with commercial brazing filler (Nicrobraz 30, Wall Colmonoy Co.). After brazing, the vanadium/Y2 O3 -stabilized ZrO2 composite membrane of this assembly was polished to remove the contamination of brazing filler. Then the assembly disc was sealed in stainless steel membrane reactor using pressure exerted on the disk steel plate by screws of reactor. It can be observed here that prior to permeation test, the cell test equipment was vacuumed to remove the air in permeation cell tube and the leakage test for the outer parts of apparatus such as line, fitting and valve was conducted with supplying nitrogen gas at feed side and helium gas at permeation side. Hydrogen permeation study was performed within the temperature range from 500 to 300 ◦ C. It was well known that the time required to reach steady state for the hydrogen permeation flux varies from temperature to temperature. Because the timedependence of permeation flux test is not perfectly constant in all measured temperature range, the hydrogen permeation data at each temperature were determined from last item of data after about ∼2–3 h from the starting time (t0 ). The temperature of the cell was increased with a heating rate of 2 ◦ C min−1 and held at desired temperature (step = 50 ◦ C) over ∼2–3 h to attain steady state condition of gas flow. Feed gases included various concentrations of hydrogen balanced nitrogen and 100% hydrogen and pure argon (99.9999%) were introduced as sweep gas, which were controlled by the mass flow controller (MKS 247C, USA). Hydrogen partial pressure on the feed side was maintained from 0.2 to 1 atm while hydrogen partial pressure on the opposite side was fixed at lower value by argon sweep. The feed and sweep flow rate was kept at 20 ml min−1 , respectively. The hydrogen content in the permeate stream was carried away by

sweep gas and analyzed using a gas chromatograph (GC-TCD, Agilent 7890, Hewlett–Packard). The leakage was measured and the hydrogen permeation fluxes were corrected based on the measured leakage in the case of 20% H2 feed condition. Typical leakage was about 0.006 ml/cm2 min. 3. Results and discussion 3.1. Thermal analysis of the vanadium/Y2 O3 -stabilized ZrO2 precursor mixture under air and helium For reference, thermal analysis of metallic vanadium and Y2 O3 stabilized ZrO2 ceramic phase was carried out. The DTA plots and TG analysis were taken with different gas flow rates to determine the stability properties of the oxidation process and phase transformations for the single membrane components. Fig. 2 shows the typical thermoanalytical curves (DTA and TG curves) of metallic vanadium, Y2 O3 -stabilized ZrO2 ceramic phase and vanadium/Y2 O3 -stabilized ZrO2 precursor mixture as described in Section 2 under different gas flux conditions. The weight gain in air (20 ml/min) is primarily attributed to the oxidation reaction of metallic vanadium with the formation of V2 O3 (inset, Fig 2A). From Fig. 2A, one can observe that the shape and height of lower part of TG/DTA curves below about 315 ◦ C for vanadium samples is independent of the helium gas flow rates. Starting at 305 ◦ C an exothermic effect can be distinguished for the metallic vanadium at low gas flow rate (20 ml/min). These effects can be due to the formation of the VOx structures. Under these conditions, as temperature increase, it can be seen two different TG linear relationships obtained depending on gas flow (20 and 100 ml/min), along with increasing amounts of VOx crystallites. In contrast, based on the TGA results in Fig. 2A, it was noted that the Y2 O3 -stabilized ZrO2 thermogram remained almost constant up to about 1000 ◦ C with a mass loss amounts to 1.0%. The Y2 O3 -stabilized ZrO2 ceramic phase is quite stable to heat, no significant variation was observed in the DTA data and no phase transformation was detected by thermal analysis. The thermogravimetric behavior of the vanadium/Y2 O3 stabilized ZrO2 precursor mixtures at different helium flow rates (20, 50 and 100 ml min−1 ) is shown as thermograms in Fig. 2B. The

152

vanadium

106

He 20 ml/min He 100 ml/min

30 5 o C 37 5 o C

a

J.H. Park et al. / Journal of Membrane Science 370 (2011) 149–157

20

175

103

150

0 vanadiun in air (20 ml/min)

Y2O3-stabilized ZrO2 oxide

125

102

-20

DTA /a.u.

104 TG /%

TG /%

105

vanadium 100

101

500

100

1000

Y2O3-stabilized ZrO2 oxide

o

315 C

-40

99 200

400 600 o Temperature / C o

5 o

C 68

140 o

C 58 0

4

48 5

o

C

0 31

130 120

0 0

-2 400

800 o

C

110

0

PVB in He (100 ml/min)

50

5

-4

40

100

Vanadium/Y2O3-stabilized ZrO2 precursor mixture

90

DTA /a.u.

2 100 TG /%

TG /%

-60 1000

6

C

b

800

20 ml/min

200

the present work. To this end, the decomposed reaction mechanism of PVB binder in the vanadium/Y2 O3 -stabilized ZrO2 precursor mixture is not simple because PVB is a copolymer and then the decomposed products are very complicated on the precursor mixture containing negligible amounts of organic materials. The burn-out of PVB from green bodies of various oxide ceramics and different metal powders was studied by Masia et al. [33] and Angermann et al. [34,35], respectively. As demonstrated in detail in the above-mentioned previous papers, the authors reported that the decomposition of the polymer, in air or in argon, was enhanced by the presence of oxide powders [33], whereas metal powders catalyze the copolymer degradation process [34,35]. Furthermore, Seo et al. [36] studied the decomposition of PVB in the matrix-green sheet and electrode-green sheet. It was reported that most of thermal degradation of pure PVB takes place under 400 ◦ C, whereas in the presence of metal components, the decomposition temperature of the green body is lowered to around 200 ◦ C. Our thermoanalytical results (see inset in Fig. 2B) are consistent with the temperature data compiled by Seo et al. [36]. At the end, on the basis of thermal analysis under varying gas flow conditions (Fig. 2), here it is argued that the vanadiumbased metal ceramic composite membrane must be pre-calcined under helium gas at relatively low-temperature (T ∼ = 300 ◦ C) and long-time treatment to minimize the formation of VOx at interface between vanadium and Y2 O3 -stabilized ZrO2 ceramic phase and, at the same time, to maximize the decomposition of PVB from vanadium/Y2 O3 -stabilized ZrO2 precursor mixture.

50 ml/min

400 600 o Temperature / C

100 ml/min

800

-6 1000

Fig. 2. TG–DTA curves of (a) Y2 O3 -stabilized ZrO2 ceramic phase (inset: metallic vanadium under air conditions) and (b) vanadium/Y2 O3 -stabilized ZrO2 precursor mixture according to gas conditions (inset: PVB under helium conditions).

TG and DTA curves up to 1000 ◦ C can roughly be divided into two well-delimited sections over the gas flow range investigated: one region of a nearly constant mass loss is followed by a region where the mass continuously increases. The second mass gain, ranging from 400 up to 1000 ◦ C, is mostly attributable to the VOx formation by partially oxidation of metallic vanadium. Under this assumption, the vanadium oxidation rates for the vanadium/Y2 O3 -stabilized ZrO2 precursor mixture under low gas flow rate are higher than those under high gas flow rates at a given reaction temperature of 400 to 1000 ◦ C. The results in Fig. 2 clearly show that the evaluation of the precursor membrane stability, via DTA–TG analysis, is thermally stable below 300–400 ◦ C. It is well known that, as a class, vanadium has a strong affinity for oxygen [30,31]. In fact, a major restriction on the use of metallic vanadium composite membrane at elevated temperatures is their susceptibility to oxygen embrittlement by (1) impurities (i.e., air, oxygen trace) and (2) through the Y2 O3 -stabilized ZrO2 ceramic phase that potentially incorporates the effect of oxygen diffusion along phase boundaries. Moreover, a study of formation of vanadium oxides between vanadium and Y2 O3 -stabilized ZrO2 interface in different atmosphere is not mentioned in literature, but it is mentioned that platinum oxides (PtOx ) on the Pt/Y2 O3 -stabilized ZrO2 interface were formed on the ceramic component between 400 and 750 ◦ C with the maximum amount of oxide occurring near 600 ◦ C in air [32]. In addition, PVB was used as organic binder in preparing the dense vanadium/Y2 O3 -stabilized ZrO2 composite membranes in

3.2. Characterization of the vanadium/Y2 O3 -stabilized ZrO2 precursor mixture and sintering under high vacuum conditions In this section, using the two-step sintering process, the authors obtain pure vanadium/Y2 O3 -stabilized ZrO2 composite membranes from a ceramic metal precursor mixture. As described above, the vanadium/Y2 O3 -stabilized ZrO2 precursor mixture was pre-calcined in the temperature region of 300 ◦ C under helium atmosphere for 10 h. This process is effective to decompose the organic polymer phase from the vanadium/Y2 O3 -stabilized ZrO2 precursor mixture and thus preventing the formation of the impurity phases in the ceramic metal composite membrane. The final stage of a two-step sintering process was based on a sintering process in high (∼10−5 Torr) vacuum conditions at high temperature. Preliminary investigations were made by examining the XRD patterns of dense vanadium/Y2 O3 -stabilized ZrO2 composite membranes after sintering at 1600 ◦ C for 2 h under three different gas conditions: helium flux, low (<10−3 Torr), and high (∼10−5 Torr) vacuum conditions. Some XRD patterns of vanadium/Y2 O3 -stabilized ZrO2 composite membranes according to sintering condition are shown in Fig. 3A–C. XRD measurements in these ceramic metal composite membranes (i) indicate a strong phase dependence of vanadium with atmosphere conditions, which is consistent with the TG behaviors of vanadium/Y2 O3 -stabilized ZrO2 precursor mixture (Fig. 2); furthermore, the result shows that (ii) there are no peaks of vanadium oxide when the ceramic metal composite membranes are mechanical polished (Fig. 3D), and at the same time (iii) vanadium of ceramic metal composite membranes is partially converted into vanadium oxide, under helium atmosphere (Fig. 3A). Previous works have shown that vanadium metal can react also with small amount of oxygen in helium atmosphere [37–39]. In the second sintering stage under low vacuum condition (<10−3 Torr), based on Fig. 3B, can be note that vanadium peak was weak while peaks of various impurities existed. Based on the XRD measurements it can be concluded that both preparation methods used (helium atmosphere and low vacuum conditions, respectively)

J.H. Park et al. / Journal of Membrane Science 370 (2011) 149–157

VO2 , 44-0253

YSZ, 48-0224

VC, 07-0257

V, 22-1058

ing. This analysis has confirmed the presence of oxide phases evenly distributed in the resultant vanadium metal matrix. This was further verified by the dot map of vanadium distribution (Fig. 4D) in the area of the membrane indicated in Fig. 4C. Ceramic metal composite membranes are composed of vanadium, in the form of metal (red), in which the second phase, Y2 O3 -stabilized ZrO2 ceramic phase (green), is suspended. In the present section, on the basis of micrographic observation and X-ray diffraction analysis, it is argued that the second-step annealing at high temperature under high vacuum conditions was required to prevent the formation of second phase compounds in a vanadium-nonmetallic impurity systems (i.e., oxides, carbides). More specifically, dense vanadium/Y2 O3 -stabilized ZrO2 composite membrane without any carbide impurity can be prepared by simple mechanical polishing. In conclusion, the sintering results reported here are interesting because it was concluded that is it possible to achieve a well-defined homogeneous distribution of the vanadium metal within the Y2 O3 -stabilized ZrO2 ceramic phase for possible use as hydrogen separation membrane.

Unknown

Intensity (arbitrary unit)

(a)

(b)

(c)

3.3. Hydrogen permeability of the vanadium/Y2 O3 -stabilized ZrO2 composite membrane

(d)

20

30

40

50

60

70

153

80

o

2θ( ) Fig. 3. XRD patterns of vanadium/Y2 O3 -stabilized ZrO2 composite membranes according to three different sintering conditions: (a) helium, (b) low vacuum, (c) high vacuum samples before mechanical polishing, and (d) high vacuum composite membrane after polishing.

caused detrimental effects to vanadium/Y2 O3 -stabilized ZrO2 composite membranes. Fig. 3C and D shows the room temperature XRD patterns of dense vanadium/Y2 O3 -stabilized ZrO2 composite membranes after high vacuum sintering process (1600 ◦ C; 2 h) and after mechanical polishing its surface with 600 grit SiC abrasive paper. For the vanadium/Y2 O3 -stabilized ZrO2 composite membranes, an appreciable amount of metallic vanadium remained after sintering at 1600 ◦ C under high vacuum conditions and, from the same Fig. 3C, it can be observed also the presence of vanadium carbide phase. The characteristic reflections of the vanadium carbide compounds can be seen clearly in Fig. 3C. Within the detection limit of the XRD method, the characteristic reflections of vanadium carbide compounds disappear, and the reflections of vanadium metal appear on the surface of dense vanadium/Y2 O3 -stabilized ZrO2 composite membranes, after mechanical polishing. The observed results could be explained by formation of vanadium carbide compound throughout the surface of vanadium/Y2 O3 -stabilized ZrO2 composite membranes. This vanadium carbide (VC) was formed by the reaction between vanadium of composite membrane and carbon originated from sample holder (which is made of carbon) during sintering process. These observations are consistent with SEM/EDX analysis and EDX dot-map analysis of ceramic metal composite membrane (Fig. 4) before mechanical polishing with SiC abrasive paper, which showed carbon element to become progressively more important from the bulk to surface of membrane. In fact, Fig. 4A and B shows a cross section SEM micrograph and EDX spectra, respectively, of surface membrane. Marker bar represents 6 ␮m. The thickness of the revealed carbon element layer is on the order of about 10 ␮m (Fig. 4B). Moreover, Fig. 4C shows representative SEM images of the ceramic metal composite membrane after mechanical polish-

Hydrogen flux through a 380 ␮m-thick vanadium/Y2 O3 stabilized ZrO2 composite membrane as a function of temperature and time using 100% H2 as feed gas is presented in Fig. 5. The hydrogen fluxes were measured from isothermal experiments from 500 ◦ C to 300 ◦ C. Apparently, as the temperature was decreased from 500 to 300 ◦ C, the hydrogen permeation flux of dense vanadium/Y2 O3 -stabilized ZrO2 composite membrane partially increased from 0.92 to 1.08 ml min−1 cm−2 . There is no existing work, to the best of our knowledge, concerning hydrogen permeation flux of dense vanadium/Y2 O3 -stabilized ZrO2 composite membranes in different gas conditions. However, in the previous papers it was reported that a flux of ∼0.8 and ∼2 cm3 min−1 cm−2 was obtained for a nickel/Ba(Zr0.1 Ce0.7 Y0.2 )O3 (266-␮m-thick) [40] and palladium/Y2 O3 -stabilized ZrO2 (210␮m-thick) [41] composite membrane, respectively, using 100% hydrogen as the feed gas at 900 ◦ C. It is important to note that a hydrogen flux of ∼1 ml min−1 cm−2 was obtained in our work at low temperature (300–500 ◦ C), and a thick of approximately 380 ␮m. The inset of Fig. 5 shows the characteristic dependence of logarithm of hydrogen flux of vanadium/Y2 O3 -stabilized ZrO2 composite membrane on reciprocal temperature. It can be noted that the vanadium/Y2 O3 -stabilized ZrO2 composite membrane obtained in this study showed the high hydrogen permeation flux at any operating temperature (∼1 ml min−1 cm−2 ). Interestingly, the high temperature (i.e., 500 ◦ C) had a much lower hydrogen flux than the membrane tested at low temperature (i.e., 300 ◦ C). Fig. 5 also summarizes the transient response behaviors of hydrogen permeation at the temperature range of interest. This means that time-normalized curves in Fig. 5 represent the transient response behaviors of the dense vanadium/Y2 O3 -stabilized ZrO2 composite membrane at the start of each measure (t0 ). This figure shows that, in first approximation, (i) the hydrogen permeation rate of initial states at each isothermal process was lower than the final hydrogen flux data (I, II, III, IV and V), (ii) the transient response rate from t0 was decelerated by the decrease of temperature from 500 to 300 ◦ C, and (iii) the hydrogen permeation flux does not perfectly reach to the attended steady state value at 450 ◦ C (II) and 400 ◦ C (III) in the range of time here considered (130 min). It should be stressed here that, just on the basis of transient response behaviors of hydrogen permeation, this temperature range may be important for maintaining the stability of vanadium/Y2 O3 -stabilized ZrO2 composite membrane. An extensive structural study of the inter-

154

J.H. Park et al. / Journal of Membrane Science 370 (2011) 149–157

Fig. 4. (a) Cross-section SEM micrograph (marker bar = 6 ␮m), and (b) results of the EDS linear analysis of carbon element distribution on cross-sections of the sample before mechanical polishing. (c) Representative SEM image of the same ceramic metal composite membrane after mechanical polishing with SiC abrasive paper and (d) corresponding dot map localization of vanadium (red) obtained by energy-dispersive X-ray microanalysis of (c) sample (marker bar = 100 ␮m).

actions in the above temperature range could help us define the nature of events in the vanadium/Y2 O3 -stabilized ZrO2 composite membrane structure. Additionally, (iv) hydrogen permeation flux slightly decreased with time at 300 ◦ C (Fig. 5). This last effect

V IV

300 C

1.05

350 C

III

t0 1.00

400 C

II 0.95

450 C

I

500 C

0.90

log (hydrogen flux)

2

Hydrogen flux /ml/cm min

1.10

1

1.2

1.4

1.6

1.8 -1

1000/T /K

0

60

120

180

240

time /min Fig. 5. Hydrogen flux normalized (t0 = 0 s) through 380 ␮m-thick vanadium/Y2 O3 stabilized ZrO2 composite membrane as a function of temperature and time using 100% H2 as feed gas. Error bars denote the range of final flux in the temperature range from 500 to 300 ◦ C. Inset: dependence of logarithm of hydrogen flux on reciprocal temperature.

may be due to the deactivation of dense vanadium/Y2 O3 -stabilized ZrO2 composite membrane at the end of experimental protocol, as will be shown in the next section on the basis of structural studies using XRD to explain the phase changes occurring in hydrogenpermeability processes. As a result, the exact nature of all these experimental hydrogen flux data and observations are still not known but, for instance, the average hydrogen flux data of experiment at 500 ◦ C (see I in Fig. 5) is nearly 6/7 of the corresponding V isothermal experiments at 300 ◦ C. Hydrogen permeation fluxes through dense vanadium/Y2 O3 stabilized ZrO2 composite membrane according to feed concentration and temperature are shown in Fig. 6. For clarity, the inset of Fig. 6 shows the dependencies of the hydrogen flux of the process on reciprocal temperature. A comparison of the two graphs indicates that (I) a direct correlation between hydrogen feed concentration and the measured hydrogen flux data was observed. Moreover, hydrogen permeation fluxes through vanadium/Y2 O3 stabilized ZrO2 composite membrane (II) depends strongly on gas feed concentration at given temperature, but (III) the average flux was found to be independent of temperature, at given hydrogen feed concentration. The preliminary results reported above are very interesting since they are the first experimental data for hydrogen permeability through a dense vanadium/Y2 O3 -stabilized ZrO2 composite membrane under different gas conditions. It has been found that vanadium/Y2 O3 -stabilized ZrO2 composite membrane can work as hydrogen separation membranes at very low temperature.

J.H. Park et al. / Journal of Membrane Science 370 (2011) 149–157

0.8

100 vol %

log (hydrogen flux)

1.0

2

Hydrogen flux /ml/cm min

1.2

1

60 vol % 40 vol %

0.1

0.6

1.3

1.4

1.5

1.6

1.7

1.8

-1

1000/T /K

0.4

300 C 350 C 400 C 450 C 500 C

0.2 0.0 20

40

60

80

100

Hydrogen feed concentration /vol% Fig. 6. Hydrogen flux through 380 ␮m-thick vanadium/Y2 O3 -stabilized ZrO2 composite membrane according to hydrogen feed concentration and temperature. Inset: dependence of logarithm of hydrogen flux on reciprocal temperature and feed composition.

In the other hand, our preliminary results on vanadium/Y2 O3 stabilized ZrO2 composite membrane are quite different from those of the another composite membranes as nickel/Ba(Zr0.1 Ce0.7 Y0.2 )O3 (266-␮m-thick) [40] and palladium/Y2 O3 -stabilized ZrO2 (210␮m-thick) [41] whose hydrogen flux increased with increasing temperature. In the case of hydrogen permeable metal such as palladium and nickel composite membrane, hydrogen solubility increases with temperature but in the refractory group V metals such as vanadium decrease of hydrogen solubility more rapidly than the increase of the diffusion coefficient at high temperature [17]. Moreover, the steady-state hydrogen flux increased with increasing temperature as was expected since hydrogen permeation through palladium increases exponentially with temperature [42]. The reasons are not immediately clear, but it may be the particular blocking effect of vanadium-nitrogen compound on active site of vanadium [21,22]. In the other hand, surface impurity layers were found to reduce the hydrogen permeability of vanadium by a factor of several hundred over the temperature range from 200 to 850 ◦ C [19,43]. A similar sensitivity of hydrogen permeability to surface impurity has been also observed for niobium-based sample by Levin and Stickney [44]. In first approximation, the mechanism of the flux decrease at high temperature (see Fig. 5) has been attributed to formation of N-based compound and other impurities from the surface during the flux test periods. Further testing is necessary to confirm this hypothesis. However, we will continue to investigate the mechanism of the non-Pd-based membrane alloys, such as V-based alloys, during the next years work. 3.4. Ex situ structural characterization of the vanadium/Y2 O3 -stabilized ZrO2 composite membrane As we saw in Section 1, vanadium/Y2 O3 -stabilized ZrO2 composite membranes have been preliminary characterized with respect to their chemical stability during the hydrogen-permeability measurements. The results of obtained vanadium/Y2 O3 -stabilized ZrO2 composite membrane after permeation test are shown in Fig. 7. The XRD spectra represent the sum of the Bragg contribution generated by each crystalline phase present in the surface of ceramic metal composite membrane. In particular, the XRD pattern of the vanadium/Y2 O3 -stabilized ZrO2 composite membrane possessed

Intensity (arbitrary unit)

1.4

155

(a) VN0.2, 08-0380

YSZ, 48-0224 V, 22-1058

(b)

20

40

60

80

2θ(o) Fig. 7. XRD patterns of vanadium/Y2 O3 -stabilized ZrO2 composite membrane after permeation test: (a) permeate and (b) feed side.

some additional weak peaks marked by solid dots due to vanadiumnitrogen alloys (i.e., VN0.2 ) at the surface membrane. Based on XRD results obtained after permeation tests, our study confirms that the nitrogen presence always lowers the stability of vanadium/Y2 O3 -stabilized ZrO2 composite membrane thus favoring the formation of vanadium-nitrogen alloy [45]. There are several reasons that cause ceramic metal composite membranes to deteriorate, but this result would indicate that one cause of membrane deterioration in V-based composite membrane, that affects hydrogen permeation performance at high temperature [21,22,30,31,43,46], is the formation of vanadium-nonmetallic compounds. Furthermore, formation of the vanadium nitride in permeation side membrane suggests that vanadium/Y2 O3 stabilized ZrO2 composite membrane can react with nitrogen gas. This observation is producible and shed some light on the origin of instability in vanadium-based composite membrane in presence of nitrogen gas.

3.5. High temperature characterization and stability under different atmospheres As explained in previous section, the chemical stability of hydrogen separation membranes is a critical issue. In order to investigate the brittleness and reactivity under hydrogen and nitrogen, the vanadium metal was analyzed by high temperature XRD in the temperature range of interest. Fig. 8 shows high temperature XRD results of vanadium powder as a function of temperature. This figure shows clearly that the XRD pattern changes significantly under different atmospheres according to temperatures. As temperature increased, intensities of vanadium peaks decreased gradually; whereas the impurity concentration is proportional to the temperature. In particular, the diffraction peak intensities of vanadium decreased above 300 ◦ C in hydrogen atmosphere (Fig. 8A). The XRD measurement results shown in Fig. 8B indicate that the stability of vanadium decreased above 400 ◦ C under nitrogen atmosphere. In the thermodynamic context, this means that the rate of reaction with nitrogen is much less at a given temperature than the corresponding hydrogenation reaction. These results are consistent with those of other studies [47]. In addition, a sluggish transformation between 500 and 600 ◦ C in some heterogeneous vanadium-nitrogen alloys was observed earlier [48].

156

J.H. Park et al. / Journal of Membrane Science 370 (2011) 149–157

vacuum conditions has been proved to be effective in fabricating high-quality ceramic metal composite membranes. In this experimental work the influence of different gases on the hydrogen permeation through a dense vanadium/Y2 O3 -stabilized ZrO2 composite membrane was studied for the first time. In summary, our preliminary results suggested that the maximum hydrogen flux through 380 ␮m thick membrane exposed to flowing 100% H2 as feed gas was 1.08 ml min−1 cm−2 at low temperature, T = 300 ◦ C. Experiments show the promising result of the proposed dense vanadium/Y2 O3 -stabilized ZrO2 composite membranes in comparison with some existing ceramic metal composite membranes. Deterioration of vanadium/Y2 O3 -stabilized ZrO2 composite membrane is originated from vanadium nitride formed after permeation test using various content H2 /balanced N2 as the feed gas. The tendency of the transformation of the vanadium/Y2 O3 stabilized ZrO2 composite membrane structure was evaluated by the appearance and the amount of the vanadium-nitrogen alloys (VNx ) and vanadium hydride (VHx ) solid solutions under nitrogen (100 ml min−1 ) and hydrogen (100 ml min−1 ) gas conditions, at above 300 and 400 ◦ C, respectively. These results suggest that we can further increase the hydrogen flux by protective coating layer and/or by reducing the thickness of vanadium/Y2 O3 -stabilized ZrO2 composite membrane. Moreover, effects of alloying elements on the vanadium-based composite membrane stability for hydrogen production, separation and related processes will be an important area of future study. Acknowledgement This work was supported by Energy & Resource R&D program (2008-C-CD11-P-09-0-0000) under the Ministry of Knowledge Economy, Republic of Korea. References

Fig. 8. High temperature XRD results of vanadium powder as a function of temperature and different gas atmosphere: (a) hydrogen atmosphere (100 ml min−1 ) and (b) nitrogen atmosphere (100 ml min−1 ).

4. Conclusions The outstanding preparation and high temperature stability of the vanadium/Y2 O3 -stabilized ZrO2 composite membrane constitutes a key point for the development of this new composite material. Thus, the objective of this work was for preparation, characterization and stability test of dense vanadium/Y2 O3 -stabilized ZrO2 composite. Novel preparation process of vanadium/Y2 O3 -stabilized ZrO2 composite membrane for hydrogen purification membrane was investigated and sintering conditions were optimized for preparation of new high-permeability ceramic vanadium-based composite membranes. Based upon the XRD analysis at room temperature, significant differences were observed in the crystalline structure of the vanadium/Y2 O3 -stabilized ZrO2 composite membranes when they were prepared in different atmospheres (vacuum, air, hydrogen, nitrogen, and helium). A complex sintering process under

[1] G.Q. Lu, J.C. Diniz da Costa, M. Duke, S. Giessler, R. Socolow, R.H. Williams, T. Kreutz, Inorganic membranes for hydrogen production and purification: a critical review and perspective, J. Colloid Interface Sci. 314 (2) (2007) 589–603. [2] S. Adhikari, S. Fernando, Hydrogen membrane separation techniques, Ind. Eng. Chem. Res. 45 (3) (2006) 875–881. [3] X. Wei, J. Kniep, Y.S. Lin, Hydrogen permeation through terbium doped strontium cerate membranes enabled by presence of reducing gas in the downstream, J. Membr. Sci. 345 (1–2) (2009) 201–206. [4] H. Matsumoto, T. Shimura, H. Iwahara, T. Higuchi, K. Yashiro, A. Kaimai, T. Kawada, J. Mizusaki, Hydrogen separation using proton-conducting perovskites, J. Alloys Compd. 408–412 (2006) 456–462. [5] S.J. Song, J.-H. Moon, T.H. Lee, S.E. Dorris, U. Balachandran, Thickness dependence of hydrogen permeability for Ni-BaCe0.8 Y0.2 O3−ı , Solid State Ionics 179 (33–34) (2008) 1854–1857. [6] J. Sunarso, S. Baumann, J.M. Serra, W.A. Meulenberg, S. Liu, Y.S. Lin, J.C. Diniz da Costa, Mixed ionic-electronic conducting (MIEC) ceramic-based membranes for oxygen separation, J. Membr. Sci. 320 (1–2) (2008) 13–41. [7] E. Kikuchi, Membrane reactor application to hydrogen production, Catal. Today 56 (1–3) (2000) 97–101. [8] M.T. Ravanchi, T. Kaghazchi, A. Kargari, Application of membrane separation processes in petrochemical industry: a review, Desalination 235 (1–3) (2009) 199–244. [9] J. Shu, B.P.A. Grandjean, A. Van Neste, S. Kaliaguine, Catalytic palladium-based membrane reactors: a review, Can. J. Chem. Eng. 69 (5) (2009) 1036–1060. [10] R. Bredesen, K. Jordal, O. Bolland, High-temperature membranes in power generation with CO2 capture, Chem. Eng. Process. 43 (9) (2004) 1129–1158. [11] H. Yoshida, S. Konishi, Y. Naruse, Effects of impurities on hydrogen permeability through palladium alloy membranes at comparatively high pressures and temperatures, J. Less Common. Met. 89 (2) (1983) 429–436. [12] J.N. Armor, Applications of catalytic inorganic membrane reactors to refinery products, J. Membr. Sci. 147 (2) (1998) 217–233. [13] R.E. Buxbaum, A.B. Kinney, Hydrogen transport through tubular membranes of palladium-coated tantalum and niobium, Ind. Eng. Chem. Res. 35 (2) (1996) 530–537. [14] R.E. Buxbaum, T.L. Marker, Hydrogen transport through non-porous membranes of palladium-coated niobium, tantalum and vanadium, J Membr. Sci. 85 (1993) 29–38. [15] S. Uemiya, Brief review of steam reforming using a metal membrane reactor, Top. Catal. 29 (1) (2004) 79–84.

J.H. Park et al. / Journal of Membrane Science 370 (2011) 149–157 [16] M.V. Mundschau, X. Xie, C.R. Evenson, Superpermeable hydrogen transport membranes, non porous inorganic membranes, Chapter 4, in: A.F. Sammells, M.V. Mundschau (Eds.), Nonporous Inorganic Membranes, Wiley VCH Verlag GmHB, Weinheim, 2006, pp. 107–130. [17] Review of hydrogen isotope permeability through materials, in: S.A. Steward (Ed.), National Information Service, Laurence Livermore National Laboratory Report, Springfield, VA, USA, 1983. [18] P.J. Grimmer, Scale-up of hydrogen transport membranes for IGCC and futureGen plants, in: U.S. DOE Hydrogen Annual Merit Review Meeting, Arlington, VA, USA, 2006. [19] E.H. Van Deventer, T.A. Renner, R.H. Pelto, V.A. Maroni, Effects of surface impurity layers on the hydrogen permeability of vanadium, J. Nucl. Mater. 64 (1977) 241–248. [20] M. Yamawaki, T. Namba, T. Yoneoka, K. Shida, M. Kanno, Effect of argon ion sputtering of surface on hydrogen permeation through vanadium, J. Nucl. Sci. Technol. (Tokyo Jpn.) 20 (5) (1983) 405–413. [21] T. Namba, H. Miyaguchi, M. Yamawaki, M. Kanno, Hydrogen permeation through vanadium and the effect of surface impurity layer on IT, J. Nucl. Mater. 105 (2–3) (1982) 318–325. [22] T. Namba, M. Yamawaki, M. Kanno, Surface effects on hydrogen permeation through fusion reactor materials – a new model and its application to vanadium membrane, J. Facul. Eng. (The University of Tokyo) (B) 37 (1984) 491–516. [23] R.A. Strehlow, H.C. Savage, Formation and use of oxide films to impede outgassing of metals, J. Nucl. Mater. 53 (1974) 323–327. [24] V.I. Pokhmurs’kyi, O.R. Sokolovs’kyi, V.V. Fedorov, High-temperature hydrogen permeability of vanadium and niobium, Mater. Sci. 30 (4) (1995) 410–418. [25] M.V. Mundschau, Hydrogen transport membranes, Patent No US 6,899,744 B2, Eltron Research, Inc., USA, May 31, 2005. [26] H. Hayashi, T. Saitou, N. Maruyama, H. Inaba, K. Kawamura, M. Mori, Thermal expansion coefficient of yttria stabilized zirconia for various yttria contents, Solid State Ionics 176 (5–6) (2005) 613–619. [27] B. Lin, S. Wang, X. Liu, G. Meng, Simple solid oxide fuel cells, J. Alloys Compd. 490 (1–2) (2010) 214–222. [28] V.V. Kharton, A.V. Kovalevsky, A.P. Viskup, A.L. Shaula, F.M. Figueiredo, E.N. Naumovich, F.M.B. Marques, Oxygen transport in Ce0.8 Gd0.2 O2−ı -based composite membranes, Solid State Ionics 160 (3–4) (2003) 247–258. [29] J.H. Park, J.P. Kim, H.T. Kwon, J. Kim, Oxygen permeability electrical property and stability of La0.8 Sr0.2 Co0.2 Fe0.8 O3−␦ membrane, Desalination 233 (1–3) (2008) 73–81. [30] G. Busch, A. Tobin, Oxidation of vanadium and vanadium alloys in gaseous helium coolants containing water vapor impurities, J. Nucl. Mater. 141–143 (1986) 599–603. [31] J. Bethin, A. Tobin, Application of vanadium and vanadium alloys to a fusion reactor blanket, J. Nucl. Mater. 122–123 (1984) 864–868.

157

[32] T. Chao, K.J. Walsh, P.S. Fedkiwa, Cyclic voltammetric study of the electrochemical formation of platinum oxide in a Pt/yttria-stabized zirconia cell, Solid State Ionics 47 (3–4) (1991) 277–285. [33] S. Masia, P.D. Calvert, W.E. Rhine, H.K. Bowen, Effect of oxides on binder burnout during ceramics processing, J. Mater. Sci. 24 (1989) 1907–1912. [34] H.H. Angermann, F.K. Yang, O. Van der Biest, Catalytic effects during binder removal from metal powder compacts, Int. J. Powder Metall. 28 (1992) 361–367. [35] H.-H. Angermann, O. Van der Biest, Binder removal mechanisms from powder compacts during thermal debinding, in: Proc 1993 Powder Metallurgy World Congress, Kyoto, Japan, 1993. [36] J.J. Seo, S.T. Kuk, K. Kim, Thermal decomposition of PVB (polyvinyl butyral) binder in the matrix and electrolyte of molten carbonate fuel cells, J. Power Sources 69 (1–2) (1997) 61–68. [37] J. Choi, J. Ha, J. Hong, Synthesis and catalytic properties of vanadium interstitial compounds, Appl. Catal. A 168 (1998) 47–56. [38] H. Preiss, D. Schultze, K. Szulzewsky, Carbothermal synthesis of vanadium and chromium carbides from solution-derived precursors, J. Eur. Ceram. Soc. 19 (1999) 187–194. [39] P.K. Tripathy, On the thermal decomposition of vanadium nitride, J. Mater. Chem. 11 (2001) 1514–1518. [40] C. Zuo, T.H. Lee, S.E. Dorris, U. Balachandran, M. Liu, Composite NiBa(Zr0.1 Ce0.7 Y0.2 )O3 membrane for hydrogen separation, J. Power Sources 159 (2) (2006) 1291–1295. [41] U. Balachandran, T.H. Lee, L. Chen, S.J. Song, J.J. Picciolo, S.E. Dorris, Hydrogen separation by dense cermet membranes, Fuel 85 (2) (2006) 150–155. [42] S.N. Paglieri, K.Y. Foo, J.D. Way, A new preparation technique for Pd/alumina membranes with enhanced high-temperature stability, Ind. Eng. Chem. Res. 38 (1999) 1925–1936. [43] R. Hayakawa, Y. Hatano, K. Watanabe, Influence of surface oxide film on hydrogen permeation through vanadium membrane, in: Annual Report of Hydrogen Isotope Research Center, vol. 19, Toyama University, Toyama, Japan, 1999, pp. 54–64. [44] R.L. Levin, R.E. Stickney, Permeation of hydrogen isotopes through fusion reactor materials, in: Proc. Fifth Symposium on Engineering Problems of Fusion Research, Princeton University, 1974. [45] O.N. Carlson, J.F. Smith, R.H. Nafziger, The vanadium-nitrogen system: a review, Metall. Mater. Trans. A 17 (10) (1986) 1647–1656. [46] K. Takata, T. Suzuki, Permeation of hydrogen in vanadium. I. experimental details, Mater. Sci. Eng. A 163 (1993) 91–97. [47] E.A. Gulbransen, K.F. Andrew, The kinetics of the reactions of vanadium with oxygen and nitrogen, J. Electrochem. Soc. 97 (11) (1950) 396–404. [48] P. Muije, R. Horton, S. Duran, Kinetics of solid vanadium-nitrogen reactions at high temperatures, Metall. Mater. Trans. B 5 (1) (1974) 97–104.