ceramic composite membrane by electroless plating

ceramic composite membrane by electroless plating

Colloids and Surfaces A: Physicochemical and Engineering Aspects 179 (2001) 229 – 235 www.elsevier.nl/locate/colsurfa Preparation of asymmetric Ni/ce...

224KB Sizes 0 Downloads 18 Views

Recommend Documents

No documents
Colloids and Surfaces A: Physicochemical and Engineering Aspects 179 (2001) 229 – 235 www.elsevier.nl/locate/colsurfa

Preparation of asymmetric Ni/ceramic composite membrane by electroless plating Xia Changrong a,*, Guo Xiaoxia a, Li Fanqing b, Peng Dingkun a, Meng Guangyao a a

Department of Materials Science and Engineering, Uni6ersity of Science and Technology of China, Hefei 230026, PR China b USTC Structure Research Laboratory, Uni6ersity of Science and Technology of China, Hefei 230026 PR China

Abstract Although a lot of work has been done on electroless nickel coating, work on preparing asymmetric nickel membranes on porous ceramic supports is rarely reported. In this work, microfiltration nickel membranes are prepared on porous alumina supports with electroless plating. A new approach technique, sol – gel process is used to activate the alumina supports. The coated membranes were investigated by means of scanning electron microscopy and gas permeation test. Membrane thickness and amount of nickel increase with plating time. Membrane pore size decreases greatly as electroless coating is processed for 15 min. After 15-min coating the pore size decreases slightly. For a 90-min electroless plated membrane, the thickness reaches 4.5 mm and the mean pore radius is 0.13 mm with a narrow distribution. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Composite membrane; Inorganic membrane; Nickel; Asymmetric membrane; Electroless plating

1. Introduction Preparation of inorganic membranes has received much attention since 1980. Especially, asymmetric inorganic membranes have been of interest. These membranes consist usually of a thin separation layer and a porous carrier, which serves as a support and has pore size larger than the separation layer. Sol – gel technique is considered to be the most practical method for fabrication of porous ceramic membranes [1,2]. * Corresponding author. Tel.: + 86-551-3603234; fax: +86551-3631760. E-mail address: [email protected] (X. Changrong).

Examples are porous alumina, zirconia and titania membranes, which are produced at commercial level nowadays. In addition to the sol–gel technique, slip-casting, track-etch and chemical vapor deposition have also been applied for the fabrication of porous asymmetric inorganic membranes [1,2]. Electroless plating plays an important role as a metal coating technique on glass, ceramic, plastic or metal substrates [3–6]. This technique provides distinct advantages such as uniformity of deposits on complex shapes, hardness, low cost as well as using very simple equipment. It is based on the controlled auto-catalytic reduction of metastable metallic salt complexes on target surface. Re-

0927-7757/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 7 - 7 7 5 7 ( 0 0 ) 0 0 6 4 2 - 7

230

X. Changrong et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 179 (2001) 229–235

cently, electroless plating has been used to prepare metal membranes on porous supports. Examples are dense palladium membranes and alloys membranes of palladium and silver [3 – 5]. However, as far as we know, no work has been done on the fabrication of porous nickel membrane on a porous ceramic support by the electroless plating. In addition to separation application with a pressure gradient, porous nickel membranes also offer potential applications in solid oxide fuel cells for collecting electron at anode. In order to initiate electroless plating, it is necessary to activate the insulating surface. In general, palladium metal is used as a catalyst center, and hence, the following two methods have primarily been utilized to deposit palladium particles oil insulator surfaces [6]. One technique is a two-step treatment, which uses an acidic tin chloride solution as a sensitizer and an acidic palladium chloride solution as an activator, respectively. In this process, Sn(II) ions are initially adsorbed onto the substrate surface after being immersed in SnCl2 solution. Next, palladium metal particles with zero valence are deposited around the tin ions adsorbed onto the substrate surface as the following reaction Sn2 + + Pd2 + “Sn4 + +Pd0 Since the deposited palladium particles act as catalyst centers, the reaction of electroless plating can proceed. Another technique is a one-step treatment, which uses a mixed solution of tin chloride and palladium chloride. In this process, the colloidal particles of Pd – Sn alloy act as catalyst centers and are adsorbed onto the substrate surface by immersion in a mixed PdCl2 –SnCl2 solution. A new activation process based on a sol–gel process was reported recently [3]. In this process, a composite film of palladium nuclear on alumina particles is deposited onto a porous substrate instead of sensitizing the substrate surface in a conventional method. The process might take the advantage of resulting in a membrane with good adhesion. To prepare the composite film, a mixed solution consisting of boehmite (AlOOH) sol and acidic palladium chloride is used as the source of alumina and palladium metal, respectively. In the

solution, Pd2 + is absorbed onto the surface of boehmite particles. The porous substrate is then brought into contact with the solution, a layer consisting of boehmite particles with Pd2 + is formed at the boundary of support and solution due to capillary forces. After heating and reduction, a composite film with Pd particle is formed, and the palladium metal serves as the catalyst center. In this work porous alumina supports are activated with this new process. Microfiltration nickel membranes are then electroless-deposited on the activated substrate. And the membranes are characterized by means of gas permeate and scanning electron microscopy. The pore size and its distribution are also investigated.

2. Experiments As shown in Fig. 1, the first step is the synthesis of Pd–Al2O3 composite membranes on porous alumina substrates by a sol–gel process. The precursor of Al2O3 composite membrane is boehmite sol obtained by peptization of boehmite precipitates, which results from the hydrolysis of aluminum tri-sec-butanol with distilled water. Palladium chloride, which is mixed with boehmite sol, is used as the source of palladium. With this mixed sol, PdO–Al2O3, composite membrane is prepared on a-alumina supports by a dipping– drying–calcining procedure. The composite oxide is then activated in a hydrogen stream to get an activated substrate. The second step is electroless nickel plating on the activated substrates. And finally, nickel membranes are characterized in terms of thickness, gas permeate and active pore size.

2.1. Preparation of Pd–Al2O3 composite membrane Boehmite (g-AlOOH) sols were prepared by adding dropwise aluminum tris-sec-dutoxide(B2400, Geel, Belgium) to water, which was heated to 85°C and stirred at high speed. About 2.0 1 of water was used per mole alkoxide. After addition of the alkoxide, it was kept boiling until most of

X. Changrong et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 179 (2001) 229–235

the butanol evaporated. The resulting precipitates were then peptized by adding 0.10 mole HNO3 per mole butoxide to get a colloidal suspension. It was refluxed for 12 h at 85°C to form a 0.40 M stable boehmite sol. Polyvinyl alcohol (PVA) solutions were added to boehmite sol under reflux-

231

ing to get a 0.2 wt.% PVA composition. PVA was used here as drying control chemical additives. Palladium chloride solution was made by adding 0.034 g PaCl2 to 10 ml concentrated chloride acid. It was diluted to 150 ml and concentrated to 10 ml again by heating to evaporate hydrogen chloride. The solution was diluted to 50 ml dripped to 100 ml boehmite sol, and agitated for 4 h at 85°C to get a dip-solution, a colloid stable for more than a week at room temperature and red–yellow in color. The supported Pd–Al2O3 composite films were prepared by dipping home-made a-alumina supports (diameter, 12 mm; thickness, 2 mm; porosity, 40%; active pore size, 0.90 mm) for 10 s in the dip-solution, drying for 24 h at room temperature and 70% relative humidity, calcining at a rate of 20°C h − 1 to 900°C, where they were kept for 5 h, and finally, heating to 400°C in a stream of mixed gas consisted of 50 vol.%, hydrogen and 50 vol.% argon. After activating in hydrogen stream, the substrates were kept in a 2 g/100 ml sodium hypophosphite solution before nickel plating.

2.2. Electroless plating

Fig. 1. Schematic route for preparing nickel membrane. Table 1 Bath composition and operating conditions of electroless nickel plating Chemical NiSO4 · 6H2O NiCl2 · 6H2O EDTA-2Na NaH2 PO2 Na3C5H5O7 2H2O pH Bath temperature (°C)

Concentration (g/100 ml) 1.2 0.15 1.5 0.5 0.1 9.0 50

For electroless nickel plating, a hypophosphitebased bath was used. The composition of the bath is given in Table 1, in which 2Na salt of ethylenediaminetetraacetic acid (EDTA-2Na) was used as a complexing agent to maintain the level of nickel ion in solution. The pH of the bath was adjusted to 9.0 with ammonium hydroxide. Bath solution of 40 ml cm − 2 substrate was used for the electroless nickel plating, which was carried out at 50°C. Several pieces of substrates were used to investigate the amount of nickel as a function of plating time.

2.3. Characterization of nickel membranes The nickel film thickness and surface morphology were determined by scanning electron microscopy (SEM, Hitachi X-650). The active pore size and pore size distribution of the membranes were obtained by a so-called modified bubble point method [7] with a home-made device [8] and hydrogen permeation properties were measured

232

X. Changrong et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 179 (2001) 229–235

decreases as the plating processes, since 40 ml bath per square centimeter substrate is used. This might be one of the reasons for the decrease of plating rate.

3.2. Membrane characterization

Fig. 2. The amount of deposited nickel as a function of electroless plating time.

by a gas permeation system established in our laboratory. The electronic resistivity is 25–26 V cm at room temperature for membranes plated for 70, 90, and 100 min.

3. Results and discussions

3.1. Plating rate As it is well known that an electroless plating is a complex chemical reaction, which is affected not only by the composition of plating bath, but also temperature, pH and catalytic reaction on the substrate surface. It should be noted that the substrates were activated by a sol – gel technique. Thus, the plating process is activated with Pd nuclear around alumina grain on the substrate. When the activated substrate was immersed in the electroless-plating bath, blisters could be observed on the surface of the substrate after a brief delay. This shows the reaction processes after the substrate is brought in contact with the plating bath. The amount of nickel plated is plotted in Fig. 2 as a function of plating time, from which it can be seen that the amount of plated nickel increases with plating time. Fig. 2 also shows that plating rate is much faster at the initial stage of plating than after 1-h plating. For a period of 100-min plating about 16.0 mg cm − 2 nickel is plated during the first 50 min, and in the next 50 min, the amount is only 43 mg cm − 2. The concentration

After 15 min of electroless nickel plating, the thickness reached an average 2.2 mm, although the coating is not uniform. The entire surface of the substrate is covered by nickel, which is formed by small crystals of dimensions around 0.2 mm, with low dense packing. As coating proceeds with time up to 100 min, the thickness of the deposit and the crystal size progressively increase, leading to a more compact coating. The SEM also shows (Fig. 3) that the nickel deposits are formed by very fine component crystallites. The fine crystallites are packed in such a way so as to give hemispheric grains. The grain size increases with plating time as shown in Fig. 3. The isotropic growth seems to indicate that the growth front is homogenous and the grains merge to form a continuous film. Nickel is also observed to be deposited on the face, which is not activated. No protection is made on this side during plating. However, it was hardly observed that nickel was deposited inside the substrates by investigating the cross-section of the substrate with SEM. Hydrogen permeation data were measured as a function of plating time to test whether the nickel membrane is porous. The transmembrane pressures were set at 640 and 500 mmHg with a vacuum pump. Fig. 4 shows the flux versus plating time. The flux data decrease with plating time. Flux through a porous membrane, according to Hagen–Poiseuille equation, can be described as, J=

Dr 2o 8Ltm

Where J is the flux; DP, the transmembrane pressure; r, the mean pore radius; o, the porosity of membrane; m, the dynamic viscosity of tested gas; t, the torturity of pore; and L, the thickness. The equation shows that flux is inversely proportional to the membrane thickness. It has been shown above that the amount of nickel increases

X. Changrong et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 179 (2001) 229–235

with plating time, and consequently, which indicates that the thickness of the nickel membrane increases with plating time, and consequently, flux decreases according to the equation. The membrane thickness, which is shown in Table 2, does make contribution to the flux decrease. Another factor, which affects flux, is the mean pore radius. The pore size and its distribution were measured by a modified bubble-point method [7] This method allows detecting the available pores, which are connected from one side of a membrane

233

Fig. 4. Dependence of nickel membrane flux on electroless plating time at transmembrane pressure of 640 and 500 mmHg.

to the other. The mean pore size of the membranes is listed in Table 2. The pore size distribution curves of the substrate and nickel membranes have almost the same characteristics. Fig. 5 shows the pore size distribution of the membrane plated for 90 min and the substrate. From Table 2, it can be seen that pore size decreases with plating time. Smaller the pore size, lower is the flux. As compared with the thickness increase of 2.3 mm, the pore radius decrease is very small, from 0.17 to 0.13 mm. Thus, the increase in membrane thickness is the major reason for the flux decrease.

3.3. Formation of porous nickel membranes

Fig. 3. SEM micrographs of the surface of nickel membranes plated for (a) 15; (b) 25; (c) 35; (d) 45 ; (e) 55; (f) 70 min. The bar corresponds to 5 mm.

The reason for the uncompleted covering of the pore with nickel even increasing the plating time may be due to the fact that the activation process of the alumina surface does not produce a uniform distribution of the palladium in the surface of the alumina substrate. The amount of palladium on the alumina external surface as well as on the wall of pores is important but it is very small in the pores. The inhomogeneous distribution of palladium is caused by the slip-casting process, which is applied to activate the substrate. In this process, the dry alumina support is brought into contact with the dip-solution. Capillary forces are present inside the support pores and water of the dip-solution is sucked into the support. At the boundary of the support and dip solution, a layer is formed by concentration of the boehmite particles on which surface Pd2 + is ad-

234

X. Changrong et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 179 (2001) 229–235

sorbed. So palladium is deposited preferentially at the edge of pores and not uniformly on the surface. As a consequence, nickel is preferentially plated at the edge of pores. In geometry, nickel can be deposited in two directions. One is radial deposition, which causes a reduction in the pore size. The other is vertical deposition resulting in the increase of membrane thickness. In the beginning, the rate of radial deposition might be comparable with that of vertical deposition. As shown in Table 2, in 15 min after the reaction is processed, the membrane thickness is 2.2 mm and the pore radius decreases 0.26 mm. It is reasonable that there is a point before which the rate of radial deposition is equal to vertical. However, further research needs to be done to confirm if there is such a point and how long it is. After the reaction is processed, the radial deposition may be suppressed since the solution composition in the pore should be lower than in the main bath. The reason is obvious,

nickel ion and other chemicals take more time to diffuse into the pores than onto the surface. As a result, vertical deposit rate is much faster than radical. This is coincident with the SEM investigation and pore size measurement. For an example, the membrane thickness increases 3.0 mm while the pore radius decreases only 0.30 mm, from 0.43 to 0.13 mm in 90-min plating. The formation of asymmetric nickel membrane on ceramic support is shown in Fig. 6. The suppression of radial deposition might be the other reason for the decrease of plating rate with time as shown in Fig. 2. One reason for not completely covered porosity is, as explained above, that the surface activation of the alumina is not homogeneous. The other may be that hydrogen is liberated during the process. There are many blisters on the surface, most probably caused by localized growth of the nickel deposit, which increases with plating time. Some blisters release from the surface several

Table 2 The effect of plating time on the thickness and pore radius Plating time (min) Membrane thickness (cm) Pore radius (mm)

15 2.2 0.17

25 2.4 0.15

45 3.0 0.14

55 4.0 0.14

70 4.4 0.13

90 4.5 0.13

100 4.5 0.13

Fig. 5. Pore area distribution vs. pore diameter for (a), porous alumina substrate; (b), nickel membrane electroless plated for 90 min.

X. Changrong et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 179 (2001) 229–235

235

Fig. 6. Schematic mechanism for electroless nickel membrane deposition (a), porous alumina support; (b), porous alumina substrate with palladium preferentially deposited on the surface; (c), at first, nickel is equally deposited inside the pore and on the surface; the pore radius decreases and membrane thickness increases at the same rate; (d), after that, nickel is deposited on the surface but rarely in the pore; pore size remains almost unchanged and the thickness increases.

minutes after the reaction is processed. Unfortunately, we cannot identify if there are blisters in the pore during plating. Assumption that there are blisters in them should be reasonable, since it is the same reaction on the wall of pores as at the surface. When a blister is formed in the pore, it is difficult to leave under the experimental conditions. The in-pore blister prevents the chemicals from diffusing into the pore. As a result, the pore cannot be covered fully by nickel particles.

4. Conclusion Nickel membranes with pore radius of 0.13 mm are prepared with electroless plating on porous alumina support with a pore radius of 0.43 mm. The membrane thickness is lower than 5 mm. No cracks or pinholes are observed in the membrane. It takes only 15 min to get a 2.2-mm-thick membrane. However, more than 90 min is needed to plate a 4.5-mm-thick membrane. The pore size decreases from 0.43 to 0.17 mm after 15-min plating. But it only decreases by 0.04 mm in the next 85 min. The preferential deposition of palladium on alumina support is one of the reasons for the formation of porous nickel membrane. The other is preferential nickel deposition on the substrate due to long-term diffusion of chemicals from bath to the pore and the formation of bristles in the pore. .

Acknowledgements This research work was supported by National Natural Science of China under Contract No. 29631020 and 29801003.

References [1] A.J. Burggraaf, L. Cot, Fundamentals of Inorganic Membrane Science and Technology, Elsevier, Amsterdam, 1996. [2] R. Bhave, Inorganic Membranes, Synthesis, Characterization and Properties, Van Rostrand Reinhold, New York, 1991. [3] A. Li, G. Xiong, J. Gu, L. Zheng, Preparation of Pd/ceramic composite membrane, 1. Improvement of the conventional preparation technique, J. Membr. Sci. 110 (1996) 257–260. [4] J. Shu, B.P.A. Grandjean, E. Ghali, S. Kaliaguine, Simultaneous deposition of Pa and Ag on porous stainless steel by electroless plating, 77 (1993) 181 – 195. [5] E. Kikuchi, S. Uemiya, Preparation of supported thin palladium – silver alloy membranes and their characteristics for hydrogen separation, Gas Sep. Purif. 5 (4) (1991) 261 – 266. [6] H. Yoshiki, K. Hashimoto, A. Fujishima, Reaction mechanism of electroless metal deposition using ZnO thin film (I): process of catalyst formation, J. Electrochem. Soc. 142 (2) (1995) 428 – 432. [7] P. Mikulasek, P. Dolecer, Characterization of ceramic membranes by active pore size distribution, Sep. Sci. Tech. 29 (9) (1994) 1183 – 1192. [8] X. Changrong, M. Guangyao, Y. Pinghua, P. Dingkun, Preparation and structure characterization of nanofiltration alumina membrane, Chin. J. Mater. Res. 12 (2) (1998) 135 – 138.