A performance study on the electrocoating process with a CuZnAl nanocatalyst for a methanol steam reformer: the effect of time and voltage

Abstract

Coated microreactors provide an alternative strategy to enhance the performance of desired objects especially portable devices and in fuel cell applications. To successfully employ a catalyst in microreactors, study of coating methods is essential. In this paper, a high potential electrophoretic deposition technique was used to coat a CuZnAl catalyst on stainless steel. The prime aim of this examination is to determine the effects of the time and voltage of the coating on the catalyst performance. First, the CuZnAl catalyst was synthesized by a co-precipitation method to catalyze the methanol steam reforming. The powder was characterized by X-ray diffraction, ICP-OES and BET analyses. Second, based on the deposition time, methanol conversion was improved by an increase in the time of the coating to 4 minutes followed by a bit of a fall of both the conversion and deposition weight for a time of 5 minutes. Third, it was found that an increase in voltage from 60 to 140 V enhanced conversion while no deposition was achieved for higher voltages. The optimum conditions of 140 V and 4 minutes resulted in nearly full methanol conversion and a CO selectivity of 4.91% at 270 °C. Finally, the microreactor possessed a higher activity than that of a packed bed one. The coated layers were analyzed by field-emission scanning electron microscopy.

---

1 Introduction

Nowadays, industrial applications of hydrogen including oil desulphurization, methanol, ammonia synthesis, the water gas shift reaction, and hydrogenation of chemical agents highlight its situation as a pivotal agent.^1,2 In regards to reforming as a critical approach of hydrogen production, liquid energy carriers like methanol, dimethyl ether, and gasoline are appropriate candidates for on-board applications.^3,4 Methanol is broadly used for hydrogenous fuel processors especially in mobile applications because it can be reformed through a Cu-based catalyst in the range of 200–300 °C.⁵ Methanol reforming has several advantages such as easy storage, transport, and conversion with a microreformer. Moreover, the attractive and promising results of methanol steam reforming (MSR) give special insight into this reaction. It is occurs through the following reactions:

CH₃OH + H₂O → CO₂ + 3H₂, ΔH° 298 K = +49.7 kJ mol⁻¹ (1)

CO + H₂O → CO₂ + H₂, ΔH° 298 K = −41.2 kJ mol⁻¹ (2)

CH₃OH → CO + 2H₂, ΔH° 298 K = +90.7 kJ mol⁻¹ (3)

where the reactions (1), (2) and (3) represent MSR, the water gas shift (WGS) and decomposition of methanol (DM), respectively.⁶ To achieve more effective hydrocarbon reforming, optimization of all parameters attributed to both the catalyst and reactor design is crucial due to their influential physicochemical properties. Coated metal-based microreactors represent significant advantages such as a higher surface to volume ratio, smaller mean distance of the specific fluid volume to the reactor walls, and better heat and mass transfer that facilitate integration with fuel processor systems.⁷ In this way, the potential to coat thin layers of reforming catalysts onto the metal microreactors is a crucial factor for this field. Typically, the sol–gel method, electrophoretic deposition (EPD), electrochemical deposition, and electroless plating are employed for liquid phase coating, whereas chemical and physical vapor deposition, and also plasma spraying are used for gas phase coating.⁸ In this regard, liquid phase methods are preferred due to the capacity of the previously prepared powder coatings.

EPD is broadly employed to form a favorable ceramic layer,^9–11 catalyst layer,^12,13 solid oxide fuel cell membrane,^14,15 or even a porous membrane.^16,17 It is a well-known technique that is simultaneously performed by the two processes of charged particle migration in suspension contained by an electric field between two electrodes (electrophoresis), and particle deposition onto an electrode (electrocoagulation).¹³ Typically, EPD represents significant advantages such as being a simple procedure, having low cost equipment, a short deposition time, high controllable thickness of the coating, and co-deposition of various materials.^18–20 EPD generally consists of two sets of factors: (i) those based on suspension parameters, and (ii) those related to the physical parameters during the procedure such as voltage, time, etc.¹³

In this study we examined the cathodic EPD of a CuZnAl catalyst on stainless steel to produce hydrogen from the methanol steam reforming reaction. The physical properties of voltage and time of the EPD procedure were evaluated to find the methanol conversion dependency on the coating parameters. The coated plates were successfully characterized and reacted to determine the optimum case.

2 Experimental

2.1 Reactor design

A flat-plate stainless steel (SUS304) microreactor was designed. The microstructures (L = 108, W = 1, D = 0.3 mm) were machined by computerized numerical control (CNC). Both covering plates were machined to increase the surface to volume ratio, and to direct the flow path on the catalyst layer. The coated plates were sandwiched between the covering plates, and the housing structure was finally sealed by a graphite gasket and bolts. The plate and reformer design are shown in Fig. 1.

Fig. 1 The micro-channel reactor configuration.

2.2 Catalyst preparation

The CuZnAl catalyst was prepared by the co-precipitation method. First, an aqueous solution of cupric nitrate trihydrate (Cu(NO₃)₂·3H₂O), zinc nitrate tetrahydrate Zn(NO₃)₂·4H₂O, and aluminium nitrate nonahydrate Al(NO₃)₃·9H₂O with a molarity of 0.5 M was prepared. The prepared solution was heated up to 60 °C. In the next step, the Na₂CO₃ solution (0.2 M) was added dropwise to adjust the pH to around 7. After precipitation, the slurry temperature was fixed at 60 °C for a certain time of 3 h under continuous stirring. After that, the cooled mixture was filtered and washed with warm DI water to remove the ions. The prepared cake was dried at 110 °C for 18 h and calcined at 350 °C for 4 h in air atmosphere with a heating rate of 3 °C min⁻¹.²¹ The chemical composition of the synthesized catalyst was determined by inductively coupled plasma-optical emission spectrometry (ICP-OES, ARL, model 3410 apparatus) after dissolution of the catalyst in hydrochloric acid, and is given in Table 1.

Table 1 Structural properties of the CuZnAl catalyst in the powder and coated layer state

Catalyst	CuZnAl	CuZnAl	CuZnAl
a Scherrer’s equation was used.
Sample	Powder	Coated fresh	Coated used
Chemical composition (ICP-OES) (wt%)	38.69Cu, 51.47Zn, 9.83Al	—	—
Surface area (m^2 g^−1)	121	—	—
Pore volume (cm^3 g^−1)	0.488	—	—
Pore size (nm)	16.1	—	—
Crystallite size^a (nm)	8.69	6.15	7.39
Lattice parameters (Å)	a = 3.2452	a = 3.2348	a = 3.2384
	b = 3.2452	b = 3.2348	b = 3.2384
	c = 5.2951	c = 3.0713	c = 3.0749

---

2.3 Slurry preparation

First, the synthesized catalyst was milled for 10 minutes (due to the low initial particle size, a short time of milling was selected) at a speed of 250 rpm and then was sieved via a fine screen mesh (no. 400), for a particle size of less than 38 μm. It effectively improves the suspension stability. Each suspension included isopropanol (organic medium), the nanocatalyst of CuZnAl (powder content), and an additive agent of polyethylenimine (PEI). It contained 75 g L⁻¹ of nanocatalyst in isopropanol, and 0.6 wt% of the additive agent. PEI was used as a polymeric agent to create strict bonding and to increase the stability, control of deposition, and drying rate. The procedure was completed by 15 minutes of strong stirring, and then with 15 minutes of sonication in an ultrasonic bath (1200M-Soltec). Several physical–chemical properties of the slurry and the coating process were assessed to obtain optimal conditions of a desirable case. All the cases are illustrated in Table 2.

Table 2 Slurry composition for various EPD coatings of the CuZnAl catalyst

Catalyst	EPD			Slurry			Coated layer
	Voltage (V)	Time (minutes)	Electrode distance (mm)	Powder (%)	Binder (%)	Solvent (cm^3)	Deposition (mg)
CuZnAl	100	2	15	3	0.6	40	21
CuZnAl	100	4	15	3	0.6	40	37
CuZnAl	100	5	15	3	0.6	40	37
CuZnAl	140	4	15	3	0.6	40	48
CuZnAl	60	4	15	3	0.6	40	18
CuZnAl	180	4	15	3	0.6	40	0

---

2.4 Substrate treatment

Stainless steel (SUS304) plates were cut to 32 (L) and 22 (W) mm in dimension, with an exposed area of 7.04 cm². Prior to the deposition, the substrate plates were mechanically polished from 800 to 1200 grit emery papers. They were then washed with a rinsing agent and DI water followed by ultrasonic treatment with acetone for 5 minutes. Consequently, they were dried at room temperature followed by being kept in a desiccator.

2.5 Electrophoretic deposition methodology

The coating was made using a constant voltage power supply unit (SPS-900NP-Navasanpardaz). The electric current was also recorded using a digital multimeter (Haoyue, China). Electrophoretic deposition was conducted in a 120 mL beaker with a conventional three-electrode cell system. The electrodes were mounted at a distance of 15 mm. The stainless plates with the same cathode size were considered for the anode as well. An overview of the coating procedure and the EPD setup are shown in Fig. 2.

Fig. 2 Preparation steps of the electrophoretic CuZnAl slurry and EPD setup.

2.6 Catalytic layer characterization

The crystalline catalyst structure was characterized by XRD using a diffractometer (PANalytical X’pert-Pro) and Cu Kα monochromatized radiation as the X-ray source with a Ni filter in the 2θ range of 10–80°. The crystallite sizes could be calculated from the widths of the X-ray diffraction peaks using the Debye–Scherrer equation:where D is the crystal size of the powder, λ is the X-ray wavelength (0.15405 nm), β is the full width at half maximum (FWHM) of the diffraction peak (radian), and θ is the diffraction angle at the peak maximum. The crystalline phases were recognized according to the joint committee on powder diffraction standards (JCPDS) database. The catalyst layer surface morphology was analyzed by field-emission scanning electron microscopy (FESEM, MIRA3 TESCAN). An accelerating voltage of 15 kV and a working distance of 25 mm were used. A gold coating was applied to the specimens to avoid ionization under the electron beam in FESEM. The Brunauer–Emmett–Teller (BET) method was used to measure the surface area of the prepared catalysts. The lattice parameters (a, b and c) were calculated from the diffraction pattern at around 2θ = 32° and 44° through the relating expression for a hexagonal structure.

2.7 Methanol steam reforming

MSR over the CuZnAl catalyst was conducted in a flat-plate microreactor, as shown in Fig. 1. The microreactor was located in an electrical heater to supply the reaction heat. Besides, a quartz tube-shaped reactor (16 mm outer diameter) was used to compare the microreactor performance. It was also placed into a cylindrical furnace. A k-type thermocouple was attached inside the covering plate to control the reaction temperature. Prior to MSR, catalyst activation was performed as follows: the catalyst-coated plates were reduced at 300 °C for 4 h under H₂ atmosphere with a flow rate of 30 mL min⁻¹ to reduce copper oxide into Cu⁰.²² In the case of reforming, a methanol–water mixture with a ratio of 1 : 1.3 was fed to the evaporator section, containing a tubular segment filled with inert alumina granules with a temperature of 150 °C, and then supplied into the reaction section by a syringe pump (Japan, jp500). A cold trap was located after the output to condense and separate the residual methanol and water. Gas chromatography (GC, Shimadzu 8A) equipped with a thermal conductivity detector and a carbosieve column was used to analyze the reformed compositions. Moreover, the output flow rate was measured by a soap bubble flow meter. In all experiments, depending on the deposition weight, flow rates of 0.1 to 0.6 mL h⁻¹ and 7 to 20 mL min⁻¹ were considered for the pump and Ar, respectively. It is required to make an equal gas hourly velocity (GHSV) of 23 000 h⁻¹ and a ratio of argon/pump = 33.33 for all cases. In this paper, the range of 230–270 °C was selected as the operating temperature. The methanol conversion (X_MeOH) and yield of hydrogen (Y_H2) were calculated by the following equations, respectively:where F (mol s⁻¹) is the normal flow rate of the effluent gas. ϑ_H2 and ϑ_MeOH,in denote the molar flow rate of H₂ obtained in the output and the methanol molar flow rate fed into the microreactor, respectively.

3 Results and discussion

3.1 XRD analysis

Fig. 3 shows the XRD patterns of the powder and coated layer in different states at 2θ = 10–70°. A strong peak at 2θ = 36.15° and a low intensity peak at 2θ = 38.76° have appeared which represent CuO (the monoclinic phase of CuO, JCPDS 01-080-1268). The mentioned peaks include the (−1, 1, 1) and (1, 1, 1) structures, respectively. Moreover, the peaks at 2θ = 31.81, 34.64, 36.15, and 56.76° are related to ZnO (the hexagonal phase of ZnO, JCPDS 01-079-0207). The main peak of ZnO at 2θ = 36.15° overlaps with the monoclinic phase of CuO. No strong peak attributed to CuO was separately detected, although a high Cu loading had been selected. It could be related to the microcrystalline structure or coverage by the strong and wide peaks of ZnO. In the case of Al₂O₃, no considerable peak was found. It might be because of the low crystallinity or low temperature of calcination. The amorphous phase of Al₂O₃ causes a higher BET surface area and greater catalyst stability against Cu particle sintering. A temperature of 350 °C was selected as both the calcination temperature of the synthesis and coating.²³ According to the results in Table 1, it could be observed that crystallinity is present for the used case. The growth of ZnO crystallinity might be attributed to the structural changes that will affect the interaction between copper and zinc oxide.²⁴ Moreover, in the case of Cu sintering, the tendency depends on the interaction with the ZnO matrix and the Cu/Zn ratio.²⁵ Moreover, for the coated cases, the patterns show additional peaks of stainless steel 304 which mainly include Fe and Fe–Ni at 2θ = 44.35 and 64.52°, and 2θ = 43.49 and 74.54°, respectively.

Fig. 3 XRD patterns of the coated layers of the CuZnAl catalyst at 140 V, 4 minutes and 0.6 wt% PEI.

3.2 Time of EPD

Fig. 4 shows the top-view FESEM images of the coated layer for different times of EPD (2, 4 and 5 minutes). The results display surface morphology variation with the EPD time. Although no high differences are shown, side effects like crack formation influence the final efficiency, and it is required to investigate the effect of the time. An almost homogenous layer containing low intensity cracks was observed for a time of 2 minutes. The electric field is decreased during the time of deposition due to the increase in electric resistivity of the layer. Therefore, smaller particles could deposit when the shear stress is decreased at a long time of EPD. Local agglomeration on the surface might contribute to the smaller particles, as shown in Fig. 4. In the case of crack formation, an increase in the time could raise the layer thickness, which reasonably affects crack formation. An increase in the resistivity of the layer by growth of the thickness influences the rate of evaporation and drying.²⁶

Fig. 4 FESEM images and reconstructed 3D view of the coated layers for different times of EPD.

Fig. 5 shows the weight of the coated layers against the time of EPD within the range of 2 to 5 minutes. At the initial time of deposition, both the methanol conversion and weight distribution are proportional to the deposition time. As it continues to 4 minutes, a linear appearance is reached, while further time of EPD leads a slight fall of both the conversion and deposition weight. Fig. 6 shows that raising the temperature results in higher methanol conversion. The conversion reached 98.8% for a time of 4 minutes at 270 °C. It is increased from 83.5 to 98.8% at 270 °C, as the deposition time is differed from 2 to 4 minutes. Besides, a packed bed reactor was also compared to the coated cases and its details are illustrated in Table 3. It could be found that the weight of deposition broadly influences the conversion. More active sites are reached with higher deposition. However, after a time of 4 minutes, a decrease in the conversion is revealed. In the case of CO formation, a suggested mechanism for CO formation is a consecutive production by a reverse water gas shift reaction. It is obvious that the reverse water gas shift is more active at a higher temperature.^27–29 The CO concentration depends on the contact time in this mechanism.³⁰ The higher methanol conversion leads to a greater CO content in the output.³¹ The CO selectivity could be indirectly linked to the time of EPD through this investigation.

Fig. 5 Weight deposition and methanol conversion vs. time and voltage.

Fig. 6 Microreactor performance vs. time of EPD. S/C = 1.3 and GHSV = 23 000 h⁻¹.

Table 3 Comparison of packed bed and coated configuration for the MSR reaction

Parameter	Microreactor	Fixed bed reactor
Reaction temperature (°C)	270	270
Ar/pump	33.33	33.33
S/C	1.3	1.3
GHSV (h^−1)	22 000	22 000
MeOH conversion (%)	100	86
H2 yield (%)	99.9	99.9
CO2 selectivity (%)	95.08	96.59
CO selectivity (%)	4.91	3.40
Catalyst loading (mg)	96	100
Particle size (μm)	Less than 37	250–420

---

Fig. 7 demonstrates the current density against time of deposition. The electrophoretic current is crucial for determination of the deposition rate.³² The plot is correspondingly decreased during the time of deposition which indicates the rate of deposition loss. This is strongly related to the resistivity of the formed layer.

Fig. 7 Current density and weight distribution vs. deposition time under 100 V.

3.3 Effect of voltage

Fig. 5 displays the deposition weight as a function of applied voltage. It is noticeable that the increase in applied voltage up to 140 V raises the weight of deposition. However, the weight of deposition suddenly fell at voltages higher than 140 V for the time of 4 minutes, so that no deposition occurred. This evaluation could be described by the following equation. It expresses that a higher voltage intensifies the average velocity of the CuZnAl particles:³²

V = μ_eE (8)

where V expresses the particle velocity (m s⁻¹), μ_e the electrophoretic mobility (m² s⁻¹ v⁻¹) and E the applied electric field (V m⁻¹). According to the equation, the particles move fast so that a regular arrangement could not form. Fig. 8 shows the surface of the CuZnAl layer with different voltages in the range of 60 to 180 V for 4 minutes. The level of agglomeration is raised with an increase in voltage. More inhomogeneity is reached with a higher particle velocity during deposition.¹⁹ As shown in Fig. 9, the methanol conversion is raised with an increase in voltage from 60 to 140 V. It is clear that a higher voltage effectively results in a greater weight of deposition in the range of 60 to 140 V. Full methanol conversion was achieved at 270 °C for 140 V. Besides the weight of deposition, both the porosity and cracking could result in a higher conversion due to the developing effective diffusion factor.³³ In contrast, the performance of 60 V and the packed bed resulted in a lower conversion of 92 and 86%, respectively, at 270 °C. In the case of CO selectivity, 60 V and the packed bed represented 4.06 and 3.4%, respectively, at 270 °C. The H₂ yield ranged from about 66 to 99.9% and it is increased as the applied voltage is raised. Finally, with all voltages 99.9% was reached at 270 °C.

Fig. 8 FESEM images and reconstructed 3D view of the coated layers for different voltages of EPD for 4 minutes.

Fig. 9 Microreactor performance vs. voltage of EPD. S/C = 1.3 and GHSV = 23 000 h⁻¹.

Fig. 10 shows the current density and weight of deposition for different voltages during 4 minutes. As expected, the current density is increased from 60 to 140 V due to the raising of the electrical field. Moreover, the slope of the current density is raised with an increase in the voltage which represents the force to the particles during the time.

Fig. 10 Current density and weight distribution vs. voltage for 4 minutes.

3.4 Effect of GHSV and S/C

The effect of gas the hourly space velocity (GHSV) for MSR over the catalytic coated plates is shown in Fig. 11. The conditions of a molar ratio of steam to methanol of 1.3 : 1 (S/C), and a temperature of 270 °C, were considered. The GHSV was varied from 12 000 to 36 000 h⁻¹. As shown in Fig. 11, the methanol conversion decreased with an increase in the GHSV, which is attributed to a decrease in the residence time of reactants over the catalytic bed which consequently lowers the methanol conversion. The conversion is varied from 82.63 and 100% in the considered range. However, there is no significant change for the H₂ and CO₂ fractions but it strongly varies for CO. It is obvious that an increase in the GHSV slightly lowers the H₂ yield from 99.9 to 97.73%, while the GHSV is differed from 12 000 to 36 000 h⁻¹. The effect of S/C over the coated CuZnAl catalyst was investigated. As shown in Fig. 12, the methanol conversion significantly increases with a rise in S/C below 1.3, at 270 °C. In the case of the fractions, CO is decreased with an increase in S/C. The volumetric fraction of CO is about 3.14% when S/C is 0.7, and then is decreased to 0.64% with a rising S/C to 2. According to the results, a higher S/C is more favorable to reduce the CO content. Eqn (1) shows that a S/C of 1.0 will be stoichiometrically optimal for MSR. However, according to eqn (1) and (3), excess steam develops methanol conversion and decreases the CO content through shifting the water gas shift equilibrium to the right. Moreover, the system will be more cost-effective with less steam due to heating problems.

Fig. 11 Microreactor performance for 140 V and a time of 4 minutes vs. GHSV (h⁻¹).

Fig. 12 Microreactor performance for 140 V and a time of 4 minutes vs. steam to methanol molar ratio.

4 Conclusions

In regards to catalyst coating onto metal-based microreactors, further study takes into account alternatives to improve the procedure to increase efficiency. In an overview of this study, the interesting method of EPD was successfully performed to coat a CuZnAl layer onto stainless steel for methanol steam reforming. The optimal conditions of the coating procedure were achieved by variation of the physical-based parameters of time and voltage. The results showed that a higher time of EPD gave more deposition to a time of 4 minutes and then the methanol conversion was decreased with an increase in the time of deposition. At the next step, regarding a time of 4 minutes as optimal, the variation of voltage (60–180 V) was evaluated. The results showed that an increase in voltage to 140 V caused more deposition weight and roughness of the surface. After that, no deposition was found with voltages higher than 140 V. Full methanol conversion was practically achieved at 140 V and a time of 4 minutes. Coated and packed bed types were compared, of which the performance of the microreactor was better than the quartz fixed bed.

List of symbols

Subscripts

CO	Carbon monoxide
CO2	Carbon dioxide
cat	Catalyst
e	Electrophoretic
hlk	Miller index
H2	Hydrogen
in	Inlet
MeOH	Methanol
out	Outlet

Abbreviations

BET	Brunauer–Emmett–Teller
cm	Centimeters
DI	Deionized
FESEM	Field-emission scanning electron microscopy
GHSV	Gas hourly space velocity
h	Hours
JCPDS	Joint committee on powder diffraction standards
kJ	Kilojoules
L	Length
M	Molarity
mL	Milliliters
min	Minutes
mm	Millimeters
nm	Nanometers
rpm	Revolutions per minute
rWGS	Reverse water gas shift
V	Voltage
W	Width
Wt	Weight

References

1. S. Hočevar and W. Summers, in, Hydrogen Technology ed. A. Léon, Springer, Berlin Heidelberg, 2008, pp. 15–79.
2. F. Meshkani and M. Rezaei, RSC Adv., 2015, 5, 9955–9964.
3. C. Song, N. Brandon and D. Thompsett, Fuel Cells Compendium, Elsevier, Oxford, 2005.
4. C. C. Shen, T. Y. Jian and Y. T. Wang, Fuel Cells, 2013, 13, 965–970.
5. A. Qi, B. Peppley and K. Karan, Fuel Process. Technol., 2007, 88, 3–22.
6. B. A. Peppley, J. C. Amphlett, L. M. Kearns and R. F. Mann, Appl. Catal., A, 1999, 179, 21–29.
7. V. Hessel, S. Hardt and H. Löwe, Chemical micro process engineering: fundamentals, modelling and reactions, John Wiley & Sons, 2006.
8. V. Meille, Appl. Catal., A, 2006, 315, 1–17.
9. H. Hadraba, J. Klimes and K. Maca, J. Mater. Sci., 2007, 42, 6404–6411.
10. H. Hadraba, D. Drdlik, Z. Chlup, K. Maca, I. Dlouhy and J. Cihlar, J. Eur. Ceram. Soc., 2012, 32, 2053–2056.
11. H. Akbulut, G. Hatipoglu, H. Algul, M. Tokur, M. Kartal, M. Uysal and T. Cetinkaya, Surf. Coat. Technol., 2015, 284, 344–352.
12. R. Nedyalkova, C. Torras, J. Salvadó and D. Montané, Fuel Process. Technol., 2010, 91, 1040–1048.
13. R. Nedyalkova, A. Casanovas, J. Llorca and D. Montané, Int. J. Hydrogen Energy, 2009, 34, 2591–2599.
14. S. Majhi, S. Behura, S. Bhattacharjee, B. Singh, T. Chongdar, N. Gokhale and L. Besra, Int. J. Hydrogen Energy, 2011, 36, 14930–14935.
15. J. Cherng, C. Wu, F. Yu and T. Yeh, J. Power Sources, 2013, 232, 353–356.
16. C. Felix, T.-C. Jao, S. Pasupathi, V. M. Linkov and B. G. Pollet, J. Power Sources, 2014, 258, 238–245.
17. Q. Chen, U. P. de Larraya, N. Garmendia, M. Lasheras-Zubiate, L. Cordero-Arias, S. Virtanen and A. R. Boccaccini, Colloids Surf., B, 2014, 118, 41–48.
18. A. Boccaccini, S. Keim, R. Ma, Y. Li and I. Zhitomirsky, J. R. Soc., Interface, 2010, 7, S581–S613.
19. L. Besra and M. Liu, Prog. Mater. Sci., 2007, 52, 1–61.
20. O. O. Van der Biest and L. J. Vandeperre, Annu. Rev. Mater. Sci., 1999, 29, 327–352.
21. C.-C. Chang, C.-T. Chang, S.-J. Chiang, B.-J. Liaw and Y.-Z. Chen, Int. J. Hydrogen Energy, 2010, 35, 7675–7683.
22. S. R. G. Carrazán, R. Wojcieszak, R. M. Blanco, C. Mateos-Pedrero and P. Ruiz, Appl. Catal., B, 2015, 168, 14–24.
23. K.-S. Lin, S. Chowdhury, H.-P. Yeh, W.-T. Hong and C.-T. Yeh, Catal. Today, 2011, 164, 251–256.
24. Y. Matsumura and H. Ishibe, Appl. Catal., B, 2009, 91, 524–532.
25. M. M. Günter, T. Ressler, R. E. Jentoft and B. Bems, J. Catal., 2001, 203, 133–149.
26. K. B. Singh and M. S. Tirumkudulu, Phys. Rev. Lett., 2007, 98, 218302.
27. J. Agrell, H. Birgersson and M. Boutonnet, J. Power Sources, 2002, 106, 249–257.
28. J. K. Lee, J. B. Ko and D. H. Kim, Appl. Catal., A, 2004, 278, 25–35.
29. H. Purnama, T. Ressler, R. E. Jentoft, H. Soerijanto, R. Schlögl and R. Schomäcker, Appl. Catal., A, 2004, 259, 83–94.
30. X. Yu, S.-T. Tu, Z. Wang and Y. Qi, J. Power Sources, 2005, 150, 57–66.
31. R.-Y. Chein, Y.-C. Chen, Y.-S. Lin and J. Chung, Int. J. Therm. Sci., 2011, 50, 1253–1262.
32. L. Yang, X. Wu and D. Weng, Colloids Surf., A, 2006, 287, 16–23.
33. V. Novák, P. Kočí, T. Gregor, J.-S. Choi, F. Štěpánek and M. Marek, Catal. Today, 2013, 216, 142–149.

---