Chi Ma, Le Sang, Xiaonan Duan, Jiabin Yin, Jisong Zhang

State Key Laboratory of Chemical Engineering, Department of Chemical Engineering, Tsinghua University, Beijing 100084, China

Keywords:Foam Alumina Catalyst support Substrate pretreatment Adhesion mechanism Micropacked bed

A B S T R A C T Methods of coating Al2O3 on nickel micro-foam were compared and screened, aiming to overcome the capillary force and prepare the micro-foam monolithic catalyst coatings. The surface of micro-foam substrate was pretreated by a chemical etching method to improve the adhesion of the coatings on the substrate. The results showed that the slurry circulation at 162 ml·min-1 was evaluated as the optimal method. The pore size on the substrate surface can be controlled by changing the pretreatment conditions. An empirical correlation was also proposed, showing an excellent practicality for predicting the pore size.The adhesion of the coatings with substrate pretreatment was significantly better than that without substrate pretreatment. The minimum value of mass loss after ultrasonic vibration was 3.9%.This mainly attributes to the squeezing of Al2O3 particles in the pores of substrate surface. The coatings on nickel micro-foam are hopefully used in micropacked beds for catalytic reactions.

1. Introduction

Coatings of monolithic catalysts are significant for heterogeneous catalytic reactions [1-3], which can provide a large specific surface area and uniformly disperse the active components [4,5].Metal foam is regarded as a new class of structural material due to the large porosity,small bulk density,high mechanical strength and low pressure drop [6-8], which has shown considerable promise for the applications of the catalyst substrate[9,10].Therefore,it presents a new route for incorporating catalyst nanoparticles with metal foam to prepare monolithic catalyst coatings.

In recent years,coatings on metal foam used in gas-liquid-solid micropacked beds have attracted extensive attentions [11-14],mainly because micropacked beds have the small dimensions,large surface-to-volume ratio, high mass and heat transfer performance, and excellent operating safety [15-17]. By further loading active components to prepare metal foam catalysts, the micropacked beds can be applied in highly exothermic or mass transfer limited processes, such as hydrogenation [18], oxidation [19], catalyst screening [20,21] and so on. Xionget al.[22] investigated a monolithic catalyst of Pd/carbon nanotube/nickel foam for hydrodechlorination of chlorophenols in micropacked beds. Yuet al.[23] compared the properties of the Cu-Zn-Al-Zr/Al2O3catalysts on different metal foams for methanol steam reforming in micropacked beds. The above metal foam monolithic catalysts demonstrated the effective intensification of catalytic reactions.However,these catalysts had the poor adhesion between the coatings and the metal foam substrates.The adhesion of catalyst coatings is particularly important for the catalysts and the reactors.Low adhesion and mechanical stability lead to the shedding of the coatings and the blocking of the reactor, which can significantly increase the pressure drop and even damage the whole system [24].

Coating adhesion can be improved by substrate pretreatment,which can increase the surface roughness of substrates [25-27].The preparation of coatings on metal foam with substrate pretreatment has been studied by a few researchers.Liet al.[28]developed a Ni-Al2O3/nickel foam catalyst with chemical etching of Ni foam,showing a high stability in the reaction process. Mendezet al.[29] prepared a NiPd/(CeO2-Al2O3)/aluminium foam catalyst and the substrates were previously anodized to produce a rough surface,by which the adhesion between catalyst particles and metallic surface was improved. Eganaet al.[30] compared the adhesion of catalyst coatings on aluminum foam pretreated by chemical etching and FeCrAl foam pretreated by thermal treatment. The results showed that the adhesion on FeCrAl foam was better than that on aluminum foam. In the studies above, the adhesion between the coatings and the metal foams was significantly promoted by substrate pretreatment. However, there are few reports on controlling the pretreatment conditions to obtain a controllable rough surface and enhance the adhesion of coatings on this surface.Besides, the form of coatings and the adhesion mechanism between the coatings and the pretreated substrates are still unclear.

The pore size of metal foam used in micropacked beds is usually kept smaller than 500 μm for creating the microchannels and cavities.At such a small pore size of metal micro-foam,capillary force rather than gravitational force becomes the dominant force [31-34].The slurry of catalyst coatings is difficult to flow into the inner pore structures of metal foam,and the distribution of catalyst coatings is nonuniform when the traditional preparation method of sol-gel slurry impregnation is used[16].Therefore,it is extremely necessary to develop an efficient method for coating the slurry on the metal micro-foams, which can effectively overcome the capillary force and obtain the catalyst coatings with strong adhesion and good uniformity. These studies can provide basic data for the efficient design, manufacturing, and optimization of monolithic catalysts and further application in the micropacked beds.

This work aims to prepare the coatings on the metal microfoam surface with good properties. Al2O3powders were used for the coatings due to the high specific surface area, corrosion resistance,thermal and chemical stability[35,36].To overcome the capillary force effectively, different coating methods of static impregnation(SI), ultrasonic impregnation(UI)and slurry circulation(SC)by sol-gel were compared and screened.For the improvement of adhesion between the coatings and the metal foam substrate, the surface of substrate was pretreated by a chemical etching method. An empirical correlation was also proposed to predict the pore size on the surface of substrate.Moreover,characterization analyses of scanning electron microscopy and nitrogen adsorption-desorption were performed to evaluate the properties of coatings and reveal the adhesion mechanism.

2. Experimental

2.1. Substrate pretreatment

Commercial nickel foam (Kunshan Jiayisheng Electronic Co.,Ltd., China) with 130 pores per inch (PPI) was wire-electrode cut into cylindrical blocks with the appropriate size (the diameter of 3.8 mm and the height of 5.0 mm) for the packings in the micropacked beds.All the reagents were analytically pure and used as received.The nickel foams were ultrasonically cleaned in deionized water (self-prepared) for 30 min, and then in alcohol for 30 min to remove dirt and oil. The ultrasonic vibration was provided by an ultrasonic cleaner (KQ5200DE, Kunshan Ultrasonic Instrument Co., Ltd., China) with the frequency of 40 kHz and power of 200 W.

Subsequently, the nickel foams were pretreated by a chemical etching method with a mixed acid solution of HCl, HNO3(Beijing Lanyi Chemical Products, Co., Ltd., China) and deionized water.The total concentration of mixed acid and etching time were changed with the objective of obtaining different sizes of pores on the nickel foam substrate surface. In this work, HCl:HNO3:deionized water(volume ratio)=3:1:45,40,35,30,25,20,15,10 and the corresponding concentration of about 1.03,1.15,1.30,1.49,1.75,2.11,2.66, 3.62 mol·L-1were chosen with the etching time of 60 min and the etching temperature of 20°C.At the concentration of about 1.15 mol·L-1and temperature of 20 °C, the etching time was chosen as 15,30,60,90,120,150,180 min,and at about 2.66 mol·L-1and 20 °C, the etching time was chosen for 5, 15, 30, 45, 60, 75,90 min,respectively.After etching,the nickel foams were ultrasonically cleaned in deionized water for 30 min to remove the remaining acid solution.

2.2. Slurry coating

Commercial γ-Al2O3powders (Jiangsu Xianfeng Nanomaterials Technology Co.,Ltd.,China),polyvinyl alcohol(PVA,Beijing Anxinkang Technology Co.,Ltd.,China),HNO3and deionized water were employed for the preparation of slurry by a sol-gel method. The PVA and HNO3were applied for binder and stabilizer,respectively,and the concentrations of Al2O3, PVA, HNO3, and deionized water were 12.0%, 1.8%, 1.2%, and 85.0% (mass), respectively. The PVA and deionized water were magnetically stirred and mixed for 2 h at 85 °C to dissolve the PVA. The Al2O3powders and HNO3were then added into the PVA solution and continuously stirred for 2 h,and the slurry was obtained.After that, the slurry was kept at room temperature (25℃) for 24 h to make it stable.

Fig.6. SEM micrographs of substrates after pretreatment at 1.15 mol·L-1 for(a) 60 min,(c) 120 min,(e) 180 min and at 2.66 mol·L-1 for(b) 30 min,(d)60 min,(f)90 min.

The slurry was firstly coated on the surface of nickel foams without pretreatment by three coating methods of SI, UI and SC for comparison and screening, and the schematic diagrams of the slurry coating setups are shown in Fig. 1. For the methods of SI and UI, the nickel foams were deeply impregnated in the slurry without and with ultrasonic vibration for 30 min, respectively.The ultrasonic power was set as 80, 100, 120, 160, 200 W during the ultrasonic impregnation processes. For the method of SC, the nickel foams were packed right in a tube with the inner diameter of 3.8 mm. The slurry was pumped and circulated for 30 min by a peristaltic pump (BT00-300M, Baoding Lange Constant Flow Pump Co., Ltd., China), and the slurry flow rate of 18, 54, 90, 126,162,198 ml·min-1was chosen for the experiments.Then the nickel foams were retained in the tube with the circulation of no slurry(the air circulation) for 10 min. After coated by three methods,the nickel foams were dried in still air at room temperature for 24 h, at 120℃ for 6 h, and calcined in an oven at 600℃ for 2 h(ramping up at 5℃·min-1). The schematic diagram of the substrate pretreatment and slurry coating processes is shown in Fig.2.

Fig. 1. Schematic diagrams of the slurry coating setups by (a) SI, (b) UI, and (c) SC methods.

Fig. 2. Schematic diagram of substrate pretreatment and slurry coating processes.

Fig. 3. Effects of (a) ultrasonic power and (b) slurry flow rate on coating mass.

Fig. 4. SEM micrographs of (a) nickel foam and coatings by (b) SI, (c) UI, and (d) SC with low magnification.

At last, the nickel foams with pretreatment were coated by the optimal method with the optimal condition after screening. The difference in adhesion was investigated between the coatings on the nickel foams without and with pretreatment.

2.3. Characterization

The average particle size of the Al2O3powders was determined by a nano particle size and zeta potential analyzer (Delsa Nano C,Beckman Coulter Inc.,U.S.).The surface morphologies of the nickel foam substrates and the coatings were observed by a scanning electron microscope(SEM,Sirion200,FEI,Netherlands).The textural properties of the substrates and the substrates after coating were characterized at 77 K by a dynamic nitrogen adsorption apparatus (Autosorb-1-C, Quantachrome Instruments, U.S.) with the samples pretreated at 200℃ for 6 h, and the specific surface area was evaluated by the Brunauer-Emmett-Teller (BET) method.

The adhesion of the coatings on the nickel foam substrates was determined by an ultrasonic vibration method with the same instrument in Section 2.2. As a common method [23,29,37], the samples were impregnated in deionized water and submitted to an ultrasonic vibration treatment with a power of 200 W for 30 min. After that, the samples were dried at 120℃ for 6 h. Each experimental condition was repeated for three times and the average value of the mass loss of coatings was calculated by the mass difference of the samples before and after ultrasonic vibration.

3. Results and Discussion

3.1. Comparison of coating methods

3.1.1. Coating mass

According to the results by nano particle size analysis,the average particle size of Al2O3powders is 234 nm,which can be further confirmed by the SEM micrograph,as shown in Fig.S1,(see Supplementary material). The coating mass on the nickel foam substrate by SI was 0.41 g·(g nickel foam)-1, and the effects of ultrasonic power by UI and slurry flow rate by SC on the coating mass are shown in Fig. 3. The coating mass increased with the increase of ultrasonic power, and decreased with the increase of slurry flow rate. This is because the turbulence of slurry was more violent,and the slurry was easier to flow into the inner pore of the nickel foam with a higher ultrasonic power. At a larger slurry flow rate,the slurry flow had a higher pressure and more violent turbulence,but most of the slurry can flow out of the inner pores with the air circulation, resulting in a lower coating mass. Conversely, there were large amounts of slurry on the substrate by SI,and it is difficult for the slurry to leave the substrate with no external force because of the viscosity of the slurry.

3.1.2. Surface morphology

Fig. 4 displays the SEM micrographs of nickel foam and the coatings on the nickel foam substrates by SI, UI (with ultrasonic power of 160 W),and SC(at slurry flow rate of 162 ml·min-1)with low magnification(×100),in which the large range of coatings and the loading effects can be observed. The surface of substrate was flat and smooth, and turned to rough and irregular after coating.As shown in Fig. 4(b) and (c), parts of the pores of nickel foams were blocked by the coatings in the SI and UI processes, because the slurry cannot flow out of the pores easily when it was in the pores. These coatings may cause an increase in pressure drop. In contrast, the coatings by SC in Fig. 4(d) seem more uniform with no blocking in the pores, attributing to the flow and circulation of slurry, which was consistent with the results of coating mass.

Fig. 8. Comparison of average pore size between experimental and predicted values.

The SEM micrographs of nickel foam and coatings with high magnification(×5000)are shown in Fig.5.The coatings by SI were nonuniform with the blocky structures of Al2O3,resulting from the agglomeration of Al2O3particles during the coating process. From Fig.5(c),(e), (d),and(f), it is found that the uniformity of coatings can be improved by increasing the ultrasonic power and the slurry flow rate.This suggested that the agglomeration of Al2O3particles can be eliminated by ultrasonic impregnation and slurry circulation.Comparing Fig.5(e)with(g),and(f)with(h),the surface morphologies of coatings were similar, indicating that the improvement was little when the ultrasonic power was too high(200 W) and the slurry flow rate was too large (198 ml·min-1).In addition, there were a large number of pore structures on the coatings by SC in contrast to the coatings by UI. The reason is that the bubbles were generated and mixed in the slurry when the slurry was circulated through the pores of the nickel foams. These pore structures of coatings are beneficial to the enhancement of the specific surface area of coatings and the loading of catalyst active components in future.

Fig.5. SEM micrographs of(a)nickel foam and coatings by(b)SI,(c)UI with 80 W,(e)UI with 160 W,(g)UI with 200 W,(d)SC at 18 ml·min-1,(f)SC at 162 ml·min-1,and(h)SC at 198 ml·min-1 with high magnification.

Fig. 9. SEM micrographs of Al2O3 coatings after ultrasonic vibration (a) without substrate pretreatment and (b) with substrate pretreatment, and (c) before ultrasonic vibration with substrate pretreatment.

3.1.3. Textural property

According to the results of surface morphologies of coatings,the optimal conditions of the ultrasonic power and the slurry flow rate were screened as 160 W and 162 ml·min-1, respectively. The textural properties of the coated nickel foam substrates under these conditions were employed for the comparison.Table 1 lists the textural properties of the uncoated substrate and the coated substrates by SI, UI, and SC, as well as the corresponding coating mass. After coated with Al2O3, the specific surface area and the pore volume of substrate increased, and the pore size of substrate decreased by reason of the addition of Al2O3particles with porous structure and small pore size. The specific surface area and pore volume of the coated substrate by SI were smaller than those by UI and SC because of the agglomeration of Al2O3particles.Although the specific surface area and pore volume of the coated substrate by UI were the largest,the coating mass was much larger than that by SC, resulting in a lower utilization of coatings and blocking of substrate pores. The pore size of the coated substrate by SC was larger than that by SI and UI. This is because the new pore structures on the Al2O3coatings formed during the slurry circulation, which was consistent with the SEM micrographs. The coatings by SC can provide a relatively large specific surface area at a low coating mass.

Table 1 Textural properties and corresponding coating mass of uncoated substrate and coated substrates by SI, UI, and SC methods

In conclusion,the SC at 162 ml·min-1was evaluated as the optimal method and condition for coating Al2O3slurry on the nickel foam substrate, considering the uniformity, surface structure and specific surface area of the coatings. This method is effective to overcome the capillary force of small pores for the preparation of nickel micro-foam monolithic catalyst coatings used in the micropacked beds.

3.2. Characterization of pretreated substrate

3.2.1. Pore size on the substrate surface

To improve the adhesion between Al2O3coatings and nickel foam substrate,the substrate was pretreated by a mixed acid solution,which can increase the surface roughness of substrate.When the nickel foam substrate was impregnated in the mixed acid solution,a number of pores formed on the surface of the substrate.The surface morphologies of the substrates after pretreatment were observed by SEM.For determining the average pore size on the surface of the substrates,the average values of 50 pores from the continuous area on the SEM micrographs were calculated.Parts of the SEM micrographs were chosen as examples displayed in Fig.6,and the results of average pore size under different acid concentrations and etching time are shown in Fig. 7.

Fig.7. Effects of(a)acid concentration and(b)etching time on average pore size of substrate surface.

The acid concentrations for substrate pretreatment were about 1.03,1.15,1.30,1.49,1.75,2.11,2.66,3.62 mol·L-1,and the etching time were 15, 30, 60, 90, 120, 150, 180 min at about 1.15 mol·L-1and 5,15,30,45,60,75,90 min at about 2.66 mol·L-1,respectively.The average pore size on the substrate surface increased with the increase of acid concentration (from 1.15 to 2.66 mol·L-1) and etching time (from 30 to 150 min at 1.15 mol·L-1and from 15 to 75 at 2.66 mol·L-1). The textural properties of the pretreated substrates are shown in Table S1. The specific surface area, pore volume and pore size increased in the same trend as the average pore size. When the acid concentration was too low and the etching time was too short,there was no pore formed on the substrate surface.When the acid concentration was too high and the etching time was too long, the pores covered and overlapped with each other, and even cannot be observed, which may destroy the strength of the substrates, as shown in Fig. 6(e) and (f). Consequently, the average pore size under these conditions cannot be adequately measured and calculated.As shown in Fig. 7, the average pore size ranged from 242 to 742 nm.

Fig. 11. Schematic diagrams of Al2O3 coatings on the substrates (a) without pretreatment and with pretreatment at average pore size of (b) 269 nm, (c) 504 nm, and (d)742 nm.

3.2.2. Correlation of pore size

An empirical correlation of average pore size on the surface of substrate, considering acid concentration and etching time was proposed based on the data in Fig. 7, as shown in Eq. (1).

whereD,C,andtrefer to the average pore size on the substrate surface, acid concentration, and etching time, respectively. This correlation was applicable when the average pore size was between 242 and 742 nm.The average pore size was proportional to the acid concentration’s power of 0.69 and the etching time’s power of 0.60.Both acid concentration and etching time affected the average pore size.The predicted values agreed well with the experimental values,and the errors of predicted values were within ±10% of the experimental values,as shown in Fig.8.This indicated that the correlation has an excellent practicality to predict the average pore size on the surface of substrate, and the pores with specified size can be obtained and controlled by changing the pretreatment conditions.

Fig.12. SEM micrographs of Al2O3 thin layers on the substrates(a)(b)without pretreatment and with pretreatment at average pore size of(c)(d)269 nm,(e)(f)504 nm,and(g) (h) 742 nm with different magnifications.

3.3. Adhesion of coatings

3.3.1. Comparison of adhesion

The adhesion of the coatings on the substrate surface was evaluated by the ultrasonic vibration. The coatings on the substrate without pretreatment,which were prepared by SC at 162 ml·min-1in Fig. 5(f) were destroyed after ultrasonic vibration, as shown in Fig. 9(a). The mass loss of the coatings was 20.0%. To reduce the mass loss of the coatings, the slurry was coated on the substrates with pretreatment. The mass loss of the coatings on the substrate surface with different pore sizes after ultrasonic vibration were measured and are shown in Fig. 10. The mass loss of the coatings with substrate pretreatment ranged from 3.9% to 18.0%. The minimum value of mass loss was obtained at the average pore size of 504 nm on the substrate surface, and the corresponding acid concentration and etching time were 2.66 mol·L-1and 45 min,respectively. Fig. 9(b) shows the morphology of the coatings on the substrate with the average pore size of 504 nm after ultrasonic vibration, and the adhesion of these coatings was much better compared with that without pretreatment. This suggested that the nickel micro-foam substrate with an average pore size of 504 nm and Al2O3particles with an average particle size of 234 nm (in Section 3.1.1) adhered best and exhibited a good mechanical stability.

Fig.10. Mass loss of coatings on the substrate surface with different pore sizes after ultrasonic vibration.

The morphology of the coatings on the substrate with the average pore size of 504 nm before ultrasonic vibration is displayed in Fig. 9(c), which was similar with the coatings without substrate pretreatment. The coating mass was 0.08 g·(g nickel foam)-1, and the specific surface area, pore volume, and pore size were 48.77 m2·g-1, 0.20 cm3·g-1, 14.72 nm, respectively. The specific surface area was slightly higher,and the pore volume and pore size were lower than those of the coatings without substrate pretreatment.This is because the coating mass can be slightly enhanced by the substrate pretreatment,and the textural properties of coatings were consistent with the coating mass.The above results indicated that the substrate pretreatment had little effect on the morphologies and textural properties of the coatings. By substrate pretreatment, the adhesion of the coatings on the substrate can be significantly improved.

3.3.2. Adhesion mechanism

Fig. 11 presents the schematic diagrams for the speculation of the adhesion mechanism between the coatings and the substrates.As shown in Fig.11(a)and(b),there was no pore,or the pores were small and shallow on the surface of the substrate.The Al2O3particles were deposited on the surface of substrate, and the coatings had a weak adhesion because of the property difference between metal materials and Al2O3coatings.In contrast,when the substrate had the pores with appropriate size(the pore size/the Al2O3particle size of about 2.15) on the surface, the particles were packed in the pores and tightly squeezed with each other and together with the inner wall of the pores.The particles in the pores also adhered with the particles out of the pores,and a good adhesion of the coatings can be obtained,as shown in Fig.11(c).When the pores were large and deep,the inner wall of these pores was like a large curved surface. Although the Al2O3particles can access into the pores,there was almost no squeezing between the particles and between the particles and the wall, resulting in a decrease in the adhesion,as displayed in Fig. 11(d).

The substrate was covered by coatings, therefore, it is difficult to observe the adhesion form of the coatings on the substrate surface even after the ultrasonic vibration,as shown in Fig.9.To verify the above speculation of the adhesion mechanism, the same coating method as SC was employed with the Al2O3, PVA, HNO3, and deionized water concentrations of 0.01%, 2.13%, 1.36%, and 96.50% (mass), respectively. The Al2O3concentration of the slurry is low, and Al2O3thin layers can be obtained on the substrates using this slurry, as shown in Fig. 12(a), (c), (e) and (g). In these SEM micrographs with magnification of ×5000, the thin layers on the substrates without pretreatment and with pretreatment at average pore size of 269 nm, 504 nm and 742 nm, respectively,can be clearly observed.The corresponding SEM micrographs with further magnification of ×50,000 are shown in Fig. 12(b), (d), (f)and (h). It can be seen that the adhesion forms of the coatings on the substrates with these different pore sizes are quite consistent with the speculation of the adhesion mechanism in Fig. 11. The adhesion mechanism can also be referred and verified in literature[24].

These thin layers may be regarded as the underlayers of coatings connecting to the substrates. The form of thin layers on the substrates directly affects the adhesion of the whole coatings.Consequently, according to the particle size of the Al2O3, by screening the appropriate pore size on the substrate surface and calculating the corresponding pretreatment conditions using the equation,the adhesion of coatings can be improved and the mass loss of coatings after the ultrasonic vibration can be reduced, which are beneficial to the enhancement of reaction efficiency and operating safety in the micropacked beds.

4. Conclusions

In this work, three coating methods were compared and screened. The slurry circulation at 162 ml·min-1was evaluated as the optimal method and condition for coating Al2O3slurry on the nickel foam substrate. The substrate surface was pretreated by a chemical etching method, and the pore size on the substrate surface can be controlled by changing the pretreatment conditions.An empirical correlation was proposed to predict the pore size on the substrate surface, and the predicted values agreed well with the experimental results(at the errors within±10%).The adhesion of the coatings can be significantly improved by substrate pretreatment.The minimum value of mass loss of coatings after ultrasonic vibration was 3.9%,and the average pore size on the substrate surface was 504 nm.When the substrate had the pores with appropriate size (the pore size/the Al2O3particle size of about 2.15), the Al2O3particles can be packed in the pores and tightly squeezed to provide a good adhesion and mechanical stability. By the substrate pretreatment and the slurry circulation method, the monolithic catalyst coatings can be effectively prepared and hopefully used in the micropacked beds.

Acknowledgements

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We gratefully acknowledge the supports of National Key Research and Development Program of China (2019YFA0905100),and National Natural Science Foundation of China (21978146,21991103, 22008138) on this work.

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.05.022.

Nomenclature

Cacid concentration, mol·L-1

Daverage pore size on the substrate surface, nm

tetching time, min