Zeolite incorporation in chip-based microreactors

Zeolite incorporation in chip-based microreactors

Microporous and Mesoporous Materials 226 (2016) 424e432 Contents lists available at ScienceDirect Microporous and Mesoporous Materials journal homep...

2MB Sizes 0 Downloads 44 Views

Microporous and Mesoporous Materials 226 (2016) 424e432

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso

Zeolite incorporation in chip-based microreactors L.A. Truter a, V. Ordomsky b, J.C. Schouten a, *, T.A. Nijhuis a a b

Laboratory of Chemical Reactor Engineering, Eindhoven University of Technology, P.O. Box 513, 5600 MB, Eindhoven, The Netherlands ^t. C3, Universit Unit e de Catalyse et de Chimie du Solide, UMR 8181 CNRS, Ba e Lille, ENSCL, Ecole Centrale de Lille, 59655, Villeneuve d'Ascq, France

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 December 2015 Received in revised form 2 February 2016 Accepted 3 February 2016 Available online 13 February 2016

A method for the incorporation of zeolite coatings in glass chip-based microreactors has been developed. The use of a fluoride-based hydrothermal pretreatment method provides a microchannel surface with high roughness, good wettability, and a silica nutrient source in a single pretreatment step. This pretreated glass surface is ideal for the subsequent in-situ zeolite hydrothermal synthesis which enables the formation of zeolite coatings within the chip-based microreactor. The optimization of the zeolite precursor suspension in terms of aluminium and silica source and amount, allows for the formation of zeolite coatings which are uniform, have high surface area and high Brønsted acidity. © 2016 Elsevier Inc. All rights reserved.

Keywords: Microreactor Lab-on-a-chip Zeolite Coating

1. Introduction The application of microreactors for use in the fine chemical and pharmaceutical industries has received much attention in recent years [1e4]. The ability to produce these high-value chemicals onsite and on-demand through continuous processing gives microreactor technology a specific advantage in comparison to conventional batch-wise production of such chemicals [5e7]. Although borosilicate glass is an ideal construction material for microreactors for use in the fine chemical and pharmaceutical industries, it has been far less applied in comparison to stainless steel. Borosilicate glass microreactors have the advantage of being compact, chemically inert, corrosion resistant, and operable at high temperature and pressure [8]. However, the ability to incorporate a catalyst into such devices is challenging mainly due to the difficulty to activate the glass surface to allow for a suitably adherent coating and for the bonding between the two microchannel glass plates to not be affected by the pretreatment and coating processes [9,10]. Zeolites are advantageous materials for catalyst incorporation in microreactors since the catalyst coating can be grown onto the microchannel surface resulting in improved coating uniformity and adherence [11e17]. Currently, most methods which incorporate zeolites in glass microreactors use multiple surface pretreatments to improve the glass roughness, hydrophilicity and zeolite coverage [12,18,19]. In order to carry out these surface pretreatments, the

* Corresponding author. Tel.: þ31 40 2472850; fax: þ31 40 2446653. E-mail address: [email protected] (J.C. Schouten). http://dx.doi.org/10.1016/j.micromeso.2016.02.016 1387-1811/© 2016 Elsevier Inc. All rights reserved.

glass plate must first be pretreated, followed by the growth of the zeolite coating on the microchannel surface, and, thereafter, assembled together to form the microreactor. However, the more difficult bonding of the glass plates together after these pretreatment and coating procedures greatly affects the sealing which ultimately limits the temperature and pressure ratings of such devices. An alternative approach is to initially form the closedmicrochannel (i.e. chip microreactor) by bonding together the two glass plates and, thereafter, perform an in-situ hydrothermal synthesis where the glass is used as a nutrient source to reduce the number of pretreatment steps. The transformation of porous glass monoliths to form hierarchical ZSM-5 and TS-1 zeolite structures have effectively demonstrated the ability of the glass to act as a nutrient source and provide the scaffold for these structures [20e23]. The advantage of using the support as a nutrient source is to favour zeolite nucleation and growth on the substrate as opposed to the bulk solution [24]. However, for the support to act effectively as a nutrient source, a balance between the partial dissolution of the support to supply the silica precursors and the inclusion of these precursors into the zeolite structure needs to be found [25]. A fluoride-based hydrothermal synthesis has been demonstrated by Louis et al. to produce ZSM-5 coated filamentous glass supports [26]. The partial dissolution of the glass in the presence of the fluoride ions enabled complete coverage in less synthesis steps in comparison to the conventional hydroxide-based synthesis. Furthermore, in microreactor applications where mass transfer is very fast and the reaction kinetics are often rate limiting, it is

L.A. Truter et al. / Microporous and Mesoporous Materials 226 (2016) 424e432

highly advantageous for the coating to have a high surface area. In conventional zeolite coatings such as membrane application, seeding is often used as a pretreatment procedure to promote the intergrowth of the zeolite crystals to form a continuous layer [27]. However, in microreactor application, a coating which consists of individual nano-sized zeolite crystals would be more advantageous. The use of a nutrient source to promote the formation of a high surface-area zeolite coating has been demonstrated for the insitu hydrothermal synthesis of ZSM-5 in a microchannel by the transformation of a porous silica coating [25]. In this study, a different approach is taken for the incorporation of zeolites in chip-based microreactors in which the two glass plates are initially sealed together to form the chip-microreactor, and thereafter followed by an in-situ glass surface pretreatment and zeolite hydrothermal synthesis step. A glass pretreatment method is described which allows the glass microchannel surface to have a good surface roughness and wettability by a single pretreatment step. Furthermore, the main factors involved in the formation of a zeolite coated chip-microreactor are described. The composition of the zeolite precursor suspension and the in-situ hydrothermal synthesis conditions are optimized in order to obtain a ZSM-5 coating which has a high Brønsted acidity, good uniformity, and high surface-area coating. 2. Materials and methods 2.1. Support Borosilicate glass chip-microreactors were purchased from Chemtrix (The Netherlands) with outer dimensions of 22.5  22.5  1.4 mm (Fig. 6a). The microchannel was 420 mm diameter, 190 mm depth, and 30 cm in length. To further characterize the zeolite coating on the glass, borosilicate glass plates or bulk synthesised powder were used in some cases for comparison. 2.2. Glass pretreatment The glass pretreatment procedure was conducted before the zeolite hydrothermal synthesis within the confines of the chip microchannel. A 0.7 M NH4F solution was inserted into the chip and the chip inlet and outlet sealed. The hydrothermal treatment was done at 150  C for 48 h. After the pretreatment, the excess solution was removed from the microchannel, washed with distilled water, and dried for 24 h at 120  C. 2.3. Catalyst preparation The fluoride precursor suspension was prepared according to the method described by Louis et al. [26] Tetrapropylammonium bromide (TPABr, 99% Merck) and ammonium fluoride (NH4F, 99%, VWR) were dissolved in distilled water. In the case when tetraethylorthosilicate (TEOS, 99%, Merck) was used in the synthesis, the solution was left for 4 h to hydrolyse. Finally, the pH was adjusted to 7 by adding some drops of concentrated hydrofluoric acid (48 wt %, Merck). A typical suspension had the following composition 1 SiO2: 0.07 TPABr: 0.9 NH4F: 0.012 NaAl2O: 80 H2O. Two methods were used for the preparation of the hydroxide precursor suspension. In the first method, aluminium isopropoxide (99.9%, SigmaeAldrich) and TEOS were used. The aluminium isopropoxide and tetrapropyl ammonium hydroxide (TPAOH, 40% aq. sol., Merck) were mixed together at 0  C for 4 h. After the solution turned clear, TEOS was added dropwise. After 4 h a solution of sodium hydroxide (NaOH, 99%, SigmaeAldrich) and distilled water was added dropwise. After the completion of the hydrolysis, when

425

the solution turned clear, the solution was filtered with 0.2 mm syringe filters (0.2 mm PTFE membrane filters, VWR). In the second method, sodium aluminate (NaAlO2, Merck) was used as the aluminium source and colloidal silica (40% colloidal suspension in water, Ludox AS-40, SigmaeAldrich) as the silica source. The NaAlO2 was added together with the NaOH solution. The TPAOH was added to the colloidal silica and stirred, followed by the addition of a solution containing the NaAlO2, NaOH, and water. The use of colloidal silica resulted in the precursor suspension to form a gel which prevented the final suspension being filtered. 2.4. Hydrothermal synthesis After preparation of the precursor suspension, hydrothermal synthesis was conducted. This was carried out to pretreat the glass microchannel surface and to form a zeolite coating. In the in-situ hydrothermal treatments, the precursor suspension was injected into the chip-microreactor and the inlet and outlet sealed by use of a PEEK chip-holder (Chemtrix, The Nehterlands). The ZSM-5 hydrothermal synthesis was conducted to form (i) coating in the chip microreactor, (ii) coating on the borosilicate glass plate, (iii) bulk-synthesised powder. For coating of the borosilicate glass or preparation of the bulk-synthesised powder, the precursor suspension and the glass plate were inserted into teflonlined autoclaves. Hydrothermal synthesis was conducted for 24e48 h at 150e175  C. After the hydrothermal synthesis, the excess solution was removed from the chip-microreactor, washed with distilled water, and dried at 120  C for 24 h. Calcination was conducted in air at 500  C for 2 h at a heating rate of 1  C/min to remove the organic template. In the fluoride-based synthesis, the ZSM-5 was in the NH4- form, however, for the hydroxide-based synthesis, an ion exchange step was necessary to convert the Na-ZSM-5 to NH4-ZSM-5. A 0.5 M NH4NO3 solution was either flowed through the chip-microreactor or stirred with the ZSM-5 powder for 2 h at 80  C. Thereafter, the calcination step was repeated. 2.5. Incorporation of other zeolites by washcoating method To ascertain the extendibility of the coating method to other zeolites and heterogeneous catalysts, titanium silicalite-1 (TS-1) was incorporated by the hydrothermal synthesis method, as well as ZSM-5 by a washcoating method. The ZSM-5 washcoating was incorporated into the chipmicroreactor by the use of the bulk-synthesised ZSM-5 powder described earlier. After the fluoride pretreatment step was applied to the chip-microreactor to improve the surface properties of the glass microchannel surface, a suspension containing 30% ZSM-5, 10% pluronic F127 (BASF), in ethanol was inserted into the chipmicroreactor. The ZSM-5 suspension was displaced from the chipmicroreactor using a 50 ml/min N2 flow with a 5 bar back pressure. The back pressure was necessary to ensure that the N2 linear velocity was constant throughout the length of the chipmicroreactor. The TS-1 coated chip-microreactor was prepared using the hydrothermal synthesis method. Initially, the glass surface was pretreated using the fluoride pretreatment, followed by a TS-1 seeding coating. A suspension containing 30% TS-1 seeds and 10% pluronic F127 in ethanol was washcoated into the chip-microreactor. The TS-1 seeds were prepared using a precursor suspension of 0.43 TBOT:10 TEOS:5 TPAOH:7 IP: 697 H2O for 12 h at 150  C. Thereafter, the in-situ hydrothermal synthesis method was conducted using the same methodology as described previously [28]. A precursor suspension was prepared by mixing anhydrous isopropyl alcohol

426

L.A. Truter et al. / Microporous and Mesoporous Materials 226 (2016) 424e432

(IP, anhydrous, 99.5þ %, Alfa Aesar) with titanium butoxide (TBOT, 97%, Aldrich). Afterwards, deionized water and aqueous tetrapropyl-ammonium hydroxide (TPAOH, 40% aq. sol., Merck) was mixed together and added dropwise to the TBOT mixture, followed by the addition of tetraethylorthosilicate (TEOS, 98%, Aldrich). The TS-1 precursor suspension was inserted into the chipmicroreactor and hydrothermal synthesis was conducted for 48 h at 150  C. 2.6. Characterization The elemental composition of the catalyst was determined by inductively coupled plasma optical emission spectrometry (ICPOES) using a Spectra CirosCCD system. The ZSM-5 powder was dissolved by cold acid digestion with a mixture of HF, HCl and HNO3 at room temperature. The coated chip-microreactor was examined with a XL30-ESEM-FEG scanning electron microscope (SEM) operated at 10 kV, working distance of 10 mm and various magnifications. The glass wettability was evaluated by contact angle measurements of water on the glass surface using an automatic contact angle meter (DataPhysics OCA 30). A 1 mL droplet of demineralized water was dropped on the surface of the substrate. The contact angle of the water droplet with the substrate was determined with SCA 20 software before and after the NH4F hydrothermal treatment. An average contact angle was determined. AFM images were obtained using Solver NEXT equipment from NT-MDT using tapping mode (semicontact) and analysed with Nova_Px software. Nitrogen adsorption isotherms were obtained on an ASAP-2020 Micromeritics instrument with a standard procedure after vacuum pretreatment at 300  C for 12 h up to a residual pressure below 0.1 Pa. The surface area of the bulk synthesised powder was determined from the nitrogen adsorptionedesorption isotherms at 196  C. Temperature Programmed Desorption of ammonia (Autochem, Micrometrics) was performed to determine the amount of acid sites present. The powder zeolite was outgassed by heating to 500  C under a helium flow. Thereafter, the sample was cooled to 120  C and a 3% NH3 flow of 50 ml/min was brought into contact with the sample for 120 min, followed by helium flow for 60 min. The temperature was increased to 800  C at 10  C/min while recording the NH3-desorption using a thermal conductivity detector. 3. Results and discussion 3.1. Fluoride-based synthesis A fluoride-based hydrothermal synthesis method was implemented in order to directly form a zeolite coating on the glass surface whereby the glass could act as a nutrient source, reducing the number of pretreatment steps typically required for

functionalising the glass surface. Table 1 summarises the synthesis conditions at which the fluoride-based hydrothermal synthesis was implemented on borosilicate glass plates. The ability of the borosilicate glass to act as a nutrient source was investigated in sample F-1 where no TEOS was added to the precursor suspension. Fig. 1a shows the formation of a ZSM-5 coating on a borosilicate glass plate indicating that the glass is able to act as a nutrient source to provide the necessary silica required for the zeolite synthesis. With the addition of TEOS to the precursor suspension, the crystal size increased and star-like zeolite crystal structures formed (Fig. 1b,c). The formation of these star-like zeolite crystal structures were described by Louis et al. to be caused by the self-organization of F and TPAþ ions around a so-called reservoir of silica nutrients [29]. This results in a high nucleation rate where multiple zeolite crystals grow from the silica nutrients. The incorporation of aluminium into the ZSM-5 structure of sample F-2 was not exceptionally good (Table 1). The actual Si/Al ratio of 72, determined by ICP, indicated the presence of aluminium. However, the TPD-NH3 results indicated a low Brønsted acidity of 0.04 mmol Hþ/gzeolite which is indicative that a considerable amount of the aluminium is not incorporated into the zeolite framework (theoretical amount on the basis of ICP of 0.23 mmol Al/ g). The poor Brønsted acidity is consistent with previous findings where it is thought that the fluoride ions combine with aluminium to form AlFx compounds thus reducing the amount of Brønsted acid sites [30,31]. Although the fluoride-based synthesis provided sufficient dissolution of the borosilicate glass to provide the necessary silica precursors for ZSM-5 formation, the crystals were large and of poor Brønsted acidity. This is not ideal for microreactor applications with the intention to use the in-situ hydrothermal method since these large zeolite crystals could easily block the microchannel, as well as result in internal mass transfer limitations when applied in chemical reactions. In an effort to reduce the ZSM-5 crystal size, the synthesis temperature was decreased to 150  C (F-3) which resulted in an amorphous porous layer covering the glass surface. Without the addition of TPABr and NaAlO2 (F-4), this porous layer still formed (Fig. 1e,f). The formation of this porous layer is thought to be due to the combination of NH4F etching and the hydrothermal treatment. The use of a hydrothermal treatment under supercritical conditions has been shown to form porous glass at temperatures of 200e400  C [32e35]. A mechanism describing the formation of porous glass from pyrex glass has been described elsewhere [33]. The hydroxyl ions from the solution attack the silanol bonds in the glass resulting in hydration of the glass therefore causing a decrease in the glass transition temperature. In the case of fluoride-based solutions, the reactive fluoride groups (HF2, H2F2) in the solution aid in the formation of SieF units which further react with F species resulting in the removal of silica from the glass matrix [36]. The formation of a porous layer in the NH4F/HF-based solutions at room temperature has been demonstrated on less chemically resistant glass and

Table 1 Fluoride-based synthesis conditionsa and coating properties on borosilicate glass plates. Sample

Precursor suspension molar compositionb TEOS

NH4F

TPABr

NaAlO2

F-1 F-2 F-3 F-4 F-5

0 1 0 0 0

0.9 0.9 0.9 0.9 0.9

0.07 0.07 0.07 0 0

0.012 0.012 0.012 0 0

a b c

Hydrothermal synthesis for 48 h. H2O concentration kept constant at 80 mol. Determined from bulk synthesised powder.

T ( C)

Crystal size (mm)

Brønsted acidityc (mmol Hþ/gzeolite)

Si/Alb (ICP)

175 175 150 150 175

11 27 e 0.4 82

e 0.04 e e e

e 72 e e e

L.A. Truter et al. / Microporous and Mesoporous Materials 226 (2016) 424e432

427

Fig. 1. SEM images of the ZSM-5 coatings on glass plates after fluoride-based synthesis (A) F-1; (B), (C) F-2; (D) initial untreated borosilicate glass; (E) F-4 treated glass; (F)F-4; (G) F5.

silicon substrates [37e39]. It is therefore thought that the combination of a NH4F etching solution and mild hydrothermal treatment conditions enable the formation of a porous layer on the, more chemically resistant, borosilicate glass surface. The ability to change the size of the pores was investigated by changing the NH4F concentration and hydrothermal treatment temperature. However, at higher NH4F concentrations and higher temperature (175  C), the susceptibility of (NH4)3(SiF6)F crystals forming on the surface increased (Fig. 1g) [38]. In contrast, lower NH4F concentrations and temperatures resulted in no porous layer to form on the surface. The use of a fluoride-based hydrothermal synthesis method to form zeolite coatings directly on the glass microchannel's surface is thus not effective in forming coatings with a high surface area and Brønsted acidity. However, the use of a fluoride-based hydrothermal synthesis pretreatment could be an effective means to providing a suitable surface for the subsequent hydroxide-based hydrothermal synthesis step. In the following sections, the surface properties of the glass after the fluoride-based hydrothermal synthesis pretreatment are investigated, as well as its application as a surface pretreatment for chip-based microreactors. 3.2. Characterization of NH4F pretreated glass In order for a substrate to be a suitable surface for the growth of a zeolite coating on the surface, it should have a good surface wettability and roughness [10]. For this reason, the surface properties of the resulting NH4F hydrothermally treated glass were investigated with the intention to use the treated glass in a subsequent in-situ zeolite hydrothermal synthesis. Atomic force microscopy (AFM) and surface contact angle measurements were done on the borosilicate glass before and after the NH4F hydrothermal treatment to determine the surface roughness (Fig. 2) and wettability (Fig. 3). Fig. 2 depicts the glass surface after the NH4F hydrothermal treatment, where a significant surface roughness is observed. AFM

of the untreated glass showed a highly smooth surface with no visible surface defects or observable roughness (image not shown for brevity). The root-mean square (rms) roughness of the glass plate before and after the NH4F treatment is 0.9 nm and 185 nm, respectively, which indicates that the NH4F treated glass has a significant improvement in the surface roughness. It has been demonstrated in other studies that with an increase in surface roughness, a higher amount of surface imperfections occur on the surface which aids in the creation of nucleation sites during the zeolite synthesis [18,40,41]. The effects of the surface treatment on the zeolite coating are further discussed in the preceding section. In Fig. 3 the surface contact angle of the glass was initially 45⁰ indicating the surface to be relatively hydrophobic. After the NH4F hydrothermal treatment the surface contact angle decreased to 1.7 indicating that the pretreated surface becomes very hydrophilic and thus the wettability is greatly improved (Fig. 3). This low surface contact angle of less than 5⁰ corresponds to a high amount of SieOH groups (>15 OH nm2) [10]. The improvement in surface wettability is thought to enable better contact with the zeolite precursor suspension which enables better zeolite coverage during the hydrothermal synthesis step [18]. Conventionally, the pretreatment of the glass microchannel surface requires various steps such as an initial etching or micropowder blasting step to roughen the surface; the deposition of a protective film to prevent glass dissolution; coating with a TiO2 film followed by UV-irradiation to improve the surface wettability; and finally, seeding the surface with zeolite crystals to improve zeolite coverage [18,19]. However, the use of the fluoride-based hydrothermal treatment provides a single-step pretreatment procedure for providing a surface with high surface roughness and wettability, and in addition provides a silica nutrient source. Due to the improved surface properties of the glass, the NH4F hydrothermal treatment was used as a surface pretreatment before the second zeolite-forming hydrothermal synthesis step was conducted. This pretreatment step was conducted within the glass

428

L.A. Truter et al. / Microporous and Mesoporous Materials 226 (2016) 424e432

Fig. 2. Atomic force microscopy (AFM) image of the glass plate after the fluoride surface pretreatment.

Fig. 3. Surface contact angle of the glass plate (A) before pretreatment; (B) after pretreatment.

chip-microchannel (after the sealing of the two glass plates together). 3.3. Hydroxide-based synthesis The conditions used for the hydroxide-based hydrothermal synthesis and coating properties are summarised in Table 2. The effect of the silica and aluminium source, and Si/Al ratio were varied in order to optimize the ZSM-5 coating in terms of crystal size, coating thickness and Brønsted acidity. 3.4. The influence of the NH4F pretreatment The effect of the pretreatment method previously described is demonstrated in Fig. 4 where silicalite coatings are formed in the chip-microreactor by use of an in-situ zeolite hydrothermal synthesis. Without the pretreatment step (Fig. 4a), the zeolite coverage is poor with large zeolite crystals growing on the glass microchannel surface. The smooth surface, poor wettability and high chemical resistance of the borosilicate glass makes the formation of a zeolite coating difficult. However, if the pretreatment step is initially performed, followed by the zeolite hydrothermal synthesis, the zeolite

coverage is greatly improved (Fig. 4b) where the coating thickness is approximately 8 mm consisting of 3 mm crystals. The formation of the porous layer during the NH4F pretreatment aids in increasing the nucleation rate at the glass microchannel surface during the zeolite hydrothermal synthesis step favouring the formation of a zeolite coating over bulk crystallization in the microchannel. This increase in nucleation rate can be attributed to the (i) improved roughness leading to more surface imperfections and thus sites for nucleation; (ii) better surface wettability enabling better contact with the precursor suspension; (iii) higher supersaturation at the surface due to the dissolution of the porous layer to provide silica precursors. The use of the fluoride-based hydrothermal treatment is thus an effective pretreatment method for obtaining zeolite coatings with good coverage. It should be noted that the partial dissolution of the glass surface during the pretreatment method could result in various impurities in the glass being incorporated into the zeolite framework which could affect the zeolite catalytic activity and/or zeolite nucleation and growth. Although, Fig. 4, seems to indicate that the zeolite nucleation and growth is not significantly affected by the impurities possibly arising from the prior surface pretreatment. In the subsequent sections, the in-situ fluoride-based hydrothermal treatment is used for the pretreatment of the glass chip-microchannel surface.

Table 2 Hydroxide-based synthesis conditionsa and coating properties.

MFI-1 S-1b AL-1 AL-2 a b c d

Precursor compositionb

Al source

Si/Al (theory)

Si/Al (ICP)d

Brønsted acidityc (mmolHþ/gzeolite)

Crystal size (mm)

Coating thick-ness (mm)

BETb (m3/g)

1:0.22:0.13:65.5:0 1:0.22:0.13:40:0.07 1:0.22:0.13:65.5:0.034 1:0.22:0.13:65.5:0.033

e NaAl2O NaAl2O C9H21 O3Al

e 15 29.9 29.1

>10000 16 20.7 28.1

e 1.1 0.8 0.7

3 2.5e5.5 1.6 0.5

8 23 2.6 5.5

359 307 364 382

Hydrothermal synthesis for 48 h. Molar ratio (SiO2: TPAOH: NaOH: H2O: NaAlO2/C9H21O3Al). Colloidal silica as silica source. Determined from bulk synthesised powder.

L.A. Truter et al. / Microporous and Mesoporous Materials 226 (2016) 424e432

429

Fig. 4. Cross section of silicalite coated chip-microreactor (A) without NH4F pretreatment step, (B) with initial NH4F pretreatment step (MFI-1).

3.5. The influence of silica and aluminium source and amount In order to further optimize the zeolite coating, the effects of the silica and aluminium source in the precursor suspension were investigated. Colloidal silica was tested as a silica source in the precursor suspension (Sample S-1). This polymeric silica source was chosen in

an attempt to match the dissolution of the glass silica porous layer and the silica from the precursor suspension in order to limit bulk crystallization and instead favour the formation of the zeolite coating. The resulting ZSM-5 coating in the chip-microreactor is depicted in Fig. 5a,b where a thick coating of >10 mm with crystal size ranging between 2 and 6 mm was formed (Table 2). It is thought that due to the slow dissolution of the colloidal silica from the

Fig. 5. Cross-section of ZSM-5 coated chip microreactor (A, B) S-1; (C, D) AL-1; (E, F) AL-2.

430

L.A. Truter et al. / Microporous and Mesoporous Materials 226 (2016) 424e432

precursor suspension, more of the silica from the glass porous layer is able to act as a nutrient source, thereby increasing the supersaturation at the microchannel surface and thus increasing the nucleation and growth rate. This enables the formation of a thicker gel-layer to form on the glass microchannel surface which enables the formation of thicker coatings. The coating also showed good reproducibility along most of the channel length (Supporting Information). However, due to the precursor suspension being a gel, it could not be filtrated before inserting it into the chipmicroreactor which made periodic blockages inevitable in some coated chip-microreactors. Due to the difficulty of avoiding these blockages, the preparation of clear precursor suspensions with lower viscosity and particle size with the use of TEOS as a silica source was required. With the use of TEOS, as a monomeric silica source, the susceptibility of bulk crystallization increases. Therefore, in order to prevent bulk crystallization from occurring which could potentially lead to microchannel blockage, the water content in the precursor suspension was optimized. Once sufficient dilution of the precursor solution was obtained, no blockages in the chip-microreactor were encountered due to the rate of zeolite nucleation and growth on the microchannel surface exceeding bulk crystallization. The crystal size of the resulting coatings which used TEOS as a silica precursor were smaller (<3 mm) with high BET surface areas (>350 m3/g). It should be noted that due to the low zeolite loading in the chipbased microreactor (~0.4 mg), it was not possible to directly measure the BET surface area and Brønsted acidity of the zeolite coated chip-based microreactor. Instead the same precursor suspension was used to prepare the bulk synthesised powder and used for characterization. The incorporation of aluminium into the ZSM-5 coating was investigated by variation of the aluminium source and Si/Al ratio. The effects on the crystal size, Brønsted acidity and coating coverage are provided in Table 2. NaAlO2 (S-1 & AL-1) and aluminium-isopropoxide (AL-2) were used as aluminium sources. The aluminium source is seen to have an effect on the ZSM-5 crystal size where the NaAlO2-containing, AL-1 sample, was 1.6 mm in comparison to the aluminium-isopropoxide AL-2 sample crystal size of 0.5 mm. The use of aluminium-isopropoxide has previously been shown to produce nano-sized ZSM-5 powder of 10e100 nm [42,43]. In this study, the crystal size is larger which can be attributed to the higher amount of Naþ ions present which lead to the agglomeration of silica ions resulting in a larger crystal size. However, the addition of NaOH was necessary for the dissolution of the porous glass layer, as well as for the inclusion of aluminium into the zeolite structure. The Naþ ions act as charge compensators

aiding in the inclusion of aluminium into the zeolite framework [44e46]. Both aluminium sources, provide good Brønsted acidity with the AL-1 sample containing a slightly higher amount of Brønsted acid sites (0.8 mmol Hþ/gzeolite) in comparison to 0.68 mmol Hþ/gzeolite in sample AL-2. This is mainly due to the lower Si/Al ratio of sample AL-1. Fig. 6 shows a comparison between the chip-microreactor before and after the coating procedure. While a zeolite coating has been successfully incorporated into the chip microreactor, the catalyst loading was too low to enable quantifiable results to be obtained with its application in a catalytic reaction (~0.4 mg) or to perform detailed characterization of the specific coatings. For this reason the in-situ hydrothermal NH4F treatment and zeolite synthesis methodologies have been applied to similar chip-microreactors of 1 m in length to obtain sufficient catalytic activity in subsequent work [47].

3.6. Zeolite incorporation by other methods/zeolites The direct hydrothermal synthesis of silicalite and ZSM-5 zeolite was shown to be an effective method for incorporation into chipmicroreactors. The use of the complete chip-microreactor, after the assembly of the two glass plates together, overcomes the difficulty of sealing and possible contamination often encountered during the pretreatment and coating steps; while the use of a fluoride hydrothermal synthesis pretreatment provides conducive surface properties for the subsequent coating step. It is therefore of interest to ascertain whether these coating techniques and the approach of using the complete chip-microreactor, and fluoride pretreatment method could be applied to other zeolites and heterogeneous catalysts. For this reason, some preliminary experiments were performed to ascertain whether other zeolites could be incorporated using a similar hydrothermal synthesis method, or alternatively, heterogeneous catalysts in general could be incorporated by the use of a washcoating method. The incorporation of titanium silicalite-1 (TS-1) zeolite into chip-based microreactors was shown to be possible by the application of the in-situ hydrothermal synthesis method (Fig. 7a,b). However, due to the TS-1 being more hydrophobic, it required both the fluoride pretreatment, as well as an initial seeding step to induce the formation of a TS-1 coating during the hydrothermal synthesis. The resulting TS-1 coating consisted of 900 nm crystals with a coating thickness between 4 and 14 mm. The difference in the TS-1 coating thickness was thought to be attributed to the initial washcoating of the TS-1 seeds being non-uniform where the coating was thicker at the bottom of the channel due to the effects of gravity.

Fig. 6. Chip microreactor (A) original uncoated chip microreactor; (B) ZSM-5 coated chip microreactor (AL-1).

L.A. Truter et al. / Microporous and Mesoporous Materials 226 (2016) 424e432

431

Fig. 7. Preliminary experiments for the incorporation of (A,B) TS-1 coatings by an in-situ hydrothermal synthesis method; (C,D,E) ZSM-5 coatings by a washcoating method, in a chip microreactor.

Alternatively, a washcoating method was applied for the incorporation of ZSM-5 into the chip-microreactor. Without the use of the fluoride pretreatment method, the coating thickness was minimal and the adherence was poor. However, with the use of the initial fluoride pretreatment step for the activation of the microchannel surface, the washcoating of the chip-microreactor became possible. Fig. 7cee shows the ZSM-5 washcoated chipmicroreactor. The uniformity of the coating varied substantially and the susceptibility of blockage was sometimes present. The microchannel corner's coating was significantly thicker than the top or bottom coatings which could result in mass transfer limitations when applied in chemical reactions. The uniformity of the coating obtained by the in-situ hydrothermal synthesis method is therefore better in comparison to the washcoating method. However, these preliminary TS-1 and washcoated ZSM-5 coatings give some indication that the ability to incorporate other zeolites and heterogeneous catalysts into chip-microreactors is possible.

4. Conclusions In this work a method for the incorporation of zeolites into chipmicroreactors has been developed. A versatile glass pretreatment method has been developed which, in a single step, provides a surface with high roughness, good wettability, and a silica nutrient source. The combination of a NH4F solution under hydrothermal conditions, enables the surface of the glass microchannel to be transformed to a porous layer on the glass microchannel surface.

This pretreated glass surface is effective for the incorporation of zeolites by both a washcoating and an in-situ hydrothermal synthesis method. The in-situ hydrothermal synthesis step has, however, an improved uniformity and overcomes the susceptibility of channel blockage. Through the implementation of both a fluorideand hydroxide-based hydrothermal synthesis method, it was concluded that the fluoride-based synthesis enabled the formation of a zeolite coating by the partial dissolution of the glass surface but resulted in large zeolite crystals with a low Brønsted acidity. Alternatively, the use of the hydroxide-based hydrothermal synthesis enabled a superior zeolite coating by the optimization of the ZSM-5 precursor solution in terms of the crystal size, coverage, and Brønsted acidity. The optimized ZSM-5 coating consisted of 0.5 mm crystals forming a uniform 5.5 mm coating with a relatively high Brønsted acidity. The extension of the hydrothermal synthesis coating method to other zeolites such as TS-1 was also shown to be feasible. Acknowledgements The authors would like to thanks NRSC-Catalysis for the financial support. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.micromeso.2016.02.016.

432

L.A. Truter et al. / Microporous and Mesoporous Materials 226 (2016) 424e432

References [1] A. de Mello, R. Wooton, Lab. Chip 2 (2002) 7N. €we, Chem. Eng. Sci. 59 (2004) 4789. [2] H. Pennemann, V. Hessel, H. Lo [3] D.M. Roberge, B. Zimmermann, F. Rainone, M. Gottsponer, M. Eyholzer, Org. Process Res. Dev. 12 (2008) 905. [4] K.S. Elvira, X. Casadevall, R.C.R. Wootton, J. Andrew 5 (2013) 905. [5] J. Wegner, S. Ceylan, A. Kirschning, Chem. Commun. 47 (2011) 4583. e, P.H. Seeberger, Chem. A Eur. J. 12 (2006) 8434. [6] K. Geyer, J.D.C. Code [7] E.V. Rebrov, Catal. Ind. 1 (2010) 322. [8] M. Brivio, W. Verboom, D.N. Reinhoudt, Lab. Chip 6 (2006) 329. [9] T. Frank, Microreactors Org. Chem. Catal (2013) 53e80. [10] V. Valtchev, S. Mintova, M. Tsapatsis, Ordered Porous Solids, 2009. [11] E.V. Rebrov, G.B.F. Seijger, H.P.A. Calis, M.H.J.M. de Croon, C.M. van den Bleek, J.C. Schouten, Appl. Catal. A Gen. 206 (2001) 125. [12] Y.S.S. Wan, J.L.H. Chau, A. Gavriilidis, K.L. Yeung, Microporous Mesoporous Mater 42 (2001) 157. [13] M.J. den Exter, H. van Bekkum, C.J.M. Rijn, F. Kapteijn, J.A. Moulijn, H. Schellevis, C.I.N. Beenakker, Zeolites 19 (1997) 13. [14] B.J. Schoeman, A. Erdem-Senatalar, J. Hedlund, J. Sterte, Zeolites 19 (1997) 21. [15] J. Lik, H. Chau, Y. Shan, S. Wan, A. Gavriilidis, K. Lun, Chem. Eng. J. 88 (2002) 187. [16] J.L.H. Chau, A.Y.L. Leung, K.L. Yeung, Lab. Chip 3 (2003) 53. [17] E. Mateo, R. Lahoz, G.F. De La Fuente, A. Paniagua, J. Coronas, J. Santamaría, Chem. Mater 16 (2004) 4847. [18] M. Mies, E.V. Rebrov, J.C. Jansen, M.H.J.M. de Croon, J.C. Schouten, J. Catal. 247 (2007) 328. [19] O. Muraza, E.V. Rebrov, T. Khimyak, B.F.G. Johnson, P.J. Kooyman, U. Lafont, M.H.J.M. de Croon, J.C. Schouten, Chem. Eng. J. 135 (2008) S99. [20] W.-Y. Dong, Y.-J. Sun, H.-Y. He, Y.-C. Long, Microporous Mesoporous Mater 32 (1999) 93. [21] F. Scheffler, W. Schwieger, D. Freude, H. Liu, W. Heyer, F. Janowski, Microporous Mesoporous Mater 55 (2002) 181. [22] T. Peppel, B. Paul, R. Kraehnert, D. Enke, B. Lücke, S. Wohlrab, Microporous Mesoporous Mater 158 (2012) 180.

[23] C. Shen, Y.J. Wang, C. Dong, G.S. Luo, Chem. Eng. J. 235 (2014) 75. [24] E.E. Mcleary, J.C. Jansen, Top. Catal. 29 (2004) 85. [25] L.A. Truter, V.V. Ordomsky, T.A. Nijhuis, J.C. Schouten, J. Mater. Chem. 22 (2012) 15976. [26] B. Louis, C. Tezel, L. Kiwi-Minsker, A. Renken, Catal. Today 69 (2001) 365. €lsch, D. Creaser, J. Caro, J. Sterte, J. Memb. Sci. 159 [27] J. Hedlund, M. Noack, P. Ko (1999) 263. [28] L.A. Truter, D.M. Perez Ferrandez, J.C. Schouten, T.A. Nijhuis, Appl. Catal. A Gen. 490 (2015) 139. [29] J. Arichi, B. Louis, Cryst. Growth Des. 8 (2008) 3999. [30] B. Louis, L. Kiwi-Minsker, Microporous Mesoporous Mater 74 (2004) 171. [31] R.J. Gorte, Catal. Lett. 62 (1999) 1. [32] S.D. Stookey, United States Patent, US 3,498,803, 1970. [33] F.A. Sigoli, Y. Kawano, M.R. Davolos, M. Jafelicci, J. Non, Cryst. Solids 284 (2001) 49. [34] F.A. Sigoli, M.V. Giotto, M.R. Davolos, M.J. Ju, 201 (2003) 1196. [35] R. Leboda, E. Mendyk, V.A. Tertykh, Mater. Chem. Phys. 42 (1995) 7. [36] D.M. Knotter, J. Am. Chem. Soc. 122 (2000) 4345. [37] R. Houbertz, J. Vac, Sci. Technol. B Microelectron. Nanom. Struct. 12 (1994) 3145. [38] H. Totsuka, J. SID 10 (2002) 381. [39] T. Dittrich, S. Rauscher, V.Y. Timoshenko, J. Rappich, I. Sieber, H. Flietner, H.J. Lewerenz, Appl. Phys. Lett. 67 (1995) 1134. [40] V. Valtchev, S. Mintova, Zeolites 15 (1995) 171. [41] V. Valtchev, S. Mintova, L. Konstantinov, Zeolites 15 (1995) 679. €urer, B. Kraushaar-Czarnetzki, Microporous Mesoporous [42] G. Reding, T. Ma Mater 57 (2003) 83. ndez, J.A. Melero, Microporous Meso[43] R. Van Grieken, J.L. Sotelo, J.M. Mene porous Mater 39 (2000) 135. [44] A.E. Persson, B.J. Schoeman, J. Sterte, J.E. Otterstedt, Zeolites 12 (1995) 611. [45] R. de Ruiter, J.C. Jansen, H. van Bekkum, Zeolites 12 (1992) 56. [46] J.H. Koegler, H. van Bekkum, J.C. Jansen, Zeolites 19 (1997) 262. [47] L.A. Truter, V. Ordomsky, T.A. Nijhuis, J.C. Schouten, Lab Chip submitted (n.d.).