Zeolite molecular sieves: preparation and scale-up

Zeolite molecular sieves: preparation and scale-up

Microporous and Mesoporous Materials 82 (2005) 217–226 www.elsevier.com/locate/micromeso Zeolite molecular sieves: preparation and scale-up John L. C...

185KB Sizes 7 Downloads 76 Views

Microporous and Mesoporous Materials 82 (2005) 217–226 www.elsevier.com/locate/micromeso

Zeolite molecular sieves: preparation and scale-up John L. Casci Johnson Matthey Catalysts, P.O. Box 1, Billingham, Cleveland TS23 1LB, UK Received 25 April 2004; received in revised form 15 November 2004; accepted 20 November 2004 Available online 5 May 2005

Abstract This paper is aimed at new researchers, and more experienced scientists, who are intending to embark on programmes of work involving zeolite synthesis. The paper describes, from an industrial perspective, aspects of preparation and scale-up. Following some initial comments on environment, health and safety the paper provides an overview of some basic principles covering thermodynamics and kinetics. Thereafter, the paper ‘‘de-constructs’’ a typical zeolite synthesis mixture describing the role of individual components and reaction conditions. Finally zeolite scale-up is examined touching on economic considerations, raw material sources and the basic principles of mixing/agitation.  2005 Elsevier Inc. All rights reserved. Keywords: Zeolite preparation; Scale-up; Mixing; Thermodynamics

1. Introduction If this paper had the subtitle ‘‘An Industry Perspective’’ then the first aspect that would be considered in ‘‘Preparation’’ would be––for what purpose? Zeolite molecular sieves find extensive use in catalytic and sorption–separation processes and ion-exchange applications––see the other papers in this series and Proceedings from the 14th International Zeolite Conference itself. The application then drives aspect of scale of operation (and scale-up), the optimisation and development required and the final physical form. This last aspect can be of crucial importance since, for most applications, the zeolite powder prepared by synthesis has to be formed into a larger physical body, typically extrudates. If the aspects concerning application dominate the initial considerations then what immediately follows is a review of the EHS (environment, health and safety) issues. Such considerations feature rarely in the literature, although recently there has been a useful summary provided by Robson [1]. Table 1 attempts to capture some basic considerations. EHS policies in individual organi1387-1811/$ - see front matter  2005 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2005.01.035

sations will dictate how a programme or project proceeds and how issues are identified and tackled but, in general, EHS matters should be dealt with by using ‘‘engineering controls’’ as a first measure then backedup, if necessary, by PPE (personal protective equipment). So, for example, dust handling and containment facilities should be employed as a first measure rather than dust masks (which may be required to augment any dust handling capability). In terms of environmental impact, aspects of solid, liquid and gaseous effluent should be considered. These considerations should cover waste raw materials, waste products and solids that result from equipment cleanout and those that may be present in wash-water. Issues from vent-lines and spillages must also be considered in addition to off-gas from calcination furnaces etc. Of particular note are the odour problems that can be associated with organic templates. Many of these templates are quaternary ammonium compounds and these can readily undergo Hoffmann degradation by the action of base and in the process may liberate an amine. Many amines have high volatility and pose a very significant odour problem.

218

J.L. Casci / Microporous and Mesoporous Materials 82 (2005) 217–226

Table 1 Some environmental, health and safety issues associated with zeolite synthesis Environment

Effluent-solid

Effluent-liquid

Health

Waste zeolite (product) from: • Failed syntheses • Vessel clean-out Residual raw materials Waste-liquor, including wash-water (filtration), vessel Clean-out and ion-exchange containing: • Dissolved silica • Salts (nitrate, chloride, bromide, fluoride etc.) • Organic templates and their degradation products at extremes of pH: • Strongly alkaline (synthesis) • Strongly acidic (ion-exchange)

Effluent-gaseous

Vent streams (e.g. pressure relief valves from synthesis) to off-gas (e.g. from calcination) including: • Organic templates or degradation products––odour issues! • VOCs––see above • Nox • Volatile halides Some streams may contain entrained dust

Dust and liquids

• Dust or splashed/spilled solutions – Respiration of dusts or aerosols or – absorption through skin • Raw materials e.g. silica • Product e.g. zeolite(s) • Organic templates and/or degradation products

Safety

Physical Chemical

Trapped pressure in autoclaves Reagents • Alkaline, HF • Organic templates (degradation products) • Heteroatoms e.g. Be

a review of the products morphology and toxicity should be considered as part of any EHS study. At this point is worth noting the wider EHS context: sustainability, which may be summarised as . . .Meeting the needs of the current generation without compromising the needs of future generations. . . One of the implications is, obviously, waste minimisation but the wider context of energy and raw material usage should also be considered. Such a review is beyond the scope of this paper. Zeolites are a group of frameworks or ordered porous solids that cover a wide range of pore sizes. Fig. 1 summarises the different structure types from zeolites themselves to pillared solids to mesoporous materials. To this listing may be added ‘‘hybrid’’ structures such as the silsesquioxane structures and other materials that are featuring in recent literature reports. Even restricting discussion to zeolites (and zeotypes), however, still covers a wide variety of framework types and compositions. For scientists involved in zeolite synthesis an essential source of information is the atlas of topologies that may be found in hard copy [3] or via the web [4]. These sources provide an excellent overview of the framework structures and their compositions. Table 2 summarises the current status of zeolite framework types using the 3-letter code formalism that is part of the accepted nomenclature for microporous and mesoporous solids [5]. Of the 150 different framework topologies over 60% exist as silicate types (some of these also exist as phosphates)––thus silica-based materials are still the dominant material type. In addition to those in Table 2 there are a number of frameworks based on germanates and arsenates. A significant number of these framework topologies are naturally occurring and this number is growing as minerals are found

Mesoporous Molecular Sieves

Two features in Table 1 are worthy of further comment: Pillared Layered Solids

(i) Templates. These are widely employed in zeolite synthesis and the materials cover a wide spectrum of toxicity. Consequently, before any work commences a full review of their EHS properties must be considered. Additionally, it should be appreciated that many templates degrade during synthesis and an assessment of the degradation products should be undertaken as part of the initial work. (ii) Zeolites. Certain naturally occurring zeolites have been associated with mesothelioma, see for example the early paper by Elmes [2] and consequently

Zeotypes

Zeolites

1

2

3

4

10

Pore Diameter /nm Fig. 1. Representation of ordered porous solids of varying pore sizes.

J.L. Casci / Microporous and Mesoporous Materials 82 (2005) 217–226

Silicates ABW AFG BEA BIK BOG BRE CAS CHI DAC DDR DOH EAB EDI EMT EPI EUO FER GME

GOO HEU KFI JBW LAU LIO LOS LOV LTL LTN MAZ MEI MEL MEP MER MFI MFS MON

MOR MTN MTT MTW NAT NES NON OFF PAR PAU PHI ROG SGT STI THO TON WEN YUG

Silicates and phosphates

Phosphates

AFI ANA AST BPH CAN CHA SRI FAU GIS LEV LTA RHO SOD

AEI AEL AET AFO AFR AFS AFT AFY APC APD ATN ATO ATS ATT ATV AWW CLO VFI

that represent the natural counterpart of materials that were first seen in the laboratory (e.g. Gottardiite has the framework topology of NES). Examination of the atlas shows a number of the framework types have very similar ‘‘compositions’’ (e.g. MOR and LTL) and the significance of this in terms of composition will be discussed later.

2. Fundamental principles There are a variety of means by which thermodynamic information can be obtained (or determined) for zeolites, and Table 3 attempts to summarise some of these approaches together with the relevance and applicability. The computational approaches are generally used in structure solution but have been used to ‘‘rank’’ materials in terms of lattice energies or enthalpies of formation [11]. This is frequently part of a programme aimed at enumerating theoretical or hypothetical frameworks and assessing their viability. The ‘‘mineralogical’’ approach of mapping the stability field has been carried out relatively rarely [7] most likely because of the intense experimental work required. It is interesting to note that the ‘‘highthroughput’’ experimental methods are, in some cases, now being employed to carry out this mapping [12] hence it is likely that this method will increase in prominence. The equilibrium model of Lowe [9] while being limited to high-silica zeolites (it could be extended to other systems) has tremendous value in providing an insight to the overall mechanism of formation and also provides practical value. At the heart of the equilibrium model is the transformation of an amorphous gel into crystalline zeolite. Thus the overall free energy change is given by:

Table 3 Examples of approaches and applicability of thermodynamic information Approach

Information

Applicability (limitations)

Ref.

Mineralogical

Full stability field: temperature, pressure, pH, component activities

Experimental: all materials But time consuming

[7]

Computational

Lattice energy (or free energy of formation)

Theoretical: all materials Provides relative energies

[8]

Equilibrium

Reactant (silica) yield as a function of pH

Theoretical: limited materials Currently limited to high-silica zeolites

[9]

Arrhenius equation

Activation energy (actual or notional)

Experimental: all materials Estimating crystallisation times But theoretical basis uncertain

[10]

DG ¼ RT lnðK s;gel =K s;zeolite Þ assuming that: (a) The solid phases are in equilibrium with the solution phase. (b) The SiO2/Al2O3 ratio is high (i.e. >20/1) such that the system is, effectively, dominated by silica, silicate. Thus Ks,gel and Ks,zeolite are the solubility products of the amorphous gel and zeolite respectively. The equilibrium model can be used in a variety of ways––most obviously to calculate pH changes and yields (on silica) as a function of reaction composition. A simple example is provided in Fig. 2––an assessment

Yield (%Silca in product zeolite)

Table 2 Zeolite framework topologies by 3-letter code and framework type

219

100 90 80 70 60 50 0

2

4

6

8

10

12

x in 60 SiO2 : (x + 1)Q2O : Al2O3 : 3000 H2O

Fig. 2. Equilibrium model [9] yield (% silica in product zeolite) for a reaction mixture of composition.

J.L. Casci / Microporous and Mesoporous Materials 82 (2005) 217–226

of the yield on silica (% silica in product zeolite) for a particular composition: 60SiO2 : ðx þ 1ÞQ2 O : Al2 O3 : 3000H2 O In Fig. 2, it can be seen that the yield ranges between 75% and 98% over the range of base levels employed. This is an excellent reminder that, in high-silica zeolite syntheses, the yield on silica is incomplete and that most (all) syntheses have silica remaining in solution at the end of the crystallisation. The immediate implication of this is that on completion of crystallisation the liquor is saturated with silica and is in (quasi-) equilibrium with the solid zeolite. If that system is perturbed then the equilibrium shifts. Two such perturbations can/do occur: (i) On cooling the reaction mixture to ambient (i.e. at the end of the reaction) the solution becomes supersaturated (with respect to silica) and this supersaturation will be discharged, until re-equilibration has taken place. This discharge of supersaturation can (usually) amount to a few %w/w silica being deposited with (or onto the zeolite). (ii) If the pH is perturbed such as by template degradation (downwards perturbation) then this, again, can lead to deposition of silica although in this case this may be as part of the zeolite framework. This can have significant impact on the properties of the material [13]. A simpler concept and useful tool is the calculation of an apparent activation energy (Ea) for a particular zeolite. In the Arrhenius equation: k ¼ A  expðEa =RT Þ where k is the reaction rate and for zeolites this should be determined from linear crystal growth rate measurements. Few such studies have been carried out and what is often used is 1/(crystallisation time). While lacking rigour, such that the values for the activation energy have limited meaning, it is a convenient tool for allowing crystallisation times to be interpolated from the available information. Care should be taken when extrapolating the data since it is possible that one can move from one crystallisation field to another––such that the target zeolite is no longer prepared. A good point to complete any discussion on the thermodynamics of zeolite synthesis is with Breck [6] who notes that zeolites are metastable species and if left in their synthesis magma will convert to more stable materials. An example of this is provided in Fig. 3 for the synthesis of MFI (ZSM-5) from the templatefree or ‘‘inorganic’’ system. The plot shows XRD intensity as a function of time for a reaction of composition: 60SiO2 : Al2 O3 : 10Na2 O : 3000H2 O at 160  C

XRD Intensity

220

0

50

100

150

200

250

300

Time(hours)

MFI

Kenyaite

alpha-Q

ANA

Fig. 3. Illustration of reaction over-run for MFI (ZSM-%) prepared from a template-free synthesis.

It can be seen that if a ‘‘snap-shot’’ was taken at about 50 h then the products identified would have been MFI + Kenyaite (a layer silicate) whereas after about 250 h only alpha-quartz and ANA (analcime) are present. In this case the MFI has effectively disproportionated into an aluminium-rich phase (ANA) and a silica-rich phase (alpha-quartz) both of which are more stable.

3. Reaction composition and key variables 3.1. General comments Zeolite synthesis is conventionally represented by a schematic such as that shown in Fig. 4. Such a schematic is used to represent the mixing of the raw materials to provide a ‘‘gel’’ that may be aged for a period (temperature, T1 for a time, t1) before reacting (temperature, T2 for a time, t2) to provide the crystalline zeolite. The ‘‘gel’’ can vary dramatically in appearance from a clear solution [14,15] to a stiff, apparently dry, paste. Most gels, however, have an intermediate structure that allows pumping and stirring. Gel ageing can be important as can crystallisation temperature (and time), order of mixing of components and agitation.

Raw Materials

Gel T1, t1, (ageing)

T2, t2, (crystallisation)

Fig. 4. Zeolite synthesis schematic.

Zeolite

J.L. Casci / Microporous and Mesoporous Materials 82 (2005) 217–226

3.2. Reaction composition Zeolite reaction mixtures are usually written as the mole ratios of the oxide components: SiO2 : TO2 : M2 O : R2 O : H2 O þ Seed þ Salt Table 4 attempts to summarise these components and their role(s). The framework components are Si, T. These components impact on the framework composition (and product type) and on the properties of the material. A wide range of T-atoms may be used from Si (i.e. generating an uncharged silica polymorph) through Al, Ga, B, Ti etc. These heteroatoms impact on the framework charge and applications (from ionexchange to catalysis). A ‘‘mineralisation’’ component is essential in zeolite synthesis and this is usually base although there has

Table 4 Zeolite reaction mixtures: components and roles SiO2:TO2:M2O:R2O:H2O + Seed + Salt Component

Role

SiO2, TO2

Framework components

Comments • Si may be replaced by Ge • T may be Si, Al, B, Fe, Ga, Ti etc. • Ratio Si/T affects crystallisation rate • Ratio Si/T affects framework charge and properties

M2O

Mineraliser or base

• Base (OH) is routine but fluoride may be employed • M is usually an alkali metal or a quaternary ammonium compound • Na ! K ! Rb ! Cs can alter product but also crystal morphology and size • Concentration can affect product purity, crystallisation rate and framework composition (gradient)

R2O

Template

• Structure direction • Structure blocking • Framework stabilisation (to over-run) • Location of T-atoms • Crystal morphology and size

H2O

Solvent

• Limited work with other solvents: – Alcohols/diols – Ammonia

Seed

Seed

• Structure directing (rare) • Increase rate (overall) of crystallisation

Salt

Direct/Indirect

• Direct addition • Crystal size and morphology • Indirect addition––through use, for example, of Al2(SO4)3 as Al source • Salts may act as framework stabilisers cf. templates

221

been good success with fluoride ions [16,17]. In general the base is added via an alkali metal or quaternary ammonium compound. Of the alkali metals Na dominates. While the other alkali metals can be used these tend to slow the crystallisation but can provide interesting changes in crystal size and morphology. The term ‘‘template’’ is usually applied to amines or quaternary ammonium compounds. Such materials are usually thought of as structure directing agents but they also play crucial roles in structure stabilisation and have even been shown to be effective (in some circumstances) as structure blocking agents [18]. The use of water as the solvent in zeolite synthesis is widespread but there have been notable attempts to employ other solvents [19a] from glycol to (liquid) ammonia [19b]. One pitfall when exploring non-aqueous solvents is the addition of small quantities of water through the raw materials (base, silica, alumina etc.). These low levels of water can result in zeolite crystallisation via the aqueous phase and lead the unwary into believing that the ‘‘solvent’’ has been responsible. Seed crystals are often present inadvertently if the reaction vessel has not been thoroughly cleaned (see later) but they can also be added deliberately. Addition, at 1–5%w/w based on the framework components, is usually carried out in order to reduce the overall crystallisation time [19c]. Similar to ‘‘seed’’, salt (e.g. Na2SO4) is often added without realising but in this case it is because of reagent choice. Simply if Al2(SO4)3 is used as the source of Al then (through the reaction with the base, NaOH) Na2SO4 is generated in situ in the reaction gel. Salts may impact on crystal size [20] and morphology (but at high loadings) and can act as framework stabilisers––this is more common for small-pore aluminiumrich zeolites. 3.3. Key variables in zeolite synthesis Even when the reaction mixture is ‘‘fixed’’ a large number of variables can impact on zeolite crystallisation. Of these variables, the temperature and time employed are the most obvious––this is because of the metastable nature of zeolites, see Fig. 3. However, the raw material sources (purities, reactivities), order of mixing and the agitation employed can also have major effects. A simple explanation can be provided for the impact that may be observed for variations in raw material sources and/or order of mixing. We have seen that zeolites are metastable species and that many of the known topologies actually have very similar compositions (e.g. LTL and MOR). Bearing this in mind it can be seen that if a specific raw material is employed that has a slower rate of dissolution (for example for silica) then the initial

222

J.L. Casci / Microporous and Mesoporous Materials 82 (2005) 217–226

gel may not have the ‘‘target’’ composition but may be deficient in silica thus providing a lower effective silica–alumina ratio. If the rate of dissolution of the silica is slow compared to the induction period for a phase formed from a lower silica–alumina ratio then the target phase may not be produced. Similarly, if a viscous gel is produced from the reaction mixture but mixing is inadequate or, indeed, absent then an inhomogeneous reaction mixture may result with pockets of gel having different compositions each ‘‘pocket’’ acting like a ‘‘mini-reactor’’ and generating phases corresponding to the composition in that mini-reactor. 3.4. Seeding and reactor ‘‘sterilisation’’ The purpose of this section is to discuss means of avoiding reactor ‘‘memory’’ effects. Essentially the objective is to have each reaction proceed in a manner as if a new reactor was being employed each time. Memory effects may be associated with bulk or trace residues of previous reaction mixtures or products being present in the reactor. These materials may reside on the walls or stirrers as bulk material or in microscopic ‘‘crevices’’ in the reactor walls. A number of solutions may be employed: • Employ new reactors (or reactor liners) for each run. • Clean/decontaminate/sterilise the reactor after use. The former approach is only really applicable to certain small-scale synthesis usually using glass or plastic reactors (at low temperatures) and, consequently, cleaning is the normal solution. The type of cleaning employed depends on the phases being prepared and the extent of the problem (some target product zeolites are more susceptible than others). In most cases a wash with caustic is sufficient (e.g. 1 M NaOH at 160 C for 1 h). However, for some products it is necessary to employ either multiple treatments or an acid wash (in addition to or instead of the caustic wash) employing HNO3 or HF. Great care must be taken with all acid and alkali solutions (especially HF). Nee et al. [21] describe an extensive cleansing procedure as part of their work on the synthesis of zeolites in the absence of alkali metal cations including the use of a ‘‘scavenging reaction’’.

• Scale of operation. Applications can vary from 10 kg in some fine chemical batch syntheses to 100 + tes in a bulk refining/petrochemical process. • Final physical form. Zeolites may be provided as the ‘‘as synthesised’’ powder or a complex shaped particle containing a binder. • Product sophistication. This can vary dramatically with some applications demanding, for example, control on crystal size and morphology (so that some channels are running in a particular direction) and composition (both bulk and in terms of the aluminium gradient across crystals). A complete review of these aspects is obviously outside the scope of this paper that will, instead, concentrate on more general aspects. 4.2. Process schematic Fig. 5 contains a generalised schematic for zeolite manufacture. It relates to a material that requires post-synthesis treatments (calcination and ion-exchange) and an extruded final physical form. Also highlighted in Fig. 5 are some of the environmental impact points: calciner off-gas scrubbing and aqueous effluent treatment. Overall, the schematic is not atypical for a commercial zeolite catalyst, indeed, some commercial materials require even more sophisticated treatments. The schematic does, however, illustrate that zeolite synthesis is a multi-step process only one of which is the crystallisation of the zeolite. 4.3. Raw materials Kuhl [22] provides a good introduction to source materials for zeolite synthesis. For the generalised reaction composition: Raw Materials Gel make-up Crystallise Filter/Wash

Off-gas scrubbing

Dry Calcine Ion Exchange

Filter/Wash

4. Scale-Up

Effluent Treatment

Dry

4.1. General comments and context

Binder/Plasticiser Form (Extrude)

At the outset of this paper comments were made on the end-use or purpose of the (zeolite) material under study. This impacts most strongly when scale-up is considered. Examples of the key drivers are:

Dry Calcine

Fig. 5. Generalised schematic of zeolite (catalyst) manufacture.

J.L. Casci / Microporous and Mesoporous Materials 82 (2005) 217–226

SiO2 : Al2 O3 : M2 O : R2 O : H2 O þ Seed The dominant component, by weight, is water. For most syntheses de-mineralised water is used to avoid the complicating effect of ions such as calcium. Good practice dictates that, whether at lab-scale or full commercial scale, the water added in other components is taken into account in the overall composition. This water may be ‘‘added’’ as part of a hydrate e.g. Al(NO3)3 Æ 9H2O or as physisorbed water on, for example, silica. For the sources of silica and alumina there are four key features that dominate choice of raw materials: (i) (ii) (iii) (iv)

Cost. Purity. Availability. Ease of handling/use.

For aluminium the choice tends to be between: • an aluminium salt e.g. aluminium nitrate (Al(NO3)3 Æ 9H2O), • sodium aluminate: solid (as a hydrate) or solution. Aluminium nitrate has the disadvantage of being deliquescent and the fact that its use means that a salt (e.g. sodium nitrate) is effectively being added to the reaction mixture. The tendency to pick-up water means that the composition is uncertain and must be checked before use to ensure that the correct amount of aluminium is being added. Sodium aluminate can be obtained as a solid and a solution. There is a tendency for Fe contamination (brownish colour to solutions) and the materials can age generating insoluble alumina species. It is suggested that batches, once obtained, are analysed for water content and Na/Al ratio (this is not 1.0 but tends to be about 1.3) so that the water and sodium can be taken into account when calculating the reagent quantities. There are a variety of silica sources available and these are summarised in Table 5. These materials vary in their cost, reactivity, purity and their ease of handling. In the view of the author, colloidal silica (sodium stabilised) provides the best combination for larger scale synthesis (of templated) zeolites for catalytic applications.

223

Base, particularly NaOH, is readily available and will not be discussed further. Templates, when used, can be problematic. If used as the hydroxide (rather than halide) care should be taken to analyse for alkali metals (used to regenerate the ionexchange resin employed in the manufacture). While a number of templates are available in commercial quantities, many are not and have to be manufactured. Some of the organic syntheses (quaternisations) are in aqueous solution and it is worth considering whether the resulting material can be taken ‘‘as is’’ rather than being separated and purified––since these are expensive steps. 13C NMR is a useful means of characterising products and identifying impurities. Seeds, when employed, usually takes the form of ‘‘as synthesised’’ material. Note the use of ‘‘as synthesised’’ is particularly important for templated zeolites since template-free materials may well react in the alkaline solution and become ineffective as seed. Since the seeds is added at 1–5%w/w of the framework components (SiO2 and Al2O3), the source of the seeds may actually be from a preparation carried out at smaller scale. 4.4. Economic considerations––volume–time yield Zeolite syntheses tend to be batch wise processes with limited work describing continuous crystallisations [23]. Many such syntheses require elevated temperatures (and pressures) using expensive autoclaves. There is a desire to maximise output per reactor and in any analysis the total cycle time (of the batch process) must be considered, in which the following steps are included: • • • • • •

reagent charging, warm-up (to temperature), time at temperature (reaction time), cool-down, discharge, sterilisation or wash-out.

There is complexity involved in any such optimisation and individual plants will have different ‘‘optima’’ depending on availability of equipment, space and the specifics of the materials being manufactured. For example, gel make-up could be carried out inside the

Table 5 Examples of typical silica sources Silica source

Formula

Comments

Sodium water glass Precipitated silica Colloidal silica

0.3Na2O Æ SiO2 Æ 7H2O SiO2 SiO2

Fumed silica TEOS

SiO2 Si(OC2H5)4

Cheap. Low levels of Al Cheap. Contains NaCl and low levels of Al More expensive. Usually available with SiO2 contents 30–40%w/w. Stabilised with Na or NH4 More expensive. Very pure. Very low levels of Al and Cl More expensive. Very pure. Must be hydrolysed generating an aqueous ethanol solution

224

J.L. Casci / Microporous and Mesoporous Materials 82 (2005) 217–226

reactor (by adding components and using the reactors agitation system to mix and form the gel) or in an external vessel. To increase the output for a given reactor there are two basic principles: (i) Decrease the crystallisation time. (ii) Increase yield per syntheses. To decrease the crystallisation time (bearing in mind comments above concerning total cycle time) there are three main methods: • Increase temperature (see previous comments on Arrhenius equation). • Use ‘‘seed’’––this shortens or removes the induction period for many zeolites. • Tune the composition (silica–alumina ratio, base level, template level etc.). Of these, temperature and seeding are most commonly applied since there are often other drivers for not altering the composition. To increase the yield per run it is most usual to look to reduce the overall water content. It should be noted that altering the base level (pH) can increase the yield (see previous comments on the equilibrium model of Lowe [9]) but there may be other reasons why this would not be employed (silica–alumina ratio of bulk and/or aluminium gradients through the crystal). Reduction in water content can be achieved for many syntheses–– ultimately it impacts on the product purity either directly (compositional field changed such that another zeolite is favoured) or indirectly through the viscosity of gel and the ability to stir/mix and achieve a homogeneous reaction mixture. At the limit of water reduction so-called ‘‘dry syntheses’’ can be carried out [24–26]. 4.5. Engineering aspects 4.5.1. Materials of construction Materials of construction (MOC) is an important consideration in any plant design and specialists in engineering and metallurgy will provide the necessary guidance and advice. The importance of MOC cannot be overstated––at the limit it impacts on reactor integrity where major failures can occur due to, for example, stress corrosion cracking. Long term corrosion is more than just an inconvenience because of the effect on: • plant operability/reliability, • product contamination. The solution to such issues is to employ the correct metallurgy and laboratory scale work can play a major part in this by providing (long term) operating experi-

ence using specific metals (usually via so-called ‘‘coupon testing’’). The correct metallurgy is heavily dependent on the nature of the mineraliser (and solvent) with OH and HF requiring different materials of construction. The presence of salts (at these pH values) must also be taken into account. Obviously at the large scale some laboratory ‘‘solutions’’ are not practical––from the use of disposable reactors to disposable reactor liners. While lined reactors are employed (glass to rubber to Teflon to ceramic) across industry (not specifically for zeolites) the intention is to provide a ‘‘permanent’’ barrier rather than a transient and disposable reactor type. Finally, when considering materials of construction the requirements of reactor ancillaries also need to be considered from reactor internals to lines to pumps. 4.5.2. Mixing/agitation While many laboratory scale synthesis (especially ‘‘micro-synthesis’’) experiments are carried out without agitation for large scale synthesis mixing is essential in all but extreme cases. While a recent paper [27] explores a new mixing process to minimise shear, in general studies on mixing are poorly reported. Broadly, the role of agitation in zeolite synthesis is not well understood, in part, because mixing has to perform a number of different functions throughout the course of a crystallisation: • • • • • • •

Reagent dissolution. Initial gel formation. Maintaining a homogeneous gel. Assisting with gel structure break-up. Maintaining uniform temperature across the reactor. Transferring ‘‘nutrients’’ to growing crystals. Keeping the zeolite crystals in suspension on completion of the reaction.

It is evident that the large viscosity difference that (usually) occurs during the course of a zeolite crystallisation is at the centre of this issue: the reactive gel may be very viscous whereas on completion of the crystallisation the liquor is much closer to water in terms of its viscosity. While agitation may play a number of different roles, in almost all literature reports it is standard practice for a single agitator configuration to be used (including stirrer speed) for all aspects of the reaction. In most reports the choice of agitator, and configuration, is often arbitrary and has not been studied even to the extent of knowing whether the configuration results in vortex formation in the reactor. There is a wide diversity of impeller types available from helical ribbons to slow paddles to high shear mixers. Indeed a recent paper [27], describes a helical ribbon or Archimedes screw type agitator for zeolite synthesis. In ideal circumstances the choice of agitator for lab-scale work should be based on the configuration to be involved at the full-scale. For the student wishing

J.L. Casci / Microporous and Mesoporous Materials 82 (2005) 217–226

Vd Baffles

Vh Id

X

where Vd Vh Id X

= = = =

And

= 0.3 – 0.4Vd = 0.4 Vh = Vh i.e. an impeller for every Vd in height (such that a vessel with Vh/Vd of 2 would have 2 impellers)

Id X Vd

Vessel diameter Vessel height Impeller diameter Impeller height above reactor base

The impeller should be a four-blade 45o pitched-paddle rotated to give down-thrust. The baffles should have a diameter ca. 0.05 – 0.1Vd

225

preparation and scale-up. The large volumes of current literature should not blind researchers to the value, and need, of looking at seminal texts and exploring other related disciplines such as chemical engineering and geology. Information on the thermodynamics of zeolites is readily available and the various approaches all provide useful information: some fundamental so aiding understanding and some practical. Deconstructing a zeolite synthesis mixture allows an appreciation of the roles of individual components and the key reaction variables. Above all each project is influenced by two key factors: (i) the application for which the zeolite is being prepared, (ii) the Environmental, Health and Safety issues.

Fig. 6. Suggested impeller configuration for zeolite synthesis.

Acknowledgment to know more about agitation they can be presented with a considerable amount of complex literature with a useful text being provided by Harnby et al. [28]. The purpose of this section is to provide some guidance on agitator configuration. It is suggested that a useful starting point is a configuration as shown in Fig. 6. No specific stirring speed can be suggested since it depends on the actual reaction mixture being studied. In general it is useful (if the reactor can be observed) to increase stirring speed until vortexing is observed then reduce it until the surface is ‘‘flat’’. This configuration can employed over a wide range of scales. For scale-up, however, while the geometry can be scaled it is (usually) not practical to simply employ the same agitator speed at different scales of operation. To scale agitator speed the simplest method is to keep agitator tip-speed constant. A more rigorous approach [28] is to keep the power input constant (i.e. power/volume constant). Power input may be determined from the agitator directly or calculated. The calculation is somewhat complex in that it is a function of the agitator (dimensions, speed), vessel (dimensions) and fluid (density, viscosity), see below: Power inputðper unit volumeÞ ¼ fnðN ; IdÞ  ðVd; VhÞ  ðq; lÞ where N, Id, Vd and Vh are as defined in Fig. 6 and l is the viscosity and q is the density of the fluid.

5. Conclusions This paper has attempted to demonstrate the complex and multi-disciplinary approach required for zeolite

The author acknowledges various colleagues especially Jose Lopez-Merino, Ian Robson, Alex Antonini and Steve Huntley and thanks Johnson Matthey for permission to publish.

References [1] H. Robson, Micropor. Mater. 22 (1998) 525. [2] P.C. Elmes, Chem. Soc. (Special publication) 33 (1980) 423. [3] Ch. Baerlocher, W.M. Meier, D.H. Olson, Atlas of Zeolite Framework Types, fifth revised ed., Elsevier, Amsterdam, 2001. [4] Available from: . [5] L.B. McCusker, F. Liebau, G. Engelhardt, Micropor. Mesopor. Mater. 58 (2003) 3. [6] D.W. Breck, Zeolite Molecular Sieves, Wiley-Interscience, New York, 1974. [7] T.S. Bowers, R.G. Burns, Am. Miner. 75 (1990) 601. [8] For simulated annealing––see W.H. Press, B.P. Flannery, S.A. Teukolsky, W.T. Vetterling, Numerical Recipes, Cambridge University Press, Cambridge, 1986; For coordination sequence––see D.E. Akporiaye, G.D. Price, Zeolites 9 (1989) 321. [9] B.M. Lowe, Zeolites 3 (1983) 300. [10] (a) See for example: K.-J. Chao, T.C. Tasi, M.-S. Chen, I. Wang, J. Chem. Soc. Faraday Trans. 77 (1981) 547; (b) R. Mostowicz, L.B. Sand, Zeolites 2 (1982) 143. [11] E.C. Moloy, L.P. Davila, J.F. Shackelford, A. Navrotsky, Micropor. Mesopor. Mater. 54 (2002) 1. [12] (a) See for example D.E. Akporiaye, I.M. Dahl, A. Karlsson, R. Wendelbo, Angew. Chemie. Int. Ed. 37 (1998) 609; (b) R. Wendelbo, D.E. Akporiaye, A. Karlsson, M. Plassen, O. Vistad, EuropaCat VI, 2003. [13] J.L. Casci, A. Stewart, in: A. Corma, F.V. Melo, S. Mendioroz, J.L.G. Fierro (Eds.), 12th International Congress on Catalysis, Studies in Surface Science and Catalysis, vol. 130, Part A, Elsevier, Amsterdam, 2000, p. 851. [14] A.E. Persson, B.J. Schoeman, J. Sterte, J.E. Otterstedt, Zeolites 14 (1994) 557.

226

J.L. Casci / Microporous and Mesoporous Materials 82 (2005) 217–226

[15] C.S. Tsay, A.S.T. Chiang, Micropor. Mesopor. Mater. 26 (1998) 89. [16] J.L. Guth, H. Kessler, R. Wey, New Developments in Zeolite Science and Technology, Elsevier, Amsterdam, 1986, pp. 121–129. [17] J.E. Hazm, P. Caullet, J.L. Paillaud, M. Soulard, L. Delmotte, Micropor. Mesopor. Mater. 43 (2001) 11. [18] J.L. Casci, in: J. Weitkamp, H.G. Karge, H. Pfeifer, W. Ho¨lderich (Eds.), Zeolites and Related Microporous Materials: State of the Art 1994, Studies in Surface Science and Catalysis, vol. 84, Part A, Elsevier, Amsterdam, 1994, p. 133. [19] (a) D.M. Bibby, L.P. Dale, Nature 317 (1985) 157; (b) D.M. Millar, M. Garces, in: M.M.J. Treacy, B.K. Marcus, M.E. Bisher, J.B. Higgins (Eds.), Proceedings of the 12th International Zeolite Conference, Material Research Society, Warrendale, 1998, p. 1535; (c) C.J. Bronic´, B. Subotic´, M. Sˇkreblin, Micropor. Mesopor. Mater. 28 (1999) 73. [20] R. Kumar, P. Mukherjee, R.K. Pandey, P. Rajmohanan, A. Bhaumik, Micropor. Mesopor. Mater. 22 (1998) 23.

[21] J.R.D. Nee, B.M. Lowe, J.L. Casci, Zeolites 14 (1994) 610. [22] G. Kuhl, Micropor. Mater. 22 (1998) 515. [23] P.M. Slangen, J.C. Jansen, K. van Bekkum, G.W. Hofland, F. van der Ham, G.J. Witkamp in: M.M.J. Treacy, B.K. Marcus, M.E. Bisher, J.B. Higgins (Eds.), Proceedings of the 12th International Zeolite Conference, Material Research Society, Warrendale, 1998, p. 1553. [24] A.K. Patel, L.B. Sand, in: J.R. Katgen (Ed.), Molecular Sieves 2, 4th International Conference, ACS Symposium Series, vol. 40, American Chemical Society, Washington, 1977, pp. 207– 218. [25] P.R. Hari, C.A. Leony Leon, K. Ueyama, M. Matsukata, Micropor. Mesopor. Mater. 21 (1998) 305. [26] A. Arnold, S. Steuernagel, M. Hunger, J. Weitkamp, Micropor. Mesopor. Mater. 62 (2003) 97. [27] C. Bebon, D. Colson, B. Marrot, J.P. Klein, F. Di Renzo, Micropor. Mesopor. Mater. 53 (2002) 13. [28] N. Harnby, M.F. Edwards, A.W. Nienow, Mixing in the Process Industries, Butterworth Heinemann, Oxford, 2000.