Clays and Clay Minerals, Vol. 34, No. 5,597-603, 1986.

METHYL t-BUTYL ETHER (MTBE) PRODUCTION: A COMPARISON OF MONTMORILLONITE-DERIVED CATALYSTS WITH A N ION-EXCHANGE RESIN
J. M. ADAMS, K. MARTIN, R. W. MCCABE, AND S. MURRAY Edward Davies Chemical Laboratories, University College of Wales Aberystwyth SY23 lINE, United Kingdom Abstract--Montmorillonite-based catalysts were compared with an acidic ion-exchange resin of the type used industrially for the production of methyl t-butyl ether (MTBE) from methanol and isobutene or t-butanol. When 1,4-dioxan was used as solvent, Al3+-exchanged montmorillonites had about half the efficiency of the resin Arnberlyst 15 at 60~ they were, however, about twice as efficient at this temperature at Ti3+-montmorillonite or K10, a commercially available acid-treated bentonite. Montmorillonite exchanged with Chlorhydrol solutions to give interlayer [AI1304(OH)2(H20)12]7+ ions and pillared clays derived from such materials were poor catalysts, as was K306, a more drastically acid-treated bentonitebased commercial catalyst. Freeze-drying of the AP+-day before reaction to produce a more open, porous structure had no effect on its catalytic efficiency. The activation energy for the reaction of isobutene and methanol in dioxan was 44 El/mole for an AP+-clay catalyst compared with 25 kJ/mole for reactions catalyzed by Ambedyst 15. With no solvent (as in industrial processes), the rates of reaction were considerably slower for both the clay- and resin-catalyzed reactions. As has been found previously for resin-catalyzed reactions using stoichiometric amounts or an excess of methanol, the rate was proportional to the isobutene concentration, and the rate-determining step appeared to be protonation of the alkene. The performance of the AP+-clay catalyst was increased by reducing the water content of the clay. In most reactions the clay catalysts were equilibrated at 12% relative humidity. Exposure of the clay to a low vacuum (10-1 ton-) before use increased its catalytic activity from 50 to 60% of that of Amberlyst 15. Key Words--Amberlyst 15, Catalysis, Chlorhydrol, Ion-exchange resin, Methyl t-butyl ether, Montmorillonite. Moreover, methods have been described for the acidtreatment o f montmorillonite-based catalysts and the addition o f water to i m p r o v e their ability to catalyze the production of M T B E (see e.g., Gregory and Westlake, 1982, 1983). In the present work the activity of several catalysts derived from a montmorillonite are compared with that of the macroreticular sulphonic acid resin A m berlyst 15 (Anonymous, 1967) which is similar to the acidic resin used industrially for the production o f MTBE (Ancillotti et al., 1977). Ancillotti et al. (1977, 1978) and Gicquel and Torck (1983) showed that at m e t h a n o l : i s o b u t e n e ratios reasonably close to unity the pore spaces o f Ambeflyst 15 resin are filled with methanol, which acts as the solvent at the reaction site in the formation o f MTBE. The methanol molecules pick up protons from the resin and become the catalytic agents. The reaction shows first-order dependence on isobutene concentration; thus the rate determining step appears to be alkene protonation,

INTRODUCTION Methyl t-butyl ether (MTBE) is an effective octane booster for unleaded gasolines (Csikos et al., 1975; Pecci and Floris, 1977; Nishizawa et al., 1974; Watanabe et al., 1973). It also acts as a v o l u m e extender, by adding volume to the gasoline pool and by reducing the severity of the naphtha reforming and related conversion operations (Muddaris and Pettman, 1980). Industrial processes for the production o f M T B E are based on liquid phase reaction o f methanol with isobutene over an acidic ion-exchange resin (Gupta and Prakash, 1980). Most o f the processes are carried out between 35 ~ and 90~ lower temperatures favor m o r e complete conversion and longer catalyst lives but lower reaction rates. The process is conducted at the lowest pressures compatible with keeping the C4 feedstocks in the liquid state (100-200 psi). Ion-exchanged montmorillonites are acidic catalysts (Mortland and Raman, 1968), and Bylina et al. (1980) and Ad am s et al. (1982) reported that they can catalyze the production o f M T B E from methanol and isobutene or from methanol and t-butanol (Adams et al., 1981) as follows: M e O H + CH2=C(C[--I3)2 --* Me-O--(2(CH3)3 M e O H + HOC(CH3) 3 ~ Me-O-C(CH3)3 + HaP. Copyright 9 1986, The Clay Minerals Society

+ CH2=C\cH3 +/CH3 -~ CH3OH + C H 3 - C \ , CH3 to form the tertiary carbocation. 597

598

Adams, Martin, McCabe; and Murray

Clays and Clay Minerals

T h e m a j o r i t y o f t h e catalysts s t u d i e d i n t h e p r e s e n t i n v e s t i g a t i o n were r e l a t e d to A P + - m o n t m o r i l l o n i t e w h i c h is k n o w n to b e h i g h l y acidic ( A d a m s et al., 1983a, 1983b) a n d i n c l u d e A P + - m o n t m o r i l l o n i t e , montmorillonite exchanged with the polycation [ A l t 3 0 4 ( O H ) z 4 ( H z O ) l z ] 7+, a p i l l a r e d ( A 1 2 0 3 ) c l a y ( V a u g h a n a n d Lussier, 1980; V a u g h a n et aL, 1979, 1981), a n d t w o c o m m e r c i a l a c i d - t r e a t e d clays, K 1 0 a n d K306 (Sfid-Chemie AG). In addition, reactions with Ti3+-montmorillonite, a material having very high acidity for s u c h catalysts ( M o r i k a w a et aL, 1983), were also i n v e s t i g a t e d . All o f t h e s e c l a y - b a s e d c a t a l y s t s (inc l u d i n g K 1 0 a n d K 3 0 6 ) were d e r i v e d f r o m clay f r o m t h e s a m e source. EXPERIMENTAL

Catalyst characterization
The bentonite used for most of this study came from Moosburg, Federal Republic of Germany (Tonsil 13, Siid-Chemie AG). The <2-um clay fraction was prepared by sedimentation. An X-ray powder diffraction (XRD) examination showed that this fraction was ~ 98% montmorillonite with ~ 2% quartz impurity. An X-ray fluorescence (XRF) analysis gave 64.0% SiO2, 20.8% A1203, 3.86% Fe203, 0.07% TiO2, 1.5% CaO, 1.42% MgO, 0.04% K20, and 1.20% Na20. Loss on ignition was 7.1%. The cation-exchange capacity totaled 103 meq/100 g of air-dry material and consisted of 68, 33, and 2 meq/100 g of Na +, Ca 2+, and Mg 2+, respectively. Montmorillonite from Upton, Wyoming, was used to study the dependence of reaction rate on concentration ofisobutene and methanol. Here also, sedimentation was carried out to separate the <2-urn fraction.

which suggested a significant collapse of the pillars between the clay layers.~ For all catalysis experiments, material calcined at 200~ was employed. This catalyst contained 3.78% interlayer A1, compared with the 5.2% calculated if all interlayer exchange sites were filled with AI~3 polycations having 7 + charge. The surface area was 229 m2/g. To prepare Ti3+-montmoriUonite, 2.5 g of the Na+-mont morillonite was suspended in 100 cm 3 of deionized water, to which was added 5 cm 3 of 30% TIC13 solution (BDH). The suspension was stirred for 30 min, after which it was heated to 70"C and stirred for an additional 30 rain before centrifugation, resuspended i n 250 cm 3 of hot deionized water, and centrifuged again. A vacuum desiccator containing silica gel was used to dry and store the pale blue clay. Titration with permanganate indicated that 90% of the cation-exchange sites contained Ti 3+ and that the clay was stable over many weeks when not exposed to the atmosphere. Cation exchange of the Wyoming montmorillonite was carried out using 0.7 M AIC13 solution for 24 hr, followed by washing to remove excess exchanging salt in a manner similar to that employed for the Tonsil 13 clay.

Reaction conditions and analysis

In some experiments reactions were carried out under reflux; however, the majority of experiments were performed in sealed, Teflon-lined, brass vessels that were fitted with sampling ports and contained a magnetic stirrer. Known weights of catalyst, pre-equilibrated at 12% relative humidity (RH), were used for each reaction. Normally 0.5 g of catalyst was used, but for reactions involving the effect of variation in isobutene/metbanol ratios, only 0.1 g was employed. For reactions between methanol and t-butanol, the alcohols were added to the reaction vessel only a few seconds after the clay catalyst was added, so as to minimize exposure of the clay to atmospheric moisture. In most experiments using isobutene as the reagent, the clay, methanol and, if necessary, 1,4-dioxan solvent were added to the vessel, which was then sealed. The isobutene was then liquefied and transferred to the reaction vessel using a syringe. The amount of isobutene was calculated from the volume and density (0.85 g/cm3). In some experiments in which low concentrations of isohutene were employed, a saturated solution ofisobutene in 1,4-dioxan was used as the reagent. The solubility limit at room temperature and pressure was found to be ~ 1 mmole isobutene/ cm 3 of dioxan. The reactions were monitored by diluting 50-ul aliquots with 0.5 cm 3 of 1,4-dioxan and storing them in polytetrafluorethylene sealed vials prior to analysis by gas-liquid chromatography (GLC). GLC measurements were made with a 2.7-m glass column of 0.25-inch i.d. packed with 2% polyethylene glycol (PEG 20 M) plus 10% silver nitrate on Chromasorb P. RESULTS AND DISCUSSION

Catalyst preparation
To aid later cation-exchange processes, a large quantity of the Tonsil 13 clay was exchanged initially with Na +, The clay was added to an 0.1 M solution of NaCI and heated to 70~ with vigorous stirring for 1 hr. Two fractions were produced by sedimentation. The finer fraction was subjected to ultrasonic treatment for 10 min, and the <2-urn fraction was collected by sedimentation. AP+-montmorillonite was prepared by exposing some of this <2-~m fraction to an 0.3 M A1C13 solution (pH = 2.7) for 24 hr. The clay was then centrifuged and resuspended in deionized water repeatedly until a test for excess salt proved negative. At this point the pH of the super'natant liquid was 4.3. The clay was isolated by centrifugation, dried at 40~ and mechanically ground. The basal spacing of this clay was 14.8 ~ and the aluminum content was 0.98% higher than before aluminum exchange--a value correspondng to a cation-exchange capacity (CEC) of 109 meq/100 g if the aluminum is assumed to be present as A13+ monomers. This value is in reasonable agreement with the measured CEC (103 meq/100 g). Some of the <2-/zm Na+-montmorillonite was exchanged with a cationic oxyhydroxyaluminum species, thought to be [AlI304(OH)24(HzO)1217+,after the manner of OcceUi and Tindwa (1983) using Chlorhydrol (ACH) solutions (Reheis Chemicals Company). The basal spacing of this clay was 18.6/~. A pillared clay was produced by following example 1 in Vaughan et al. (1979). After the cation-exchange steps using ACH solutions the clay was dried in an oven overnight at 86~ After cooling, the basal spacing of this clay was 18.6 ~. Calcining the material at 200~ for 15 rain reduced the spacing slightly to i7.7/~, but calcining at 400"C resulted in a broad peak extending from 5 ~ to 6~ (CoKa radiation),

Reactions under reflux ( M e O H + t-BuOH)

In preliminary reactions the ratios of the two liquid reagents, m e t h a n o l a n d t - b u t a n o l , were v a r i e d w h i l e k e e p i n g t h e a m o u n t o f c a t a l y s t fixed a n d t h e t o t a l v o l u m e c o n s t a n t . T h e yield i m p r o v e d i f a n excess o f t - b u t a n o l was e m p l o y e d (Figure 1). T h i s excess w a s necessary, b e c a u s e , in a d d i t i o n to t h e f o r m a t i o n o f M T B E ,

Comparable experiments carried out using the Wyoming montmorillonite Volclay-200 showed no such collapse; the X-ray powder diffraction spacing decreased slightly to 16.8/~.

Vol. 34, No. 5, 1 9 8 6

Methyl-t-butyl ether production using a montmorillonite catalyst

O

599

2.0

2.0

1.5
O

1.5

yield (rnmole)

yie{d (mrnote)

0.5

0.5

30 time{rnin)

60

90

30

60

90

tirnelrnin)
Figure 2. Comparison of catalytic efficiency of some derivatives of the montmorillonite from Moosburg, Federal Republic of Germany (Tonsil 13). Reactions were under reflux (60~ with a mole ratio ofMeOH : t-BuOH of 1:2; 1,4-dioxan was used as solvent. 9 = A13+-montmorillonite; I = Catalyst K10; 9 = Pillared clay; X = Catalyst K306;/x = ACH-montmorillonite.

Figure 1. Yield of methyl t-butyl ether from different ratios of methanol and t-butanol over AP+-montmorillonite. Reactions were carried out under reflux. Total volume was constant, as was total number of moles of the reactions; 1,4dioxan was used as solvent. MeOH : tBuOH ratios: O = 100: 200 mmole;/x = 150:150 mmole; x7 = 200:100 mmole.

t-butanol may undergo acid-catalyzed dehydration to isobutene (and water), after which the isobutene would be lost from the reflux apparatus. A series of reactions were then carried out to c o m pare the efficiency of the montmorillonite catalysts for the reaction of alcohols to give MTBE. The a m o u n t of catalyst used in each reaction, the methanol to t-butanol ratio, and the total volume were kept constant. All catalysts were pre-equilibrated at 12% R H to ensure reproducibility. The AP+-montmorillonite was by far the best catalyst for the reaction (Figure 2). The commercial catalyst K10 had an activity which was c o m parable to, but less than that of the AP+-clay, confirming the previous work of Adams et al. (1982) who used a montmorillonite from Wyoming. The K10 catalyst is a clay which has been acid treated to some degree (Siid-Chemie, 1985) leaving exsolved AP + species in the interlayer region. Its slightly lower activity is probably due to the fact that the catalyst contains material other than montmorillonite (Stid-Chemie, 1985), i.e., no size-fractionation was performed on the original bentonite material. The K306 catalyst is derived from the Moosburg bentonite by extensive, acid treatment until almost no f montmorillonite structure remains (Hojabri, 1971). Its lack of activity is therefore not surprising (Adams et al., 1982). Figure 2 also shows that the ACH-clays and pillared clays had low activity. The ACH-clays were

probably relatively inactive because the globular polycations left little free space in the interlayer region, thereby blocking access for reactant molecules. Alternatively, the inherent acidity of the materials might simply be less. To improve the activity of the pillared clays, a procedure recommended by Vaughan et al. (1981)-- treatment with a m m o n i u m hydroxide-- was followed to increase the cation-exchange capacity of the samples. Repeat reactions showed virtually no increase in catalytic activity. This lack of activity of pillared clays in low temperature, acid-catalyzed reactions producing amines, esters, and ethers was also noted by Diddams et al. (1984).
Reactions in sealed vessels ( M e O H + t - B u O H )

Reactions were carred out using differing ratios of methanol : t-butanol and either stoichiometric amounts or an excess of methanol. The solvent was 1,4-dioxan. The rate of production of MTBE was found to be directly proportional to the concentration of t-butanol. Comparable reactions were then performed using AP +clay and Amberlyst 15 catalysts. The Amberlyst 15 resin was somewhat more efficient (Figure 3) and gave a yield of MTBE of 47% after 40 m i n at 80~ compared with 18% for the clay catalyst. The activation energy was also lower for the resin-catalyzed reaction: 23 _ 5 kJ/mole, compared with 85 _ 13 kJ/mole.

600

Adams, Martin, McCabe, and Murray

Clays and Clay Minerals

% 0
yield

A A

A A--

/
A o

_
o

rut e

(mmo[~i

3[,~
0

,
l

0
B

30
tirne(min)

B O
O O

90

2 3 isob utene (mmo[e)

o
10
yield

6
Fete

(tamale)

3
o

51

~t
2

6 MeOH(remote)

lO

30 time(rain)

60

90

Figure 3. Comparison of activity of (A) AP+-montmorillonite and (B) Amberlyst 15 for reaction of methanol and t-butanol to give methyl t-butyl ether and water. Both catalysts were pre-equilibrated at 12% RH. Reactions were carried out in sealed vessels using comparable quantifies of catalyst, methanol, and t-butanol; 1,4-dioxan was used as solvent. Temperatures: I = 62", x7 = 70", O = 81 ~ zx = 90~

Figure 4. (A) Effect on the rate of reaction of variation in concentration of isobutene at 70"C with excess methanol, 0.1 g of AP+-clay catalyst, 0.4 cm 3 methanol (= 10 mmole) and ~ 3 cm 3 1,4-dioxan were used (total volume = constant). (B) Effect on the rate of reaction of variation in concentration of methanol at 70"C. 0.1 g of AP catalyst was used with 2.5 mmole isobutene and 0.5-10 mmole methanol Total volume was kept constant; ~ 3 cm 3 of 1,4-dioxan was used as solvent. In both parts of figure the rates are in arbitrary units.

Reactions in sealed vessels ( M e O H + isobutene)

T o study the d e p e n d e n c e o f the rate o f f o r m a t i o n o f M T B E on m e t h a n o l and isobutene concentrations, reactions were p e r f o r m e d with A P + - e x c h a n g e d m o n t morillonites at 70~ In each e x p e r i m e n t , d i o x a n was used as solvent. W i t h excess m e t h a n o l , the rate was directly p r o p o r t i o n a l to the isobutene c o n c e n t r a t i o n (Figure 4a) as has been found for resin-catalyzed reactions (Ancillotti et al., 1978; G i c q u e l a n d Torck, 1983). A t a fixed c o n c e n t r a t i o n o f isobutene, the rate o f M T B E p r o d u c t i o n rose with the m e t h a n o l concentration, but, after this c o n c e n t r a t i o n reached that o f the isobutene, the rate b e c a m e constant (Figure 4b). T h i s b e h a v i o r is distinctly different f r o m that o b s e r v e d for A m b e r l y s t 15 resin (Ancillotti et al., 1978; G i c q u e l and Torck, 1983), where the rate increased rapidly until the m e t h a n o l c o n c e n t r a t i o n was a b o u t one tenth o f t h a t o f the isobutene. A t higher m e t h a n o l concentrations, the rate decreased rapidly until, with an excess o f m e t h a nol, it b e c a m e i n d e p e n d e n t o f m e t h a n o l c o n c e n t r a t i o n (Ancillotti et al., 1978). In the present work, it was

found that for b o t h types o f catalyst, w h e n m e t h a n o l was in excess, as it is in industrial processes, the rate was i n d e p e n d e n t o f m e t h a n o l concentration. A t l o w e r concentrations o f m e t h a n o l the rate was e n h a n c e d w h e n the A m b e r l y s t 15 catalyst was used. A n c i l l o t t i et aL (1978) suggested that u n d e r these c o n d i t i o n s the alkene is p r o t o n a t e d directly by the SO3H groups o f the resin, which are m o r e acidic t h a n s o l v a t e d p r o t o n s (Gates and Rodriguez, 1973). F o l l o w i n g the type o f p r o c e d u r e a d o p t e d by Purnell and Ballantine (1984) a reaction s c h e m e can be constructed, as s h o w n below:
KI

A(1) = A(i)
K2

MeOH(1) -~ M e O H ( i )
K3 +

M e O H ( i ) + H + = MeOH2(i)
+ K4

MeOH2(i) + A(i) ~ M e O H ( i ) + AH+(i)

K5

M e O H ( i ) + AH+(i) ~- MTBE(i) + H +
K6

M T B E ( i ) ~ MTBE(1),

Vol. 34, No. 5, 1 9 8 6

Methyl-t-butyl ether production using a montmorillonite catalyst

601

15 10
yield

(mmote~
U ~

15IB
yield 10

2O 40 time(rain)

(mm~ t
r

time(rain) Figure 5. Comparisonof the activity of (A) A13+-mont morilloniteand (B) Amberlyst 15 for reaction of methanol and isobutene. Reactions were carried out in sealed vessels using comparable quantities of catalyst, methanol, and isobutene; 1,4-dioxan was used as solvent. Temperatures: + = 52~ I = 63", x7 = 72~ O = 82~ zx = 93~

2'0

&O

rates were only about one third of those found for A1~+clays at ~ 60~ Two further modifications were also made to the basic A13+-montmorillonite-catalyzedreactions. First, an attempt was made to improve the reactivity of the clay by addition of concentrated HC1 at the rate given by Gregory and Westlake (1982), i.e., 2% of the weight of the reactants. The rate of the reaction was reduced (rather than increased) to about two thirds of that produced by AP+-clays alone. Second, the A13+-clay was freeze dried rather than dried in an oven before equilibration at 12% RH, in order to produce a material with a much more open, porous macrostructure (Weiss et al., 1952; Norrish and Raussell-Colom, 1962) which might, therefore, produce increased reaction rates? No improvement in efficiency was noted, however, which confirms that in this reaction, if the clay layers were held apart by solvent molecules, access of alcohol and alkene to the active site was not rate determining. These findings contrast with other higher temperature reactions where macroporosity leads to enhanced diffusion characteristics (Pinnavaia et al., 1984). Some of the freeze-dried clay was used directly (rather than after equilibration to 12% RH). This material, which had been produced using a low v a c u u m ( 1 0 - ' torr) had a 20% greater activity due to greater acidity displayed by montmorillonites of low water content (Mortland and Raman, 1968).

Solvent effects
All the reactions studied above were carried out in 1,4-dioxan, a solvent that has been found to be efficacious in m a n y clay catalyzed reactions involving tertiary carbocation intermediates. The beneficial effects are thought to be due to promotion of miscibility of reagents (Adams et aL, 1983a). To make the comparison between clay catalysts and ion-exchange resin as close as possible, 1,4-dioxan was also used in Amberlyst-catalyzed reactions, even though no solvent is used in industrial processes. The activation energy reported here for Amberlyst-catalyzed reactions is considerably lower than that reported when no solvent was employed, i.e., 25 kJ/mole compared with 70-80 kJ/mole (Ancillotti et al., 1977, 1978; Gicquel and Torck, 1983). MTBE production is apparently highly solvent dependent, regardless of the type of catalyst. In previous MTBE formations over clay catalysts, Adams et al. (1982) found no correlation between the basal spacing of a particular M "+-montmorillonite/solvent system and its catalytic ability. In the present investigation, the range of solvents was extended to include carbon tetrachloride and toluene. As can be

where A represents the alkene, isobutene, and 1 and i indicate the external liquid phase or an intercalated species respectively. Based on the results obtained using different ratios of isobutene and methanol (Figure 4), step (4) in this scheme is rate limiting. The possibility that dioxan could be the proton-carrying species involved in the protonation of the alkene in reactions where it was used as solvent can be discounted at the concentration ratios used in this work in as much as the PKA of dioxan is 2 units lower than the pKA of methanol. Reactions were carried out using A13+-clay catalysts and Amberlyst 15 resin over the temperature range of interest. Here also, the Amberlyst 15 was more efficient than the clay catalyst, e.g., it gave a 29% yield after 20 rain at 62~ compared with 14% for the clay (Figure 5). Compared with previous experiments using the alcohols, however, the relative efficiency of the clay was slightly greater. The activation energy of the clay-catalyzed reaction was again higher than that found for the resin; i.e., 44 _+ 3 kJ/mole compared with 25 + 13 kJ/mole. Reactions were also carried out under the same conditions using Ti3+-montmorillonite, but the

2 The basal spacing of this freeze-dried, equilibrated (12% RH) clay was the same as for the oven-dried equilibrated clay.

602

Adams, Martin, McCabe, and Murray

Clays and Clay Minerals

Table 1. Effectof solvent upon the rate of production of MTBE from methanol and isobutene.
Reactions reported in this paper t Solvent Rate (% yield/rain)
Reactions reported by Adams et aL (1982) 2

Solvent

Rate (% yield/rain)

None Carbon tetrachloride Toluene

0.280 0.062 0.08

1,4-dioxan

1.55

1,2-dimethoxyethane Diethylene glycol diethyl ether n-pentane Tetrahydropyran N-methylmorpholine Tetrahydrofuran 1,4-dioxan

0.013 0.007 0.002 0.018 0.002 0.003 0.252

Reactions were carried out at 62~ and used 0.5 g of AP+-montmorillonite(12% RH). Where solvents were employed, quantities of reagents used were 2 cm3 methanol, 1.7 cm 3 isobutene, and 6.3 cm 3 solvent. When no solvent was employed, 5 cm 3 methanol and 4.25 cm 3 of isobutene were used. 2 Reactions below were carried out at 60~ and used 1 g of A13+-montmorillonite (Wyoming) (12% RH). Quantities of reagents were 1 cm3 methanol, 3 cm 3 solvent, and 18-58 mmole isobutene, i.e., 1.2-3.8 cm 3, seen from the data in Table 1, 1,4-dioxan was the most efficient solvent studied. It is not immediately obvious why dioxan should be so useful a solvent. Adams et al. (1982) suggested that miscibility of reagents and solvent is important and that dioxan, water, methanol, and isobutene are m u tually soluble in the concentration ranges employed here. This is, however, not the whole story inasmuch as they also found evidence of specific cation/solvent pairing. Addition of dioxan to the reactive mixture should reduce the concentration of methanol somewhat in the interlayer space of the clay. As reported above, however, if methanol was in excess, the most significant factor affecting the reaction rate was the concentration ofisobutene in the interlayer of the clay. This concentration in turn depended on the isobutene concentration in the solvent or in methanol if no solvent was employed. The solubility of isobutene in methanol and dioxan was found to be 0.06 and 0.09 g/cm 3, respectively, at room temperature and pressure. CONCLUSIONS The catalytic activity of a range of montmorillonitebased catalysts for low-temperature MTBE production is broadly in line with their expected acidities, i.e., A13+-clay > acid treated-clay (K10) >> ACH-treated clay ~ pillared clay. The activity of A13+-montmorilloniteat 60~ is ~ 50% of that exhibited by an acidic ion-exchange resin similar to that used industrially. The reaction in the clay does not appear to be diffusion controlled, at least if 1,4-dioxan is used as the solvent; increasing the macroporosity in the clay had no effect on reaction rate. The activation energy using such a catalyst is higher than when using the resin, i.e., 44 and 25 kJ/mole, respectively, if 1,4-dioxan is used as the solvent. For both clay and resin catalysts, the use of 1,4-dioxan significantly lowers the activation energy from that observed when other (or no) solvents are employed. This decrease is probably due to the fact that 1,4-dioxan is mutually miscible with all reactants and products in the concentration range studied. Moreover, the solubility of isobutene in dioxan is greater than that in methanol. ACKNOWLEDGMENTS We thank English Clays Lovering Pochin and Co. Ltd. (ECLP) and the British Technology Group for support. Stid-Chemie AG, Munich, kindly donated the samples of Tonsil 13 and K-catalysts. ECLP also analyzed the sedimented clays by X-ray fluorescence and X-ray powder diffraction. REFERENCES Adams, J. M., Clapp, T. V., and Clement, D. E. (1983a) Catalysis by montmorillonites: Clay Miner. 18, 411-421. Adams, J, M., Clement, D. E., and Graham, S. H, (1981) Low temperature reaction of alcohols to form t-butyl ethers using clay catalysts: J. Chem. Res., $254-255. Adams, J. M., Clement, D. E., and Graham, S. H. (1982) Synthesis of methyl-t-butyl ether from methanol and isobutene using a clay catalyst: Clays & Clay Minerals 30, 129-134. Adams, J. M., Clement, D. E., and Graham, S. H. (1983b) Reactions of alcohols with alkenes over an aluminum-exchanged montmorillonite: Clays & Clay Minerals 31, 129136. Ancillotti, F., Masi Mauri, M., and Pescarollo, E. (1977) Ion exchange resin catalyzed addition of alcohols to olefins: J. Catal. 46, 49-57. Ancillotti, F., Massi Mauri, M., Pescarollo, E., and Romagnoni, L. (1978) Mechanisms in the reacti~ between ~ fins and alcohols catalyzed by ion-exchange resins: J. Mol. Catal. 4, 37-48. Anonymous (1967) Amberlyst 15 Synthetic Resin Catalyst: Rhom and Haas Co., Philadelphia, Pennsylvania, 4 pp. Bylina, A., Adams, J. M., Graham, S. H., and Thomas, J. M. (1980) Chemical conversions using sheet silicates: simple method for producing methyl t-butyl ether: J. Chem. Soc. Chem. Comm., 1003-1004. Csikos, R., Laky, J., Pallay, I., and Vajta, L. (1975) Ether additives for gasoline: Ger. Often. DE 2, 444, 528, 4 pp. Diddams, P. A., Thomas, J. M., Jones, W., Ballantine, J. A., and Purnell, J.H. (1984) Synthesis, characterization, and catalytic activity of beidellite-montmorillonitelayered silicates and their pillared analogues: aT. Chem. Soc. Chem. Comm., 1340-1342,

Vol. 34, No. 5, 1986

Methyl-t-butyl ether production using a montmorillonite catalyst

603

Gates, B. C. and Rodriguez, W. (1973) General and specific acid catalysis in sulphonic acid resin: J. Catal. 31, 27-31. Gicquel, A. andTorck, B. (1983) Synthesis ofmethyl tertiary butyl ether catalyzed by ion-exchange resin. Influence of methanol concentration and temperature: J. Catal. 83, 918. Gregory, R. and Westlake, D. J. (1982) Method of promoting the activity of cation-exchangeable layered clay and zeolite catalysts in proton-catalyzed reactions: European Patent Application 0045618A2, 17 pp. Gregory, R. and Wesflake, D. J. (1983) Use of stabilized pillared inteflayered clays as catalysts in reactions capable of catalysis by protons: European Patent Application 0083970A1, 23 pp. Gupta, J. C. and Prakash, J. (1980) MTBE technology: Chem. Eng. World, 15, 27-31. Hojabri, F. (1971) Dimerization of propylene and its uses for isoprene manufacture: J. AppL Chem. Biotech. 21, 8789. Morikawa, Y., Wang, F., Moro-oka, Y., and Ikawa, T. (1983) Conversion of methanol to low molecular weight hydrocarbons over Ti ion-exchanged form of layered silicate minerals: Chem. Letters, 965-968. Mortland, M. M. and Raman, K. V. (1968) Surface acidity of smectites in relation to hydration, exchangeable cation, and structure: Clays & Clay Minerals 16, 393-398. Muddaris, G. R. and Pettman, M. J. (1980) Now, MTBE from butane: Hydrocarbon Processing Int. Ed. 59, 91-95. Nishizawa, T., Tokumaru, T., Kamiyama, Y., and Watanabe, Y. (1974) Gasoline composition with high octane number: Japan Kokai 74, 26, 306, 44 pp. Norrish, K. and Raussell-Colom, J.A. (1962) Effect of freezing on the swelling of clay minerals: Clay Min. Bull. 5, 916.

Occelli, M. L. and Tindwa, R. M. (1983) Physiochemical properties of montmorillonite interlayered with cationic oxyaluminum pillars: Clays & Clay Minerals 31, 22-28. Pecci, G. and Floris, T. (1977) Ether ups anti-knock of gasoline: Hydrocarbon Processing Int. Ed. 56, 98-102. Pinnavaia, T. J., Tzou, M. S., Landau, S. D., and Raythatha, R . H . (1984) On the pillaring and determination ofsmectite clay catalysts by polyoxocations of aluminium: J. Mol. Catal. 27, 195-212. Purnell, J. H. and Ballantine, J. A. (1984) Competitive intercalation and intercalary kinetics: J. Mol. Catal. 27, 169178. Siid-Chemie AG (1985) K-Catalysts: Sfid-Chemie AG, Munich, Federal Republic of Germany, p. 2. Vaughan, D. E. W. and Lussier, R . J . (1980) Preparation of molecular sieves based on pillared interlayered clays (PILC): inProe. 5th Int. Conf. Zeolites, Naples, 1980, L. V. C. Rees, ed., Heyden, London, 94-101. Vaughan, D. E. W., Lussier, R. J., and Maggee, J. S. (1979) Pillared interlayered clay materials useful as catalysts and sorbents: U.S. Patent 4,176,090, 9 pp. Vaughan, D. E. W., Lussier, R. J., and Maggee, J. S. (1981) Pillared interlayered clay products: U.K. Patent Application GB 2,059,408A, 5 pp. Watanabe, Y., Kobayashi, J., and Nishizawa, T. (1973) Modifying agents raising octane number of gasoline: Japan Kokai 73,23,803, 4 pp. Weiss, A., Fahn, R., and Hofmann, U. (1952) Skeletal structure in thixotropic gels: Naturwissenschaften 89, 351-352.

(Received 31 May 1985; accepted 19 February 1986; Ms. 1491)