Alcohol Dehydrations over ZSM-5 Type Zeolites, Montmorillonite Clays and Pillared Montmorillonites

Jour nal of the Chi nese Chem i cal So ci ety, 2000, 47, 509-518 509 Alcohol Dehydrations over ZSM-5 Type Zeolites, Montmorillonite Clays and Pillared Montmorillonites Hung-en Lin ( ) and An-Nan Ko* ( ) Department of Chemistry, Tunghai University, Taichung, Taiwan, R..C. The re ac tions of aliphatic al co hols (eth a nol, 1-propanol and 2-propanol) were stud ied at 1 atm and 150-300 C by us ing ZSM-5 type zeolites, montmorillonites, and pil lared montmorillonites. With H-ZSM-5 (X) Y zeolites, the to tal num ber of acid sites in creases with a de crease of Si 2/Al 2 3 mo lar ra tio (X) and cal cin ing tem per a ture (Y). In ad di tion, ap par ent in crease in the ra tio of strong to weak acid sites oc curs with in creas ing X or de creas ing Y. The acid i ties of M-ZSM-5 (51) 600 zeolites fol low the se quence: Li Na > K > Cs. Pil lared clays ex hibit both larger sur face ar eas and more acid sites than the clays. The al co hol con v er sions de crease in the or der of 2-propanol > 1-propanol > eth a nol, in ac cor dance with the rel a tive sta bil i ties of the cor re spond ing carbenium ions. The cat a lytic ac tiv i ties are par al lel to the to tal num ber of cat a lyst acid sites and the re ac tion tem per a ture. Fa vor able for ma tions of ethers are ob served at low re ac tion tem per a ture and small con tact time on a cat a lyst with weak acid strength. Sim ple con sec u tive re ac tions and com bined path way of par al lel and con - sec u tive re ac tions are pro posed, re spec tively, for the de hy dra tion of eth a nol and for those of 1-propanol and 2-propanol. INTRDUCTIN The cat a lytic re ac tions of aliphatic al co hols have been in ten sively stud ied by us ing a va ri ety of cat a lysts pos sess ing acidic and/or ba sic cen ters. 1-20 Among the al co hols stud ied, the con ver sion of 2-propanol has re ceived con sid er able at ten - tion; it is of ten uti lized as a model re ac tion to probe cat a lyst acid-base prop er ties. 21 The de hy dra tion prod ucts, ethers and olefins, are pro duced on acid sites, whereas the dehydro - genation prod ucts, al de hydes or ke tones, are formed via ba sic sites. In the con ver sion of 2-propanol over al kali cat ion ex - changed zeolites, most of the 2-propanol was con verted to pro pyl ene and wa ter over Li and Na zeolites while both de hy - dra tion and dehydrogenation oc curred on K, Rb and Cs zeolites. 2 With metal ox ide and mixed metal ox ides, cat a lytic ac tiv ity of 2-propanol was par al lel to sur face acid ity for both ether and ole fin for ma tion. 6 The de hy dra tion of 2-propanol over a Sn 2-AlP 4 cat a lyst was cor re lated to the sur face acid den sity and was struc tur ally in sen si tive. 19 The ac tiv ity was en hanced with in creas ing sur face acid den sity and in sen si tive to the acid strength of sil ica-alumina cat a lysts. 22 In the de hy - dra tion of 2-propanol by lan tha num-y ze o lite, the ac ti va tion en er gies were 128.5 and 139.5 kj/mol for the for ma tion of diisopropyl ether and pro pyl ene, re spec tively. 17 In the re ac - tion of meth a nol, eth a nol, 2-propanol and t-butanol on AlP 4, the cat a lyst ac tiv ity in creased as a func tion of sur face acid ity as well as the -sub sti tu tion in the al co hol. 13 For the al co hol de com po si tion, the re ac tion rates were in the or der 1-propanol > eth a nol > meth a nol on Al 2 3, but the rates were al most iden - ti cal on Ni/Al 2 3. 11 En hanced cat a lytic per for mances in eth a - nol de hy dra tion were ob served by desilication and sub se quent steam treat ment of H-ZSM-5 ze o lite. 23 In the re ac tion of eth a - nol to eth yl ene and di ethyl ether over Mg-Al ba sic ox ides, pure Al 2 3 ex hib ited the high est de hy dra tion ac tiv ity. 24 Me dium pore sized H-ZSM-5 zeolites are novel cat a - lysts for their prop er ties of acid ity as well as shape se lec tiv ity. Con se quently, they have at tracted much at ten tion in both ac a - demic re search and in dus trial ap pli ca tion. 25 The acid ity of ZSM-5 zeolites can be al tered by vary ing the Si 2 /Al 2 3 mo - lar ra tio, 26 the cal cin ing tem per a ture 27 and al kali cat ion ex - change. Ac cord ing to the TPD pro files of am mo nia from H-ZSM-5 zeolites, both con cen tra tions of strong and weak acid sites in creased with the de crease of the Si/Al ra tio. 26 Ther mal treat ment of ZSM-5 type cat a lysts de creased the Brönsted and/or Lewis acid sites. 27,28 Al though mont mo ril lo - nite clays were used in the pe tro leum in dus try be fore the ap - pli ca tion of zeolites, the montmorillonites pil lared with var i - ous metal ox ides have been de vel oped dur ing the past two de - cades. Af ter cat ion ex change of so dium mont mo ril lo nite (Na-M), the acid ity is al tered. Ad di tionally, in ter ca la tion of Na-M clay with metal ox ide re sults in ap par ent in crease of sur face area, interlayer dis tance and sur face acid ity, which is fa vor able for cat a lyz ing larger mol e cules. 29 De spite the fact that many re search pro jects have been car ried out on the cat a lytic re ac tion of al co hols, there were dif fer ent re sults re ported in the lit er a ture. For the de hy dra tion of al co hols, the cat a lytic ac tiv ity and the prod uct se lec tiv ity de pend on the na ture of al co hols, the struc ture and acid ity of

510 J. Chin. Chem. Soc., Vol. 47, No. 3, 2000 Lin et al. cat a lysts, and the re ac tion con di tions. In this study, three types of cat a lysts were uti lized in clud ing ZSM-5 zeolites, mont mo ril lo nite clays and pil lared montmorillonites. Var i - ous H-ZSM-5 zeolites hav ing dif fer ent acidic prop er ties were ob tained by chang ing Si 2/Al 2 3 mo lar ra tio, cal cin ing tem - per a ture and al kali cat ion ex change. The acid i ties of clays and pil lared clays were al tered by var i ous cat ion ex change and by dif fer ent pillaring spe cies, re spec tively. Sys tem atic in ves ti - ga tion was per formed on the cat a lytic de hy dra tion of both pri - mary and sec ond ary al co hols, i.e., eth a nol, 1-propanol and 2-propanol, in or der to gain de tailed in sight into the cor re la - tions of cat a lytic re sults with cat a lyst prop er ties. In par tic u - lar, the re ac tion paths and as so ci ated mech a nisms of var i ous al co hols over H-ZSM-5 ze o lite are com pared and dis cussed. EXPERIMENTAL SECTIN Cat a lyst Preparation H-ZSM-5 zeolites (PQ corp.)were cal cined at 400, 500, 600 and 800 C. The sym bol H-ZSM-5 (X) Y de notes H- ZSM- 5 ze o lite with Si 2/Al 2 3 mo lar ra tio X that was cal - cined at Y C. Al kali ion ex changed ZSM-5(51), viz. M- ZSM-5(51)-600, were pre pared as fol lows: 10 g of H- ZSM- 5(51) were cal cined at 400 C for 24 h, cooled to room tem per - a ture in a des ic ca tor, and then added with stir ring into 200 ml of 1 M al kali metal chlo ride so lu tion at 60 C. The stir ring was con tin ued for 8 h. The so lu tion was fil tered and washed with deionized wa ter un til there was no de tect able chlo ride ion in the fil trate. Then the sam ple was air dried, heated at a rate of 6 C/min to 110 C for 12 h, then with the same heat ing rate to 600 C, and fi nally cal cined for 5 h. Cat ion ex changed mont mo ril lo nite was pre pared by add ing slowly 50 g so dium mont mo ril lo nite (Kunipia-F, Kunimine In dus tries Co., Ltd.) into 500 ml deionized wa ter and stir ring for 24 h. Then the so lu tion was di a lyzed in deionized wa ter un til its con duc tiv ity was less than 10-1 cm -1. Finally the sam ple was dried at 110 C in an air oven and ground into powder. Alu mina pil lared mont mo ril lo nite was pre pared as fol - lows: so dium hy drox ide so lu tion was added dropwise at a rate of 2 ml/min into a 0.1 M alu mi num chlo ride so lu tion. The fi - nal mo lar ra tio H /Al was 2.0. The re sult ing so lu tion was aged at 60 C for 12 h be fore the re quired amount (4 mmol Al/g Na-M) of so lu tion of the alu mi num polyoxycation [Al 13 4(H) 24(H 2) 7+ 12 ] 29 was added into the Na-M so lu tion. Af ter air dry ing in a des ic ca tor, the sam ple was dried at 110 C for 5 h and cal cined at 500 C for 5 h to ob tain alu mi num pil - lared mont mo ril lo nite (Al-PM). For the prep a ra tion of chromia pil lared mont mo ril lo nite (Cr-PM), 58.9 g Na 2C 3 were added with rig or ous stir ring into a 5.6 l so lu tion con tain - ing 222.5 g Cr(N 3) 3 9H 2. The re sult ing so lu tion was stirred for 3 h, aged at 95 C for 36 h and cooled to 35 C. Then 10 g of Na-M were added into this so lu tion. The re sult - ing so lu tion was then fil tered, dried and heated at a rate of 1 C/min to 200 C for 3 h. Cat a lyst Char ac ter iza tion The BET sur face ar eas of the cat a lysts were mea sured with a sorp tion an a lyzer (Quantachrome Quantasorb). The pow der x-ray dif frac tion pat terns of the sam ples were ob - tained from a XRD spec trom e ter (Shimadzu R3A). The catalyst acid i ties were de ter mined by the tem per a tureprogrammed desorption of am mo nia ac cord ing to the pro - ce dures re ported in the lit er a ture. 30 Cat a lytic Reaction Re ac tion runs were car ried out in a fixed-bed, in te - gral-flow re ac tor (1.1 cm I. D. 32 cm) at 150-300 C un der at mo spheric pres sure. Prior to the re ac tion, the cat a lyst was ac ti vated in flow ing air at 400 C for 1 h and then ad justed to the re ac tion tem per a ture with flow ing ni tro gen gas. Re agent grade eth a nol (E. Merck), 1-propanol (R. D. H.) or 2-propanol (R. D. H.) were fed into the re ac tor via a micropump (Razel A-99) to gether with ni tro gen gas of flow rate 100 ml/min. For the re ac tion of di ethyl ether, its sat u rated va por at 25 C was car ried with flow ing ni tro gen gas (16 ml/min) into the re - ac tor. The prod ucts were col lected with a pre heated gas sam - ple bulb and pe ri od i cally an a lyzed with a G. C. (HP 5890 se - ries II) equipped with a PNA col umn (Supelco 0.2 mm 50 m) and a flame-ionization de tec tor. The prod ucts were also iden ti fied with a GC-MS (Fiscons). RESULTS AND DISCUSSIN The catalyst prop er ties For the sam ples of H-ZSM 5 (X) Y, ei ther in crease of X or Y causes only a slight in crease in the rel a tive peak in ten - si ties of 2, i.e., 7.8 and 8.7 but no vari a tion in the peak po si - tion as ob served from the XRD di a grams. As H-ZSM-5 was ion ex changed with a al kali metal cat ion, the XRD pat terns ex - hibit a slight de crease in all peak in ten si ties, in di cat ing small dim i nu tion of crystallinity. The d 001 spac ings are 17.7 and 18.2 Å for Al-PM and Cr-PM, re spec tively. The cor re spond - ing val ues of interlayer dis tance are 8.1 and 8.6 Å. Ta ble 1 in di cates the phys i cal prop er ties of the cat a lysts. The BET sur face ar eas of H-ZSM-5 sam ples re main sta ble with re spect to ther mal treat ment. Fur ther more, only a small vari a tion of BET sur face ar eas oc curs af ter ionic ex change of

Al co hol Dehydrations over Zeolites J. Chin. Chem. Soc., Vol. 47, No. 3, 2000 511 Table. 1. Physical Properties of Various Catalysts Catalyst BET surface area (m 2 /g) Total acidity (mmol/g) Strong acidity (mmol/g) Weak acidity (mmol/g) Strong acidity Weak aciidity H-ZSM-5(31)400 416 1.12 0.54 0.58 0.90 H-ZSM-5(51)400 413 0.84 0.45 0.39 1.15 H-ZSM-5(75)400 393 0.65 0.38 0.27 1.41 H-ZSM-5(31)500 430 1.00 0.51 0.49 1.04 H-ZSM-5(51)500 423 0.81 0.42 0.39 1.08 H-ZSM-5(75)500 430 0.53 0.30 0.23 1.30 H-ZSM-5(31)600 401 0.70 0.26 0.44 0.59 H-ZSM-5(51)600 378 0.46 0.19 0.27 0.70 H-ZSM-5(75)600 342 0.42 0.22 0.20 1.10 H-ZSM-5(31)800 399 0.25 0.083 0.17 0.49 H-ZSM-5(51)800 411 0.16 0.054 0.11 0.49 H-ZSM-5(75)800 389 0.17 0.066 0.10 0.66 Li-ZSM-5(51)600 383 0.44 0.16 0.28 0.57 Na-ZSM-5(51)600 390 0.45 0.17 0.28 0.61 K-ZSM-5(51)600 396 0.26 0.083 0.18 0.46 Cs-ZSM-5(51)600 426 0.24 0.084 0.16 0.53 K-M 59 0.073 0.037 0.036 1.03 Al-M 34 0.11 0.057 0.053 1.08 Cr-M 14 0.094 0.041 0.053 0.77 Al-PM 235 0.23 0.092 0.14 0.66 Cr-PM 253 0.76 0.46 0.30 1.53 H-ZSM-5 (51) with al kali cat ions. How ever, af ter pil lared with alu mina or chromia, both Al-PM and Cr-PM ex hibit much larger sur face ar eas that are con sis tent with the lit er a - ture. 29 Fig. 1 il lus trates a typ i cal ex am ple of the tem per a - ture-programmed desorption pro file of am mo nia from H-ZSM-5 (X) 500 sam ples. Two peaks at 250 and 480 C with a di vid ing min i mum at 350 C were ob served, im ply ing ad sorp tion on weak and strong acid sites, re spec tively, as ob - served pre vi ously. 26,28,31 How ever, three peaks in the TPD pro file of am mo nia were also re ported; 32 the peak cor re spond - ing to me dium acid sites was not de tected in our ex per i ment. As the cal cin ing tem per a ture in creases up to 600 and 800 C, only one broad peak ap pears, ex cept for H-ZSM-5 (75) 600, in di cat ing ap par ent de cline of the strong acid sites. This is in ac cor dance with pre vi ous IR stud ies on H-ZSM-5 (45) cal - cined at 500 C 28 that have shown three bands due to C group vi bra tions at 3610, 3670 and 3745 cm -1. Cal ci na tion at 800 C al most caused dis ap pear ance of the band at 3610 cm -1 as - signed to acid H groups, im ply ing the di min ish ing of strong Brönsted acid sites. The M-ZSM-5 (51) 600 sam ples also dis - play a broad desorption curve. The cat a lyst acid ity was ob - tained from the peak area of TPD plots, in as so ci a tion with cal i bra tion data. For the sam ples of H-ZSM-5 (X) Y, the to tal num ber of acid sites in creases with de creas ing the val ues of Si 2/Al 2 3 mo lar ra tio (X) or cal cin ing tem per a ture (Y). Fig. 1. TPD pro files of am mo nia from H-ZSM-5 (X) 500 zeolites.

512 J. Chin. Chem. Soc., Vol. 47, No. 3, 2000 Lin et al. These find ings are in agree ment with other re ports. 26-28 The acid i ties of M-ZSM-5 (51) 600 zeolites fol low the or der: Li Na > K > Cs. With clays and pil lared clays, these de crease in the or der of Cr-PM > Al-PM > Al-M Cr-M > K-M. Pil laring of Na-M clay with alu mina or chromia re sults in re mark able in crease in cat a lyst acid ity as was also re ported else where. 29 The strong and weak acid ity were es ti mated ac cord ing to the peak area above and be low 350 C in the TPD pro file, re spec - tively. Ta ble 1 lists the to tal acid ity, the strong and weak acid - ity as well as the ra tio of strong to weak acid ity for var i ous cat - a lysts. Ap par ent de crease in the ra tio of strong to weak acid sites oc curs with in creas ing cal cin ing tem per a ture Y and de - creas ing Si 2/Al 2 3 mo lar ra tio X of H-ZSM-5 zeolites. Eth a nol de hy dra tion Un der our ex per i men tal con di tions, eth a nol was con - verted into de hy dra tion prod ucts: di ethyl ether and eth yl ene; eth a nol dehydrogenation to acetaldehyde was not de tected. Fig. 2 shows cat a lytic re sults of eth a nol re ac tion at 300 C, W/F 0.45 g h/mol and 2 h pro cess time over var i ous cat a lysts. For H-ZSM-5 (X) Y zeolites, an in crease of X or Y re sults in the low er ing of cat a lytic ac tiv ity. In the case of M-ZSM-5 (51) 600, the eth a nol con ver sions are in the or der: Li > Na > K > Cs. The pil lared montmorillonites, Al-PM and Cr-PM, ex - hibit much higher ac tiv i ties than montmorillonitee clays, K-M, Al-M and Cr-M. Fig. 3 shows the in flu ence of cat a lyst to tal acid ity on eth a nol con ver sion. A lin ear in crease of eth a - nol con ver sion with the to tal num ber of acid sites was ob - served. How ever, the mont mo ril lo nite clays and pil lared clays are much less ef fec tive in cat a lyz ing de hy dra tion re ac - tion of eth a nol as com pared to H-ZSM-5 (X) Y zeolites due to their smaller sur face ar eas and dif fer ent struc tures. Ad di - tionally, the higher acid ity of pil lared clays is re flected by the better cat a lytic ac tiv ity. Among clays and pil lared clays, K-M clay ex hib its the high est se lec tiv ity of di ethyl ether be cause of its low acid ity. The prod uct se lec tiv ity dis played in Fig. 2 re - veals that di ethyl ether is fa vor ably formed on ZSM-5 cat a - lysts with weak acid ity such as H-ZSM-5 (X) 800 and M- ZSM- 5 (51) 600 (Ta ble 1); the selectivities of di ethyl ether on these cat a lysts are larger than 83%, whereas the op po site trend was ob served for eth yl ene se lec tiv ity. Sim i lar find ings were ob served in the re ac tion of eth a nol on desilicated HZSM-5 33 and of 1-butanol on AlP 4. 15 Fig. 4 shows the ef fect of re ac - tion tem per a ture on the cat a lytic re sults over H-ZSM-5 (51) Fig. 2. Cat a lytic re sults of eth a nol re ac tions over var i - ous cat a lysts. Re ac tion con di tions: 300 C; W/F 0.45 g h/mol; Pro cess time 2 h. ( ) eth a nol con - ver sion; ( ) eth yl ene se lec tiv ity; ( ) di ethyl ether se lec tiv ity. Fig. 3. In flu ence of cat a lyst to tal acid ity on eth a nol con ver sions. Re ac tion con di tions: the same as Fig. 1. (µ) H-ZSM-5 (X) 400; ( ) H-ZSM-5 (X) 500; ( ) H-ZSM-5 (X) 600; ( ) H-ZSM-5 (X) 800; ( ) M-ZSM-5 (51) 600; ( ) clays and pil - lared clays.

Al co hol Dehydrations over Zeolites J. Chin. Chem. Soc., Vol. 47, No. 3, 2000 513 500 ze o lite. As the re ac tion tem per a ture in creases from 200 to 300 C, the eth a nol con ver sion also in creases from 20 to 80% whereas the di ethyl ether se lec tiv ity de creases from 100 to 10%. Here se lec tiv ity var ies very sharply with tem per a ture. Con se quently, re ac tion tem per a ture is an im por tant fac tor in con trol ling the prod uct se lec tiv ity of this re ac tion. In the re ac - tion of eth a nol over AlP 4 cat a lysts, both eth a nol con ver sion and eth yl ene se lec tiv ity were en hanced when the re ac tion tem per a ture in creased from 250 to 400 C; how ever, di ethyl ether al ways pre dom i nated. 13 To in ves ti gate the re ac tion path ways on H-ZSM-5 (51) 500, the mo lar ra tio eth yl ene/di ethyl ether as a func tion of con tact time, W/F, is shown in Fig. 5. As the val ues of W/F ap - proach zero, the eth yl ene/di ethyl ether mo lar ra tio also be - comes close to zero. Such a re sult im plies that di ethyl ether and eth yl ene are pri mary and sec ond ary prod ucts, re spec - tively. To fur ther con firm the pro duc tion of eth yl ene from di - ethyl ether, the re ac tion of di ethyl ether was fol lowed un der the same ex per i men tal con di tions. The di ethyl ether con ver - sion and eth yl ene se lec tiv ity were 100% and 97%, re spec - tively. More over, the for ma tion of ether from an al co hol and the alkene was ex cluded by the re sults from deu te rium trace stud ies in the re ac tion of 2-butanol. 14 A S N2-type mech a nism was pro posed in the pro duc tion of ether from al co hol. 14,34 Thus, con sec u tive re ac tions of eth a nol are pro posed and shown as fol lows: 2 C 2 H 5 H -H 2 C 2 H 5 C 2 H 5 -H 2 2 C 2 H 4 ther re ports sug gested dif fer ent re ac tion paths, viz. sim ple par al lel re ac tion on AlP 4 and SAP mo lec u lar sieve 35 and a com bi na tion path way of par al lel and con sec u tive re ac tions over AlP 4 cat a lyst. 13 In the re ac tion of eth a nol over AlP 4, eth yl ene was pro posed as a pri mary plus sec ond ary sta ble prod uct 13 ac cord ing to the op ti mum per for mance en ve lope (PE) curves. 36 How ever, the ini tial se lec tiv ity of eth yl ene was much smaller than that of di ethyl ether. 1-Propanol dehydration Pre vi ous TPD re sults for 1-propanol on H-ZSM-5 (70) re ported that only 40% of the strongly ad sorbed al co hol desorbed as pro pyl ene and wa ter. 37 n Al 2 3, 1-propanol de - com posed to form pro pyl ene, dipropyl ether, ac e tone and hy - dro gen gas. 11 In this study, only de hy dra tion oc curs to pro - duce pro pyl ene, dipropyl ether and wa ter. Fig. 6 shows cat a - lytic re sults of 1-propanol re ac tion at 250 C, W/F 0.45 g h/mol and 2 h pro cess time over var i ous cat a lysts. With H-ZSM-5 (X) Y zeolites, in creas ing X or Y re sults in a de - crease of cat a lyst ac tiv ity. For the se ries of M-ZSM-5 (51) 600 zeolites, the 1-propanol con ver sion de creases in the or - der: Li > Na > K > Cs. The cat a lytic ac tiv i ties of clays and pil - lared clays are ap par ently lower than those of H-ZSM-5 (X) Y zeolites. How ever, the pil lared clays ex hibit better per for - mance than clays due to their larger sur face ar eas and higher acid i ties. Fig. 7 dis plays the re la tion of 1-propanol con ver - Fig. 4. Ef fect of re ac tion tem per a ture on cat a lytic re - sults of eth a nol over H-ZSM-5 (15) 500. ( ) eth a nol con ver sion; ( ) eth yl ene se lec tiv ity; ( ) di ethyl ether se lec tiv ity. Fig. 5. Eth yl ene/di ethyl ether mol ra tio as a func tion of con tact time over H-ZSM-5 (51) 500. Re ac tion con di tions: 300 C; pro cess time 2 h.

514 J. Chin. Chem. Soc., Vol. 47, No. 3, 2000 Lin et al. sion with cat a lyst to tal acid ity. The cat a lytic ac tiv ity is par al - lel to the to tal acid amount. In ad di tion, the ef fect of to tal acid ity on 1-propanol con ver sion is more pro nounced with a ZSM-5 type cat a lyst than with the clays or pil lared clays, prob a bly due to their dif fer ences in sur face area and cat a lyst struc ture. The prod uct selectivities are com pared among H-ZSM-5 (X) Y zeolites: the dipropyl ether se lec tiv ity in - creases gen er ally with a de crease of X and in crease of Y. For M-ZSM-5 (51) 600 zeolites, it fol lows the se quence Li < Na < K < Cs (Fig. 6). These re sults sug gest the fa vor able for ma tion of dipropyl ether on weak acid sites. There fore, the ef fects of cat a lyst acid ity on cat a lytic ac tiv ity and prod uct se lec tiv ity are sim i lar to those ob served in the de hy dra tion of eth a nol. Fig. 8 il lus trates the de pend ence of cat a lytic ac tiv ity and prod - uct se lec tiv ity on the re ac tion tem per a ture. Raising tem per a - ture re sults in re mark able in crease of both 1-propanol con - ver sion and pro pyl ene se lec tiv ity with con com i tant de crease of dipropyl ether se lec tiv ity. Fig. 9 shows the de pend ence of pro pyl ene to dipropyl ether mo lar ra tio on the con tact time W/F. The mo lar ra tio is Fig. 7. In flu ence of cat a lyst to tal acid ity on 1-propanol con ver sions. Re ac tion con di tions: the same as Fig. 6; Sym bols: the same as Fig. 3. Fig. 6. Cat a lytic re sults of 1-propanol re ac tions over var i ous cat a lysts. Re ac tion con di tions: 250 C; W/F 0.45 g h/mol; pro cess time 2 h. ( ) 1- propanol con ver sion; ( ) pro pyl ene se lec tiv ity ; ( ) dipropyl ether se lec tiv ity. Fig. 8. Ef fect of re ac tion tem per a ture on cat a lytic re - sults of 1-propanol over H-ZSM-5 (51) 500. ( ) 1-propanol con ver sion; ( ) pro pyl ene se lec - tiv ity; ( ) dipropyl ether se lec tiv ity.

Al co hol Dehydrations over Zeolites J. Chin. Chem. Soc., Vol. 47, No. 3, 2000 515 1.78 by ex trap o lat ing to zero value of W/F; it re fers to the ini - tial se lec tiv ity ra tio of pro pyl ene to dipropyl ether. More over, the ra tio in creases with in creas ing con tact time, im ply ing pro - pyl ene is a pri mary and sec ond ary prod uct. The re ac tion net - work is thus shown as fol lows: 2 C 3 H 7 H -H 2-2H 2 C 3 H 7 C 3 H 7 C 3 H 6 -H 2 a lysts, in agree ment with lit er a ture re ports. 2,10,12 Fig. 11 il lus - trates the cor re la tion of 2-propanol con ver sion with cat a lyst to tal acid ity. En hance ment of cat a lytic ac tiv ity with an in - crease of cat a lyst to tal acid ity is ob served, which is con sis tent with other stud ies. 19,22 Fig. 12 re veals the ef fect of re ac tion tem per a ture on cat a lytic ac tiv ity and prod uct se lec tiv ity for re ac tions over H-ZSM-5 (51) 500. Raising the tem per a ture from 150 to 250 C causes a sharp in crease of 2-propanol con - ver sion from 4.5 to 99% but only a small in crease of pro pyl - ene se lec tiv ity, i.e., 90-94%, in di cat ing only slight vari a tion of prod uct se lec tiv ity with re ac tion tem per a ture. The re ac tion path way is pro posed as fol lows: The above re ac tion path way is a com bi na tion of par al lel and con sec u tive re ac tions. 2 (CH 3 ) 2 CHH -H 2 (CH 3 ) 2 CHCH(CH 3 ) 2 -H 2 2-Propanol de hy dra tion In the con ver sion of 2-propanol, only the de hy dra tion prod ucts of pro pyl ene and diisopropyl ether are formed. Sim - i lar re sults were re ported for re ac tions over metal ox ides, 6 mixed metal ox ides, 6,12 AlP 4, 13 Sn 2 -AlP 19 4 and LaY ze o - lite. 17 Fig. 10 shows cat a lytic re sults of 2-propanol re ac tion over var i ous cat a lysts at 200 C, W/F 0.45 g h/mol and pro - cess time 2 h. n H-ZSM-5 (X) Y and M-ZSM-5 (51) 600 zeolites, the in flu ences of X or Y and al kali cat ion on the cat a - lytic ac tiv ity are less ap par ent as com pared to those in the re - ac tions of eth a nol and of 1-propanol. How ever, the pil lared clays ex hibit higher ac tiv ity than the clays. It is note wor thy that the pro pyl ene selectivities are larger than 91% for all cat - Fig. 9. Pro pyl ene/dipropyl ether mol ra tio as a func tion of con tact time over H-ZSM-5 (51) 500. Re ac - tion con di tions: 250 C; pro cess time 2 h. - 2H 2 2 C 3 H 6 Such a re ac tion net work is con sis tent with a pre vi ous study. 13 It is also sim i lar to that of 1-propanol. How ever, the main path in this re ac tion is the di rect de hy dra tion of 2-propanol to pro - pyl ene. Com par i son between cat a lytic re sults of var i ous al co hols In the com par i son of Figs. 4, 8 and 12, it is found that the cat a lytic ac tiv i ties for the de hy dra tion of al co hols fol low the or der 2-propanol > 1-propanol > eth a nol. Sim i lar re sults are also re ported else where. 11,13 To ra tio nal ize our cat a lytic re - sults of var i ous al co hols, the re ac tion mech a nisms of de hy dra - tion, partly sim i lar to ref. 34, are pro posed in Scheme I. The ad sorp tion of al co hols (HRH) on the Brönsted acid sites of ZSM-5 zeolites (ZH) yields an oxonium ion (I)-frame work an ion pair via pro ton trans fer (step 1). Such a com plex may con vert back to free al co hol (step-1) or de hy - drate via a carbenium ion in ter me di ate (step 2) to an alkene (step 3). The in ter ac tion of oxonium ion (I) with a free al co hol mol e cule, fol lowed by de hy dra tion and pro ton trans fer forms the ether (step 4). The ad sorp tion of ether on zeolites then gives an oxonium ion (II) (step 5) with sub se quent de com po - si tion to the alkene (step 6). The dif fer ences in the cat a lytic ac tiv ity among al co hols are well ex plained by the com pe ti tion be tween step -1 and step 2 due to the sta bil ity of carbenium ion (HR + ), that fol lows the se quence 2-C 3H 7 + > 1-C 3H 7 + > C 2H 5 +. Ac cord ingly, the same or der of cat a lytic ac tiv ity is found for the cor re spond ing al co - hols. The selectivities of alkene and ether de pend on the com - pet i tive re ac tion rates of step 2 and 4. For al co hol re ac tions over H-ZSM-5 (51) 500 zeolites, the rel a tive rates de crease in the or der 2-propanol > 1-propanol > eth a nol, due to the rel a - tive sta bil ity of the cor re spond ing carbenium ions. Hence, the

516 J. Chin. Chem. Soc., Vol. 47, No. 3, 2000 Lin et al. Scheme I 1 ( ¹ ^ + HRH + ZH HRH 2 + Z - 1-2 - HR + + H 2 + Z 3 R + H 2 + ZH - HRH + HRH 2 + Z 4 HRRH + H 2 + ZH 5 + HRRH 2 + H 2 + - ( ¹ ^ + Z 6 2 R + 2 H 2 + ZH R = C 2 H 4, (CH 2 ) 3, CH 2 CHCH 3 alkene selectivities ex hibit the same trend (Fig. 4, 8 and 12). The dif fer ences of re ac tion path ways among var i ous al co hols can be in ter preted as fol lows: With eth a nol, the re ac tions pro - ceed through steps 1, 4, 5 and 6 due to the un fa vor able for ma - tion of pri mary carbenium ion C 2H 5 + (step 2). Con se quently, this is a con sec u tive re ac tion. In the re ac tion of 1-propanol and 2-propanol, all steps 1 to 6 are pos si ble that lead to the com bined path ways of par al lel and con sec u tive re ac tions. For 2-propanol, how ever, step 2 and 3 are much more pre ferred as com pared to step 4 due to the higher sta bil ity of the sec ond ary carbenium ion 2-C 3H 7 +. In creasing tem per a ture causes a dra - matic in crease in the con ver sions of all al co hols be cause the re ac tion rates of var i ous steps in Scheme I are en hanced. The pos i tive ef fect of re ac tion tem per a tures on the alkene select - ivities is much more pro nounced in the case of eth a nol and 1-propanol; this re sult dem on strates that the ac ti va tion en ergy for alkene for ma tion is sig nif i cantly larger than that for ether pro duc tion. Fur ther more, the dif fer ence of these two ac ti va - tion en er gies be comes smaller in the de hy dra tion of 2- propanol. With LaY ze o lite, these two Ea val ues dif fer by 10 kj/mol. 17 Fig. 10. Cat a lytic re sults of 2-propanol re ac tions over var i ous cat a lysts. Re ac tion con di tions: 200 C; W/F 0.45 g h/mol; pro cess time 2 h. ( ) 2- propanol con ver sion; ( ) pro pyl ene se lec tiv ity; ( ) diisopropyl ether se lec tiv ity. Fig. 11. In flu ence of cat a lyst to tal acid ity on 2-propanol con ver sions. Re ac tion con di tions: the same as Fig. 10; sym bols: the same as Fig. 3.

Al co hol Dehydrations over Zeolites J. Chin. Chem. Soc., Vol. 47, No. 3, 2000 517 1-propanol and 2-propanol dehydrations. ACKNWLEDGEMENTS We thank the Na tional Sci ence Coun cil of the Re pub lic of China for fi nan cial sup port. We also thank Kunimine In - dus tries Co., Ltd. for sup ply ing Kunipia-F. Re ceived No vem ber 1, 1999. Key Words Al co hol de hy dra tion; Clays; Eth a nol; Pil lared clays; 1-Propanol; 2-Propanol; ZSM-5 zeolites. REFERENCES Fig. 12. Ef fect of re ac tion tem per a ture on cat a lytic re - sults of 2-propanol over H-ZSM-5 (51) 500. ( ) 2-propanol con ver sion; ( ) pro pyl ene se - lec tiv ity; ( ) diisopropyl ether se lec tiv ity. CNCLUSINS For the sam ples of H-ZSM-5 (X) Y, the to tal num ber of acid sites in creases with a de crease of X or Y val ues; ap par ent in crease in the ra tio of strong to weak acid sites oc curs with in - creas ing X or de creas ing Y. The acid i ties of M-ZSM-5 (51) 600 fol low the trend: Li Na > K > Cs. With clays and pil - lared clays, these de crease in the or der of Cr-PM > Al-PM > Al-M Cr-M > K-M. Un der our re ac tion con di tions, only de hy dra tion prod - ucts, ethers and alkenes, are formed. The cat a lytic ac tiv ity and the prod uct se lec tiv ity de pend on the kinds of al co hol, the cat a lyst prop er ties and the re ac tion con di tions. The al co hol con ver sions fol low the se quence: 2-propanol > 1-propanol > eth a nol, in ac cor dance with the rel a tive sta bil i ties of the cor re - spond ing carbenium ions. Lin ear re la tion ships are ob served for the cat a lytic ac tiv i ties with the to tal num ber of cat a lyst acid sites. How ever, clays and pil lared clays are less ef fec tive in cat a lyz ing al co hol de hy dra tion due to their smaller sur face ar eas and dif fer ent struc tures. Raising the re ac tion tem per a - ture en hances the al co hol con ver sions. Higher selectivities of ethers are gen er ally at tained at low re ac tion tem per a ture and small con tact time by us ing cat a lysts of weak acidic strength. n H-ZSM-5 zeolites, sim ple con sec u tive re ac tions are pro - posed for eth a nol de hy dra tion, whereas com bined path ways of par al lel and con sec u tive re ac tions are pro posed for 1. Venuto, P. B.; Landis, P. S. Adv. Catal. 1968, 18, 259. 2. Yashima, T.; Suzuki, H.; Hara, N. J. Catal. 1974, 33, 486. 3. Knozinger, H. Adv. Catal. 1975, 25, 184. 4. Noller, H.; Kladnig, W. Catal. Rev. 1976, 13, 149. 5. Wu, M. M.; Kaeding, W. W. J. Catal. 1984, 88, 478. 6. Luy, J. C.; Parera, J. M. Appl. Catal. 1986, 26, 295. 7. Tanabe, K.; Misono, M.; no, Y.; Hattori, H. in New Solid Acids and Bases, Their Cat a lytic Prop erties, Delmon, B.; Yates, J. T. Eds. Elsevier: Am ster dam, 1989. Vol. 51. 8. Gervasini, A.; Auroux, A. J. Catal. 1991, 131, 190. 9. Noller, H.; Andreu, P.; Hun ger, M. Angew. Chem. Int. Ed. Engl. 1991, 10, 172. 10. uqour, A.; Coudurier, G.; Vedrine, J. C. J. Chem. Soc. Far a day Trans. 1993, 89, 3151. 11. Chen, B.; Fal coner, J. L. J. Catal. 1993, 144, 214. 12. Lahousse, C.; Mauge, F; Bacheller, J.; Lavalley, J.-C. J. Chem. Soc. Far a day Trans. 1995, 91, 2907. 13. Campelo, J. M.; Gar cia, A.; Herencia, J. F.; Luna, D.; Ma - rinas, J. M.; Romero, A. A. J. Catal. 1995, 151, 307. 14. Shi, B.; Da vis, B. H. J. Catal. 1995, 157, 359. 15. Bautista, F. M.; Delmon, B. Appl. Catal. A 1995, 130, 47. 16. Balkus, K. J. Jr.; Khanmamedova, A. K. J. Catal. 1995, 151, 10. 17. Rudham, R.; Spi ers, A. I. J. Chem. Soc. Far a day Trans. 1997, 93, 1445. 18. Larsen, G.; Lotero, E.; Petkovic, L. M.; Shobe, D. S. J. Catal. 1997, 169, 67. 19. Mostafa, M. R.; Youssef, A. M. Ma ter. Lett. 1998, 34, 405. 20. Arias, D.; Colmenares, A.; Cubeiro, M. L.; Goldwasser, J.; Lopez, C. M.; Machado, F. J.; Sazo, V. Catal. Lett. 1997, 45, 51.

518 J. Chin. Chem. Soc., Vol. 47, No. 3, 2000 Lin et al. 21. Krylov,. V. in Ca tal y sis by Non-Metals, Ac a demic Press: New York, 1970, p. 118. 22. El-Nabarawy, Th.; Khalil, L. B.; Ha ma da, M. A.; Nawar, N. Adsorpt. Sci. Technol. 1997, 15, 125. 23. Le Van Mao, R.; Borsuk, N.; hayon, D.; Ramsaran, A.; Le, S. T.; Denes, G. Ma ter. Res. Soc. Symp. Proc. 1994, 454, 73. 24. Di Cosimo, J. I.; Diez, V. K.; Xu, M.; Iglesia, E.; Apesteguia, C. R. J. Catal. 1998, 178, 499. 25. Jacobs, P. A. Carboniogenic Ac tiv ity of Zeolites, Elsevier: Am ster dam, 1977. 26. Inui, T.; Suzuki, T.; Inoue, M.; Murakami, Y.; Takegami, T. in Struc ture and Re ac tiv ity of Mod ified Zeolites, Jacobs, P. A. et al. Eds. Elsevier: Am ster dam, 1984, pp. 201-216. 27. Rudham, R.; Winstanley, A. W. J. Chem. Soc. Far a day Trans. 1994, 90, 3191. 28. Kurschner, U.; Parlitz, B.; Schreier, E.; hlmann, G.; Volter, J. Appl. Catal. 1987, 30, 159. 29. Burch, R. Ed. Pil lared Clays, Ca tal y sis To day 1987, 2, pp. 185-367. 30. Ko, A.-N.; Yang, C.-L.; Zhu, W.-D.; Lin, H.-E. Appl. Catal. A 1996, 134, 53. 31. Reschetilowski, W.; Unger, B.; Wendlandt, K.-P. J. Chem. Soc. Far a day Trans. 1 1989, 85, 2941. 32. Topsoe, N.; Pedersen, K.; Derouane, E. G. J. Catal. 1981, 70, 41. 33. Le Van Mao, R.; Borsuk, N.; hayon, D.; Ramsaran, A.; Le, S. T.; Denes, G. Ma ter. Res. Soc. Symp. Proc. Ma te - rials Re search So ci ety, 1997, p. 73. 34. Sun, Q.; Herman, R. G.; Klier, K. J. Chem. Soc., Chem. Commun. 1995, 1849. 35. Arias, D.; Colmenares, A.; Cubeiro, M. L.; Goldwasser, J.; Lopez, C. M.; Carmen, M.; Machado, F. J.; Sazo, V. Catal. Lett. 1997, 45, 51. 36. Ko, A.-N.; Wojciechowski, B. W. Prog. Re act. Kinet. 1983, 12, 201. 37. Aronson, M. T.; Gorte, R. J.; Farneth, W. E. J. Catal. 1986, 98, 434.