вход по аккаунту


Dehydration of Alcohols on Aluminum Oxide.

код для вставкиСкачать
acetic acid to an acetic acid-polyvinyl acetate solution r26bI.
A guess at a reasonable upper limit for the solvent reorganization rate derives from the study of the dissociation of c-aminocaprolactam dimers, which occurs at a rate between lo9 and 1010 sec-1 in various
solvents [40,411. Since the reorganization of solvent
clusters under consideration in connection with
memory effects would require several such elementary
steps, the rate would be substantially lower than this
12. Conclusions
Although it is plausible to attribute some of the memory
effects to the occurrence of conformationally isomeric
cations, there are many cases in which the conformational barriers between the “twisted” and symmetrical
forms are so low that non-classical or extramolecular
causes for the observed effects must be considered.
The systems exhibiting memory effects provide uniquely suitable devices for revealing the existence of sub[40] K . Bergman, M. Eigen, and L. De Mneyer, Ber. Bunsenges.
species. The problem of defining the nature of these
intermediates is at the heart of carbonium ion chemistry. We hope that the observations reported here
will provide some stimulus to the more intensive
study of the microscopic structure of solutions, which
may provide the key to an understanding of the extramolecular group of memory effects.
The physical phenomena underlying memory effects
are very probably present whenever carbonium ions
are formed in solution. In the most general terms, one
can conclude that a description of the “structure” of a
carbonium ion in solvolytic medium is incomplete
without a specification of the means by which it is
This work has been supported in part by the National
Institute of Arthritis and Metabolic Diseases (Grant
A M 07505), the Petroleum Research Fund, the Wisconsin Alumni Research Foundation, and the National
Science Foundation, to whom we are grateful. All of the
experimental observations and many of the concepts
have been contributed by the dedicated and enthusiastic
group of collaborators whose names are given in the text
references. To work with them has been both challenge
and reward.
physik. Chern. 67, 819 (1963).
Received: December i t . 1967
[A 661 I € )
G e r m a n version: Angew. Chem. 80. 765 (1968)
[41] M . Eigen, Discuss. Faraday SOC.39, 130 (1965).
Dehydration of Alcohols on Aluminum Oxide
The dehydration of alcohols on y-aluminum oxide, which yields water, olefins, andlor
ethers, was studied with the aid of kinetic methods and I R spectroscopy. The unimolecular
olefin formation probably proceeds via a surface compound, in which an alcohol molecule
is joined by I W O angular H-bonds l o an OH group and an oxygen ion on the surface. The
ether formation (bimolecular reaction), on the other hand, requires OH groups and oxygen and aluminum ions on the surface. The ether is formed f r o m a surface alkoxide group
and molecularly adsorbed alcohol.
1. Introduction
The heterogeneous alumina-catalyzed dehydration of
alcohols has been known since the end of the 18th
century. Nevertheless, it is only recently that efforts
have been made to explain the elementary processes in
heterogeneously catalyzed eliminations, whereas the
elucidation of reaction mechanisms in the liquid phase
has reached a much more advanced stageC11. In the
[*I Priv.-Doz. Dr. H. Knozinger
Physikalisch-Chemisches Institut der Universitat
8 Miinchen 2 , Sophienstr. I t (Germany)
[l] I). V . Banfhorpe: Elimination Reactions. Elsevier, Amsterdam 1963.
Angew. Chem. internat. Edit. / VoI. 7 (1968) 1 Xu. 10
dehydration on aluminum oxide (which has been
mainly studied with ethanol as the alcohol), efforts
have often been made to find analogies with the reaction in the liquid phase. An excellent review of developments up to 1960 has been published by Winfield [21.
Widely differing views are still held concerning the
mechanism of the heterogeneously catalyzed dehydration of alcohols on aluminum oxide. In the case of aliphatic and some other alcohols the situation is complicated, particularly by the simultaneous occurrence
of a bimolecular reaction leading to ether formation.
[ 2 ] M. E. Winfield in P . H . Emmett: Catalysis. Reinhold, New
York 1960, Vol. VII, p.93.
79 1
Kinetic studies are then meaningful only if the reaction scheme is known.
An oxonium-carbonium ion mechanism similar to that
assumed for the liquid-phase reaction in acidic media
is also supported by several authorsr3-61 for the dehydration of alcohols on solid surfaces. A proton from
the surface is assumed to add to the alcohol in the adsorption step. This process can be explained by the
formation of more or less strongly polarized Hbonds 121. Taking into account the surface properties
of aluminum oxide, Pines and Huugr7J showed that
this mechanism is also valid for the dehydration of
2-butanol (cf. however, [8J). The participation of surface protons or OH groups is supported by the need
for water to be present at the contact surface[;
however, there appears to be no unambiguous relationship between the dehydrating activity and the
acidity of the surface'71. For an oxonium ion to be
formed by proton addition, the surface of the aluminum oxide must contain acid centers of the Brmsted type; Purryilll, however, found only centers of
the Lewis type in I R studies on aluminum oxide with
adsorbed pyridine. Any Brmsted acids present are
certainly only weakly acidic 1121, so that ionic addition
of protons to alcohols is not very likely.
On the basis of the need for water to be present at the
catalyst surface 19,101, and taking into account the
principle of the least structural change, Eucken and
Wicke r 1 0 , 1 3 , 1 4 1 derived a hydrogen exchange mechanism, in which the substrate molecule is adsorbed in
such a way that it already approximates as closely as
possibly to the structure of the product molecule. In an
oxonium-caibonium ion mechanism and in some other
mechanisms, hydrogen must be exchanged with the
catalyst surface.
In opposition to ionic mechanisms, a covalently bound
adsorption complex in the form of a surface alkoxide
has been postulated for the dehydration of ethanol [Is- 171. The dissociation of this surface compound
yields ethylene, whereas the condensation of two adjacent alkoxide groups leads to ether formation.
Topchievu et al. "71 believe the active centers to be the
(31 F. C. Whitmore, J. Amer. chem. SOC.54, 3274 (1932).
[4] W . S . Brey j r . and K . A . Krieger, J. Amer. chem. SOC.71,3637
[ S ] J . G. M . Bremner, Research I , 281 (1948).
[6] R . A. Ross and D . E . R . Bennett, J. Catalysis 8, 289 (1967).
[7] H . Pines and W. 0.Haag, J. Amer. chem. SOC.83,2847 (1961).
[S] H. Pines and J. Manassen, Advances Catalysis related Subjects 16, 49 (1966).
[9] L . A . Munro and W . R. Horn, Canad. J. Res. I 2 , 707 (1935).
[lo] A. Eucken and E. Wicke, Naturwissenschaften 32,161 (1944).
[l 11 E. P . Parry, J. Catalysis 2, 371 (1963).
[I 21 H . KnBringer and H . Spannheimer, Ber. Bunsenges. physik.
Chem. 70, 575 (1966).
[13] A . Eucken, Naturwissenschaften 34, 374 (1947).
[14] E. Wicke, Z. Elektrochem. angew. physik. Chem. 52, 86
[15] J. B. Senderens, Bull. SOC.chim. France [4] 1 , 692 (1907).
1161 V. N . Ipatieff: Catalytic Reactions at High Pressures and
Temperatures. Macmillan Co., New York 1936.
[17] K . V . Topchieva, K . Yun-Pin, and 1. V. Smirnova, Advances
Catalysis related Subjects 9, 799 (1957).
surface OH groups. Though alkoxide structures have
been detected in the IR spectra of some alcohols adsorbed on aluminum oxidell*. 191, this does not provide
conclusive proof of the mechanism described.
Heiba and Lundis [201 compared the products obtained
in the dehydration of alcohols with those obtained on
pyrolysis of the corresponding alkoxides. However,
since the authors certainly did not obtain the primary
products above 300 "C, the results cannot be taken as
conclusive proof of a given mechanism.
Vusserberg 1211 suggested a free-radical mechanism,
and free radicals of compounds having extremely
strong electron donor properties have in fact been detected on aluminum oxide surfaces by visible and UV
adsorption spectroscopy and by ESR studies 122-251.
These donor properties are not so pronounced in alcohols as in the aromatic compounds studied. Moreover, since free-radical mechanisms proceed preferentially in apolar media, whereas the surfaces of metal
oxides must be iegarded as polarr261, this route is
rather unlikely.
Some authors regarded the dehydration (and the dehydrogenation) of alcohols as electron donor-acceptor
reactions in the electronic theory of heterogeneous
catalysis. Thus Wolkenstein 127,281 and Garner [291
postulate a rate-determining electron acceptor step for
the dehydrogenation and a rate-determining donor
step for the dehydration. Haufkr301, on the other hand,
regards the rate-determining step in the dehydrogenation of ethanol as a donor step, i.e. the desorption of
acetone. It seems very doubtful, however, whether the
dehydration of alcohols is a donor-acceptor reaction,
since it presumably involves the displacement, not
of single electrons, but of electron pairs. Thus Meye 1311
showed that neither p- nor n-doping of a-AlzO3 had
any definite effect on the rate and activation energy of
the dehydration of methanol to dimethyl ether.
Different conclusions were reached by Pines et al., who
studied the effect of the structure of the substrate on
the distribution of the primary products. Secondary
isomerization of the primary products, which generally
[18] R. G. Greenler, J. chem. Physics 37, 2094 (1962).
[19] D. Treibrnann and A . Simon, Ber. Bunsenges. physik. Chem.
70, 562 (1966).
[20] E . Heiba and P. S . Landis, J . Catalysis 3,471 (1964).
[21] V . E . Vasserberg, J . R. Davydova, and T . V . Georgievskaya,
Kinetika i Kataliz 2, 773 (1961).
1221 J. J. Rooney and R. C . Pink, Trans. Faraday SOC.58, 1632
[23] D. M. Brouwer, J. Catalysis 1, 372 (1962).
[24] A . E. Hirschler and J. 0.Hudson, J. Catalysis 3, 239 (1964).
[25] B. D. Flockhart, J. A . N . Scott, and R. C . Pink, Trans.
Faraday SOC.62, 730 (1966).
[26] G.-M. Schwab and L. Lassak, Kolloid-Z., Z . Polymere 206,
37 (1965).
[27] F. F. Wolkenstein, Advances Catalysis related Subjects 12,
223 (1960).
I281 F. F. Wolkenstein: The Electronic Theory of Catalysis on
Semiconductors. Pergamon Press, Oxford 1963.
(291 W. E. Garner, Advances Catalysis related Subjects 9, 169
[301 K . Hauffe, Advances Catalysis related Subjects 7,213 (1955).
[311 W. Meye, Diplomarbeit, Universitat Miinchen, 1967.
Angew. Chem. internat. Edit.
1 Vol. 7 (1968) 1 No. I0
occurs on pure aluminum oxide 1321, was suppressed by
modification of the catalysts with ammonia, pyridine,
or piperidine.
Pines a n d Pillai (331 found preferential trans-elimination in
the dehydration of menthol an d neomenthol. To explain this,
they suggested neighboring-group effects an d reaction at
preferred sites on the catalyst, e.g. o n steps or in sufficiently
narrow pores a nd cracks; this view had already been advanced by Schwab (341. T h e heterogeneously catalyzed dehydrochlorination of menthyl an d neomenthyl chlorides also
proceeds by rrans-elimination [ 3 5 1 . T h e principal product in
the dehydration of 2-endo- an d 2-exu-bornanol is camphene.
Pines et al.[361 interpret this reaction as resulting from a
synchronous frans-elimination in which t h e hydroxyl group
is attracted by a n acidic surface center (A) and t h e proton by
a basic center (8). They also reached th e same conclusion
for the dehydration of endu- an d exu-norbornanol, which is
assumed t o proceed in pores o f a suitable size. f in es et a/.
also assume tha t trans-elimination is preferred in th e dehydration of primary aliphatic alcohols. This conclusion is
based on the distribution o f primary products in th e dehydration of l-butanol171, isobutanol, an d 2-phenyl-1-propanoIr371 with a n d without “T-labeling[sl, t h e effects of neighboring groups also being taken into account. T h e authors
rule o u t a n oxonium-carbonium ion mechanism in these cases.
To clarify t h e question of t h e order in which th e elementary
steps take place, the dehydration o f primary and secondary alcohols was compared with th at of tertiary alcohols. Pines
e t a / . concluded from this comparison that t h e dehydration of
tert-pentyl alcohol proceeds by a carbonium ion mechanism,
whereas the hydroxyl group an d th e y-proton ar e removed
simultaneously from t h e primary neopentyl alcohol (3x1.
trans-Elimination is also claimed for th e pairs 3,3-dimethyl-2butanol and 2,3-dimethyl-2-butanol (381 an d 3,3-dimethyI-2the secondary alpentanol a nd 2,3-dimethyl-2-pentanol[381,
cohol being preferentially dehydrated by a synchronous
mechanism, while t h e tertiary alcohols react via a carbonium
ion. Neighboring-group effects leading to reorientation of t h e
carbon skeleton a r e also possible, particularly in th e case o f
alcohols of the neopentyl type. On dehydration of 2-butanol
( 1 ) a n d of 2- a nd 3-pentano1, 1- an d cis- an d trans-2-alkenes
were obtained a s primary productsI7~381.This cis-alkene is
always favored in relation t o t h e trans-alkene t o a greater degree than tha t corresponding t o thermal equilibrium. Pines
er al. explain this result by a synchronous trans-elimination.
T he intermediate proton-olefin complex (2) is thought t o b e
responsible for the unusually high cis:trans ratio.
Pines and Manassen 181 concluded that acidic and
basic centers take part in the dehydration of alcohols
on aluminum oxide. A carbonium ion mechanism appears to be favored in the dehydration of tertiary alcohols and a synchronous mechanism in the case of
primary or secondary alcohols. If possible, the reaction proceeds as a simple 1,2-trans-eIinlination. Neighboring-group effects can lead to the migration of
methyl groups. The reaction is thought to take place
preferentially on steps, in pores, and in cracks in the
oxide. Thus the aluminum oxide must surround the
alcohol molecule in such a way that the acidic centers
can act as proton donors (electron acceptors) and the
basic centers as proton acceptors (electron donors).
The catalyst is therefore regarded as a pseudosolvent,
and the strong analogy with eliminations in the liquid
phase is stressed.
Using these views as a basis, Juin an d Pillair391 proposed a
mechanism (adsorbed phase substitution-elimination mechanism) that also embraces t h e ether formation.
If t h e aluminum oxide is assumed t o have electrophilic (A)
and nucleophilic (B) sites o n its surface, there should also be
two types of adsorbed alcohol molecules, i.e. carbonium
ions ( 3 ) and alkoxide ions ( 4 ) . A nucleophilic attack by ( 4 )
o n the positively charged C atom of (3) leads to ether forma~
while t h e abstraction
tion (cf. homogeneous S N reactions),
of a 9 proton from t h e carbonium ion ( 3 ) by t h e alkoxide (41
(or by the basic site on t h e surface, as is now assumed) leads
t o the olefin. According t o this view, therefore, the catalyst
gives t h e reacting molecules t h e required polarity an d fixes
them in t h e correct arrangement. T h e ad so r b ei phase thus
resembles a polar medium in which nucleophilic substitution
an d elimination compete with each other.
R-?. ,H
The simultaneous dehydration of phenol and aliphatic alcohols, as well as poisoning experiments,
carried out by the same authors [39al lends support to
the concept of two different species taking part in
ether formation.
2. The Reaction Scheme in the
Dehydration of Alcohols
B + BH
[32] H . Pines and W. 0.Haag, J. Amer. chem. SOC. 82, 2471
[33] H. Pines and C . N. Pillai, J. Amer. chem. SOC.83, 3270
[34] G . - M . Schwab and E. Schwab-Agallidis, J. Amer. chem. SOC.
71, 1806 (1949).
[35] P. Sndrdu, E. Bussmann, H . Nuller, and S. K. Sim, Z . Elektrochem., Ber. Bunsenges. physik. Chem. 66, 739 (1962).
[36] K. Wafanabe, C. N. Pillai, and H . Pines, J. Amer. chem. SOC.
84, 3934 (1962).
[37] J. Herling and H . Pines, Chem. and Ind. 1963, 984.
[38] C. N. Pillai and H . Pines, J. Amer. chem. SOC.83, 3274
Angew. Chem. internut. Edir. 1 Vol. 7 (1968) 1 Nu. 10
The dehydration of simple alcohols on aluminum oxide
leads mainly to ethers, olefins, and water, which may
be formed in successive or parallel steps or else sirnultaneously. For the dehydration of methanol and
ethanol up to about 200 O C , Balaceanu and Jungers [4OJ
assume primary ether formation with possible equilibrium. In the dehydration of ethanol, ethylene is formed as a secondary product with re-formation of
ethanol; it is also formed directly at higher temper[39] J. R. Jain and C. N . Pillai, Tetrahedron Letters 11, 675
139aI J . R. Jain and C. N. Pillai, J. Catalysis 9, 322 (1967).
[40] J . C. Balaceanu and J . C. Jungers, Bull. SOC.chim. belges 60,
476 (1951).
atures. Isagulyants et al. 141,421 reached the same conclusions on the basis of investigations on [14C]ethanol but assumed that the olefin is formed as a
secondary product by the dehydration of ethers.
Stauffer and Kranich 1433, on the other hand, assumed
that the primary aliphatic alcohols up to n-hexanol
are dehydrated to the ether and the olefin in independent parallel steps. The results obtainable for other
alcohols are not very extensive.
A good method for the determination of the reaction
schemes of complex reactions is the study of the product distribution as a function of the catalyst temperature and the contact time. We used this method to
study the reaction scheme for the dehydration of some
unbranched and branched aliphatic alcohols (44,451.
y-Al~O3was used as the catalyst for all the kinetic
studies described.
Since the dehydration of methanol up to 320 "C yields only
ether by the bimolecular reaction
2 CH30H
alcohol, E
ether, 0
olefin, W
between 240 and 250 ' C . The quantities of ethanol and ethylene in Figure 2 are initially equimolar
(C2Hs)zO + CzHsOHi- CzH4
Above 260 O C , the ethylene concentration increases more
rapidly than the ethanol concentration, and at the same time
water begins to appear as a result of secondary dehydration
of ethanol in accordance with
Even with long contact times, n o ethylene is observed below
240 O C , while small concentrations that increase with the
contact time appear at 256 "C (Fig. 3).
ether, 0 = olefin, W = water.
temperature. A t low temperatures, only equimolar quantities of ether and water are obtained; the reaction therefore
proceeds in accordance with:
Ethylene is formed only above 240 ' C . The maximum in the
ether-concentration curve points t o secondary decomposition of the ether. Since ethylene formation begins in the immediate vicinity of the point of inflection in the ether curve,
ethylene should be formed by the secondary decomposition
of the ether. Diethyl ether on this catalyst does in fact give
ethanol and ethylene (cf. Fig. 2). Both diagrams show that
the secondary decomposition of ether on y-AlzO3 begins
[41] G. W. Isagulyants, A. A. Balandin, E. J. Popow and Yu. J . ,
Derbenzew, 2. fiz. Chim. 38,20 (1964).
[42] G. W. Isagulyants and A . A. Balandirr in: Radioisotopes in
the Physical Sciences and Industry. Intern. Atomic Energy
Agency, Vienna 1962, p. 245.
[43]J . E. Stauffer and W . L. Kranich, Ind. Engng. Chem., Fundamentals I , 107 (1962).
I441 H . Knozinger and R. Kohne, J. Catalysis 3, 559 (1964).
[45] H . Knozinger and R . Kohne, J. Catalysis 5,264 (1966).
T I"C1 -z
Fig. 1. Product distribution in the dehydration of ethanol on y-AlzO,
as a function of the temperature 1451.
A = alcohol, E
+ HzO
with possible equilibration, the reaction scheme will be discussed for the case of ethanol. Figure 1 shows the partial
pressures of the components occurring as a function of
Fig. 2. Product distribution in the decomposition of diethyl ether o n
a function of the temperature [45].
~ - A 1 2 0 1as
i g sec mi-']
Fig. 3. Reaction isotherms for the dehydration of ethanol on y-Al2O3
at 256 O C 145).
A = alcohol, E
ether, 0
olefin, W = water, U
Since the ethylene concentration does not exhibit the induction period (convex curvature with respect to the abscissa) characteristic of pure secondary reactions, direct
olefin formation from ethanol must be superimposed on the
secondary ethylene formation from the ether. Two steps (3)
and (4) leading to ethylene thus evidently proceed simuttaneously and begin on y-AlzO3 in approximately the same
temperature range (240-250 "C).
At about 300 OC and above, the ethanol is not only dehydrated,
but also dehydrogenated t o acetaldehyde and hydrogen (451.
The unbranched primary alcohols up t o n-hexanol behave in
exactly the same manner. However, the temperature range
for pure ether formation decreases as the number of carbon
atoms increases, so that even for n-butanol, the reactions
start almost simultaneously. Direct kinetic studies cannot
therefore be carried out on any of these reactions for primary
alcohols higher than n-propanol.
In the case of the higher homologs of the primary alcohols,
isomerizations also occur in the initially formed I-olefins.
In agreement with Pines and HaagL71, we found isomerizaAngew. Chem. internat. Edit. 1 Vol. 7 (1968)
No. I0
tion of I-butene t o cis- and trans-2-hutene (with a clear preference for the cis product) in the dehydration of n-butanol,
though no skeletal isomerization was detected 1451.
Thus the steps corresponding to reactions ( I ) to (4) are
observed for the primary alcohols containing two to
six carbon atoms, and the ether formation can be
clearly isolated for ethanol and n-propanol at low
temperatures. As the temperature is raised, the primary olefin formation from alcohol and the secondary
olefin formation from ether start in the same temperature range and steadily become more important. It is
not certain whether the secondary olefin formation
proceeds with regeneration of alcohol in accordance
with step (2) or with dehydration of the ether in accordance with step (3). Step (2) should however be
favored at low temperatures at least, since it is known
from experience that the probability of the simultaneous participation of several bonds in the reacting
molecule is very small. On the other hand, step (3)
cannot be ruled out at about 300 "C (cf. f 1 7 9 . The possible reaction directions for primary aliphatic alcohols
are shown in Scheme 1, which is in agreement with
the results obtained by Buiaceanu and Jungersf4oJ as
well as by Bliss, Butt et al. [46,47,47aI
study of ether formation, and tert-butanol, isobutanol,
and cyclohexanol for the study of olefin formation.
Since the catalytic activity of a surface is strongly dependent on its structure and properties as well as on
the surface compounds of the substrates and reaction
products, the y-AIzO3 surface and the adsorption structures of alcohols and their dehydration products will
be considered before the kinetic studies are described.
3. The Surface Structure of y-A12O3
In the simplest model, the surface centers of aluminum
oxide are regarded as incompletely coordinated
aluminum and oxygen ions, the surface being assumed to be completely free from OH groups. Other
models take into account cation defects and strained
sites, which result f r o m i h e dehydration of the originally hydratedIsurface [50,511. By investigation of
the water-absorption and analysis of the IR spectra of
water adsorbed on y-AI203, Perit521 developed a
more detailed model based on the following considerations (cf. Fig. 4a).
2 0 t 2 W
Scheme 1 .
Isopropanol, 2-butanol, isobutanol, and tert-butanol
have also been investigated [44,451. The dehydration of
isopropanol leads to the formation of propylene, and
yields only small quantities of isopropyl ether. The
dehydration of 2-butanol yields I-butene and cis- and
trans-Zbutene, which, according to Pines and Haag [32J
are formed as primary products in parallel steps. Isobutanol and tert-butanol give isobutylene in yields of
more than 95 % (based on reacted alcohol).
Benzyl alcohol gives only dibenzyl ether up to
200 "C 1481, while cyclohexanol gives cyclohexene [7,49,941. A slow secondary skeletal isomerization
of cyclohexene into methylcyclopentene was observed
by Pines and Haag [71 only above 300 "C on strongly
acidic oxides.
The alcohols suitable for kinetic studies are those in
whose dehydration individual primary steps can be
definitely isolated at least in narrow temperature
ranges. Methanol, ethanol, n-propanol, and benzyl
alcohol can be used as representative examples for the
[46] J. B. Butt, H. Bliss, and C. A . Walker, A. I. Ch. E. JournalB,
42 (1962).
[47] H . J. Solomon, H . Bliss, and J. B. Butt, Ind. Engng. Chem.,
Fundamentals 6, 325 (1967).
[47a] R. Mezoki and J . B. Butt, Ind. Engng. Chem., Fundamentals 7, 120 (1968).
1481 P . Sabatier and P . Mailhe, Ann. Chim. physique 181 20, 298
[49] M . Kraus, K . Kochloefl, L. Beranek, and V. Baiant, Roc.
third int. Congr. Catalysis, Amsterdam 1964, p. 577.
Angew. Chem. internat. Edit. J Vol. 7 (i968) / No. I0
Fig. 4. Vertical section through an ideal y-AlzO3 surface.
(a) hydrated surface, (b) dehydrated surface,
= Al (after
The oxide is in the form of a defect lattice of the spinel
type. Ionic surfaces are generally bounded by anions,
which are more readily polarizable than cations.
Though many authors (e.g. f 5 3 9 believe the [lll]face
of the oxygen sublattice t o be energetically favored,
Peri concludes that y-AI203 is bounded by the [loo]
face. To reduce the surface energy, oxides of this type
characteristically adsorb water, which leads whenever
possible to the formation of OH groups. The liberation
of water from a hydrated surface by condensation of
two OH groups must expose oxygen ions. The re[50] E. B. Cornelius, T . H . Milliken, G . A. Mills, and A. G .
Oblad, J. physic. Chem. 59, 809 (1955).
[51] S. G. Hindin and S . W. Weller, Advances Catalysis related
Subjects 9, 70 (1957).
[52] J. B. Peri, J. physic. Chem. 69, 211, 220 (1965).
1531 J. H. de Boer, G . M . M . Houben, B. C. Lippens, W. H. Meijs
and W . K. A. Wulruve, J. Catalysis I, 1 (1962).
maining OH groups acquire different properties,
which Peri explains by the change in their immediate
surroundings. At the same time, aluminum ions in the
second atomic position lose their full coordination
and become accessible from the surface through
oxygen vacancies (cf. Fig. 4b).
Aluminum oxide retains a part of its unimolecular
surface coating of water even on dehydration for
several hours under vacuum, the proportion retained
depending on the heating temperature [54,553. This
retained water is described as irreversibly adsorbed,
and is partly in the molecular state and partly in the
form of OH groups [52,54,561.Adsorption isobars for a
water partial pressure of zero, i.e. for the irreversibly
adsorbed water, show that pure water does not give a
unimolecular coating of OH groups in the temperature range for the catalytic dehydration of alcohols 1551. According to the Peri model, OH groups and
oxygen ions as well as incompletely coordinated
aluminum ions will therefore be accessible at the surface in this range. Since alcohols are also irreversibly
adsorbed [57,583, the distribution of these three types
of centers will be modified by the adsorption of the
It has been shown by gas chromatography that irreversibly adsorbed water on aluminum oxide is
partly displaced by ethanol, and irreversibly adsorbed
ethanol by waterl591. The aluminum ions whose coordination is disturbed by the condensation of two
O H groups with formation and desorption of water
may be regarded as acidic centers of the Lewis
type [GO, 611, which can be very selectively poisoned by
organic bases [11 121. The OH groups are extremely
weakly acidic Bransted centers [11,621. Protons in cation vacancies have also been suggested as proton
donors, but these would become effective only at
temperatures above 400 "C because of their low activation energy [631.
4. IR Studies on the Adsorption Structures
4.1. Adsorption of Alcohols
O n adsorption of alcohols, particularly ethanol, 1R
studies showed three surface compounds:
[54] J. B. Peri and R. B. Hannan, J. physic. Chem. 64, 1526
I551 H. Spannheimer and H. Knozinger, Ber. Bunsenges. physik.
Chem. 70, 570 (1966).
[56] J. H. de Boer, J. H. M . Fortuin, B. C. Lippens, and W . H.
Meijs, J. Catalysis 2, 1 (1963).
1571 J. J. Kipling and D. B. Peakall, 3. chem. SOC.(London) 1957,
[58] H. Knozinger and H . Stolz, Kolloid-Z., Z. Polymere 223,
42 (1968).
[59] H . Knozinger, H . Spannheimer, and G. Kinshofer, unpublished.
[60]J . B. Peri, J. physic. Chem. 69, 231 (1965).
[61] J. B. Peri, Actes Congr. intern. Catalyse, 2e, Paris 1960,
Vol. 1, p. 1333.
1621 E. Koberstein, Z. Elektrochem., Ber. Bunsenges. physik.
Chem. 64, 906 (1960).
I631 S. E. Tung and E. Mclninch, J. Catalysis 3,229 (1964).
1. molecular aIcoho1 adsorbed via H bonds [67,701;
2. alkoxide groups formed by dissociative adsorption [18,19,64-67,75,761;
3 . carboxylate groups formed by oxidation of the
alcohol I18,19,67,76,80,811.
4.1.1. M o l e c u l a r A d s o r p t i o n of A l c o h o l s
Whereas Babushkin et a/. [64,65Jand Boreskov et aE. [661
thought that they could still detect molecularly adsorbed ethanol even at high desorption temperatures,
(methanol and ethanol), Treibmann and
Simon 1191 (ethanol, n-propanol, and isopropanol) and
KagelL671 (n-propanol and n-butanol) found that alcohols are readily desorbed, even at room temperature. NMR studies showed that the hydroxyl group
of the alcohols interacts directly with the oxide surface 168,691;it must therefore be assumed that H bonds
are formed. KagelI671 assumes without direct spectroscopic evidence, that the structures ( 5 ) and (6) are
both present, i.e. that the alcohol forms both active
and passive H bonds.
However, the temperature dependence of the OH
stretching bands on adsorption of methanol, ethanol,
tert-butanol, isobutanol, cyclohexanol, and benzyl
alcohol aIso indicates the presence of structures with
two angular H bonds per alcohol molecule[701.
The absorption o f the free OH groups of alcohols should be
expected to appear at 3600 to 3650 cm-1; the OH stretching
vibrations of the chain associates formed by linear H bonds
in the liquid phase occurs at 3350 cm-1 for all the alcohols in
question. The shift of the OH stretching band is a measure
of the strength of the H bonds. Linear H bonds are always
stronger than angular bonds [711. The pronounced frequency
shift in liquid alcohols is thus explained by the formation of
Iinear H bonds. All adsorbed alcohols give strong OH stretching bands centered around 3500 cm-1. At low temperatures
the bands are asymmetric, with a shallow gradient on the low
wave-number side. As the temperature is raised, the bands
decrease in intensity, starting at lower wave numbers, though
the center still remains at 3500 cm-1.
Thus at low temperatures there must be adsorption structures with relatively strong, probably substantially linear H
bonds as well as weaker angular bonds, since the OH stretching band extends down to 3200 cm-*. As the temperature is
raised, the strong hydrogen bonds are broken whiie the weak[64] A. A. Babushkin, A . V. Uvarov, and L . A. Ignateva, Materialien des x. Allunionskongresses iiber Spektroskopie 1957,
p. 161.
[65] A . A. Babushkin and A. V. Uvarov, Doklady Akad. Nauk
SSSR 110, 581 (1956).
I661 G . K . Boreskov, Yu. M . Shtschekotschichin, A. D . Makarov,
and V. N . Filimonov, Doklady Akad. Nauk SSSR 156,901 (1964).
[67] R. 0. KageL J . physic. Chem. 71, 844 (1967).
[681 G. De la Hardrouyere, Arch. Sci. 13, 1 (1960).
[69] D. Geschke and H . Pfeifer, Z. physik. Chem. 232,127 (1966).
[70] H . Knozinger, E. Ress, and H . Biihl, Naturwissenschaften
54, 516 (1967).
[71] W. Luck, Naturwissenschaften 52, 25, 49 (1965).
Angew. Chem. internat. Edit. J VoI. 7 (1968)
1 No. I0
er bonds (absorption at 3500 cni-1) remain intact. Since the
absorption band of cyclic alcohol dimers with two nonlinear H bonds in dilute solutions 1721 and in N2 matrices 1731
was localized at 3500cm-1 similar structures, such as (7)
should be favored on adsorption of alcohols on aluminum
A structure such as (7) will be more stable and more
heat-resistant than the singly H-bonded structures ( 5 )
and (6) if the energy of formation of each of the two
angular H bonds is greater than half the energy of
formation of singly H-bonded structures. Such energy
relations have been found for the cyclic dimers of
alcohols in dilute solutionr741. In the adsorption of
alcohols on surfaces, steric factors must also be taken
into account since, owing to the geometry of the alcohol molecules, the formation of a linear active H bond
brings the alkyl group closer to the surface, this effect being particularly pronounced in the case of tertbutanol. The steric factors thus favor the formation of
adsorption complexes of structure (7), in which the
alkyl groups are pushed farther away from the surface.
An adsorption via two passive H bonds .to the surface
by way of the two free electron pairs of the hydroxy
oxygen atom cannot be dismissed categorically. In
this case, the frequency of the free OH vibration of the
alcohol should occur; however, this absorption would
hardly be recognizable in the steep high-frequency
flank of the band due to the perturbed OH-groups.
4.1.2. Alkoxide S t r u c t u r e s
A comparison of the IR-spectra of adsorbed phases
with those of solid alkoxides shows the presence of
surface alkoxide groups on adsorption of methanol 1181, ethanol [18, 19,64-66,751, n-propano1[19.671,
isopropanol[191, n-butanolf671,and benzyl alcohol [761.
In the Peri model of the aluminum oxide surface,
there are two possible routes for the formation of
alkoxide by dissociative adsorption:
1. A hydroxyl proton may be split off to form a surface OH group, while the alkoxide residue occupies
an oxygen vacancy in the surface (Lewis center);
2. The C - 0 bond of the alcohol may be broken, the
resulting hydroxyl group filling an oxygen vacancy on
the surface, while the alkyl residue adds to an adjacent oxygen ion.
Though the somewhat lower energy of rupture of the
C - 0 bond in the free alcohol[771 favors the second
[72] L. P. Kuhn, J. Amer. chem. SOC.74, 2492 (1952).
[73] M. van Thiel. E. D. Becker, and G. C. Pimentel, J. chem.
Physics 27, 95 (1957).
[74] W . Luck, Ber. Bunsenges. physik. Chem. 69, 626 (1965).
[75] H . Knozinger, Habilitationsschrift, Universitat Miinchen
I761 H . Knozinger, H . Biihl, and E. Ress, J . Catalysis, in press.
[77] H . Preuss: Quantentheoretische Chemie Bibliographisches
Institut, Mannheim 1863.
Angew. Chem. internat. Edit.
Vol. 7 (1968) 1 No. I0
possibility, the lattice energy liberated on incorporation of the alkoxide or hydroxyl group decides in
favor of the rupture of the 0 - H bond. Treibmann and
Simon "91 concluded from displacement reactions that
the alkoxide formation involves incompletely coordinated aluminum ions ; Kage/[671 reached the same
conclusion on the basis of mass-spectroscopic residual
gas analyses after adsorption.
While dissociative adsorption with formation of a surface
alkoxide has been substantially verified for the alcohols
discussed so far, no analogous surface compound has been
detected o n adsorption of rert-butanol, isobutanol, and
cyclohexanol[75,761. This can be explained on the basis of
the thermal stability of the corresponding alkoxides. The
thermal stability of zirconium alkoxides decreases in the order
Zr(OCzH5h 9 Zr[OCH(CH3)Z]4 > Zr[OC(CH3)314 M
Zr[OC(CH3)2C2H5]4 (781. The methoxides of aluminum and
zirconium are stable up t o 240 and 280°C respectively,
whereas fert-butoxides have a n extremely poor thermal
stability [791. All the alcohols studied so far that can form a
surface alkoxide give ethers on dehydration on aluminum
oxide at low temperatures, while the others tend toward
olefin formation.
4.1.3. C a r b o x y l a t e S t r u c t u r e s
Oxidation of the alcohols during adsorption or
aluminum oxide above about 170°C to give carboxylate structures with the aid of the oxygen ions on
the surface has been observed for methanol[l8l,
ethanol [18,19,751, n-propanol, and n-butanol[67,801.
Corresponding surface compounds have recently also
been found on adsorption of isobutanol and benzyl
alcohol on aluminum oxide above about 120 "C [811.
Treibmann and Simonc191 concluded from the occurrence of an acetate structure on adsorption of isopropanol that the carbon skeleton of the molecule is
broken down. These carboxylate structures are observed up to 400 "C; the adsorption of water does not
change their adsorption bands. In view of its high
stability, this structure could hardly be important as
an intermediate in surface reactions up to 250°C.
However, the dehydrogenation of ethanol observed
above about 300°C could proceed via an acetate
structure 1821.
4.2. Adsorption of Water
The adsorption of alcohols in the catalytic temperature
range is always accompanied by the appearance of a
band at about 1600cm-1, the intensity of which increases with rising temperature, and which is not due
to the alcohol or any of its surface compounds[ssl.
Using the Nz matrix technique, van Thiel, Becker and Pimentel1831 were able t o show that the water molecules in the
H-bonded clusters have a deformation band at 1633 cm-1,
[78] D. C. Bradley and M . M. Faktor, J. appl. Chem. 9, 435
(1959); Trans. Faraday SOC.55, 2117 (1959).
[791 D. C. Bradley and M . M . Faktor, Nature (London) 184, 55
[SO] V. Corso, C . R. hebd. SBances Acad. Sci. 259, 1413 (1964).
[811 H . Knozinger and H . Biihl, unpublished.
[82] H . Knozinger, 2. physik. Chem. N.F. 48,151 (1966).
[83] M . van Thief, E. D. Becker, and C . C. Pimentel, J. chem.
Physics 27, 486 (1957).
while the corresponding band for the free monomeric molecule occurs at 1600 cm-1. Since the surface of aluminum oxide
contains OH groups and oxygen ions, it must be assumed
that molecular water is adsorbed by means of H bonds.
The deformation band is a particularly suitable aid to the
detection of such H bonds, as was shown by Glemser and
Hartert [841 for the adsorption of water on aluminum oxide at
room temperature. The temperature-dependence of the intensity and position of this band provides information about
the structure of the H bonds [851.
Figure 5 shows the region around 1600cm-1 for the rehydration of a sample of y-AI203 that has been heated under
vacuum at 520 ' C . This oxide initially shows no absorption
adjacent OH groups on the surface, a condition that
shouId be satisfied for the adsorption of pure water on
aluminum oxide between 80 and 170°C since up to
200°C the surface of y-A1203 still has a 50-80 %
covering of OH groups 152, 551. However, adsorption
on isolated OH groups as in (9) should also be possible, at least when two adjacent OH groups are not
available. This case is particularly probable when
water is adsorbed in the presence of excess alcohol, i.e.
under conditions that lead to the development of the
band at 1600cm-1 on adsorption of alcohol in the
catalytic temperature range.
H, ,H
4.3. Adsorption of Diethyl Ether
Fig. 5. Deformation bands of water adsorbed on y-A1203.
1: after evacuation for 60 min at 520"C/lO-3 torr, spectrum recorded at
80°C, 2: water adsorbed at 8O"C, 3: at 130°C. 4: at 168"C, 5: water
desorbed at 168% evacuation for 70 min at 10-3 torr.
between 1630 and 16OOcm-1. On rehydration (55 torr of
water vapor), a strong band due to polymeric water associates
with strong, substantially linear H bonds develops at 1630
cm-1. It is conceivable that multilayer adsorption occurs,
with incipient formation of liquid water in the form of
cluster-like regions, at least in pores and cracks. With rising
temperature and hence decreasing surface concentration, the
intensity of the band at 1630 cm-1 decreases, while a band
simultaneously develops a t 1600 cm-1; this latter band is
assigned to the deformation vibration in quasimonomeric
water molecules. (Molecules that are only passively involved
in H bonds via free pairs of electrons exhibit the vibration
frequencies of the free molecule 186.871.) The band observed
at 1600cm-1 o n adsorption of alcohols after catalytic dehydration has started must therefore be the deformation vibration of water molecules that are adsorbed on surface OH
groups by passive H bonds and which do not undergo any
further interactions. Yates[881 found a band at 1605 cm-1,
which he interpreted in a similar manner, on adsorption of
water o n titanium dioxide.
It is generally assumed that molecules containing oxygen and having two free pairs of electrons are preferentially adsorbed via two passive H bonds as in
(8) [891. This is thought to be the case in the adsorption
of water on silanol groups 1901. This type of adsorption
bonding via two H bonds requires two immediately
[84] 0. Glemser and E. Hurtert, Z. anorg. allg. Chem. 297, 175
[85] H. Kndzinger and E. Ress, 2. physik. Chem. N.F. 59, 49
[86] L. P. Kuhn, J. Amer. chem. SOC.74, 2492 (1952).
[87] D . P . Stevenson, J. physic. Chem. 69, 2145 (1965).
I881 D . J. C. Yates, I. physic. Chem. 65, 746 (1961).
1891 L. H . Little: Infrared Spectra of Adsorbed Species. Academic
Press, New York 1966, p. 263.
I901 M. M . Egorov, W. I. Kvilividze, A. V . Kizelev, and K . G .
Krussilnikov, Kolloid-Z., 2. Polymere 212, 126 (1966).
Diethyl ether is similar to the water molecule in that
the oxygen atom has two free pairs of electrons,
though ether cannot act as a proton donor and form
active H bonds. However, the H-bond acceptor properties of the two molecules on adsorption on oxide surfaces having OH groups should be qualitatively similar.
IR investigation of the adsorption of diethyl ether on
aluminum oxide[*5l shows that up to about 17O"C,
it is bound only in a readily desorbable form. By
analogy with the adsorption structures proposed for
silicon dioxide[90-92] and the structures (8) and (9)
for the adsorption of water on y-AIzO3, it is assumed
that diethyl ether is present in one of the forms (10)
and ( I I ) , depending on the surface concentration of
OH groups.
Bands indicating the formation of a surface ethoxide
appear in the spectrum only above 200 "C1851. This
compound may be formed directly by dissociative adsorption of the ether or by secondary adsorption of
the ethanol resulting from catalytic decomposition [*I.
[911 A. V. Kizelev and V. I. Lygin, Usp. Chim. 31, 351 (1962).
[*I Note addedinproof(Sept. 19, 1968): In contrast to the results
discussed here [85] no great shift of the OH stretching highfrequency bands of the surface hydroxyl groups after desorption
of diethyl ether was observed by Arai, Saito, and Yoneda [92a].
The authors therefore rule out adsorption via H bonds. However, this conclusion does not necessarily apply when ether is also
present in the vapor phase. On comparison of the desorption
spectra with spectra of the (C2H&O-AlC13 complex, the authors
conclude that a coordinative adsorption bond is formed between ether and incompletely coordinated aluminum ions below
100 "C. This complex undergoes thermal decomposition to
ethanol, diethyl ether, and ethylene. A surface ethoxide is observed at temperatures above 130 "C.
[92] V. Y. Davidov, A. V . Kizelev, and V. I. Lygin, Kolloidnyj 2.
25, 152 (1963).
[92a] H . Arai, Y . Saito, and Y. Yoneda,J. Catalysis 10,128 (1968).
Angew. Chem. internat. Edit. 1 VoI. 7 (1968)
No. 10
5. Kinetics and Mechanism of the Dehydration of
Only a few intensive studies have been carried out on
the kinetics of the dehydration of alcohols. The test
reaction used in most cases was the formation of
ethylene from ethanol at relatively high temperatures.
However, the product distribution (cf. Section 2) shows
no definite reaction steps. Topchieva and Ron?anovskii[931 described the kinetics of the formation of
ethylene from ethanol by a simple Langmuir formula
containing an inhibition term for water. According to
Kochloefl el al. 1941 the formation of olefins from some
aliphatic and alicyclic alcohols at about 200 "C and at
alcohol pressures of 70 t o 80 torr is a zeroth order
reaction. On the basis of kinetic data for the dehydration of 2-butanol and n-hexanol, Kittrelland Mezaki1951
proposed the participation of two types of centers.
The same conclusion was reached by de Mourges
et al.[961 for the dehydration of isopropanol, and by
Carrci et al. for the dehydration of cyclohexanol[96al.
From the kinetic analysis of reaction isotherms for the
dehydration of ethanol, on the other hand, Solomon.
Bliss, and Butt [47,47al postulated a one-center mechanism, in which the simultaneous formation of ethylene
and ether is assumed to take place via a common intermediate. The kinetics of the formation of olefins from
tert-butanol, isobutanol, and cyclohexanol and of the
bimolecular ether formation on dehydration of
methanol, ethanol, n-propanol, and benzyl alcohol, i.e.
of clearly isolable reaction steps, will now be discussed.
5.1.2. I n f l u e n c e of t h e P r o d u c t s
No decrease in the reaction rate due to the olefins
formed could be observed in the cases of isobutanol,
tert-butanol, and cyclohexanol, whereas water strongly
inhibits dehydration to the olefin. Inhibition by water
has also been observed in the formation of olefins from
ethanol 1931 and isopropanol[961. The influence of water
always decreases with increasing alcohol partial pressure and with Iising temperature. The experimental
results cannot be quantitatively described by the usual
Langmuir formula for unimolecular decompositions
with product inhibition. O n the other hand, the reaction rates found can be excellently reproduced over
the entire measuring range up to a water:alcohol
concentration ratio of 15% at all temperatures by the
empirical equation
r = ro@A = I/&
+ bpw)
(r = reaction rate, ro = reaction rate for zeroth order,
both in mole/sec per g of catalyst, pa, p w = partial
pressures of alcohol and water, b = constant). In linear
form, equation (5) becomes:
The validity of equation (6) is confirmed by Figure 6
for the formation of olefin from cyclohexanol. It has
also been checked in the same way for tert-butanoll971
and isobutanol1761.
5.1. Olefin Formation
Under the conditions chosen for the measurements,
tert-butanolC971 and cyclohexanol[761 are dehydrated
above about 130 "C, whereas the reaction begins only
above 160 "C in the case of isobutanol.
5.1.1. I n f l u e n c e of t h e A l c o h o l P r e s s u r e
At low conversions ( t 3 %), i.e. at low partial pressures of water, the reaction rate is independent of the
alcohol pressure over a wide range; the reaction is thus
zeroth order under the following conditions: Isobutanol up to 240 "C above 80 torr, tert-butanol up to
180 "C above 150 tori-, and cyclohexanol up to 180 "C
above 80 torr.
[93] K . V . Topchieva and B. V . Romanovskii, Kinetika i Kataliz 6,
279 (I 965).
1941 K . Kochloef?, M. Kraus, Chon Chin-Shen, L. Beranek, and
V . BaZanr, Collect. czechoslov. chem. Commun. 27, 1199 (1962).
1951 J . R. KittreN and R. Mezaki, Ind. Engng. Chem. 59, 28
[96] L. de Mourges, F. Peyron, Y. Trambouze, and M. Prettre,
J. Catalysis 7 , 117 (1967).
[96a] S. Carrb, S. Sanfangelo, and A . Fusi, Chim. e Ind. (Milano)
48, 229 (1966).
1971 H . Knozinger and H . Buhl, Ber. Bunsenges. physik. Chem.
71, 73 (1967).
Angew. Chem. internat. Edit.
/ Vol. 7 (1968) / No. I0
Fig. 6. Kinetics of olefin formation from cyclohexanol at alcohol
pressures of 115 (0)
and 155 torr
(155 "C in each case), and 100 (0)
and 160 torr (0)
(175 "C in each case).
The expression 0, in equation (5) may be regarded as
an adsorption isotherm. 0, then represents the
relative coverage of the surface with alcohol in the
presence of water. A similar equation with the pressure
at powers less than one was derived by Sipsl97al for
one-component adsorption exhibiting saturation properties at energetically heterogeneous surfaces; the
validity of this equation has been established experimentally for a number of systems [97b, 97~1. The
[97a] R. Sips, J. chem. Physics 16, 490 (1948).
[97b] H . Bradley, Trans. Faraday SOC.31, 1652 (1935).
[97c] E. 0.Wiig and 5'. B. Smith, J. physic. Chem. 55,27 (1951).
IR spectra had shown that under the catalytic conditions, two-point adsorption in accordance with (7)
is favored for alcohols, whereas a one-center adsorption in accordance with (9) must be assumed for
water in the presence of an excess of alcohol.
5.1.3. A c t i v e C e n t e r s
The dehydration activity has repeatedly been found to
exhibit an optimum value for a given pretreatment
temperature of the aluminum oxide 19,10,981. This optimum shows that a certain surface water content
is essential to the reaction, but that other active centers
besides OH groups must also be present on the surface.
This is suggested by the adsorption structure f7), which
requires the presence of oxygen ions. If aluminum
hydroxide (bayerite) is used as the catalyst, the dehydration of tert-butanol does not begin below 180 "C,
i.e. about 50°C higher than on the oxide[s*], Aluminum hydroxide has only O H groups on the surface,
and in this form it is evidently inactive for the dehydration of tert-butanol. Above about 150 "C, the
trihydroxide loses water to form the monohydroxide,
with the result that oxygen ions and aluminum ions
become accessible on the surface as well as O H groups,
and the dehydration of tert-butanol can gradually
A decision as to whether oxygen and/or aluminum
ions are involved as active centers in the dehydration
can be obtained by, poisoning the aluminum oxide
with pyridine. This strong Lewis base is very selectively adsorbed on the incompletely coordinated
aluminum ions at the surface, which are Lewis-acid
centers, with formation of an electron donor-acceptor
complex [111. The activity of the catalyst in olefin formation is not affected by modification with pyridine [75,99,1001. The (Lewis) acidic aluminum ions in
the surface, which are responsible inter alia for isomerization
and poIymerization [991, thus play no
part in the formation of olefins from alcohols, and
the O H groups and oxygen ions may be regarded as the
active centers [*I.
5.1.4. M e c h a n i s m of O l e f i n F o r m a t i o n
Olefin formation begins with an adsorption complex
(7). The water inhibiting the reaction blocks only the
O H groups by a passive H bond. If, in accordance with
[*I Note added in proof (Sept. 19, 1968): The participation of
basic centers has recently been established by poisoning experiments with tetracyanoethylene [fOOa]. Formation of radical
ions when this compound is adsorbed on aluminum oxide has
been observed by means of electron spin resonance. Modification of aluminum oxide catalysts by tetracyanoethylene causes a
considerable inhibition of the formation of not only ethylene
but also ether. Since steric effects can be ruled out, active participation of basic centers in dehydration is thus proved.
I981 K. V . Topchievo, E. N . Rosolovskaja, and 0 . K . Sharajev,
Vestnik Moskovskogo Univ. Ser. Math., Mech., Ast., Physik,
Chem. 14, 217 (1959).
1991 L. Berrinek, M . Kraus, K. Kochloej?, and V. Boiant, Collect.
czechoslov. chem. Commun. 25, 2513 (1960).
[loo] M . Misono, Y.Suito, and Y . Yonedu, Proc. third int. Congr.
Catalysis, Amsterdam 1964, Vol. 1, p. 408.
[100a] F. Figuerus Roca, A . Nohl, L. de Mourges, and Y . Trambouze, C . R. hebd. Stances Acad. Sci. 266, 1123 (1968).
Pevi[521, the area required for an oxygen atom in the
surface of y-Al2O3 [see structure ( 7 ) ] is assumed to be
about 8 &, the H - 0 . . .H angle between the H bonds
i s about 110 ",which is very close to the H - 0 - H angle
in the water molecule. In agreement with the principle
of least structural change [1011, therefore, the water
molecule may already be substantially formed in the
adsorption structure (7) under the conditions of
The further rearrangement could occur in two ways.
With strongly polarized H bonds, (7) may approximate to an oxonium ion and change by charge migration into a carbonium ion, which allows the formation
of an olefin by subsequent transfer of a proton to a n
adjacent surface oxygen ion. In this form the course
of the reaction corresponds to the oxonium-carbonium
mechanism of olefin formation 13-51. However, a
concerted E2-type mechanism is also possible, with
simultaneous rupture of the C-H and C - 0 bonds in
the alcohol. According to the results obtained by
Pines et aZ. 181, it should be assumed that tertiary alcohols, which form relatively stable tertiary carbonium
ions, are dehydrated to olefins by an ionic mechanism,
and the other aliphatic alcohols by a concerted
mechanism. @-Elimination always appears to occur if
the molecule contains Ij-protons [8,102J. In both
mechanisms the water molecule is preformed in the
adsorption complex, and the hydrogen exchange with
the surface observed by Eucken and Wicke [ l o , 13,141
must necessarily take place.
The rate-determining step can be established by measurement of the kinetic isotope effect. investigations of
this type have so far not been published for the dehydration of tert-butanols on y-Al2O3 L1031. No
primary isotope effect was observed in the dehydration
of (CH3)3COD, in which heavy water is formed on
deuterated surfaces; the rate cannot therefore be determined by the desorption of the water or the cleavage
of the RO-H bond. On the other hand, a high primary
alcohol effect is found in the dehydration of
(CD&COD. Below 200 "C, removal of the @-proton
must therefore be the slowest step also with tertbutanol.
Only a secondary isotope effect would be expected in
the case of a pure carbonium-ion mechanism. An E2type mechanism must therefore also be assumed for
the dehydration of tert-butanol at temperatures below
200 "C. The investigations by Pines were all conducted
at considerably higher temperatures. At low temperatures, olefin formation from aliphatic alcohoIs over
aluminum oxide should therefore proceed fundamentally by an E2-type mechanism, the transition
state - determined by the ionization potential of the
alkyl group - containing ionic contributions. These
ionic contributions should become more important
as the temperature increases, so that when the temperature is sufficiently high for olefin formation from
[loll F. 0. Rice and E. TeNer, J. chern. Physics
6, 489 (1938).
[lo21 G . - M . Schwab, 0. Jenkner, and W. Leitenberger, Z . Elektrochem. Ber. Bunsenges. physik. Chem. 63, 461 (1959).
[lo31 H . Kniizinger and A . Scheglilo, unpublished.
Angew. Chem. internat. Edit.J Vol. 7 (1968) 1 N o . I0
tertiary alcohols an El-type transition state will be
present, as was observed by Pines. This concept is
further supported by the strong temperature dependence of the isotope effect.
Olefin formation by dissociation of short-lived alkoxide
groups has also been discussed 1171. The alkoxide
would be formed by dissociative adsorption of alcohol
on incompletely coordinated aluminum ions in the
surface [19,67J. However, since poisoning of these
Lewis centers with pyridine does not decrease the
activity of the aluminum oxide catalyst in the olefin
formation [75,99,1OOJ, a mechanism of this type must be
rejected, at least at low temperatures [*I.
substituted phenylethanols. However, these effects
are less important than the steric effects. Since the
styrenes formed do not polymerize, the authors rule
out the possibility of an ionic dehydration mechanism.
A Hammett relationship between the logarithm of the
rate constant and the 6* constant can be verified for
these reactions (Fig. 7).
5.1.5. R e a c t i o n R a t e s o f t h e A l c o h o l s
Dohse [lo41 and Bork and Tolstopja-Tova [1051 found a
definite relationship between the activation energy for
olefin formation on bauxite and the methyl substitution. According to Dohse, a methyl group in the K position decreases the activation energy by 5.5 kcal/
mole, in the p position by 2.5 kcaljmole, and in the y
position by 0.5 kcal/mole. Stauffer and Kranich [43J
report a constant activation energy of 30.8 kcal/mole
for the formation of olefins from linear alcohols from
ethanol to it-hexanol on y - A I 2 0 3 . Kraus, Kochloefl et
al. [49,941 showed that steric effects are much more important than electronic effects in the dehydration of
aliphatic and cyclic secondary alcohols and of stereoisomeric alkylcyclohexanols on aluminum oxide.
The authors conclude that the rate of reaction of
organic compounds in heterogeneously catalyzed reactions is determined mainly by the steric arrangement
of the reacting molecule and particularly of the chemisorbed species.
Kraus and Kochloefl 11061 also detected electronic effects in the rate of dehydration of meta- and para[*] Note added in proof (Sept. 19, 1968): Confirmation of the
alkoxide mechanism has recently been attempted by investigations of the thermal desorption of ethanol [103a] and diethyl
ether [92a] adsorbed on aluminum oxide. In this case olefin formation IS desribed as a concerted decomposition of the alkoxidetype surface species, whereas ether formation is regarded as proceeding via condensation of two neighboring alkoxide groups.
Using the same experimental technique, Makarov and Shtschekotshichin [103b] arrived a t similar results. However, these authors
assume direct or indirect participation of the surface carboxylate
(cf. 4.1.3) in olefin formation. In view of the high stability of
these surface compounds, such a route seems questionable,
particularly since tert-butanol does not form a compound of
this kind. Moreover, it does not appear certain whether conclusions regarding the mechanism of heterogeneously catalyzed
dehydration of alcohols from the gaseous phase can be drawn
from investigations of thermal desorption, since the experimental conditions differ greatly. In the reaction from the gaseous
phase, molecules are necessarily also present that are physically
adsorbed via H bonds; such molecules can also play a role in the
[103a] H . Arai, J.-I. Take, Y. Saito, and Y . Yoneda, J. Catalysis
9, 146 (1967).
[103b] A . D . Makarov and Yu. M . Shtschekotschichin, Metody
Issled. Katal. Reakts., Akad. Nauk SSSR, Sib. Otd., Inst. Katal.
I , 20 (1965); Chem. Abstr. 66, 53400 (1967).
[lo41 H . Dohse, Z . physik. Chem. Abt. B, Bodenstein-Festband
1931, 533.
[I051 A . Rork and A . Tolstopjaiova, Acta physicochim. URSS 8,
603 (1938).
[lo61 M . Kraus and K. Kochloej4, Collect. czechoslov. chem.
Commun. 32, 2320 (1967).
Angew. Chem. internat. Edit. / Yol. 7 (1968) 1 No. 10
Fig. I . Correlation of the rate constants (200 “ C ) for the dehydration of
substituted phenylethanols with the Hammett equation (after [1061).
1 : 2-(p-tolyl)-ethanol, 2: 2-(p-tert-butylphenyI)ethanol, 3: 2-(p-Buorophenyl)ethanol, 4: 2-phenylethanol. 5: 2-(m-methox~phenyl)ethanol,
6: 2-(1n-fluorophenyl)ethanol.
The “true activation energies” for the dehydration of
aliphatic alcohols to olefins show a marked dependence o n structureflO6al in the series qprimary> qsec >
qtert,If the E2-type mechanism represented in Section
5.1.4 is assumed to be valid, this order can be interpreted in terms of the structurally dependent ionic
contributions to the transition state and of the inductive effect of the alkyl groups adjacent to the @-C
atom. A comparison of the dehydration of tertbutanol (q = 25.5 kcal/mole 1971) and that of isobutanol
( q = 30 kcal/mole [761) may be cited as an example. The
ionic contributions to the transition state are low for
isobutanol compared to tevt-butanol. Furthermore,
the partial positive charge on C, in isobutanol is compensated to some extent by the inductive effect of the
two methyl groups in the y position. Consequently,
in the transition state the acidity of the p protons in
isobutanol is much lower than in tert-butanol, so that
the activation energies for dehydration fall as the
acidity of the 3 protons increases. Comparison of the
activation energies for the different possible reactions
of alcohols that afford a number of olefinic products
o n dehydration clearly shows that not only the inductive effect but also an inductometric effect must
come into play. On dehydration, 3-methyl-2-butanol
and 2-methyl-2-butanol yield comparable amounts of
“Hofmann” and “Saytzeff” products. If the inductive
effect alone were to operate, the lower activation
energy would be expected for the “Hofmann” product.
In fact, however, the activation energy for the formation of this product from 3-methyl-2-butanol and from
2-methyl-2-butanol is lower by 2.5 kcal/mole and
about 1 kcal/mole, respectively, than for the forma[106a] H . Knozinger and H . Biihl, unpublished.
tion of the "Saytzeff" product. The relatively low
differences in the true activation energies are nevertheless well outside the limits of experimental
accuracy. Nor can they be due to lateral interactions
in the adsorbed phase, since in each case two possible
reactions of one alcohol, and not of substrates having
different structures, are being compared. This result
is in accord with the assumption of an E2-type mechanism, because inductometric effects can only play a
role when the double bond of the incipient olefin is
already present in the transition state; this is the case
with transition states of the E2 type but not with those
of the El type.
5.2. E t h e r F o r m a t i o n
The bimolecular ether formation from methanol 11071,
ethanol [1081, n-propanol[75,763, and benzyl alcohol 1761
begins ony-Al203 at about 130-140 "C. In the reaction
ROR+ H20
being an Eley-Rideal mechanism, with a smaller contribution by a Langmuir-Hinshelwood mechanism.
However, these authors worked at conversions of up
to 20 % and temperatures at which the secondary decomposition of the ether and the olefin formation
probably cannot be neglected. Moreover, at the high
conversions and relatively low alcohol partial pressures used, the inhibition of the reaction by the water
formed may be appreciable, and the equilibrium reaction may falsify the rate of formation of the ether.
5.2.2. I n f l u e n c e of t h e P r o d u c t s
No inhibition of the reactions by the ethers formed
has been observed in any case, whereas the ether formation is strongly inhibited by water. The inhibiting
action of water decreases with rising temperature and
increasing alcohol partial pressure [76,107,10*1. As in
the formation of olefins, the rate of ether formation
can be satisfactorily described by the empirical equation ( 5 ) (in the form (6)) (Fig. 8). The quotient 0, may
again be interpreted as an adsorption isotherm, which
the equilibration is fundamentally possible. However,
since the equilibrium lies on the side of ether and
water, the rate of the reverse reaction r-1 at low product concentrations, i.e. at low conversions in the dehydration, will be low in relation to the rate of ether
formation r1. This can be confirmed by thermodynamic estimations and by measurement of the rate
of reformation of the alcohol (-1) [75,107,10*1. Thus
the ether formation from the four alcohols in question
can be followed kinetically at low conversions independently of the equilibrium reaction.
5.2.1. E f f e c t of A l c o h o l P r e s s u r e
The rate of the ether formation from methanol,
ethanol, n-propanol, and benzyl alcohol becomes independent of pressure above about 160 torr at the
highest measuring temperatures (about 200 "C), i.e.
the bimolecular ether formation is a zeroth order reaction in the limiting case of high partial pressures of
alcohol [76,107,1081. This behavior must be expected if
the Langmuir-Hinshelwood mechanism is valid. The
ethers should thus be formed from alcohols over
aluminum oxide in a bimolecular reaction from the
adsorbed phase, since a reaction oder of 1 should be
obtained in the limiting case of high substrate partial
pressures if the Eley-Rideal mechanism were valid.
Balaceanu and Jungers C4OJ reached the same conclusion on the basis of kinetic studies. de Boer et
al.clo971101, on the other hand, propose the superposition of two independent mechanisms for the formation of ether from ethanol, the principal mechanism
[lo71 D. KaIId and H . Knozinger, Chemie-1ng.-Techn. 39, 676
[lo81 H. Knozinger and E. Ress, 2. physik. Chem. N.F. 54, 136
[lo91 J. H. de Boer and R. B. Fahim, Proc. Kon. nederl. Akad.
Wetensch., Ser. B 67, 127 (1964).
[I101 J. H. de Boer, R . B. Fahim, B. G. Linsen, W . J . Visseren,
and W . F . N . M . de Vleesschauwer, I. Catalysis 7 , I63 (1967).
0 75
Fig. 6. Kinetics of ether formation from methanol at 109
160 (0).
and 225 torr (0)
(1 86 "C in each case) and from benzyl alcohol at 25
50 (0).
and 76 ton (0)
(170 'C in each case).
expresses the relative coverage of the surface with alcohol in the presence of the inhibiting water. Alcohol
is molecularly adsorbed in the structure (7), while
water blocks only one surface OH group [structme
(9)]. A limiting form of equation (5) for low surface
coverages and negligible inhibition
r = y o V'PA
was also found by Meye 1311 for the formation of ether
from methanol on y-AI203. It is unusual for an equation (5) having the typical form of the kinetic equation of a unimolecular decomposition to be valid for a
bimolecular reaction. For the ether formation one
would expect an expression containing the square of
the surface concentration of the alcohol. At least one
reactive surface compound having a finite life must be
assumed for ether formation [*21, since the probability
of a bimolecular surface reaction would otherwise be
extremely small. Since a surface alkoxide has been deAngew. Chem. internat. Edit./ Vcl. 7 (1968) / No. 10
tected in the catalytic temperature range for all the
ether-forming alcohols that have been studied spectroscopically (cf. Section 4.1.2), this species must be
regarded as the intermediate in question. The rate
constant for the reaction is proportional to the surface concentration of the ethoxide [*loa]. Because of its
finite life and its relatively high stability, its concentration may be assumed to remain constant during the
reaction. The rate equation therefore only explicitly
contains the concentration of a second reactant ,@,
i.e. an alcohol molecule adsorbed as such by means of
two nonlinear H-bonds. The rate equation for the bimolecular ether formation must consequently be of
the same form as that for the unimolecular olefin
format ion.
5.2.3. A c t i v e C e n t e r s
As in the olefin formation (cf. Section 5.1.3), it can
be shown that OH groups are also necessary as active
centers for the ether formation. Activity measurements
on aluminum hydroxide, on the other hand, show that
the catalytic activity is not due to OH groups alone,
since aluminum hydroxide is at first totally inactive for
ether formation. The ether formation begins with increasing conversion of the trihydroxide into monohydroxide, though only at higher temperatures than
on y-Al203 [75,*21. In the Peri model of the oxide surface, therefore, oxygen ions andlor incompletely coordinated aluminum ions must take part in the reaction as well as OH groups. On blockage of the Lewis
centers with pyridine, in contrast to the olefin formation, formation of ether from ethanol is strongly retarded. Figure 9 shows a plot of the relative rate of
ether formation on a catalyst modified with pyridine at
320°C with respect to the rate on pure y-Al203
against the reaction time.
-" 0 2
t IhlFig. 9. Blockage of the surface of y-AlrO, in the formation ofether from
ethanol. rpy/r = relative reaction rate of a y-AlzO, catalyst modified
with pyridine at 320 "C.
After a build-up period of about 70 minutes, the relative activity remains constant. As a result of the poisoning by pyridine, however, the rate of ether formation decreases by more than 50 %, and the activity
does not regain its original value even after several
hours. This shows that the ether formation, unlike the
olefin formation, involves the active participation of
Lewis centers. The oxygen ions must also be concluded
to be involved on the basis of the structure (7) assumed for the adsorption of molecular alcohol, so that
three types of centers, i.e. OH groups, oxygen ions,
and incompletely coordinated aluminum ions, are
[I lOa1 kf. Knozinger and ff. Stolz, unpublished.
Angew. Chem. internut. Edit. / Val. 7 (1968) 1 No. I0
necessary for ether formation (cf. also the note added
in proof, p. 800).
5.2.4. M e c h a n i s m of E t h e r F o r m a t i o n
The formation of ethers from alcohols is assumed to
result from the reaction of surface alkoxide groups
with molecularly adsorbed alcohol molecules. A
mechanism of this type is suggested by the fact that all
the ether-forming alcohols examined so far can form a
surface alkoxide that can be detected by I R spectroscopy, whereas n o analogous surface compound is
found for the olefin-forming alcohols. Since the alkoxides of the ether-forming alcohols are thermally
more stable than those of the olefin-forming alcohols 178,793, the direction followed by the reaction may
be determined by the thermal stability of the alkoxide.
The temperature range for the undisturbed ether formation thus becomes narrower with decreasing thermal
stability of the alkoxide (cf. Section 2). Thus for isopropanol and n-butanol, olefin and ether formation
and secondary decomposition of the ether begin at
roughly the same temperature.
This naturally does not conclusively prove t h e active participation of surface alkoxides in t h e ether formation a n d t h e
special reaction-directing effect o f this group. However, since
n o opposing evidence has yet been found, a course o f this
type may be assumed, particularly since these views a r e indirectly reinforced by the form o f t h e kinetic equation a n d
t h e active participation o f t h e Lewis centers, o n which t h e
alkoxide formation must occur 119.671. Ether formation by
condensation o f t w o adjacent alkoxide groups has been discussed by Topchieva, Yun-Pin, a n d Smirnovu[l71, but this
would n o t explain the active participation of t h e three types
o f centers (cf. also t h e note added in proof, p. 801).
The rate determining step in the ether formation could
in principle be the desorption of the water formed.
However, since the formation of ether from deuteriomethanols on a deuterated oxide surface, which yields
heavy water, shows no primary isotope effect [1111, the
rate cannot be determined by the desorption of water.
There are two routes by which the reaction could proceed, but no definite decision as to which is actually
followed is possible at present:
1. The R-OH bond of the molecularly adsorbed alcohol and the RO-A1 bond in the alkoxide group are
broken. No primary isotope effect is to be expected in
this case, in agreement with the experimental observations[ll*J. The rate of ether formation may be determined by the cleavage of one of these two bonds,
though it is not known which. Since the separation
energies for homolysis are much lower than those for
heterolysis [112,1131, the reaction would have to be
assumed to be substantially nonionic.
2. Alternatively, the RO-H bond in the alcohol
molecule and the R-OAI- bond in the alkoxide group
11111 H . Knozinger, A . Scheglila, and A. M . Watson, J. physic.
Chem. 72, 2770 (1968).
[I121 Landoh-Bornstein: Zahlenwerte und Funktionen, Vol. 1:
Atom- und Molekularphysik, Part 2 : Molekeln 1. Springer-Verlag,
Heidelberg 1951, p. 36.
11131 V. J . Vedeneyev, L. V . Gurvich, V. N . Kondrutyev, V. A.
Medvedev, and Ye. L . Frunkevich: Bond Energies, Ionization Potentials and Electron Affinities. Arnold Publ. Comp., London
may be broken. This route could be favored on steric
grounds. The a-C atom in the alkoxide group may be
positively charged because of the electron-attracting
action of the aluminum, and, as an electrophile, attack
the hydroxyl oxygen of the alcohol. This will cleave
the passive H-bond between the hydroxyl oxygen of
the alcohol and the surface proton. Rearrangement of
the molecule then leads readily to a structure (14)
This act of desorption corresponds, in the sense of
microscopic reversibility, to the dissociative adsorption
of alcohol with formation of the required surface
alkoxide. According to these views, different intermediates would be required for the formation of
ethers and of olefins.
Brey and Krieger 141, o n the other hand, proposed the same
intermediate, i.e. a carbonium ion, for both reactions. Topchieva, Yun-Pin, and Smirnova 1171 deduced the same scheme,
but replaced the common ionic intermediate by a n alkoxide,
as did Solomon, Bliss, and Butt 1471.
If the second mechanism is correct, the cleavage of the
RO-H bond may be ruled out as the rate-determining
step for the ether formation, since no primary isotope effect is observed[1113. The slowest step in the
ether formation should therefore be the cleavage of the
A10-R bond. The rate could in principle also be determined by the combination of the two groups to
form the ether molecule, but this is unlikely in view
of the high separation energies of the bonds to be
5.2.5. R a t e of R e a c t i o n of t h e A l c o h o l s
postulated on the basis of spectroscopic observations
for the physically adsorbed etherr85J. The second reaction product, water, must be formed in this case by
condensation of two adjacent surface OH groups.
Thus hydrogen exchange with the surface must also
occur in the ether formation.
These views on mechanism can also be used to describe the equilibrium reaction (-l), i.e. the hydration
of ethers
2 2ROH
This brings us to the starting state (12) chosen in the
dehydration of alcohols. I n the following, final step,
the second alcohol molecule must be formed by condensation of an alkoxide group with an OH group.
Table 1 . Reaction rates ro and “true” activation energies of
some alcohols in ether formation on Y-AIzO~.
Methanol [lo71
Ethanol [I081
n-Propanol [761
r,ylO7 (molelg sec)
at 1 8 3 O C
25.9 f 0.6
25.9 j, 0.6
while obeying the principle of microscopic reversibility.
Starting with a surface area in which two adjacent OH
groups have been formed by dissociative adsorption
of a water molecule (i.e. reversal of the desorption of
water during dehydration), the reaction can be schematically rerpesented as proceeding via the stages
(I4),(IS), (16),and (12).
The “true” activation energies for the formation of
ethers from methanol, ethanol, n-propanyl, and benzyl
alcohol are given with their maximum errors in Table 1.
To within the limits of accuracy, the true activation
energies are the same, irrespective of the substrate
alcohol. The zeroth order rates of formation Y O of the
ether decreases from methanol to n-propanol, while
the ether formation from benzyl alcohol is 4.5 times
as fast as from methanol. These observations can be
explained by the reaction of a surface alkoxide group
with a molecularly adsorbed alcohol molecule, with
cleavage of the AIO-R bond as the rate-determining
step. No data are available for the separation energy
of this bond; the separation energy of the homolysis of
the C - 0 bond in alcohols is independent of the alkyl
residue R [77.112,1131. The separation energy of the
T i - 0 bond in titanium alkoxides is also substantially
independent of the alkyl group [114J.It may therefore
be assumed that a similar state of affairs exists for the
separation energy of the C-0 bond in aluminum
alkoxides; this would explain the constancy of the
activation energy. The rates of ether formation, on
the other hand, vary with the substrate. The relative
reaction rates based on that for methanol are plotted
in Fig. 10 against the Taft constant o*. They exhibit
good agreement with the Taft equation. The differences
in rate must therefore be due to inductive effects of
[114] D . C. Bradley and M . .
Trans. Faraday SOC. 62,
2374 (1966).
Angew. Chem. iniernat. Edii. Yol. 7 (1968) } No. 10
becomes increasingly negative with increasing ionic
character of the transition state 11151, the transition
state in the proposed ether formation mechanism
(electrophilic substitution) must be assumed to be
only weakly ionic.
'A662 10;
Fig. 10. Correlation of the relative rates of ether formation with the
Taft equation.
I : benzyl alcohol, 2: methanol, 3: ethanol, 4: n-propanol.
the alkyl residues R. The straight line in Figure 10
has a positive gradient, i.e. the reaction constant p *
of the Taft equation is positive. Since this quantity
I am grateful ro my colleagues Dr. H. Biihl, G . Clement,
G. Kinshofer, R. Kohne, W. Meye, E. Ress, A. Scheglila,
Dr. H. Spannheimer, and H . Stolz for their enthusiastic
cooperation, andto Prof. Dr. G.-M. Schwab for his interest in and encouragement of our work. Special thanks
are due to the Deutsche Forschungsgemeinschaft and the
Stiftung Volkswagen werk for their generous support.
Received: November 10, 1967
[A 662 IE)
German version: Angew. Chem. 80, 778 (19681
Translated by Express Translation Service, London
[115] K. Kochloefl, Advances Catalysis related Subjects 17, 75
C 0 M M U N I CAT10 N S
Carbonyl Olefination with x-Metalated
By U.Schollkopf and F. Gerhart[*I
Dedicated to Professor H . Brockmann on the occasion of his
65th birthday
As is shown by the success of the Wittig reactionfll, the or-
ganic chemist is in urgent need of a process for the positionally specific conversion of carbonyl compounds into olefins.
We have now found that the reaction of u-metalated isocyanides (2) with aldehydes and ketones yields olefins ( 5 )
with their double bond in the position originally occupied by
the carbonyl group. T h e olefination reagents (2) are obtained
from isocyanides ( I ) [21 and bases such as butyllithium. The
metalated oxazolines (4) are important intermediates; they
result from the cyclization of the carbonyl adducts ( 3 ) and
fragment to give the olefin ( 5 ) and metal cyanate. The readiness with which this decomposition occurs depends on t h e
nature of the groups R I , Rz. and R3. The intermediates ( 4 )
and (31, respectively, can be intercepted with protons with
formation of the oxazolines ( 6 ) 131.
I,l-Diphenylethylene [b]
I , I ,2-Triphenylethylene [c]
I-Phenyl-l,3butadiene [dl
re I
1 -Phenyl1.3-butadiene
[a] Decomposition of (41; metalation and carbonyl addition at -7OC";
tetrahydrofuran as solvent.
[b] Together with 2-(diphenylhydroxymethyl)-5,5-diphenyloxazoline.
[c] Together with small quantities of 2-(diphenylhydroxyrnethyl)-4,5.5-triphenyloxazoline.
[dl Together with 5-(~-styryljoxazoline.
[el Together with 2-(diphenylhydroxymethylj-5,5-diphenyl-4-vinyloxazoline.
C = C H - R ~i LiOCN
Angew. Chem. internat. Edit.
/ Yo[. 7 (1968) !No. I0
The metalated isocyanides (2) are more reactive than the
alkylidenetriphenylphosphoranes. For instance, cc-lithiobenzhydrylisocyanide reacts even at -70 'C smoothly with
benzaldehyde to 1,1,2-triphenyIethylene (yield 30 %), whereas
diphenylmethylenetriphenylphosphorane is unreactive towards aldehydes and ketones (11.
I , I ,2-Triphenylethylene:
A solution of 50 mmole of butyllithium in 45 ml of pentane
was diluted with 150 ml of dry tetrahydrofuran. A solution
of 5.85 g (50 mmole) of benzyl isocyanide in 30 ml of tetrahydrofuran was added dropwise at -70 'C. 9.1 g (50 mmole)
of benzophenone dissolved in 50 ml of tetrahydrofuran was
then added dropwise t o the intensely yellow-red solution at
-70 "C. The mixture was allowed to warm up to room temperature and evaporated t o dryness under vacuum o n a
Без категории
Размер файла
1 616 Кб
oxide, alcohol, aluminum, dehydration
Пожаловаться на содержимое документа