Greatly improved activity in ruthenium catalysed butanone synthesis

Article abstract of DOI:10.1039/b108328g

In situ mixing of ruthenium trichloride with one equivalent of 1,10-phenanthroline yields a highly active catalyst for synthesis of butanone from buta-1,3-diene.

Full text of DOI:10.1039/b108328g

Greatly improved activity in ruthenium catalysed butanone synthesis
Robert C. van der Drift,a Wilhelmus P. Mul,b Elisabeth Bouwman*a and Eite Drenta
a Leiden Institute of Chemistry, Gorlaeus Laboratories, Leiden University, P.O. Box 9502, 2300 RA
Leiden, The Netherlands. E-mail: bouwman@chem.leidenuniv.nl; Fax: +31 71 5274451; Tel: +31 71
5274550
b Shell International Chemicals BV, Shell Research and Technology Centre Amsterdam, Badhuisweg 3,
1031 CM Amsterdam, The Netherlands
Received (in Cambridge, UK) 17th September 2001, Accepted 15th November 2001
First published as an Advance Article on the web 6th December 2001
In situ mixing of ruthenium trichloride with one equivalent
of 1,10-phenanthroline yields a highly active catalyst for
synthesis of butanone from buta-1,3-diene.
of chloride ions in some cases had a promoting effect. Here, we
report the reinvestigation of RuCl3 as catalyst precursor in the
synthesis of butanone from butadiene.
The results of in situ experiments for the direct conversion of
butadiene to butanone are shown in Table 1.† The blank
experiments shown in entries 1 and 2 confirm the need for both
ruthenium and phenanthroline. Especially the amount of ligand
is crucial. Whereas RuCl3 with two equivalents of phen (entry
5) gives only moderate activity that is comparable to previous
results (entry 3), RuCl3 with one equivalent phen is by far
superior as is shown in entry 4. This result demonstrates clearly
that with a chloride-containing precursor high activity can be
obtained, contrary to earlier ideas. Mixed with one equivalent of
bipyridine, RuCl3 is only as active as Ru(acac)3 with one phen
(entry 6).
Not only are initial rates higher, the RuCl3–phen system
surpasses the Ru(acac)3–phen system if TONs are considered as
well. After a mere two hours, 1400 turnovers can be attained.
Importantly, the RuCl3–phen system is exceedingly selective.
The most important side products in this reaction are the Diels–
Alder dimer of butadiene (vinyl cyclohexene) and butenes
together with methyl vinyl ketone, but their formation can be
reduced to less than 1% with a suitably low initial butadiene
loading of the autoclave. Thus, high TONs can be reached with
a minimum amount of side products. This is demonstrated by an
experiment in which the amount of butadiene is fed into the
autoclave in two consecutive portions. A cumulative TON of
2750 in less than 10 h was obtained. It is to be foreseen that an
even higher TON can be reached, since no serious catalyst
deactivation occurred over the two runs.
Butanone is an industrial solvent, nowadays produced on a
Mton scale per year.1 It is presently prepared from butenes in a
three-step process according to eqn. (1).2 The use of large
(1)
quantities of concentrated sulfuric acid, which need to be
reconcentrated after the second step, make this process in
principle less desirable from an environmental and an economi-
cal point of view. Therefore, new synthetic strategies have been
developed in recent years.
Butadiene is readily obtained from naphtha cracker C4-
streams and its two double bonds make it susceptible towards
electrophilic addition. Routes to butanal using butadiene via
addition of amines3 or alcohols,4 followed by isomerisation of
the allylic double bond and hydrolysis, have been patented.
Acid-catalysed hydration of butadiene gives rise to two regio-
isomeric alcohols [eqn. (2)], but-3-en-2-ol and but-2-en-1-ol, in
(2)
an equilibrium that lies heavily on the left-hand side. The former
alcohol can be isomerised to butanone [eqn. (3)].5 In a one-pot
(3)
The unexpected high activity of the RuCl3–phen system may
be explained by the role the chloride anion plays in preventing
the ligand redistribution reaction shown in eqn. (4). Its strong
coordination may prevent extensive formation of inactive bis-
and tris-phenanthroline complexes, thereby increasing the
synthesis, the thermodynamically favourable formation of
butanone avoids an extensive (and thus costly) butadiene
recycle. Unfortunately, almost all catalysts capable of allylic
alcohol isomerisation are strongly poisoned by butadiene.
In the early 1990s, the synthesis of butanone from butadiene
catalysed by a mixture of [Ru(acac)3] (Hacac = pentane-
2,4-dione), 1,10-phenanthroline (phen) and a Brønsted acid in
water was published.6 However, the reported cumulative turn
over number (TON) of 1200 in 32 h is not sufficiently high to
be economically interesting. The limited TON is due to catalyst
deactivation via two pathways: ligand redistribution and metal
reduction, which result in inactive [Ru(phen)2(S)2]2+ (S =
solvent) and [Ru(phen)3]2+ species [eqn. (4)].6,7 Yet, despite
Table 1 Synthesis of butanone from butadiene, catalysed by in situ
generated ruthenium complexesa
Entry
Catalyst precursorb
TOF/h21c
k/h21d
1
2
3
4
5
6
7
8
—
RuCl3
0
2.5e
230
960
230
250
34e
0
N.d.f
0.11
0.67
0.19
0.22
0.64
0.41
Ru(acac)3 + 1 eq. phen
RuCl3 + 1 eq. phen
RuCl3 + 2 eq. phen
RuCl3 + 1 eq. bpy
RuCl3 + 1 eq. phen + 3 eq. AgOTs
RuCl3 + 1 eq. pheng
(4)
700
a Reactions were performed in a 250 ml Hastelloy C autoclave at 145 °C for
10 h. Substrate: 10 ml butadiene. Acid: 3.5 mmol toluene-p-sulfonic acid
(HOTs). Solvent: diglyme–water 70+30. b 0.09 mmol of metal complex was
mixed in situ with the appropriate amount of ligand. c Initial turn over
frequency (TOF) calculated with first order rate equations. Turnovers are
determined with GLC. Selectivity to butanone: > 95%. d Determined by
fitting first order curves to autoclave pressure vs. time plots. e Average value
over 10 h. f N.d. = not determined. g Acid: 1.75 mmol HCl
considerable effort,8 until now no catalyst was found to be more
active.
Strongly coordinating anions like chlorides were originally
thought6 to hamper catalysis by blocking vital coordination
sites. In fact, [RuCl3·xH2O] in combination with two equiva-
lents of 2,2A-bipyridine (bpy) gave only low activity. However,
during the course of our studies,7 it was found that the presence
2746
Chem. Commun., 2001, 2746–2747
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b108328g

concentration of catalytically active mono-phenanthroline ru-
thenium complexes. To verify this hypothesis, several samples
taken during the reaction were analysed with mass spectrometry
(MS). In the early stages of the reaction, Ru(III)- and Ru(II)-
complexes with one phenanthroline, such as [Ru(phen)Cl2]+
and [Ru(phen)(S)Cl]+ are dominant. This is the first time that
ruthenium complexes containing only one phenanthroline
ligand could be identified in this reaction.6 Two other species
TON of at least 2750 after 10 h. This superior result compared
to previously reported systems emphasises the need for catalysts
that prevent ligand redistribution. The possible use of relatively
cheap RuCl3 and HCl brings an industrial process for the direct
synthesis of butanone from butadiene within reach. Future
studies are aimed at further minimising ligand redistribution,
optimisation of the reaction conditions as well as addressing the
intriguing question on the valency of Ru in the active complex
under butadiene hydration conditions and better understanding
of deactivation reactions.
This research was supported by the Technology Foundation
STW, applied science division of NWO and the technology
programme of the Ministry of Economic Affairs. The authors
are indebted to Mr W. W. Jager (SRTCA, The Netherlands) for
his skilful technical assistance and to Dr W. J. L. Genuit
(SRTCA, The Netherlands) for recording the mass spectra. Dr J.
G. de Vries (DSM, The Netherlands) and Mr W. G. Reman
(SRTCA, The Netherlands) are thanked for stimulating dis-
cussions.
are present in minor amounts: [Ru(phen)2Cl2]+ and its Ru(II
)
analogue. Most noteworthy is the complete absence of com-
plexes with three phenanthroline ligands. During the reaction,
mono-phenanthroline complexes are converted to complexes
with two phenanthroline ligands, but not with three. At the end
of the 10 h reaction, complexes with only one phenanthroline
ligand can still be detected. These results from MS prove that
chloride ions indeed coordinate strongly to the ruthenium centre
and prevent the formation of [Ru(phen)3]n+. In the Ru(acac)3–
phen system, between 20 and 40% of the ruthenium was present
as [Ru(phen)3]2+.6
The present RuCl3–phen system is much more sensitive to the
ligand to metal ratio than the Ru(acac)3–phen system. When
two equivalents of phenanthroline are added to RuCl3, the
Notes and references
strong chloride coordination becomes
a problem. [Ru-
† In a typical experiment, a high-pressure autoclave was filled with 0.09
mmol [RuCl3·xH2O], the appropriate amount of ligand and 3.5 mmol acid
(for the acid-catalysed hydration of butadiene as shown in eqn. (2)). After
addition of the diglyme–water (70+30) solvent mixture, the autoclave was
closed and purged three times with dinitrogen. Next, buta-1,3-diene was
added (10 ml) and the autoclave was heated to 145 °C. After 10 h, the
autoclave was cooled to room temperature and the contents were analysed
with GLC.
(phen)2Cl2]+ is formed instantly and chloride dissociation is
slow.9 With Ru(acac)3, initial dissociation of acac is rate
limiting and the influence of phenanthroline concentration is
therefore much smaller.
The greater basicity and flexibility of bipyridine compared to
phenanthroline render its coordination more reversible and this
in turn makes ligand redistribution more prominent. Hence, the
lower activity of RuCl3 with one equivalent bpy can be
explained by the increased formation of [Ru(bpy)2Cl2]+, which
is confirmed with MS on a spent catalyst.
1 Kirk-Othmer Encyclopedia of Chemical Technology, 4th Edition, ed. J. I.
Kroschwitz Wiley Interscience, New York, 1998.
2 Ullmann's Encyclopedia of Industrial Chemistry, 6th Edition, Electronic
Release, ed. A. Eckerle, Wiley-VCH, Weinheim, 2000.
3 J. Kanand, M. Roper, R. Paciello and A. Thome, U.S. Patent No.
5892125, 1999 (to BASF).
Complete removal of chloride ions by silver( ) salts,
I
replacing them with the non-coordinating anions toluene-p-
sulfonate or trifluoromethane sulfonate, makes the catalyst
much less stable. After an induction period, the reaction starts
with a comparable rate (k = 0.64 h21; Table 1, entry 7), but a
TON of only 340 is reached after 10 h. These results underline
the role of the chloride ion. Removal of chlorides increases
ligand redistribution considerably, thereby reducing catalyst
lifetime and overall product yield. On the other hand, increasing
the chloride concentration by addition of NaCl or using HCl
instead of toluene-p-sulfonic acid (Table 1, entry 8), decreases
reaction rate and TON only slightly. This evidently shows that
the hitherto found low activity of RuCl3–phen systems was due
to the amount of ligand and not to the presence of coordinating
chloride anions.
4 J. Kanand, R. Paciello and M. Roper, U.S. Patent No. 6166265, 2000 (to
BASF).
5 For examples of allylic alcohol isomerisation see: ref 8; G. F. Emerson
and R. Pettit, J. Am. Chem. Soc., 1962, 84, 4591; D. Baudry, M.
Ephritikhine and H. Felkin, Nouv. J. Chim., 1978, 2, 355; K. Tani, Pure
Appl. Chem., 1985, 57, 1845; S. H. Bergens and B. Bosnich, J. Am. Chem.
Soc., 1991, 113, 958; B. M. Trost and R. J. Kulawiec, J. Am. Chem. Soc.,
1993, 115, 2027; J.-E. Bäckvall and U. Andreasson, Tetrahedron Lett.,
1993, 34, 5459; D. V. McGrath and R. H. Grubbs, Organometallics,
1994, 13, 224.
6 E. Drent, Eur. Patent No. 457387, 1991 (to Shell); F. Stunnenberg, F. G.
M. Niele and E. Drent, Inorg. Chim. Acta, 1994, 222, 225.
7 R. C. van der Drift, J. W. Sprengers, E. Bouwman, W. P. Mul, H.
Kooijman, A. L. Spek and E. Drent, manuscript in preparation.
8 R. C. van der Drift, M. Vailati, E. Bouwman and E. Drent, J. Mol. Catal.
A: Chem., 2000, 159, 163.
In conclusion, an improved catalyst system has been found
for direct synthesis of butanone from butadiene. RuCl3 in situ
combined with one equivalent of phenanthroline catalyses this
conversion with an initial TOF of 960 h21 and a cumulative
9 R. G. Wilkins, Kinetics and Mechanism of Reactions of Transition Metal
Complexes, 2nd edn, VCH, Weinheim, 1991, 400.
Chem. Commun., 2001, 2746–2747
2747