Liquid-liquid biphasic, platinum-catalyzed hydrosilylation of allyl chloride with trichlorosilane using an ionic liquid catalyst phase in a continuous loop reactor

Article abstract of DOI:10.1002/adsc.200800438

The platinum-catalyzed hydrosilylation of allyl chloride with trichlorosilane was investigated in an ionic liquid-organic biphasic reaction mode. After an ionic liquid screening and repetitive batch mode experiments, the process was realized in a continuous mode using a loop reactor concept with integrated continuous separation and recycling of the ionic liquid catalyst phase. The continuous reactor could be operated for 48 h at constant activity and selectivity without addition of platinum indicating that platinum leaching into the product phase was far below 1 ppm. Enhanced selectivity for the product trichloro(3-chloropropyl)silane (compared to the state-of-the art) and the possibility to use simple platinum tetrachloride (PtCl₄) as platinum source are further attractive features of this new ionic liquid-based process concept.

Full text of DOI:10.1002/adsc.200800438

FULL PAPERS
DOI: 10.1002/adsc.200800438
Liquid-Liquid Biphasic, Platinum-Catalyzed Hydrosilylation of
Allyl Chloride with Trichlorosilane using an Ionic Liquid Catalyst
Phase in a Continuous Loop Reactor
Norbert Hofmann,a Andreas Bauer,b Thomas Frey,b Marco Auer,b,c
Volker Stanjek,b Peter S. Schulz,a Nicola Taccardi,a and Peter Wasserscheida,*
a
Lehrstuhl fꢀr Chemische Reaktionstechnik, Universitꢁt Erlangen-Nꢀrnberg, Egerlandstraße 3, 91058 Erlangen, Germany
Fax : (+49)-9131-852-7421; phone: (+49)-9131-852-7420; e-mail: wasserscheid@crt.cbi.uni-erlangen.de
Wacker Chemie AG, Johannes-Hess-Strasse 24, 84489 Burghausen, Germany
Current address: Momentive Performance Materials Quartz GmbH, Borsigstrasse 1–7, 21502 Geesthacht. Germany
b
c
Received: July 16, 2008; Published online: November 4, 2008
Abstract: The platinum-catalyzed hydrosilylation of num leaching into the product phase was far below
allyl chloride with trichlorosilane was investigated in 1 ppm. Enhanced selectivity for the product tri-
an ionic liquid-organic biphasic reaction mode. After chloro(3-chloropropyl)silane (compared to the state-
an ionic liquid screening and repetitive batch mode of-the art) and the possibility to use simple platinum
experiments, the process was realized in a continuous tetrachloride (PtCl4) as platinum source are further
mode using a loop reactor concept with integrated attractive features of this new ionic liquid-based pro-
continuous separation and recycling of the ionic cess concept.
liquid catalyst phase. The continuous reactor could
be operated for 48 h at constant activity and selectiv- Keywords: catalyst recycling; hydrosilylation; ionic
ity without addition of platinum indicating that plati- liquids; multiphase catalysis; platinum
Introduction
interesting example being transition metal-catalyzed
hydrosilylation.
Biphasic catalysis in a liquid-liquid system provides a
powerful approach to combine the advantages of both Markovnikov addition of an Si H moiety to alkenes
homogeneous and heterogeneous catalysis. The reac- with formation of a new Si C bond.
Hydrosilylation represents the metal-catalyzed anti-
À
[2,3]
À
The mecha-
tion mixture consists of two immiscible solvents. Only nism of this reaction has been studied in detail and is
one phase contains the catalyst, allowing easy product known to follow the so-called Chalk–Harrod mecha-
separation and catalyst recycling by simple decanta- nism[4] with catalyst precursors like chloroplatinic acid
tion. However, the right combination of catalyst, cata- being frequently applied.[5,6] The reaction is of high
lyst solvent, reactants and products is crucial for the preparative value and has also found extensive use
success of biphasic catalysis.[1] The catalyst solvent has for the technical production of organosilane com-
to provide excellent solubility and immobilization for pounds and organo-modified oligosiloxanes.[7] Howev-
the catalyst complex to prevent metal leaching into er, attempts to realize liquid-liquid biphasic hydrosily-
the product phase under the operating conditions. lation failed in the past due to the reactivity of most
À
These solvent characteristics have to be achieved Si H compounds with the polar solvents that might
without the solvent competing with the substrate for be used to form a biphasic system with the products
the free coordination sites at the catalytic centre. An- or an alkane extraction phase. In fact, traditional bi-
other important precondition is given by the fact that phasic solvent mixtures such as, for example, water-
the catalyst solvent should be stable towards feed- alkane or butanediol-alkane are not applicable. As a
stock, products and process conditions in order to consequence, in the traditional industrial hydrosilyla-
provide a fully recyclable catalyst solution with a long tion processes, the expensive transition metal catalyst
lifetime.
remains homogeneously dissolved in the product or is
This latter requirement may sound trivial but it has recycled by extensive downstream processing from
been the major hurdle to realize successfully liquid- the high-boiling residues of the product distillation.
liquid biphasic reactions in many cases, a particular
Ionic liquids are salts with melting points below
1008C which have attracted much interest in both
Adv. Synth. Catal. 2008, 350, 2599 – 2609
ꢂ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2599

Norbert Hofmann et al.
FULL PAPERS
academic and industrial research laboratories during could demonstrate a simple recycling by decantation
the last decade.[8] With respect to applications in mul- as the selected supports are liquids under the reaction
tiphase catalysis, this class of innovative liquid materi- conditions and solids at room temperature.[13]
als has shown to offer unique combinations of proper-
A first publication demonstrating the industrial at-
ties, such as, e. g., the combination of extremely low tractiveness of hydrosilylation catalysis in ionic liquids
vapour pressure, low nucleophilicity, high solvation was provided by researchers from Degussa AG
properties and tunable miscibility with common or- (nowadays Evonik Industries).[14] They investigated
ganic solvents.[9]
the reaction of 1-hexadecene with oligosiloxanes
The field of ionic liquid multiphase hydrosilylation using extremely low catalyst loadings in the ionic
was pioneered by Behr and Toslu in 2000 who investi- liquid phase (5 ppm Pt relative to the total mass of
gated the hydrosilylation of 10-undecenoic acid substrates) and reached high conversions within af-
methyl ester with triethoxysilane using the Speier cat- fordable reaction times of 60 min. In repetitive batch
alyst, H2PtCl6·6H2O, dissolved in the ionic liquid 1- experiments, total turnover numbers of 1,500,000
butyl-3-methylimidazolium
hexafluorophosphate could be realized by recycling the ionic catalyst solu-
([BMIM] [PF6]).[10] By applying cyclohexane as ex- tion over ten cycles. The authors stated the impor-
traction solvent and by making use of the thermomor- tance to avoid 1,3-dialkylimidazolium ionic liquids for
phic phase behaviour of the system, these authors this application as the in situ NHC-carbene complex
were able to demonstrate the feasibility of biphasic formation appeared to be the reason for low catalyst
catalysis. A successful recycling of the ionic liquid cat- activity in some systems. With pyridinium tetrafluoro-
alyst solution was also demonstrated. An example of borate, average conversions of >86% (with K2PtCl4)
Rh-catalyzed hydrosilylation was provided by van and >82% [with PtCl
2ACHTUGNTRENN(GNU PPh3)2] could be realized.
Koeten et al. who applied a fluorinated version of In the present paper, we studied the feasibility to
Wilkinsonꢃs catalyst dissolved in an ionic liquid func- expand these approaches for ionic liquid-organic bi-
tionalized with perfluorinated alkyl chains to react 1- phasic hydrosilylation catalysis to the transformation
octene with dimethylphenylsilane.[11] Initially, the cat- of chlorosilanes. We investigated the industrially rele-
alyst proved to be more active in the ionic liquid com- vant Pt-catalyzed reaction of allyl chloride and tri-
pared to conventional solvents. However, reaction chlorosilane to form trichloro(3-chloropropyl)silane,
rates went down during recycling and dropped to only which is a multi-ton product as additive for polymers
10% activity of the same reaction in benzene in the and as intermediate for the production of other sur-
15th cycle. The authors could explain this decrease in face active-organosilanes (Scheme 1).[15] As a matter
activity with the detection of Rh and phosphine leach- of fact, this process is technically challenging due to
ing from the ionic liquid phase (respectively, 4% and the difficult handling of the highly flammable, vola-
2% per cycle, with regard to the initial added tile, very easily hydrolyzable and chemically aggres-
amounts). Another example of Rh-catalyzed hydrosi- sive trichlorosilane. Moreover, the reaction shows a
lylation was published in 2006 by Marciniec and co- complex selectivity pattern. In fact, besides the main
workers.[12] These authors demonstrated reasonable reaction leading to desired trichloro(3-chloropropyl)-
recyclability of the ionic catalyst solution in a polysil- silane, trichlorosilane and allyl chloride can be con-
oxane modification reaction via hydrosilylation using verted in a side reaction to tetrachlorosilane and pro-
N-alkylpyridinium hexafluorophosphates in Rh-cata- pylene. The propylene generated in this way under-
lyzed hydrosilylation reactions. Lai and co-workers goes consecutive hydrosilylation forming trichloropro-
Scheme 1. Hydrosilylation of allyl chloride and trichlorosilane to form trichloro(3-chloropropyl)silane – main reaction and
relevant side reactions forming tetrachlorosilane, propene and trichloropropylsilane.
2600
asc.wiley-vch.de
ꢂ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2599 – 2609

Liquid-Liquid Biphasic, Platinum-Catalyzed Hydrosilylation of Allyl Chloride
FULL PAPERS
pylsilane. The latter finds no commercial use and con- ionic liquid; b) stability of the ionic liquid against the
stitutes a waste material (Scheme 1). Thus, increasing highly reactive trichlorosilane substrate; c) stability of
the selectivity to trichloro(3-chloropropyl)silane in the desired product trichloro(3-chloropropylsilane) in
comparison to the actual state-of-the art is an impor- the ionic liquid under operative conditions. These
tant issue of the process optimization in addition to three criteria were applied to screen a number of
the aforementioned catalyst immobilization and recy- ionic liquids for their suitability for the hydrosilyla-
cling aspects.
tion of allylchloride and trichlorosilane.
Moreover, in the present paper, we report for the
The ionic liquids selected for the initial screening
first time an ionic liquid-based hydrosilylation reac- are depicted in Table 1.
tion in a continuous reaction mode including continu-
The selected Pt source, PtCl4, was found to be suffi-
ous separation and recycling of the ionic liquid cata- ciently soluble in all these ionic liquids under investi-
lyst phase. The transition from repetitive batch mode gation (Pt loading above 4000 ppm as determined by
operation to a continuous reaction mode is a very im- ICP), both at room temperature and at the reaction
portant step in the process development since long- temperature of 1008C.
term catalyst stability, leaching aspects and productiv- The stability of the ionic liquid against trichlorosi-
ity issues can be demonstrated much more reliably lane proved to be a more selective prerequisite for
from a continuous system.
the ionic liquid selection. All ionic liquids have been
dried very carefully prior to these stability experi-
ments. Nevertheless, in the ionic liquids [EMIM]
[EtSO4], [EMIM] [tosylate] and [BMIM] [BF4] de-
composition was immediate, obviously due to a reac-
tion of the ionic liquidꢃs anion with trichlorosilane
even at room temperature. This decomposition pro-
Results and Discussion
Preliminary Ionic Liquid Screening
The first important step of our study was to demon- cess was also obvious from the fact that gaseous prod-
strate that ionic liquids can indeed act as a suitable ucts and a solid formed while trichlorosilane was
reaction medium for the Pt-catalyzed hydrosilylation added to the ionic liquid. Also the ionic liquids
of allyl chloride with trichlorosilane. Initially, we fo- [EMIM] [Et2PO4], [EMIM] [CF3SO3] and [BMIM]
cused our attention on three aspects that were as- [PF6] showed some signs of silane decomposition but
sumed to be of key relevance for the successful reali- to a lesser extent. Importantly, no decomposition of
zation of the concept, namely: a) sufficient solubility HSiCl3 was observed in all [Tf2N]-based ionic liquids.
of the envisaged Pt catalyst precursor PtCl4 in the Obviously, the strongly electron-withdrawing effect of
Table 1. Ionic liquids used for the initial screening.
Name
Abbreviation
Structure of cation
1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide
ACHTUNGTRNEUN[NG EMIM] [Tf2N]
1-ethyl-2,3-dimethylimidazolium bis(trifluoromethanesulfonyl)imide
ACHTUNGTRENN[UNG EMMIM] [Tf2N]
1-butyl-3-methylpyridinium bis(trifluoromethanesulfonyl)imide
trimethyloctylammonium bis(trifluoromethanesulfonyl)imide
ACHTUNGTREN[NUNG 3-MBPy] [Tf2N]
AHCTUNGTRE[NGNUN OMA] [Tf2N]
1-ethyl-3-methylimidazolium ethyl sulfate
ACHTUNGTRENNUNG
1-ethyl-3-methylimidazolium diethyl phosphate
1-ethyl-3-methylimidazolium 4-methylbenzenesulfonate
1-ethyl-3-methylimidazolium trifluoromethanesulfonate
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
1-butyl-3-methylimidazolium hexafluorophosphate
1-butyl-3-methylimidazolium tetrafluoroborate
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
Adv. Synth. Catal. 2008, 350, 2599 – 2609
ꢂ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2601

Norbert Hofmann et al.
FULL PAPERS
the two trifluoromethyl groups leads to the fact that the possibility for in situ formation of NHC com-
the SO2 groups of the anion become inert against tri- plexes does not exert a major negative effect on the
chlorosilane. Even during prolonged heating (96 h, formation and activity of the catalytic species for this
1208C in an autoclave) these [Tf2N]À-based ionic liq- specific reaction. In fact, a potentially NHC-forming
uids proved to be stable against HSiCl3.
ionic liquid cation, such as [EMIM]+, showed similar
In order to ascertain the stability of the reaction results compared to a non-NHC-forming cation, such
product in contact with the ionic catalyst solution, a as [EtPy]+.
solution of trichloro(3-chloropropyl)silane in the ionic
The comparison of the liquid-liquid biphasic runs
liquid [EMMIM] [Tf2N] containing 2000 ppm of PtCl4 (Table 2, entries 1–3) with the homogeneous system
was heated to 1008C for 6 h. No transformation of tri- applying an original technical catalyst sample provid-
chloro(3-chloropropyl)silane, for example, by sub- ed by Wacker (Table 2, entry 4) was also very instruc-
stituent exchange at the Si atom, could be found. This tive. In fact, already the first semi-batch experiments
distinguishes the bis(trifluoromethanesulfonyl)imide provided a clear indication that – at least for this spe-
systems from chloroaluminate ionic liquids for which cific hydrosilylation reaction – the ionic liquid bipha-
the ligand exchange reaction at organochlorosilanes sic system leads to a slightly improved selectivity (by
has been described in an earlier patent of Wacker 3.0 to 4.6% in S1), a result that could be reproduced
Chemie AG.[16]
many times in the course of our study.
In this context, it is noteworthy that the difference
in S1 vs. Stotal selectivity originates from the additional
hydrosilylation reaction of the by-product propene.
While S1 takes only into account the ratio of allyl
chloride hydrosilylation vs. chloride transfer, Stotal is
Experiments in a Semi-Batch Glass Reactor at
Atmospheric Pressure
Having identified bis(trifluoromethanesulfonyl)imide additionally influenced by the consecutive hydrosily-
ionic liquids as suitable catalyst phases, our first set of lation of propene (formed as the by-product in chlo-
experiments was carried out in a semi-batch mode ride transfer). Due to the highly volatile nature of the
using a glass reactor at atmospheric pressure (for de- propene, Stotal values of different experiments can
tails see Experimental Section).
Table 2 displays the results for selected [Tf2N]À mass balance of all reactants can be fully accounted.
ionic liquid systems compared with the homogeneous, Concerning activity, it is important to note that the
only be properly compared for cases in which the
industrially applied catalyst system (here denoted as turnover frequencies given in Table 2 are not suitable
“Wacker catalyst”). The latter is rather unspecified for a proper comparison of catalyst activities as all
and predominantly consists of a mixture of dimeric Pt data were collected at full or almost full conversion.
species with cylooctadiene as ligands in such an However, the values obtained clearly indicate that
amount to ensure a Pt concentration of about also in the case of liquid-liquid biphasic operation
500 ppm in an organic matrix consisting of mixed sa- very good activities can be achieved even with low
turated hydrocarbons.
catalyst loadings.
Varying the ionic liquid cation resulted in little dif-
Another interesting result from these first semi-
ferences in selectivities and reactivities (Table 2, en- batch experiments is the fact that the ionic liquid-
tries 1–3). In contrast to reports by Dyson et al.,[13] based reactions can indeed be successfully carried out
Table 2. Hydrosilylation of allyl chloride with trichlorosilane in a pressure-less semi-batch reaction system: comparison of
different solvents and catalyst systems.[a]
[b]
Entry
Solvent
Catalyst
Catalyst concentration
ppm
Conversion
[%]
Stotal
[%]
S1[c]
[%]
TOF[d]
[hÀ1]
1
2
3
4
N
PtCl4
PtCl4
600
300
200
500
100
98.4
100
100
71.7
74.7
74.3
70.6
81.9
80.9
80.3
77.3
1430
3780
3280
1180
PtCl4
Wacker catalyst[e]
[a]
Conditions: catalyst concentration: 200–600 ppm Pt in the solvent, HSiCl3/allyl chloride: 1.25:1.0, Tstart =1008C, reaction
time: 20 min, mass balance could be accounted for all experiments within 2%.
[b]
Stotal =mol [(3-chloropropyl)trichlorosilane]/ACHTUNGTRENNUNG
trichlorosilane)}.
S1=mol [(3-chloropropyl)trichlorosilane]/CAHTUNGTRENNNUG
TOF=turnover frequency: mol allyl chloride converted per mol Pt and time.
Wacker catalyst: mixture of dimeric Pt species with cyclooctadiene as ligand in such an amount to ensure a Pt concentra-
tion of about 500 ppm in a matrix of saturated hydrocarbons.
[c]
[d]
[e]
2602
asc.wiley-vch.de
ꢂ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2599 – 2609

Liquid-Liquid Biphasic, Platinum-Catalyzed Hydrosilylation of Allyl Chloride
FULL PAPERS
using relatively cheap and readily available PtCl4 as experiments are presented in Figure 1 (not all the
the source of platinum. In contrast to the actual in- runs are depicted).
dustrial system, no addition of extra ligands and no
As a result of these recycling experiments, no indi-
pre-formation of the Pt catalyst is required which sim- cation for a systematic loss in activity or selectivity
plifies the catalyst handling and reduces catalyst costs. was found. Remarkably, all recycling runs showed a
This finding reflects the favourable solvation proper- higher selectivity compared to the monophasic experi-
ties of the [Tf2N]À ionic liquids for simple Pt salts ments using the homogeneous “Wacker catalyst”
such as PtCl4.
(entry 4, Table 2). Given this excellent recyclability,
we expected very low Pt leaching into the organic
product phase. Indeed, this was confirmed by induc-
tive coupled plasma-atom emission spectroscopy
(ICP-AES) indicating that Pt leaching was below the
detection limit of 1 ppm for all organic samples. In
Recycling Experiments in Repetitive Semi-Batch
Mode
Once having ascertained that the ionic liquid biphasic order to quantify a possible leaching of ionic liquid
concept can be successfully applied to the hydrosilyla- into the product phase, the collected organic phase
tion of technically relevant chlorosilanes, we contin- samples of the nine recycling runs were united, con-
ued investigating the recyclability of the ionic catalyst centrated by evaporation under high vacuum (1508C,
solution. In such recycling experiments both the cata- 10À3 mbar) to 5 mL of heavies and analyzed by
lyst stability under reaction/separation conditions and 1H NMR spectroscopy for the characteristic ionic
the catalyst leaching behaviour are probed.
liquid imidazolium peaks. Using this methodology, no
We chose [EMMIM] [NTf2] for these recycling ex- hint for a leaching of the ionic liquid into the product
periments mainly due to the fact that this specific phase could be obtained. The same concentrate was
ionic liquid shows rapid (within less than ten seconds) again analyzed by ICP-AES and also this residue
phase separation from the reaction mixture. As a showed no sign of Pt being present above the detec-
matter of fact, the time required for complete phase tion limit of our spectrometer.
separation is of great relevance for dimensioning the
settler unit with quicker phase separation allowing for
a smaller separator unit in continuous operation.
In our recycling experiments, the ionic catalyst
Batch Autoclave Experiments
phase was allowed to separate from the product All hydrosilylation experiments described so far have
phase by gravity after each run, isolated under inert been carried out in an open glass apparatus equipped
conditions and reused without further treatment, pu- with a very efficient condenser (cooling temperature:
rification or activation. The results of the original run À108C; see Experimental Section). This cooling
(Figure 1, run No. 0) and nine consecutive recycling system was required to minimize evaporative losses of
the trichlorosilane feedstock (bp=318C). Obviously,
this kind of open system operation is not very realistic
to mimic a potential technical application of this bi-
phasic ionic liquid-based hydrosilylation system.
Therefore we decided to perform also a set of batch
experiments in a closed autoclave reactor. This reac-
tor was equipped with an overhead stirrer to provide
optimum mixing conditions between the two liquid
phases in the reactor. Moreover, a thermocouple was
mounted to probe the temperature of the reaction
mixture during reaction (for more details see Experi-
mental Section). Table 3 presents the results of these
experiments.
Unfortunately, we were not able to operate the au-
toclave isothermally when the organic phase consisted
of pure feedstock at the start of the reaction. As can
Figure 1. Pt-catalyzed liquid-liquid biphasic hydrosilylation
be seen from Table 3, entry 1, the strongly exothermic
of HSiCl3 and allyl chloride – results of the recycling experi-
reaction in the autoclave led to an increase in temper-
ments in repetitive semi-batch mode (X=conversion %;
ature of more than 1008C after 2 min reaction time.
Stotal =mol [(3-chloropropyl)trichlorosilane]/ACTHNGUTERN{UNG mol[(3-chloro-
This behaviour could be ascribed to the limited cool-
ing capacity of the applied autoclave equipment. To
avoid such undefined thermal conditions we decided
to reduce the concentration of feedstock in the reac-
propyl)trichlorosilane]}+mol(tetrachlorosilane)+mol (pro-
pyltrichlorosilane); S1=mol [(3-chloropropyl)trichlorosila-
ne]/ACHTUNGTRENNUNG{mol[(3-chloropropyl)trichlorosilane]+mol(tetrachloro-
silane)]}.
Adv. Synth. Catal. 2008, 350, 2599 – 2609
ꢂ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2603

Norbert Hofmann et al.
FULL PAPERS
Table 3. Hydrosilylation of allyl chloride and trichlorosilane in a batch autoclave reactor – test of different feedstock concen-
trations to ensure isothermal reaction conditions.[a]
Entry
Toluene added vs. reactants (molar ratio)
Tmax (DT exotherm)
Conversion [%]
Stotal[b] [%]
S1[c] [%]
1
2
3
None
3:1
5:1
2008C (1008C)
1108C (108C)
<1038C (<38C)
99.0
94.7
74.4.
67.6
72.3
76.2
78.8
80.3
81.4
[a]
Conditions: 180 ppm PtCl4 in [EMMIM] [Tf2N], HSiCl3/allyl chloride: 0.95:1.00, Tstart =1008C, reaction time: 60 min,
mass balance could be accounted for all experiments within 2%.
[b]
[c]
Stotal =mol [(3-chloropropyl)trichlorosilane]/ACHTUNGTRENNUNG
trichlorosilane)}.
S1=mol [(3-chloropropyl)trichlorosilane]/CAHTUNGTRENNNUG
tor by adding an inert solvent to the reaction mixture. tense mixing (to maximize the exchange surface) and
Toluene was selected for this purpose as it can be a high ratio of heat exchange surface/reactor volume
easily dried and also forms a miscibility gap with the (for efficient heat removal). All these requirements
ionic liquid thus supporting rapid phase separation were fulfilled in the best way by using a loop reactor
after the reaction.
as schematically shown in Figure 2. A similar reactor
Indeed, by using a diluted reactant phase the exo- concept was formerly realized by our group and
therm could be significantly reduced while reaching proven to be effective for the Ni-catalyzed dimeriza-
approximately isothermal conditions with a molar tion of olefins using chloroaluminate ionic liquids.[17]
ratio of reactants/toluene of 1 to 5 (Table 3, entry 3).
Under these conditions, the conversion was not com- with an internal phase recirculation. Due to a high
plete after 60 min reaction time [X(silane)=74.4%], volumetric flow in the recycle loop compared to the
This rig can be compared to a mixer settler set-up
AHCTUNGTRENNUNG
as the reduced feedstock concentration resulted in a inflow of feed and the outflow of products, the reactor
decrease of the reaction rate. From the comparison of is fully back-mixed and shows the residence time dis-
entries 1–3 in Table 3, it can be seen that the exo- tribution of a continuous stirred-tank reactor. The
therm in entries 1 and 2 led to the expected increase main reactor components are a static mixer (Figure 3,
in reaction rate but to a reduced selectivity for the upper right side, providing a well defined fine disper-
formation of (3-chloropropyl)trichlorosilane. Remark- sion) and a phase separator (Figure 3, left side) pro-
ably, the selectivities in the temperature-controlled, viding enough residence time for a full phase separa-
diluted system (Table 3, entry 3) were very similar to tion. During reaction, the liquid-liquid biphasic reac-
the undiluted, intensively cooled, atmospheric pres- tion mixture is circulated in the loop with high flow
sure system (comparison with entry 2, Table 2) prov- rates (up to 7 L/min). Under these conditions the
ing that also in the autoclave experiment the ionic
liquid biphasic approach can realize very attractive se-
lectivities if proper heat removal and temperature
control are realized.
Liquid-Liquid Biphasic Hydrosilylation in a
Continuous Loop Reactor
Encouraged by these promising results we decided to
transfer the concept of a biphasic hydrosilylation of
chlorosilanes into a continuous reactor system. As
stated above, the continuous flow operation mode
permits a reliable and accurate “proof of concept”.
Especially, the long-term stability of the ionic catalyst
solution can be more reliably probed under conditions
that are close to the technical application scenario.
Realizing the continuous operation we aimed to im-
mobilize the ionic catalyst solution over the whole re-
action time in the reactor without any exchange or re-
generation. At the same time, the biphasic exothermic
Figure 2. Scheme of the applied loop reactor for the biphasic
Pt-catalyzed hydrosilylation of allyl chloride and trichlorosi-
lane using [EMMIM] [Tf2N] as catalyst immobilization
character of the hydrosilylation reaction required in- phase.
2604
asc.wiley-vch.de
ꢂ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2599 – 2609

Liquid-Liquid Biphasic, Platinum-Catalyzed Hydrosilylation of Allyl Chloride
FULL PAPERS
Figure 3. Detailed view of the continuous hydrosilylation loop reactor as applied.
static mixer (Sulzer ChemTech, TYP SMX DN10, 100 bar. The main separation of the product from the
80 mm) – acting as micromixer elements in the reac- ionic catalyst solution was realized in a gravity separa-
tor loop – provides an efficient dispersion of the ionic tor that was integrated in the reactor loop. In this
catalyst solution in the organic phase. The degree of way, most of the catalyst was always present in the re-
dispersion can be monitored by an inspection window. actor allowing for a direct investigation of the catalyst
The droplet size of the ionic liquid and the ex- reactivity by analyzing the isolated products. Addi-
change surface can be estimated using the following tionally, the system was also equipped with an exter-
correlation.[18]
nal ionic liquid separator that was used to recycle
where: 1 [kg/L]=density of organic and IL phase; small amounts of ionic liquid catalyst solution that
e [W/kg]=specific energy supply; s [N/m]=surface left the reactor loop through the separator. This exter-
tension; m [Pas]=total viscosity; Vi [dimensionless]= nal separator unit was also used to collect ionic liquid
viscosity number according to Calabrese;[19] and Wec samples for analytical purposes during reactor opera-
[dimensionless]=critical Weber number.
tion. 24 h/7 days operations were possible with the ap-
The silane and the olefin feedstock were fed from plied set-up due to a safety concept that monitored
two storage vessels with a maximum total inflow of all critical parameters of the rig leading to an auto-
2.5 litres per hour. Two cooler/heater units ensured matic safety shut-down if one critical parameter was
full temperature control in a temperature range up to out of range.
2008C. The applied loop reactor had a volume of
The following paragraphs describe the results of a
650 mL and could be operated at pressures up to typical hydrosilylation experiment in the continuous
Adv. Synth. Catal. 2008, 350, 2599 – 2609
ꢂ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2605

Norbert Hofmann et al.
FULL PAPERS
Figure 4. Results of a continuous hydrosilylation experiment in the loop reactor with varying silane/allyl chloride feed com-
positions (X=Conversion %).
loop reactor (details concerning the experimental pro- step negligibly affected the time to reach steady-state
cedure to run the rig are given in the Experimental operation.
Section). The results were obtained by applying an
The first steady-state conditions were kept for 4 h.
ionic catalyst solution of 150 g (100 mL) of [EMMIM] During this time, approximately 4 litres of organic
[Tf2N] containing 400 ppm PtCl4. This ionic catalyst feedstock was transformed reaching conversions of
solution was pumped into the loop reactor and the re- 50–70% with respect to chlorosilane and allyl chlo-
actor filling was completed by addition of toluene up ride. The fact that product analysis suggested different
to complete flooding of the reactor. After that, the degrees of conversions for both feedstock molecules
circulation pump was switched on and started the is explained by the fact that the by-product pathway
liquid-liquid biphasic mixing. With a volumetric flow to form tetrachlorosilane and trichloropropylsilane
of 3.2 LminÀ1, the mean Sauterꢃs diameter of the consumes two moles of silane, one to form the tetra-
ionic-liquid dispersed droplets was estimated to be chlorosilane and one to react with propene giving tri-
0.27 mm (according to the above-mentioned correla- chloropropylsilane.
tion). To start the reaction, a mixture of trichlorosi-
In order to study the influence of the feedstock
lane and allyl chloride was fed into the reactor. composition on the product selectivity, the silane/
Figure 4 shows the conversion with respect to both olefin ratio was varied in the subsequent hours of the
feedstocks and the selectivities obtained for a run experiment to roughly 2:1 (starting from 6.0 h TOS)
over 48 h time-on-stream (TOS) leading to a total or- and 3:1 (starting from 25.0 h TOS). Remarkably, no
ganic throughput of more than 14 kg with an average major changes in product selectivity were observed,
conversion of more than 50%.
only the specific conversions of chlorosilane and
The reaction system needed 1.5 h time-on-stream olefin changed according to the changing feedstock
(corresponding to 2.2 litres of organic throughput) to composition. The system in general proved to be very
reach a first steady state. This finding corresponds robust with respect to product selectivity, realizing
with the well established empirical statement that the total selectivities for (3-chlorpropyl)trichlorosilane in
reactor volume of a loop reactor has to be exchanged the range between 61% and 72% throughout the 48 h
three times (volume of loop reactor=0.65 litre) to TOS.
reach steady-state operation. This result clearly indi-
For a proper interpretation of the data in Figure 4,
cated a rapid in situ formation of the active Pt catalyst it is important to mention that between TOS 6 h und
in the ionic liquid since the catalyst pre-formation 23 h as well as between 29.5 h and 45 h TOS the rig
2606
asc.wiley-vch.de
ꢂ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2599 – 2609

Liquid-Liquid Biphasic, Platinum-Catalyzed Hydrosilylation of Allyl Chloride
FULL PAPERS
was continuously running overnight with greatly re- gands (that are required in the homogeneous system
duced feed throughput per hour in order to avoid to provide enough Pt solubility) thus decreasing cata-
overflow of tanks and sampling devices. Most interest- lyst costs and simplifying catalyst handling.
ingly, by re-adjusting the initial conditions of the first
After batch-mode screening experiments to identify
steady state after 48 h TOS (corresponding to a total the best ionic liquid and to optimize the reaction con-
organic throughput of 14.4 litres) it was possible to ditions, a continuous loop reactor with integrated
reach again the original conversion and selectivity ionic liquid separation and recycling was designed
values (comparison of steady state after 1.5 h TOS and tested. Using this reactor concept, operational
with conversion and selectivity after 48 h TOS). This stability of an ionic catalyst solution containing
finding proved impressively the operational stability 400 ppm PtCl4 has been achieved for 48 h. During the
of the ionic liquid-based biphasic system and the effi- continuous experiment, stable conversions (50–70%
ciency of the in situ catalyst separation and recycling. with respect to both chlorosilane and olefin for differ-
Moreover, this remarkable operational stability ent operating conditions) and very attractive, stable
provides also an impressive proof for the fact that Pt product selectivities [62–71% total selectivity to (3-
leaching into the organic phase was indeed far below chloropropyl)trichlorosilane] have been demonstrat-
1 ppm (the detection limit of our ICP-AAS analysis) ed. In agreement with this excellent operational sta-
as can be shown by a simple calculation. In fact, over bility, it was shown that Pt leaching into the organic
the 48 h TOS continuous operation, a constant Pt product phase was far below 1 ppm of Pt without the
leaching of 1 ppm into the organic phase would have need for any additional ligand to immobilize the
resulted in a total Pt leaching of 15 mg of Pt into the metal in the ionic liquid phase.
generated organic product. However, the total
We are convinced that the above-presented concept
amount of PtCl4 added to the ionic liquid was only of biphasic ionic liquid organic hydrosilylation cataly-
50 mg corresponding to a platinum mass of 29 mg. sis is of significant practical relevance and can be
Thus, a 1 ppm Pt leaching should have caused roughly easily implemented even into already existing plants.
a loss of half the added Pt load with a consequent
drop of activity which was clearly not observed.
Moreover, by analyzing the Pt amount remaining in Experimental Section
the ionic liquid after the 48 h TOS continuous opera-
tion, we could detect the same concentration of Pt as General Remarks
was initially added (within the error margin of our
Allyl chloride and trichlorosilane were obtained in technical
grades from Wacker AG, Burghausen. [EMIM] [Tf2N],
[BMIM] [Tf2N], [BMMIM] [Tf2N], [OMA] [Tf2N], [EMIM]
ICP-AAS analyses, typically Æ5%). It is very remark-
able that this excellent platinum immobilization was
realized without the presence of any specific ligand or
tailor-made ionic liquid structure.
[EtSO4], [EMIM] [Et2PO4], [EMIM] [tosylate], [EMIM]
[TfO], [BMIM] [PF6] and [BMIM] [BF4] were purchased
from Solvent-Innovation GmbH, Cologne. All other ionic
liquids were prepared from the chloride or bromide salts ac-
cording to well known procedures.[20] All chemicals and pre-
cursors for the synthesis of the ionic liquids were obtained
from commercial sources (Fluka, Aldrich, Merck) and used
without further purification. All ionic liquids were dried for
6 h under high vacuum (10À3 mbar, 508C) to remove traces
of water. The water content of all ionic liquids and allyl
chloride was confirmed to be below 20 ppm by Karl–Fischer
titration using a Metrohm 756 KF Coulometer with a Hy-
dranalꢄ Coulomat AG reagent. All experiments were car-
ried out under strictly inert conditions using Schlenk-tech-
nique and argon as inert gas.
Conclusions
In this paper, we have reported another successful ap-
plication of biphasic transition metal catalysis, namely
the Pt-catalyzed hydrosilylation reaction of allyl chlo-
ride with HSiCl3 using bis(trifluoromethanesulfonyl)-
ACHTUNGTRENNUNGimide salts as the catalyst immobilization phase. Our
results impressively demonstrated the unique features
ionic liquids can offer to homogeneous liquid-liquid
biphasic catalysis. For this example, the combination
of beneficial catalyst solvation, broad miscibility gaps
and excellent chemical stability proved to result in a
new and highly attractive approach to the Pt cata-
lyzed hydrosilylation of trichlorosilane.
Preparation and Handling of the Ionic Catalyst
Solution
Compared to todayꢃs technically applied homoge-
neous Pt catalyst (the “Wacker catalyst”), the product
selectivities in the studied biphasic ionic liquid sys-
tems proved to be 3–5% higher. Moreover, due to the
solvation properties of the ionic liquid, the ionic
liquid-based reaction can apply unmodified PtCl4 as
PtCl4 was added under an argon atmosphere to a pre-dried
sample of ionic liquid in such an amount to give a 2000 ppm
solution. The sample was stirred and complete solution was
obtained after 20 min with all ionic liquids under investiga-
tion. Later, additional ionic liquid was added to obtain the
desired catalyst concentration for the experiment. The base-
the Pt source. This fact avoids the need for specific li- line experiment with the “Wacker catalyst” was carried out
Adv. Synth. Catal. 2008, 350, 2599 – 2609
ꢂ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2607

Norbert Hofmann et al.
FULL PAPERS
by using a sample of the commercial homogeneous catalyst
obtained from Wacker AG, Burghausen.
high vacuum for 12 h prior to the experiment. The feedstock
of allyl chloride was delivered via an HPLC pump (0.4–
40 mLminÀ1). The equivalent amount of trichlorosilane was
controlled via a rotameter and conveyed by using 15 bar of
argon pressure. A back-pressure regulator ensured a con-
stant excess pressure (0–60 bar). A circulation pump (0–
8 LminÀ1) guaranteed the residence time behaviour of a
CSTR. At the gravimetric phase separator a diffuser slowed
the velocity of the ionic liquid-product dispersion down, so
that the ionic liquid settled and could be recirculated with
the main amount of the organic phase. The amount of or-
ganic products discharged from the reactor corresponded
exactly to the amount of feedstock added as the reactor was
operated at all times in a fully flooded state. All non-gas-
eous products were collected in a 1000-mL Schlenk flask.
Experiments in a Semi-Batch Glass Reactor at
Atmospheric Pressure
The hydrosilylation reactions under atmospheric conditions
were carried out in a 150-mL three-necked-flask, immersed
in an oil-bath with an automatic internal temperature con-
troller. The flask was equipped with a dropping-funnel, a
reflux condenser (cooling temperature Tcool =À108C) and a
magnetic stirrer. Prior to use, all glass-ware was dried under
vacuum at 1008C. 15 g of the ionic liquid catalyst solution
prepared from the stock solution mentioned above were in-
serted directly into the tree-necked-flask. The reactants
[12 g of trichlorosilane, 5.5 g of allyl chloride both dissolved
in 5 g of the product trichloro(3-chloropropyl)silane for
better heat control during the experiment] were introduced
into the dropping funnel. The reaction was started via the
addition of the reactants, whereby the dropping rate did not
exceed the reflux rate. After the completion of the reaction,
the content of the flask was collected with a 30-mL syringe
and the organic layer allowed to separate from the ionic
liquid.
Product Analysis and Mass Balance
All GC analyses of the reaction mixtures were carried out
using a Varian CP 3800 TCD with a 30 m Varian GS-Q
column and helium as the mobile phase (carrier gas veloci-
ty=1 mLminÀ1). The silanes were detected using a WLD
detector at a temperature of 3008C. Due to low peak inten-
sities, the samples were not quenched or diluted with addi-
tional solvents prior to analysis. The solvents were injected
at 2008C with a split of 0.01 using a temperature program
from 110–2508C. Correction factors for each reactant and
product were determined and periodically confirmed. The
mass balance was calculated from all components and could
be accounted for all reported experiments within 2%.
Recycling Experiments in Repetitive Semi-Batch
Mode
After the reaction, the ionic liquid was separated from the
product phase by mean of a 30-mL syringe. The ionic liquid
was reintroduced in the same glass reactor set-up mentioned
above and reused without further treatment like distillation
or extraction. Fresh feedstock was added via the dropping
funnel.
ICP-AES Analysis
ICP-AES analyses were performed using a Spectro Cir-
osCCP spectrometer at 1600 W at a wavelength of 214, 423
and 306, 471 nm. For sample preparation, 5 mL of organic
or ionic liquid sample were quenched with 7 mL of ethanol
in an ice-bath forming the related ethoxysilanes which are
known to be more inert and easier to handle. These samples
were then analyzed without further dilution by direct injec-
tion into the spectrometer. Prior to the measurements, the
system was calibrated using a five-point standard regression
with standard samples formed from 18 mL pure ionic liquid/
25 mL ethanol/9 mL pure quenched ethoxysilanes for ionic
liquid analysis and 25 mL ethanol/25 mL pure quenched
ethoxysilanes for organic phase analysis. The platinum stan-
dard solution was purchased by Sigma Aldrich (Specpure
1000 mgmLÀ1 in 20% HCl).
Batch Experiments in Autoclave
The hydrosilylation reactions were carried out using an 80-
mL stainless-steel autoclave with a 60-mL PTFE insert. The
temperature was controlled both on the external surface
and via an internal thermal element. The autoclave was
equipped with a magnetic stirrer and a dropping funnel for
setting a specific starting time. Prior to use the autoclave
was dried under vacuum at 1008C. All the chemicals were
charged by using Schlenk techniques and under strictly inert
conditions. The ionic liquid (2.5–7.5 mg PtCl4 in 15 g
[EMMIM] [BTA]) was inserted directly into the PTFE
insert. The reactants (12 g of trichlorosilane and 5.5 g of
allyl chloride) were introduced into the dropping funnel.
The reaction was started via the addition of the reactants.
For the concentration vs. time experiment samples were
taken via a needle valve without loosing pressure. After the
completion of the reaction, the reactor content was collected
with a 30-mL syringe and the organic layer was separated
from the ionic liquid.
NMR Analysis
1H and 19F NMR spectra of the collected organic phase for
traces of IL were recorded without further solvent using an
insert tube filled with DMSO-d6 on a Jeol ECX 400 spec-
trometer at 400 and 376 MHz, respectively.
For the recycling experiments the ionic liquid was reused
without further treatment.
Continuous Experiments using a 650-mL Loop
Reactor
Acknowledgements
150 g of ionic liquid with a total amount of PtCl4 of 400 ppm
were pumped into the loop reactor, which was dried under
We gratefully acknowledge financial support from Wacker
Silicones, a division of Wacker Chemie AG.
2608
asc.wiley-vch.de
ꢂ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2599 – 2609

Liquid-Liquid Biphasic, Platinum-Catalyzed Hydrosilylation of Allyl Chloride
FULL PAPERS
References
2nd edn., Wiley-VCH, Weinheim, 2007, pp 369–463;
c) T. Welton, Coord. Chem. Rev. 2004, 248, 2459–2477.
[10] A. Behr, N. Toslu, Chem. Eng. Technol. 2000, 23, 122–
125.
[1] B. Drießen-Hçlscher, P. Wasserscheid, W. Keim, CAT-
TECH 1998, 47–52.
[2] A. J. Chalk, J. F. Harrod, in: Organic Synthesis via
Metal Carbonyls, (Eds.: I. Wender, P. Pino) Wiley, New
York, 1977, pp 673–704.
[11] J. van den Broeke, F. Winter, B.-J. Deelman, G. van Ko-
ten, Org. Lelt. 2002, 22, 3851–3854.
[12] B. Marciniec, A. Maciejewski, K. Szubert, M. Kurdy-
kowska, Monatsh. Chem. 2006, 137, 605–611.
[13] a) J. J. Peng, J. Lee, Y. Bai, H. Qiu, K. Jiang, J. Jiang,
G. Lai, Catal. Comm, 2008, 9, 2236–2238; b) J. Jian
Peng, J. Yun Li, Y. Bai, W. Hong Gao, H. Yu Qiu, H.
Wu, Y. Deng, G. Qiao Lai, J. Mol. Cat. A. 2007, 278,
97–101.
[14] a) T. J. Geldbach, D. Zhao, N. C. Castillo, G. Laurenczy,
B. Weyershausen, P. J. Dyson, J. Am. Chem. Soc. 2006,
128, 9773–9780; b) B. Weyershausen, K. Hell, U,
Hesse, Green Chem. 2005, 7, 283–287.
[3] a) C. R. Krueger, R. Carl, Intern. Symp. Organosilicon
Chem. Sci. Commun. 1965, pp 74–76; b) B. Marciniec,
in: Applied Homogeneous Catalysis with Organometal-
lic Compounds, Vol. 1 (Eds.: B. Cornils, W. A. Herr-
mann), Wiley-VCH, Weinheim, 1996, pp 487–506; c) B.
Cornils, W. A. Herrmann, in: Applied Homogeneous
Catalysis with Organometallic Compounds, Vol. 2,
(Eds.: B. Cornils, W. A. Herrmann), Wiley-VCH, Wein-
heim, 1996, pp 575–601.
[4] A. J. Chalk, J. F. Harrod, J. Am. Chem. Soc. 1965, 87,
16–21.
[5] J. L. Speier, J. A. Webster, G. H. Barnes, J. Am. Chem.
Soc. 1957, 79, 974–979.
[6] J. L. Speer, Adv. Organomet. 1979, 17, 407–447.
[7] a) S. Stadtmueller, Polymers and Polymer Composites
2002, 10, 49–62; b) M. A. Brook, in: Silicon in Organic,
Organometallic and Polymer Chemistry, John Wiley
and Sons, New York, 2000; c) R. G. Jones, W. Ando, J.
Cojnowski, in: Silicon-Containing Polymers, Kluwer
Academic Publisher, Dordrecht, 2000.
[15] A. Bauer, T. Frey, P. Wasserscheid, P. Schulz, N. Hof-
mann, PCT Int. Appl. WO 2007-EP56210 20070621,
2008.
[16] K.
Mautner,
B.
Goetze,
German
Patent
DE10157198 A1.
[17] P. Wasserscheid, M. Eichmann, Catal. Today 2001, 66,
309–316.
[18] F. A. Streiff, P. Mathys, T. U. Fischer, Rec. Prog. Genie
Proc. 1997, 11, 307–314.
[19] R. V. Calabrese, P. D. Berkman, AIChE J. 1988, 34,
602–609.
[20] P. Bonhꢅte, A.-P. Dias, N. Papageorgiou, K. Kalyana-
sundaram, M. Grꢁtzel, Inorg. Chem. 1996, 35, 1168–
1178.
[8] a) P. Wasserscheid, W. Keim, Angew. Chem. Int. Ed.
2000, 39, 3772–3789.
[9] a) V. I. Parvulescu, C. Hardacre, Chem. Rev. 2007, 107,
2615–2665; b) P. Wasserscheid, P. S. Schulz, in: Ionic
Liquids in Synthesis, (Eds.: P. Wasserscheid, T. Welton),
Adv. Synth. Catal. 2008, 350, 2599 – 2609
ꢂ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2609