Biphasic, Homogeneous, and Heterogeneous Hydrocarbon Conversion Reactions with Novel Aluminum Chloride Based Catalyst Systems

Biphasic, Homogeneous, and Heterogeneous Hydrocarbon Conversion Reactions with Novel Aluminum Chloride Based Catalyst Systems DISSERTATION zur Erlangung des akademischen Grades eines Doktors der Naturwissenschaften (Dr. rer. nat.) im Fach Chemie der Fakultät für Biologie, Chemie und Geowissenschaften der Universität Bayreuth vorgelegt von Dipl. Chemiker Matthias Dötterl geboren in Naila Bayreuth, 2011 The following work has been carried out from May 2008 to August 2010 at the Chair of Inorganic Chemistry II of the Universität Bayreuth under the supervision of Prof. Dr. Helmut G. Alt. This thesis fulfils the requirements of the doctoral degree of the faculty of Biology, Chemistry and Geological Sciences at the University of Bayreuth. Thesis submitted: 11.01.2011 Date of Scientific Colloquium: 21.03.2011 Dean of the Faculty: Prof. Dr. Stephan Clemens Examination Committee: First Referee: Prof. Dr. Helmut G. Alt Second Referee: Prof. Dr.-Ing. Andreas Jess Third Referee: Prof. Dr. Axel H. E. Müller Chairman: Prof. Dr. Karlheinz Seifert Für Katharina Für KatharinaFür Katharina Für Katharina und Emil und Emilund Emil und Emil „Wer immer tut, was er schon kann, bleibt immer das, was er schon ist.“ Henry Ford Abbreviations br broad M n number average molecular weight BMIM 1-butyl-3-methylimidazolium M w weight average molecular weight Bu butyl MAO methylaluminoxane °C degree celsius Me methyl C quaternary carbon mg milligram(s) (for 13 C NMR spectroscopy) MHz Megahertz Cy cyclohexyl min minute(s) d doublet or day(s) ml milliliter dd duplicated doublet mmol millimol δ chemical shift (ppm) MP 2-methylpentene(s) DMB 2,3-dimethylbutene(s) MPa megapascal dme dimethoxyethane MS mass spectrometry Et ethyl N-MP N-methylpyrrole EROEI energy returned on energy invested NMR nuclear magnetic resonance g gram(s) not. det. not determined GC gas chromatography p para GC-MS gas chromatographyPDI polydispersity index mass spectrometry ( = M w / M n ) h hour(s) PE polyethylene HDPE high density polyethylene Ph phenyl HEX n-hexene(s) R alkylchain ∆H m melting enthalpy RON research octane number HT-GPC high temperature RT room temperature gel permeation chromatography s singlet Hz Hertz SHOP Shell higher olefin process IFP Institut Français du Pétrole SILP supported ionic liquid phase J coupling constant (Hz) t triplet or time K Kelvin T temperature L ligand T m melting point LPG liquefied petroleum gas TMA trimethylaluminum LLDPE linear low density polyethylene THF tetrahydrofurane LNG liquefied natural gas wt% weight percent m meta X halide m multiplet Ξ reference frequency M molecular weight (for NMR spectroscopy) (for mass spectrometry) I Table of Contents 1 Summary .............................................................................................................. 1 2 Introduction ........................................................................................................ 11 2.1 Fossil Fuels and Oil Refining ............................................................................... 11 2.2 Dimerization of Short Chain Olefins ..................................................................... 15 2.3 Cocatalysts for Transition Metal Catalyzed Olefin Dimerization and Polymerization Reactions .................................................................................... 19 2.4 References .......................................................................................................... 22 3 Overview of Thesis Results .............................................................................. 24 3.1 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids ................................... 24 3.2 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids ................................................................................................................. 25 3.3 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations ................................................................................................................ 25 3.4 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions ........................................................ 26 3.5 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes ........................................................................ 27 3.6 Solubility Behaviour of TiCl4, ZrCl4, and HfCl4 in Chloroaluminate Ionic Liquids ................................................................................................................. 28 3.7 Oxidative Coupling and Catalytic Cracking of Alkanes in Lewis Acidic Chloroaluminate Ionic Liquids Enhanced by Molecular Oxygen .......................... 28 3.8 Individual Contribution to Joint Publications ........................................................ 29 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids ........................... 32 4.1 Main Text ............................................................................................................. 32 4.2 References .......................................................................................................... 46 4.3 Supporting Information ........................................................................................ 48 5 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids ......................................................................... 52 5.1 Introduction .......................................................................................................... 53 5.2 Results and Discussion ....................................................................................... 55 II 5.3 Experimental Section ........................................................................................... 66 5.4 References .......................................................................................................... 71 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations ............................................................................................................... 73 6.1 Introduction .......................................................................................................... 74 6.2 Results and Discussion ....................................................................................... 76 6.3 References .......................................................................................................... 85 6.4 Supporting Information ........................................................................................ 88 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions ................................................. 103 7.1 Introduction ........................................................................................................ 104 7.2 Results and Discussion ..................................................................................... 108 7.3 Experimental Section ......................................................................................... 119 7.4 References ........................................................................................................ 122 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes .................................................................. 126 8.1 Introduction ........................................................................................................ 127 8.2 Results and Discussion ..................................................................................... 129 8.3 Experimental Section ......................................................................................... 140 8.4 References ........................................................................................................ 143 9 Solubility Behaviour of TiCl4, ZrCl4, and HfCl4 in Chloroaluminate Ionic Liquids .............................................................................................................. 146 9.1 Introduction ........................................................................................................ 146 9.2 Results and Discussion ..................................................................................... 147 9.3 Experimental Section ......................................................................................... 153 9.4 References ........................................................................................................ 154 10 Oxidative Coupling and Catalytic Cracking of Alkanes in Lewis Acidic Chloroaluminate Ionic Liquids Enhanced by Molecular Oxygen ................ 157 10.1 Main Text ........................................................................................................... 158 10.2 References ........................................................................................................ 161 11 List of Publications .......................................................................................... 163 12 Danksagungen ................................................................................................. 165 13 Declaration / Erklärung ................................................................................... 167 1 Summary 1 1 Summary The aim of this thesis was the development of novel cocatalysts for nickel catalyzed olefin dimerization and oligomerization reactions. For this purpose, cheap and air stable Lewis acidic chloroaluminate ionic liquids were the starting point for our investigations. Chloroaluminate melts are immiscible with hydrocarbon phases, and thus, biphasic oligomerization reactions allow a simple product separation by decantation. Scheme 1: Formation of highly Lewis acidic [Al2Cl7]– anions in chloroaluminate ionic liquids (a), their donor-acceptor interaction with a buffer additive (b) and examples for successfully employed buffers (c). Unmodified chloroaluminate melts containing excess aluminum chloride predominantly catalyze non selective, cationic olefin oligomerization reactions. We found that the addition of triphenylamine, triphenylphosphine, or triphenylbismuth donors efficiently suppressed these side reactions to yield an ideal cheap and air stable ionic liquid cocatalyst for common nickel complexes used for olefin dimerization or oligomerization reactions (Scheme 1). Especially triphenylbismuth was able to buffer slightly acidic chloroaluminate melts as well as highly acidic compositions. For nickel catalyzed dimerization reactions, we found that high buffering levels led to very high selectivities to give dimers, while systems with lower buffer contents were less selective but 1 Summary 2 extremely active. We proposed a mechanism for Ni(II) catalyzed selective olefin dimerization reactions in alkylaluminum free buffered chloroaluminate melts that explained our catalytic results. Further, the melting points of acidic chloroaluminate melts were reduced upon addition of BiPh3. Thus, room temperature ionic liquid compositions derived from 100 different organic halide salts were screened towards their performances in nickel catalyzed selective propene dimerization reactions. An Nmethylpyrrolidine hydrochloride based system maintained an excellent performance even after seven catalytic cycles. Amines and BiPh3 used for such systems can be easily recovered by acid base extraction. Subsequently, an optimized composition was successfully employed to dimerize ethene, propene, 1-butene, and 1-hexene with high activities and selectivities. The presence of sterically demanding tricyclohexylphosphine ligands in such systems led to the formation of valuable branched products. In order to minimize leaching effects and to investigate the interactions of a donor additive with the ionic liquid, a cationic para-trimethylammonium substituted triphenylphosphine derivative was synthesized. Its interaction with the ionic liquid was monitored by means of 31P NMR spectroscopy (Scheme 2). While strong P–Al interactions in highly Lewis acidic compositions resulted in a broad 31P NMR peak, a weak interaction in neutral chloroaluminate melts was reflected in a sharp 31P NMR signal. Scheme 2: 31P NMR spectra of the interaction of a cationic triphenylphosphine derivative with a highly Lewis acidic (left) and a neutral chloroaluminate ionic liquid (right). The concept of buffering highly Lewis acidic aluminum chloride centers was also transferred to binary homogeneous systems. In combination with stoichiometric amounts of BiPh3 or N-methylpyrrole buffer, aluminum chloride readily dissolved in toluene and methylene chloride to form a highly efficient, cheap and air stable 1 Summary 3 cocatalyst for nickel catalyst precursors. The catalytic results of these buffered homogeneous cocatalysts for olefin dimerization reactions were superior to common pyrophoric cocatalysts used for the activation of nickel catalyst precursors like EtAlCl2 or Et2AlCl. Furthermore, the Lewis acidities of these binary homogeneous cocatalyst solutions could be tuned precisely by the choice of the solvent and the type and amount of buffer. Also, buffer addition efficiently suppressed isomerization reactions of the αolefinic products. The interaction of BiPh3 and N-methylpyrrole with AlCl3 was monitored by 27Al NMR spectroscopy (Scheme 3). Strong interactions resulted in broad 27Al NMR signals and weakly Lewis acidic compositions. In contrast, cocatalyst systems with sharp 27Al NMR signals displayed high Lewis acidities. Scheme 3: 27Al NMR spectra of the donor-acceptor interaction of AlCl3 with triphenylbismuth (left) and N-methylpyrrole (right) in toluene. Further, the concept of buffering highly Lewis acidic aluminum species was extended to heterogeneous systems (Scheme 4). The Lewis acidities of EtAlCl2 or Et2AlCl modified silicas were reduced by the addition of BiPh3 or N-methylpyrrole buffers. Buffered surface modified silica cocatalysts were successfully employed to activate nickel complexes for highly selective olefin dimerization reactions. Also, surface modified silica proved to be an ideal substrate for the formation of supported ionic liquid phase (SILP) cocatalyst systems used in nickel catalyzed olefin dimerization reactions. 1 Summary 10 werden. Die Polymerausbeute, das Molekulargewicht und der Polydispersitätsindex (PDI) des so erzeugten Polyethylens konnte durch die Wahl des Lösungsmittels und der Art und der Menge des zugegebenen Donor-Additivs beeinflusst werden. Die Olefinpolymerisation mit Übergangsmetallkomplexen wurde bisher noch nicht mit auf AlCl3 basierenden Cokatalysatoren beschrieben, weder in homogenen Systemen noch in ionischen Flüssigkeiten. Im Rahmen dieser Arbeit wurden die ersten luftstabilen Aluminium-Cokatalysatorsysteme entwickelt, die ausschließlich auf AlCl3 basieren. Das Konzept der DonorAkzeptor-Wechselwirkung von stark Lewis-sauren Aluminiumzentren mit schwach Lewis-basischen Donor-Molekülen konnte erfolgreich auf zweiphasige, homogene und heterogene Systeme angewendet werden. Die erhaltenen Mischungen erwiesen sich als effiziente Cokatalysatoren für die selektive Dimerisierung und Polymerisation von Olefinen mit Nickelkomplexen. Durch eine geeignete Kombination des Lösungsmittels, der Art und der Menge der Donor-Komponente und der Prozessparameter können so maßgeschneiderte Cokatalysatoren für eine Vielzahl von Katalysatorvorstufen erzeugt werden. Die getesteten Systeme erwiesen sich bei der Olefin-Dimerisierung und Polymerisation als äußerst aktiv, langzeitstabil und sehr selektiv. Durch die Verwendung von preisgünstigen Komponenten und der Möglichkeit, Donor-Additive oder Halogenidsalze zurückzugewinnen, stellen derartige Systeme vielversprechende Alternativen zu den standardmäßig verwendeten Alkylaluminium-Cokatalysatoren dar. Stark Lewis-saure Chloroaluminatschmelzen wurden weiterhin für Crackingund Kupplungsreaktionen gesättigter Alkane verwendet. Beim Cracking von n-Heptan in Gegenwart von molekularem Sauerstoff erhöhte sich der Umsatz um mehr als eine Größenordnung im Vergleich zu Reaktionen unter Schutzgas. Dabei fielen vor allem gasförmige Alkane als Reaktionsprodukte an. Dagegen führte die gleiche Reaktion mit Cycloalkanen zu oxidativen Kupplungsreaktionen und damit zur Bildung von Dimeren und Trimeren. In Abwesenheit von O2 bildeten sich hingegen nur Skelettisomere. Außerdem wurde festgestellt, dass sich bis zu äquimolare Mengen an Zirkoniumtetrachlorid und Hafniumtetrachlorid in neutralen Chloroaluminatschmelzen lösten. Die daraus resultierenden ternären Mischungen waren stark Lewis-sauer. Die Verbindungen [Ti2Cl10][BMIM]2, [Zr2Cl10][BMIM]2 (BMIM = 1-butyl-3-methylimidazolium) und [Hf2Cl9][PhNMe3] konnten so kristallisiert und deren Kristallstrukturen bestimmt werden. 2 Introduction 11 2 Introduction 2.1 Fossil Fuels and Oil Refining Although renewable energy sources are just beginning to attract attention as serious alternatives, fossil fuels will still be the main source of energy in the next decades. With about 35% in 2009, oil remains the leading fuel followed by coal and natural gas.[1] Besides its use to generate heat or electricity, oil is an indispensable feedstock for the production of fuel oils, gasoline, petrochemicals, or plastic materials. However, with increasing demand and decreasing resources, oil production becomes more and more energy-intensive. EROEI (energy returned on energy invested) is the amount of usable energy gained from a certain energy resource divided by the energy, which had to be expended. In the beginning, giant oil fields close to the surface were exploited, with negligible energy expenditure. Due to an increasing demand and the high price of oil, even deepwater drilling, tar sands and oil shales have become economic today. However, the net energy gain from these resources is very low compared to conventional oil fields (Table 1). Table 1: Energy returned on energy invested (EROEI) ratios of some fossil fuels.[2] Fossil Fuel EROEI Oil (1940 – 1960) 100 Coal 50 Oil (1970 – 1980) 25 Oil (global average) 19 Natural Gas 10 Tar Sands 5.2 – 5.8 Oil Shale 1.5 – 4 2 Introduction 12 Therefore, it is necessary to increase the efficiency of oil refining processes to maximize the yield of valuable products in combination with a minimum of waste production and energy consumption. Certainly, liquid fuels like gasoline, diesel fuel, or kerosene are the most valuable and useful refinery products since they are the prerequisite for our mobility. Thus, the aim of a typical oil refinery is to maximize the yield of these products, namely the hydrocarbon fraction containing between six and 20 carbon atoms. Scheme 1 shows the processes and yields of a typical oil refinery. Scheme 1: Simplified crude oil processing in refineries (averaged crude oil compositions of an oil carrying pipeline[3] and averaged output of US refineries in 2009[4]). 2 Introduction 13 First, the crude oil is distilled to yield about 50% of hydrocarbons within the gasoline or diesel range. Subsequently, the gases are transformed to higher hydrocarbons, while the heavy oils are cracked to further increase the gasoline and diesel yield. Catalytic cracking reactions of high molecular weight hydrocarbons are usually performed at high temperatures and produce more valuable saturated or unsaturated gases or gasoline. Thus, large amounts of C1 to C4 alkanes and olefins are produced in refineries in addition to the gaseous fractions dissolved in crude oil. Further, the exploitation of tar sands and oil shales requires an extensive use of catalytic cracking reactions leading to the formation of even more light hydrocarbons. Unlike liquid fuels, gases cannot be stored or transported easily, and consequently, they have to be processed directly in or close to a refinery. Scheme 2 shows the gases with one to four carbon atoms, which are mainly produced in refineries, and their potential uses. Although α-olefins like ethene, propene, or 1butene are used as monomers for the production of poly-α-olefins, the market demand for fuels is by far larger than the size of the polymer market. Consequently, oil refining companies are interested in converting these unsaturated gaseous olefins to gasoline or diesel fuels. Besides the alkylation of olefins with isoalkanes catalyzed by anhydrous HF or H2SO4, selective dimerization and oligomerization reactions of α-olefins are a promising attempt to achieve that goal. 2 Introduction 14 Scheme 2: Gaseous C1 – C4 hydrocarbons accruing in oil refineries and their possible uses. 2 Introduction 15 2.2 Dimerization of Short Chain Olefins The aim of this work was to convert the gaseous olefins ethene, propene, and butenes to products containing six to 20 carbon atoms by selective dimerization and oligomerization reactions. Scheme 3 provides an overview over the corresponding dimerization and oligomerization products and their intended uses. Initially, the work should be focussed on selective propene dimerization reactions to give gasoline fuels or to upgrade C3 feeds to yield branched C6 olefins for the HF catalyzed alkylation with isobutane. For example, selective propene dimerization reactions yield a mixture of nhexenes, 2-methylpentenes, and 2,3-dimethylbutenes. The degree of branching in nickel catalyzed dimerization reactions can be increased by the addition of sterically demanding phosphine ligands like tricyclohexylphosphine.[5] For gasoline, a high degree of branching is desirable. The research octane number (RON) is a measure for the antiknocking capacity of a fuel. With a higher degree of branching, the RON increases. Table 2 shows the RONs of typical gasolines. Table 2: Research octane numbers (RONs) of common gasolines and additives.[6] Fuel RON US Gasoline 85 – 91 Normalbenzin 91 Super Benzin 95 SuperPlus Benzin 98 LPG[1] 103 – 111 LNG[2] 120 – 130 n-Heptane 0 Isooctane 100 Benzene 101 Toluene 110 Xylenes 117 [1] Liquefied petroleum gas. [2] Liquefied natural gas. 2 Introduction 16 Scheme 3: Possible dimerization and oligomerization products of ethene, propene (RONs taken from the literature[7]), and butenes and their intended uses in refineries. 2 Introduction 17 Gasoline with a high RON is necessary to maximize the engine efficiency of modern cars, which is proportional to the compression. With higher compression levels, the tendency of the fuel to self-ignite has to be reduced. Thus, the demand for high octane gasoline is steadily increasing. Especially 2,3-dimethylbutenes display very high research octane numbers around 100. With a boiling point of over 50°C, they can be used directly or after hydrogenation to boost octane ratings even in summer. Further, if the content of high octane aromatics like benzene, toluene, or xylenes will be further restricted by new regulations, appropriate alternatives have to be found. If the C6 olefins produced by selective propene dimerization reactions are alkylated with isobutane in existing HF alkylation units, even higher boiling, high octane saturated C10 gasoline blends could be obtained.[7d] In contrast to propene, oligomerization reactions of ethene should yield linear olefins. While linear α-olefins are used as co-monomers for the production of LLDPE (linear low density polyethylene), C8 to C20 oligomers can be blended into diesel fuel. For diesel fuels, a low degree of branching is advantageous to favour the self-ignition in the engine. Linear butene dimerization products can also be employed as diesel fuel or feedstock for the production of plasticizers. If branched products are obtained, these can as well be used to blend into gasoline. Ideally, a catalytic system would be able to isomerize and dimerize butenes. The less reactive, thermodynamically more stable 2butenes are isomerized “in situ” to give the more reactive 1-butene in small amounts, which is subsequently dimerized. In general, internal olefins are of lower value compared to α-olefins, which makes the latter a more economic feedstock for dimerization reactions. The large number of industrial processes dealing with olefin dimerization and oligomerization mirrors the economic importance of these reactions (Table 3). Also, the capacities are enormous. For example, the combined refinery capacities of all IFP (Institut Français du Pétrole) processes listed in Table 3 exceeded 3.6 million tons per year in 2007.[8] The most common transition metals applied for this type of catalysis are group IV metals, nickel, and chromium. While some processes selectively dimerize (Alphabutol®, DIMERSOL®, DIFASOL®) or trimerize (ChevronPhillips) α-olefins, others yield a broad distribution of oligomers. To some extent, the maximum of the oligomer distribution can be influenced by varying process parameters in unselective oligomerization processes. Except the Shell higher olefin process (SHOP), all catalytic 2 Introduction 18 systems require the addition of a cocatalyst. However, the SHOP is operated under more drastic conditions (80 – 140°C, 7 – 14 MPa) compared to nickel systems activated with aluminum based cocatalysts.[9] Table 3: Industrial transition metal catalyzed oligomerization processes.[8a] Process Metal Cocatalyst Products Alphabutol® (IFP) Ti AlEt3 1-Butene DIMERSOL® (IFP) Ni EtAlCl2 C6 / C8 DIFASOL® (IFP) Ni EtAlCl2/AlCl3 (Ionic Liquid) C6 / C8 SHOP (Shell) Ni None (Biphasic) 1-Olefins (C4 – C30) ChevronPhillips Cr EtxAlCl3–x 1-Hexene Idemitsu Zr EtxAlCl3–x 1-Olefins (C6 – C30) Alphaselect® (IFP) Zr EtxAlCl3–x 1-Olefins (C4 – C10) α-SabLin® (Sabic) Zr EtxAlCl3–x 1-Olefins (variable) Linear-1® (UOP) Ni NaBH4 (Biphasic) 1-Olefins (C4 – C10) Since nickel compounds are catalytically versatile, cheap and usually insensitive to air or moisture compared to group IV metals, we decided to focus on nickel based dimerization and oligomerization systems. The requirements of a catalytic dimerization system to be developed are as follows. First, the system has to be cheap to compete with established technologies. This requires high activities, lifetimes, conversions and selectivities to both dimer formation and a specific degree of branching. The tolerance towards impurities, mainly oxygen, sulfur, amines, or traces of water, would further reduce the costs of substrate purification. Also, the possibility to recycle or to regenerate spent catalysts would be desirable. 2 Introduction 19 2.3 Cocatalysts for Transition Metal Catalyzed Olefin Dimerization and Polymerization Reactions Nickel based catalysts usually have to be activated by aluminum alkyls. Common cocatalysts are methylaluminoxane (MAO) and its derivatives, ethylaluminum dichloride, diethylaluminum chloride, triethylaluminum, trimethylaluminum, or mixtures thereof. In Table 4, the most common cocatalysts for transition metal catalyzed olefin dimerization and polymerization reactions are listed.[5c, 10] In order to obtain a rough overview over the cost dimensions of these activators, the prices for one mol of cocatalyst were calculated from the cheapest price per unit given by common commercial laboratory chemical suppliers. Of course, the price structure may be completely different for large scale applications. However, the order of magnitude can be estimated. Perfluorinated organoboron cocatalysts, which are mainly applied for the activation of metallocene complexes,[10u, 10v] are extremely expensive. Although they are only used in stoichiometric amounts, their prices prevent their use for low cost dimerization catalysts. The most common aluminum cocatalyst, MAO, is also the most expensive one. The difficult synthesis und subsequently the high price of MAO is certainly also responsible for the fact that group IV metallocene based systems still have a niche existence. Although other alkylaluminum cocatalysts are less expensive than MAO, the perfect cocatalyst would be AlCl3. Aluminum chloride is among the cheapest chemicals available. Unfortunately, AlCl3 is insoluble in hydrocarbon solvents[11] and catalyzes side reactions like isomerization, cracking, disproportionation of alkanes as well as alkylation of aromatics with olefins and cationic oligomerization of alkenes.[4, 12] Furthermore, the proposed mechanism based on the insertion of monomers into Ni–H or Ni–C bonds of nickel(II) complexes requires an alkylating agent. Alkylaluminum compounds are pyrophoric and highly sensitive to water, oxygen, or polar impurities. Thus, they are mostly the cost determining factor for commercial applications. Feed streams have to be purified thoroughly prior to the reaction, and the high reactivity requires a cautious handling. Finally, the cocatalyst is destroyed upon product separation of the homogeneous reactions.[13] 3 Overview of Thesis Results 26 In the course of our search for optimized compositions, chloroaluminate melts with variable Lewis acidities and BiPh3 buffer contents were screened towards their performances in biphasic nickel catalyzed selective propene dimerization reactions. Slightly acidic, highly buffered systems yielded extremely selective propene dimerization systems. However, even low melting, highly Lewis acidic compositions were successfully buffered to give more than 90% dimers from propene. In combination with the melting point reduction of chloroaluminate melts upon addition of BiPh3, it was possible to synthesize and screen 100 different organic halide salts for their performances in nickel catalyzed selective propene dimerization reactions with highly Lewis acidic compositions. Various nickel compounds were employed as catalyst precursors to yield similar catalytic performances. While lower reaction temperatures came along with higher dimer selectivities, the presence of sterically demanding tricyclohexylphosphine ligands led to the formation of valuable branched products. 3.4 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions After the promising results of the biphasic experiments, we tried to transfer the principle of donor-acceptor interactions between weak organic bases and AlCl3 in ionic liquid cocatalysts to homogeneous systems. In combination with stoichiometric amounts of BiPh3 or N-methylpyrrole buffer, aluminum chloride dissolved in toluene and methylene chloride to form a highly efficient and air stable cocatalyst for nickel complexes. Depending on the ligand structure of the nickel catalyst precursor, the resulting catalyst compositions were able to either polymerize ethene or to selectively dimerize various αolefins. Further, the Lewis acidities of these binary homogeneous cocatalyst solutions 3 Overview of Thesis Results 27 could be tuned precisely by the choice of the solvent and the type and amount of buffer. The properties of the produced polyethylenes could be influenced by the cocatalyst compositions. Also, the buffer sufficiently suppressed isomerization reactions of the αolefinic products and hindered the ligand abstraction from the active nickel catalyst by AlCl3 or the Lewis acidic ionic liquid. The donor-acceptor interaction between BiPh3 or N-methylpyrrole and AlCl3 was monitored by 27Al NMR spectroscopy. Ethene polymerization also occurred in buffered chloroaluminate ionic liquids in the presence of a nickel diimine complex. 3.5 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes In combination with BiPh3 or N-methylpyrrole buffers, EtAlCl2 or Et2AlCl surface modified silicas were successfully employed as heterogeneous cocatalysts for nickel complexes to catalyze selective olefin dimerization reactions. The buffered heterogeneous cocatalysts were also able to activate a nickel diimine complex for ethene polymerization reactions. The active nickel species were immobilized at the silica surface to produce a granular polyethylene-silica composite material with a homogeneous particle size distribution. Further, surface modified silica was coated with a BiPh3 buffered chloroaluminate ionic liquid to give a supported ionic liquid phase (SILP) cocatalysts for nickel catalyzed olefin dimerization reactions. 3 Overview of Thesis Results 28 3.6 Solubility Behaviour of TiCl4, ZrCl4, and HfCl4 in Chloroaluminate Ionic Liquids In the course of the ionic liquid cocatalyst development, we found that up to stoichiometric amounts of ZrCl4 and HfCl4 dissolved in a neutral chloroaluminate melt. The high Lewis acidities of the resulting solutions accrued from the formation of [Al2Cl7]– anions from [AlCl4]– in the presence of ZrCl4 or HfCl4. From such ternary mixtures, crystals of [Ti2Cl10][BMIM]2, [Zr2Cl10][BMIM]2, and [Hf2Cl9][PhNMe3] were obtained, and their crystal structures were determined. 3.7 Oxidative Coupling and Catalytic Cracking of Alkanes in Lewis Acidic Chloroaluminate Ionic Liquids Enhanced by Molecular Oxygen While experimenting with our highly Lewis acidic chloroaluminate ionic liquid cocatalysts for biphasic olefin dimerization reactions, we observed that saturated hydrocarbons like n-heptane were efficiently cracked in the presence of molecular oxygen. In dry air, the conversion of n-heptane to give mainly gaseous alkanes was more than one order of magnitude higher compared to systems with inert atmosphere. Further, cyclic alkanes were oxidatively coupled in the presence of O2 under the consumption of two hydrogen atoms, while the same reactions only yielded skeletal isomers in argon atmosphere. 3 Overview of Thesis Results 29 3.8 Individual Contribution to Joint Publications The results presented in this thesis were obtained in collaboration with others and are to be submitted as indicated below. In the following, the contributions of all the co-authors to the different publications are specified. The asterisk denotes the corresponding author. Chapter 4 This work is to be submitted with the title “Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids” Matthias Dötterl and Helmut G. Alt* I synthesized all compounds, ionic liquids and performed the catalytic experiments presented in this work. I also wrote the publication. Christian Görl was involved in mechanistic discussions. Helmut G. Alt corrected and commented the manuscript. Chapter 5 This work is to be submitted with the title “Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids” Matthias Dötterl, Peter Thoma, and Helmut G. Alt* I synthesized all compounds, ionic liquids and performed the catalytic experiments. Further, I wrote the publication. I also characterized all products. Peter Thoma helped me measuring the NMR spectra of heteroatoms and was involved in scientific discussions, comments and correction of the manuscript. Helmut G. Alt corrected and commented the manuscript. 3 Overview of Thesis Results 30 Chapter 6 This work is to be submitted with the title “Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations” Matthias Dötterl and Helmut G. Alt* I synthesized all compounds, ionic liquids and performed the catalytic experiments. Further, I wrote the publication. Helmut G. Alt corrected and commented the manuscript. Chapter 7 This work is to be submitted with the title “Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions” Matthias Dötterl and Helmut G. Alt* I synthesized all compounds, ionic liquids and performed the catalytic experiments. Further, I wrote the publication. Peter Thoma helped me measuring the 27Al NMR spectra. Winfried Kretschmer did the high temperature GPC measurements. Helmut G. Alt corrected and commented the manuscript. Chapter 8 This work is to be submitted with the title “Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes” Matthias Dötterl and Helmut G. Alt* 3 Overview of Thesis Results 31 I synthesized all compounds, ionic liquids and performed the catalytic experiments. Further, I wrote the publication. Winfried Kretschmer did the high temperature GPC measurement. Helmut G. Alt corrected and commented the manuscript. Chapter 9 This work is to be submitted with the title “Solubility Behaviour of TiCl4, ZrCl4, and HfCl4 in Chloroaluminate Ionic Liquids” Matthias Dötterl, Isabelle Haas, and Helmut G. Alt* I synthesized all compounds, ionic liquids and performed the catalytic experiments. Further, I wrote the publication. Germund Glatz, Tobias Bauer and Isabelle Haas performed one X-ray analysis each and solved the crystal structures. Isabelle Haas helped me utilizing the X-ray data. Helmut G. Alt corrected and commented the manuscript. Chapter 10 This work is to be submitted with the title “Oxidative Coupling and Catalytic Cracking of Alkanes in Lewis Acidic Chloroaluminate Ionic Liquids Enhanced by Molecular Oxygen” Matthias Dötterl and Helmut G. Alt* I synthesized all compounds, ionic liquids and performed the catalytic experiments. Further, I wrote the publication. Helmut G. Alt corrected and commented the manuscript. 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 32 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids Matthias Dötterl[a] and Helmut G. Alt*[a] [a] Lehrstuhl für Anorganische Chemie II, Universitätsstraße 30, NW I, 95440 Bayreuth, Germany. E-mail: [email protected] Keywords: ionic liquids, chloroaluminates, nickel, biphasic catalysis, dimerization Manuscript to be submitted (Communication) 4.1 Main Text Since ionic liquids received broader attention as a new class of reaction media in the 1980s,[1] when AlCl3 based systems were thoroughly investigated, many generations of air and water stable ionic liquid systems have been developed.[2] At present, there are numerous ionic liquid based systems and processes that have found applications in industry.[2b, 3] However, many of the newly developed systems with improved stabilities and properties are based on ionic liquids containing uncommon anions. These ionic liquids are very expensive, especially when fluorine atoms are involved. Thus, it is not a big surprise that the largest industrial processes still rely on cheap first generation chloroaluminate systems. 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 33 PetroChina’s Ionikylation® process uses a composite AlCl3 / CuCl ionic liquid, which catalyzes the alkylation of isobutane with butenes. The process is currently operated in an 65000 t/a alkylation unit in China.[4] Chauvin et al. developed a process for biphasic catalytic dimerization reactions of short chain olefins in the 1990s.[5] The ionic liquid made from N,N’-alkylmethylimidazolium chloride mixed with aluminum chloride and ethylaluminum dichloride plays a dual role as cocatalyst and solvent for the catalytically active nickel complex. The Institut Français du Pétrole (IFP) brought this so called DIFASOL® process to industrial application by retrofitting it to their existing DIMERSOL® units.[6] In the year 2009, 35 DIMERSOL® units have been licensed[7] for selective dimerization reactions of C3 and C4 olefinic cuts.[6c] Typically, these processes are operated with capacities between 20000 and 90000 t/a. The biphasic DIFASOL® process provides several advantages over its homogeneous DIMERSOL® analogue, which is based on nickel complexes activated with EtAlCl2 in solution.[8] The nickel catalyst and most of the alkylaluminum species remain within the ionic liquid phase. Thus, the nickel and aluminum consumption is greatly reduced. DIFASOL® shows higher selectivities to give dimers and can be operated with streams containing smaller amounts of olefins. In combination with a much smaller reactor size, this leads to greatly improved process economics.[6b, 6c] Besides DIFASOL®, which uses ethylaluminum groups to suppress uncontrolled cationic olefin oligomerization reactions in solely AlCl3 based ionic liquids,[9] Wasserscheid et al. developed a system buffered by weak organic bases.[10] Pyridine, quinoline and pyrrole derivatives were added to slightly acidic chloroaluminate ionic liquids. The interaction of the Lewis bases with AlCl3 or the [Al2Cl7]– species present in such liquids[1a, 11] reduced the “latent acidity”.[12] The interaction prevented uncontrolled cationic olefin oligomerization reactions. A chloroaluminate liquid buffered with weak organic bases was successfully employed to activate nickel complexes for dimerization reactions of 1-butene. In a recent article, Wasserscheid et al. extended the system to propene and 1-hexene dimerization reactions and studied the kinetics of these biphasic reactions in detail.[13] They found that the reaction rates are strongly limited by mass transfer of the substrate olefins. Following the concept of Lewis base buffered systems we found that triarylamine, triarylphosphine and triarylbismuth compounds efficiently buffer ionic liquids derived 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 34 from AlCl3. The dissolution of small amounts of these buffering substances in a slightly acidic 1-butyl-3-methylimidazolium (BMIM) chloroaluminate ionic liquids (AlCl3 / BMIMCl = 1.20) yielded the ionic liquid cocatalysts. The air stable bis(imino)pyridine complex of nickel(II)bromide (Figure 1) was used as nickel source. It readily dissolved in the buffered ionic liquids and the resulting ionic liquid catalysts dimerized propene in batch experiments with high productivities and selectivities (Table 1). Figure 1: Catalyst precursor used for biphasic propene dimerization reactions. We substituted the term “activity” with “productivity” following the results of Wasserscheid et al.[13] They showed that above a certain concentration, the activities of the systems did not depend on the amount of catalyst anymore due to mass transport limitations. The activities were then even independent from the temperature. For our system we chose a catalyst concentration of 10–5 molcatalyst / gionic liquid. A data set of a catalyst concentration series can be found in the Supporting Information. The minimum catalyst concentration that still influenced the overall activity in the investigated systems was found to be around 10–6 molcatalyst / gionic liquid. Thus, the productivities could be calculated as the amount of products formed by one gram of active liquid in one hour under the assumption that the productivities of the systems did not depend on the catalyst concentration. For a typical catalytic experiment, 2 – 4 ml ionic liquid were mixed with the corresponding amount of buffer followed by the addition of the catalyst precursor (1). Between 2.5 and 4 g of the resulting homogeneous solutions were used for the experiments. 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 35 Table 1: Nickel catalyzed biphasic propene dimerization reactions in buffered acidic 1butyl-3-methylimidazolium (BMIM) chloroaluminate ionic liquids.[1] No. Buffer [Buffer] / [BMIMCl] Productivity[2] Dimers [%] 1 NPh3 0.12 4.0 81.0 2 PPh3 0.12 > 15.0[3] 93.6 3 AsPh3 0.12 0 – 4 SbPh3 0.12 6.8 Oil[4] 5 BiPh3 0.07 > 10.8[3] 89.2 6 BiPh3 0.12 > 10.8[3] 93.1 7 BiPh3 0.30 6.7 96.0 [1] Reaction conditions: 2 ml of ionic liquid with [AlCl3] / [BMIMCl] = 1.20; catalyst precursor 1; [cat] = 10–5 mol / gionic liquid; T = 25°C; stirring rate = 1200 min–1; t = 60 min; 300 ml glass autoclave; 40 – 60 ml liquid propene. [2] gproduct / gionic liquid x h. [3] Complete conversion. [4] Higher oligomers derived from a cationic oligomerization reaction. While the triphenylamine buffered system produced only 81% dimers (1), the triphenylphosphine based system with the same composition gave 94% dimers (2). However, the solubility of PPh3 strongly depended on the temperature. At 25°C, PPh3 did not dissolve completely in the ionic liquid. Due to its nonpolar character, triphenylphosphine systems suffered from heavy leaching into the organic product phase. Thus, the results strongly depended on the amount of olefin or the reaction temperature and, therefore, displayed a bad reproducibility. The addition of AsPh3 deactivated the system (3), while SbPh3 (4) showed no ability to prevent cationic olefin oligomerization reactions induced by the highly Lewis acidic aluminum centers.[9] Unexpectedly, the triphenyl compound of the last element in the row, bismuth, proved to be a perfect buffer. Starting from a concentration of 0.07 equivalents of BiPh3, the BiPh3 buffered systems produced dimers with high selectivities (5 – 7). It was completely soluble in chloroaluminate ionic liquids, even at higher ratios. There are three reasons for its high solubility. First, aromatic compounds readily dissolve in chloroaluminate ionic 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 42 phase is a major problem of DIFASOL® systems. Therefore, they are operated in combination with DIMERSOL® reactors, which totally purify the DIFASOL® feed.[6c] However, the absence of any alkylaluminum compounds arose questions about the dimerization mechanism. In the literature, a mechanism is generally accepted for twovalent nickel complexes that involves Ni–H or Ni–C bonds.[7, 18] In all of these cases it was assumed that the nickel center is alkylated by the aluminum cocatalyst to form nickel hydride or nickel alkyl species. Alternatively, Ni–C bond containing species like η3-allyl nickel compounds were employed.[18g] The olefins coordinate to a free coordination site at the square planar nickel center and are inserted into the Ni–H or Ni–C bond. Yet, this mechanism could not be applied to our system. Besides a nickel salt, buffered chloroaluminate melts only contained AlCl3, an organic halide salt and the buffer. None of these is able to form either Ni–H or Ni–C bonds from nickel halide salts. Also, the reduction of Ni(II) to give Ni(I) or Ni(0) species was unlikely in these systems. For nickel compounds in lower oxidation states than two, alternative mechanisms were also discussed. For example, nickel(I) compounds are involved in selective olefin oligomerization reactions with nickel impregnated metal oxides. For this reaction, a nickelacyclopentane was proposed, which is in equilibrium with the corresponding diolefin complex. The rate determining step was assumed to be the decomposition of the metallacycle to yield the dimer upon reductive elimination.[19] For Ni(0), metallacycles are also well known.[20] Since the product distribution resulting from selective propene dimerization reactions with catalyst precursor 1 matched the results for ligand free nickel salts,[13] we expected that the ligand was abstracted. Nickel halides were found to form Ni(AlCl4)2[21] or [Ni(AlCl4)3]–,[22] if they were mixed with aluminum chloride or acidic chloroaluminate ionic liquids, respectively. In buffered chloroaluminate melts used for olefin dimerization reactions, these weakly coordinating tetrachloroaluminate anions are expected to be easily replaced by other ligands like buffer or olefin molecules (Figure 2). Considering our catalytic results, we propose a new mechanism for nickel catalyzed olefin dimerization reactions in buffered, alkylaluminum free chloroaluminate ionic liquids using the example of ethene (Scheme 3). Similar to the mechanism proposed for Ni(I),[19] we expect the formation of a metallacycle, which is in equilibrium with a Ni(ethene)2 complex. Decomposition of the metallacycle through β-H elimination yields 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 43 an ω-alkenyl stabilized nickel hydride species. Elimination followed by the coordination of a new ethene molecule results in a nickel di-olefin complex with ethene and 1-butene being coordinated to the nickel center. At this point, there are two possible pathways. For weakly buffered, more acidic compositions, 1-butene remains at the nickel center and forms an ethyl branched nickelacyclopentane again. This route would lead to the formation of branched, higher oligomers from ethene. Yet, if the system is highly buffered, we expect the buffer to compete with 1-butene for the coordination site at the active nickel center. Since the coordination strength of 1-butene is weaker compared to ethene, 1-butene is replaced by a buffer molecule. The also weakly coordinating buffer molecule can be substituted again by an ethene molecule to form the initial Ni(ethene)2 species. Figure 2: Formation of [Ni(AlX4)3]– from nickel halide salts in acidic haloaluminate ionic liquids (a) and exchange of the weakly coordinating tetrahaloaluminate ligands by competing donor molecules (b). Although we have no experimental evidence for the oxidation state of the active nickel catalyst, we propose that the active species is Ni(II). Due to the formation of weakly coordinating tetrahaloaluminate anions in acidic ionic liquids,[22] the Ni(II) center is easily accessible. The mechanism conveniently includes a metallacycle as well as a Ni–H species. Also, our results indicated that the original product from ethene dimerization reactions was 1-butene. However, subsequent isomerization reactions disguised the real product distributions. This became clear since the selectivity to give 1-butene increased from 28% (19) to 47% (20) with twice the amount of buffer. Thus, the buffer did not only buffer Lewis acidic aluminum species, it also favoured the formation of dimers and suppressed subsequent isomerization reactions. 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 44 Scheme 3: Proposed mechanism for nickel catalyzed ethene dimerization reactions in buffered chloroaluminate ionic liquids (L = any ligand, B = buffer molecule). Finally, an EtAlCl2 based chloroaluminate ionic liquid similar to a commercial DIFASOL® composition[23] was modified with our BiPh3 buffer and tested in batch propene dimerization reactions (Table 5). The unmodified system gave about 80% dimers (27) with catalyst precursor 1. If small amounts of BiPh3 were dissolved in that system, the dimer yield jumped to almost 97%. The addition of 0.03 equivalents (30) gave the best result in combination with a high productivity. If too much buffer was added, the productivities were reduced (28, 29), while no effect could be observed with only 0.01 equivalents (31) of BiPh3. 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 45 Table 5: Nickel catalyzed biphasic propene dimerization reactions in typical DIFASOL® systems modified with BiPh3. [1] No. [EtAlCl2] / [BMIMCl] [BiPh3] / [BMIMCl] Productivity[2] Dimers [%] 27 0.20 0 > 18.1[3] 79.6 28 0.20 0.12 2.0 96.8 29 0.20 0.06 7.1 94.6 30 0.20 0.03 > 19.4[3] 94.1 31 0.20 0.01 > 13.0[3] 79.4 32 0 0.12 > 10.8[3,4] 93.1 [1] Reaction conditions: 2.5 – 4.0 g buffered ionic liquid; [AlCl3] / [BMIMCl] = 1.20; catalyst precursor 1; [cat] = 10–5 mol / gionic liquid; T = 25°C; stirring rate = 1200 min–1; t = 45 min; 300 ml glass autoclave; 40 – 60 ml liquid propene. [2] gproduct / gionic liquid x h. [3] Complete conversion. [4] Reaction time 60 minutes. The combination of EtAlCl2 and BiPh3 with AlCl3 based ionic liquids provides a promising, highly active and selective nickel catalyzed dimerization system for commercial applications. The fact that BiPh3 reduces the melting temperatures of said systems opens the door for the application of a wide range of cheap cations. The possibility to tune the Lewis acidities and to suppress subsequent isomerization reactions simply by varying the BiPh3 content should make BiPh3 buffered chloroaluminate ionic liquids interesting for many other types of reactions like Friedel Crafts alkylations, acylations, isomerizations, hydrogenations,[2b, 24] and many more. Supporting Information Available Detailed experimental procedures, the preparation of the ammonium salts, ionic liquids and complex 1 as well as a catalyst concentration series are described in the Supporting Information. 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 46 Acknowledgements Financial support by ConocoPhillips, USA, is gratefully acknowledged. M.D. thanks Dr. C. Görl for the helpful discussions and the Elite Network of Bavaria (ENB) within the graduate program “Macromolecular Science”. 4.2 References [1] a) J. S. Wilkes, J. A. Levisky, R. A. Wilson, C. L. Hussey, Inorg. Chem. 1982, 21, 1263-1264; b) C. L. Hussey, Pure Appl. Chem. 1988, 60, 1763-1772. [2] a) P. Wasserscheid, W. Keim, Angew. Chem., Int. Ed. 2000, 39, 3772-3789; b) H. Olivier-Bourbigou, L. Magna, D. Morvan, Appl. Catal., A 2010, 373, 1-56. [3] N. V. Plechkova, K. R. Seddon, Chem. Soc. Rev. 2008, 37, 123-150. [4] Z. C. Liu, R. Zhang, C. M. Xu, R. G. Xia, Oil Gas J. 2006, 104, 52-56. [5] a) Y. Chauvin, B. Gilbert, I. Guibard, J. Chem. Soc.-Chem. Commun. 1990, 1715-1716; b) Y. Chauvin, S. Einloft, H. Olivier, Ind. Eng. Chem. Res. 1995, 34, 1149-1155; c) S. Einloft, F. K. Dietrich, R. F. De Souza, J. Dupont, Polyhedron 1996, 15, 3257-3259; d) Y. Chauvin, H. Olivier, C. N. Wyrvalski, L. C. Simon, R. F. de Souza, J. Catal. 1997, 165, 275-278. [6] a) F. Favre, A. Forestiere, F. Hugues, H. Olivier-Bourbigou, J. A. Chodorge, Oil Gas Eur. Mag. 2005, 31, 83 - 91; b) B. Gilbert, H. Olivier-Bourbigou, F. Favre, Oil & Gas Science and Technology - Rev. IFP 2007, 62, 745-759; c) H. OlivierBourbigou, F. Favre, A. Forestière, F. Hugues, Ionic Liquids and Catalysis: the IFP Biphasic Difasol Process, in Handbook of Green Chemistry, Wiley-VCH, 2010, pp. 101-126. [7] A. Forestière, H. Olivier-Bourbigou, L. Saussine, Oil & Gas Science and Technology - Rev. IFP 2009, 64, 649-667. [8] a) Y. Chauvin, J. F. Gaillard, D. V. Quang, J. W. Andrews, Chem. Ind. 1974, 375378; b) D. Commereuc, Y. Chauvin, G. Leger, J. Gaillard, Oil & Gas Science and Technology - Rev. IFP 1982, 37, 639-649. [9] a) M. Goledzinowski, V. I. Birss, J. Galuszka, Ind. Eng. Chem. Res. 1993, 32, 1795-1797; b) O. Stenzel, R. Brüll, U. M. Wahner, R. D. Sanderson, H. G. Raubenheimer, J. Mol. Catal. A: Chem. 2003, 192, 217-222. [10] a) P. Wasserscheid, WO 9847616, 1998; b) B. Ellis, W. Keim, P. Wasserscheid, Chem. Commun. 1999, 337-338; c) P. Wasserscheid, M. Eichmann, Catal. Today 2001, 66, 309-316; d) D. S. McGuinness, W. Mueller, P. Wasserscheid, K. J. Cavell, B. W. Skelton, A. H. White, U. Englert, Organometallics 2001, 21, 175181. [11] C. L. Hussey, T. B. Scheffler, J. S. Wilkes, J. A. A. Fannin, J. Electrochem. Soc. 1986, 133, 1389-1391. 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 47 [12] a) I. C. Quarmby, R. A. Mantz, L. M. Goldenberg, R. A. Osteryoung, Anal. Chem. 1994, 66, 3558-3561; b) I. C. Quarmby, R. A. Osteryoung, J. Am. Chem. Soc. 1994, 116, 2649-2650. [13] M. Eichmann, W. Keim, M. Haumann, B. U. Melcher, P. Wasserscheid, J. Mol. Catal. A: Chem. 2009, 314, 42-48. [14] a) Y. Chauvin, S. Einloft, H. Olivier, Ind. Eng. Chem. Res. 1995, 34, 1149-1155; b) A. A. Fannin, L. A. King, J. A. Levisky, J. S. Wilkes, J. Phys. Chem. 1984, 88, 2609-2614; c) J. Robinson, R. C. Bugle, H. L. Chum, D. Koran, R. A. Osteryoung, J. Am. Chem. Soc. 1979, 101, 3776-3779. [15] a) F. H. Hurley, J. T. P. Wier, J. Electrochem. Soc. 1951, 98, 203-206; b) A. A. Fannin, D. A. Floreani, L. A. King, J. S. Landers, B. J. Piersma, D. J. Stech, R. L. Vaughn, J. S. Wilkes, L. Williams John, J. Phys. Chem. 1984, 88, 2614-2621. [16] R. M. Stephenson, J. Chem. Eng. Data 1993, 38, 625-629. [17] F. Peruch, H. Cramail, A. Deffieux, Macromolecules 1999, 32, 7977-7983. [18] a) W. Keim, Ann. N.Y. Acad. Sci. 1983, 415, 191-200; b) U. Müller, W. Keim, C. Krüger, P. Betz, Angew. Chem., Int. Ed. 1989, 28, 1011-1013; c) S. A. Svejda, M. Brookhart, Organometallics 1998, 18, 65-74; d) D. P. Gates, S. A. Svejda, E. Onate, C. M. Killian, L. K. Johnson, P. S. White, M. Brookhart, Macromolecules 2000, 33, 2320-2334; e) M. D. Leatherman, S. A. Svejda, L. K. Johnson, M. Brookhart, J. Am. Chem. Soc. 2003, 125, 3068-3081; f) F. Speiser, P. Braunstein, L. Saussine, Acc. Chem. Res. 2005, 38, 784-793; g) D. Roy, R. B. Sunoj, Org. Biomol. Chem. 2010, 8, 1040-1051; h) J. M. Brown, G. D. Hughes, Inorg. Chim. Acta 1996, 252, 229-237. [19] a) F. X. Cai, C. Lepetit, M. Kermarec, D. Olivier, J. Mol. Catal. 1987, 43, 93-116; b) C. Lepetit, M. Kermarec, D. Olivier, J. Mol. Catal. 1989, 51, 95-113. [20] a) R. H. Grubbs, A. Miyashita, J. Am. Chem. Soc. 1978, 100, 7416-7418; b) J. Cámpora, P. Palma, E. Carmona, Coord. Chem. Rev. 1999, 193-195, 207-281; c) R. H. Grubbs, A. Miyashita, M.-I. M. Liu, P. L. Burk, J. Am. Chem. Soc. 1977, 99, 3863-3864; d) R. H. Grubbs, A. Miyashita, J. Am. Chem. Soc. 1978, 100, 1300-1302. [21] J. Brynestad, H. L. Yakel, G. P. Smith, Inorg. Chem. 1970, 9, 686-686. [22] a) A. K. Abdul-Sada, A. M. Greenway, K. R. Seddon, T. Welton, Org. Mass. Spectrom. 1992, 27, 648-649; b) U. Keber, R. Müller, Z.Naturforsch.(B) 2007, 62, 1052-1058. [23] H. Olivier-Bourbigou, E. Pellier, A. Forestiere, EP 1968918, 2008. [24] T. Welton, Chem. Rev. 1999, 99, 2071-2084. 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 48 4.3 Supporting Information General Remarks All chemical manipulations were carried out using standard Schlenk techniques under argon atmosphere. n-Heptane was distilled from Na/K alloy under an atmosphere of argon. The products of the dimerization experiments were characterized by gas chromatography (Agilent 6850) and GC-MS (FOCUS DSQ™ Thermo Scientific). Mass spectra were recorded on a Varian MAT CH7 instrument (direct inlet system, electron impact ionization 70 eV). Elemental analyses were performed with a VarioEl III CHN instrument. Acetanilide was used as a standard. Ethene (99.9%), propene (99.3%), and 1-butene (99.3%) were purchased from Riessner Gase, Lichtenfels, and were dried over a column packed with P4O10. 1-Hexene and EtAlCl2 (0.9 M in heptane) were purchased from Acros and used without further purification. Triphenylamine, triphenylphosphine, triphenylarsine, triphenylantimony, triphenylbismuth, and cyclohexylamine hydrochloride were purchased from ABCR, 1-butyl-3-methylimidazolium chloride, 1ethyl-3-methylimidazolium chloride, 1-pyrrolidino-1-cyclopentene, tributylamine, Nmethylpyrrolidine, and AlCl3 (ReagentPlus®) were purchased from Sigma-Aldrich and used as received. Chloroethane (Chloraethyl, Dr. Henning) was purchased in a local pharmacy. Synthesis of Complex 1 The ligand precursor was synthesized from 2,6-diacetylpyridine and 4-fluoroaniline according to a literature procedure.[1] The ligand precursor (1.19 g) was dissolved in THF (100 ml) and 1.05 g of (dme)NiBr2 were added. The mixture was stirred for 2 h at 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 49 ambient temperature. After most of the solvent had been removed in vacuo, the complex was precipitated by the addition of n-pentane (100 ml). After filtration and washing with n-pentane, complex 1 was obtained in quantitative yield as an orange brown solid. EI-MS data: 567 (M•+) (1), 488 M –Br (33), 407 M –2Br (22), 349 M –NiBr2 (65), 136 FC6H4-N=C-CH3 (100). Elemental analysis (calculated, found for C21H17Br2F2N3Ni): C (44.42, 44.51), H (3.02, 3.56), N (7.40, 7.05). Synthesis of the Ionic Liquids The ionic liquids were synthesized by directly mixing the corresponding amount of AlCl3 and chloride salt in a cooled Schlenk tube. In the case of EtAlCl2 containing systems, a 0.9 molar solution of EtAlCl2 in heptane was added to the liquid and the solvent was removed in vacuo. The obtained ionic liquids were stored in Schlenk tubes. Procedure for Dimerization Reactions Using Buffered Ionic Liquids Prior to the experiments, 2 to 4 ml of the ionic liquid were filled into a Schlenk tube and mixed with the buffer. After the buffer had dissolved upon stirring, the nickel complex 1 was added. The obtained homogeneous solution was syringed into the reaction vessel. Its amount was determined by the weight difference of the syringe. For the screening experiments with propene, a 300 ml glass autoclave equipped with a stirring bar and a magnetic stirrer with a stirring rate of 1200 min–1 was used. The glass autoclave was kept in a drying oven at 150°C for several hours before an experiment. After the addition of the active ionic liquid, propene (40 to 60 ml) was condensed into the glass autoclave by liquid nitrogen cooling. Then, the autoclave was placed in a metal box for safety reasons, and the temperature was regulated by an external water bath. After the experiment, the pressure was slowly released by opening a valve. The product fraction was filtered through a short plug of silica and analyzed by gas chromatography. For the screening experiments in Table 4 and Table 6, a stirred 300 ml Parr stainless steel autoclave was used. The steel vessel was kept in a drying oven at 150°C for 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 50 several hours before an experiment, and it was filled with argon. Then, the active ionic liquid was added under argon counter flow. Ethene: n-Heptane was added to the reactor vessel under argon counter flow. The system was put under vacuum, and the temperature was raised to 40°C. Then, an ethene pressure of 10 bar was applied. Cooling of the highly exothermic reaction was achieved manually by an external liquid nitrogen cooled acetone bath. After the experiment, the vessel was quickly cooled to –20°C. The pressure was released, the weight difference of the vessel was determined, and a cooled sample was directly analyzed by gas chromatography. Propene: The vessel was evacuated, and liquid propene (200 ml) was soaked into it at 77 K. The reaction was started by quickly heating the vessel to 40°C with boiling water. Cooling of the highly exothermic reaction was achieved manually by an external liquid nitrogen cooled acetone bath. After the experiment, the vessel was cooled to 0°C, and the pressure was slowly released. After reaching ambient temperature, the weight difference of the vessel was determined, and the product phase was analyzed by gas chromatography. 1-Butene: The vessel was evacuated, and liquid 1-butene (200 ml) was soaked into it at –20°C. The reaction was started by quickly heating the vessel to 40°C with boiling water. Cooling of the exothermic reaction was achieved manually by an external ice bath. After the experiment, the vessel was cooled to –20°C, and the organic phase was analyzed by gas chromatography. Productivities were calculated from the GC spectra. 1-Hexene was added to the reactor vessel, and the reactor was stirred at 40°C. After the experiment, the organic phase was analyzed by gas chromatography. Productivities were calculated from the GC spectra. Synthesis of the Hydrochlorides of Tributylamine, 1-Pyrrolidino-1-cyclopentene and N-Methylpyrrolidine The corresponding amines were dissolved in dry diethylether, and HCl gas was bubbled through the flask until no more gas was absorbed. The white, precipitated ammonium salts were filtered, washed three times with diethylether and n-pentane, dried in vacuo and used without further purification. 4 Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids 51 Synthesis of N-Ethyl-N-methylpyrrolidinium Chloride N-ethyl-N-methylpyrrolidinium chloride was synthesized from chloroethane and Nmethylpyrrolidine in acetonitrile following the procedure described for triethylamine in the literature.[2] Catalyst Concentration Series Table 6: Catalyst concentration series of nickel catalyzed biphasic propene dimerization reactions in BiPh3 buffered acidic chloroaluminate ionic liquids.[1] No. Catalyst Concentration[2] Productivity[3] “Activity”[4] Dimers [%] 32 10–5 18.5 1.9 93.7 33 10–6 18.2 18.2 92.9 34 10–7 9.7 97.1 92.1 35 10–8 1.9 192.1 92.2 [1] Reaction conditions: 3.5 – 4.5 g buffered ionic liquid; [BiPh3] / [N-methylpyrrolidinium]+[Al2Cl7]− = 0.30; catalyst precursor 1; T = 40°C; stirring rate = 600 min–1; t = 60 min; stirred 300 ml Parr stainless steel autoclave; 200 ml liquid propene. [2] molcatalyst / gionic liquid. [3] gproduct / gionic liquid x h. [4] tproduct / molcatalyst x h. References [1] C. Qian, F. Gao, Y. Chen, L. Gao, Synlett 2003, 1419-1422. [2] N. W. Smith, J. McCloskey, US 6444846, 2002. 5 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids 58 Scheme 2: Triarylphosphines screened towards their buffering ability in the Lewis acidic ionic liquid [BMIM]+[Al2Cl7]−. The propene dimerization reactions were carried out batchwise employing a simple bis(imino)pyridine complex of nickel(II)bromide (15) as the catalyst precursor.[17] Figure 1: Catalyst precursor used for biphasic propene dimerization reactions. Table 1 summarizes the results of the propene dimerization reactions with different buffering substances. All dimerization reactions catalyzed by complex 15 yielded a similar dimer distribution consisting of about 25% n-hexenes, 69% 2-methylpentenes, and 6% 2,3-dimethylbutenes (± 2%). Obviously, the ligand had no effect on the product distributions, which were similar to the distribution of a ligand free nickel salt.[16] As described in our recent publication,[17] the term “activity” was replaced by the term 5 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids 59 “productivity”. Wasserscheid et al. showed that above a certain concentration, the activities of the systems did not depend on the amount of catalyst anymore due to mass transport limitations.[16] Thus, the productivities can be calculated as the amount of products formed by one gram of active liquid in one hour under the assumption that the productivities of the systems do not depend on the catalyst concentration. Table 1: Nickel catalyzed biphasic propene dimerization reactions in buffered acidic 1butyl-3-methylimidazolium (BMIM) chloroaluminate ionic liquids.[1] Buffer Productivity[2] Dimers [%] NPh3 > 8.6[3] 68.0 9 > 12.2[3] 57.4 10 2.9 53.5 11 6.3 78.6 12 3.7 Oil[4] 13 > 12.8[3] 69.3 14 0.4 Oil[4] 8 6.1 84.1 5 11.6 Oil[4] BiPh3 > 9.3[3] 85.5 [1] Reaction conditions: 2 – 3 g buffered ionic liquid; composition [buffer] / [BMIM]+[Al2Cl7]− = 0.30; catalyst precursor 15; [cat] = 10–5 mol / gionic liquid; T = 25°C; stirring rate = 1200 min–1; t = 60 min; 300 ml glass autoclave; 40 – 60 ml liquid propene. [2] gproduct / gionic liquid x h. [3] Complete conversion. [4] Higher oligomers derived from a cationic oligomerization reaction. Triphenylamine certainly coordinated too strongly to the aluminum centers resulting in a reduced buffering ability coming along with a low dimer yield of only 68%. The triphenylphosphine system lacked selectivity due to its low solubility in the ionic liquid. Most of the 0.30 buffer equivalents did not dissolve in the ionic liquid phase. Tris(mchlorophenyl)phosphine (11) gave a higher dimer yield of 79% due to its better solubility 5 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids 60 in the melt compared to its para-substituted equivalent (10). The 2-imidazoliumdiphenylphosphine (12), which beared a cationic 1-butyl-3-methylimidazolium chloride moiety instead of a phenyl ring,[3b] showed no buffering ability. Only higher oligomers from uncontrolled cationic olefin oligomerization reactions initiated by the highly Lewis acidic aluminum centers were obtained.[10] In contrast, diphenylphosphinoferrocene (13) sufficiently buffered the system yielding 69% dimers. The addition of the sodium salt of diphenylphosphinobenzene-3-sulfonic acid (14) almost deactivated the system. Only small amounts of higher oligomers were formed during the reaction. Probably the oxygen of the sulfonate group reacted with chloroaluminate species deactivating the system and reducing the mobility of the buffer. As expected, the iodide salt of the synthesized ionic phosphine 8 gave the best result of all tested phosphines providing 84% dimers. It was completely soluble in the ionic liquid and almost caught up with the BiPh3 system (86% dimers[17]). However, it had to be taken into account that the additional iodide ions of the buffer reduced the overall acidity of the system, which also favoured the formation of dimers. Surprisingly, the cationic phosphine with a tetrafluoroborate anion 5 did not dissolve in the ionic liquid at all. Hence this system only produced higher oligomers. Effects of 8 on EtAlCl2 Buffered Chloroaluminate Melts Additionally, the effect of the ionic phosphine 8 on the performance of a typical EtAlCl2 buffered system was investigated (Table 2). For example, this composition is applied in the commercial DIFASOL® process. In batch propene dimerization reactions, the unmodified EtAlCl2 buffered composition yielded about 80% dimers with catalyst precursor 15. Addition of 0.03 equivalents of the cationic phosphine 8 slightly increased the C6 selectivity to 84%. The addition of 0.06 equivalents of 8 resulted in 89% dimers. The combination of EtAlCl2 buffered systems and an ionic phosphine buffer improved the overall dimer selectivity about 10% compared to a standard DIFASOL® system while maintaining a high productivity. Any phosphine leaching into the product phase should be prevented by its ionic character. A stable system with a high lifetime is therefore expected. 5 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids 61 Table 2: Nickel catalyzed biphasic propene dimerization reactions in EtAlCl2 buffered systems modified with the cationic phosphine 8.[1] [8] / [BMIMCl] Productivity[2] Dimers [%] 0 > 18.1[3] 79.6 0.03 > 22.5[3] 83.6 0.06 > 13.5[3] 89.1 [1] Reaction conditions: 2 – 3 g buffered ionic liquid; [BMIMCl] / [AlCl3] / [EtAlCl2] = 1.00 / 1.20 / 0.20; catalyst precursor 15; [cat] = 10–5 mol / gionic liquid; T = 25°C; stirring rate = 1200 min–1; t = 45 min; 300 ml glass autoclave; 40 – 60 ml liquid propene. [2] gproduct / gionic liquid x h. [3] Complete conversion. NMR Investigations Since triarylphosphines were efficient buffers for Lewis acidic chloroaluminate ionic liquids, the phosphorus center makes them predestined for 31P NMR investigations. Systems with the even more effective BiPh3 buffer are difficult to investigate by nuclear magnetic resonance experiments. The naturally occurring 209Bi isotope displays a spin of 9/2. First, a [BMIM]+[Al2Cl7]− ionic liquid was saturated with hardly soluble PPh3 and mixed with a few drops of C6D6 in an NMR tube. Scheme 3 shows the 31P and 27Al NMR spectra. The “buffering” interaction can be seen clearly from the broad peak (line width at half maximum: 560 Hz) at δ = −13 ppm in the 31P NMR spectrum due to fast intermolecular exchange reactions. The PPh3 signal, which usually appears at δ = −6 ppm, is shifted to higher frequencies due to the interaction of triphenylphosphine with the Lewis acidic aluminum centers. Between δ = 4 − 10 ppm, another signal appears, which indicates a new species. It obviously consists of more than one P nuclei and gives an AB2 spin pattern. The constitution of this compound is not clear yet. The 27Al NMR spectrum displays a line width at half maximum of 2450 Hz, which is typical for an acidic [imidazolium]+[Al2Cl7]− ionic liquid.[23] Analogously to the literature, a signal with this line width arises from [Al2Cl7]− species. These anions are bigger than [AlCl4]− anions and possess a less symmetrical surrounding of the Al center. Thus, the signal is broadened by efficient nuclear quadrupole relaxation. 5 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids 62 Scheme 3: 31P and 27Al NMR spectrum of PPh3 in the ionic liquid [BMIM]+[Al2Cl7]− (C6D6, 25°C). A second sample consisting of [BMIM]+[Al2Cl7]− with 0.30 equivalents of the more basic, sterically hindered tri-tert-butylphosphine was prepared. Surprisingly, after the addition of small amounts of C6D6, the system underwent a phase separation. A clear phase (Scheme 4) separated on top of a turbid phase (Scheme 5). Upon addition of more C6D6, the phases combined again. Both phases were investigated by means of NMR spectroscopy. The peak at δ = 63.6 ppm belongs to excess tri-tert-butylphosphine, which does not interact with the ionic liquid.[24] Unlike using PPh3, the interaction of tri-tertbutylphosphine with the Lewis acidic aluminum species in the ionic liquid generates an unusual sextet between δ = 22 and 32 ppm. The P–Al interaction is stronger leading to a splitting of the 31P signal by the spin 5/2 nucleus 27Al, and exchange reactions are slow in NMR time scale. The different intensities within the sextet are normal for couplings of 1/2 spin nuclei with a quadrupole nucleus. A simple tri-tert-butylphosphine adduct of AlCl3 in organic solvents did not show a sextet[25] but broad signals. In toluene, the adducts of PMe3 or PEt3 with AlCl3 only generated a singlet in the 31P NMR spectrum, while a doublet was observed in the 27Al NMR spectrum.[26] This could be explained with fast exchange reactions in solution. In chloroaluminate ionic liquids, the exchange rates may be reduced. The π-interactions of toluene probably reduced the Lewis acidity of AlCl3 in solution leading to a weaker donor-acceptor interaction with the phosphine. Similarly to the PPh3 system above, the spectrum shows an AB2 spin pattern between δ = 56 and 60 ppm. This spin system must be derived from a different phosphine species present in the liquid. Unfortunately, the nature of that species could not be determined 5 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids 63 yet. Due to fast chlorine / alkyl exchange in the reaction mixture, ring or chain like oligophosphines could be formed. This behaviour should be discussed in another paper. Scheme 4: 31P and 27Al NMR spectrum of the upper phase of the mixture [BMIM]+[Al2Cl7]− / tri-tert-butylphosphine (0.30 equivalents) in C6D6 (25°C). Scheme 5: 31P and 27Al NMR spectrum of the bottom phase of the mixture [BMIM]+[Al2Cl7]− / tri-tert-butylphosphine (0.30 equivalents) in C6D6 (25°C). The spectra of both the upper and the lower phase, display the same peak patterns. However, the concentrations are different. The P–Al interacting species is enriched in the lower phase with a ratio of approximately 83 to 17 compared to the undefined species. In the upper phase the ratio is about 55 to 45. In the 27Al NMR spectra two sharp signals at δ = 134.6 and 132.1 ppm stick out of the broad overlying signal of [Al2Cl7]−. They belong to the corresponding Al center coupling with tri-tertbutylphosphine. The signals of at least two other Al species must be present in the system indicated by two shoulders at about 112 and 122 ppm. For comparison, another sample was prepared by dissolving the ionic phosphine 8 in a neutral [BMIM]+[AlCl4]− ionic liquid (Scheme 6). The 31P NMR signal appears at δ = −5.7 5 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids 64 ppm with a line width at half maximum of 32 Hz. These signals are broader than those obtained from 8 dissolved in CDCl3 (2.2 Hz). The line width of 41 Hz in the 27Al NMR spectrum also correlates with a typical line width of 23 Hz measured for neutral imidazolium based chloroaluminate ionic liquids.[23a] Thus, there is only little interaction of the phosphorus atoms with the [AlCl4]− species. Scheme 6: 31P and 27Al NMR spectra of the cationic phosphine 8 in the ionic liquid [BMIM]+[AlCl4]− (C6D6, 25°C). Finally, three samples of the cationic phosphine 8 in [BMIM]+[Al2Cl7]− with concentrations of 0.10 (Scheme 7), 0.30 and 0.50 equivalents (Scheme 8) were prepared and investigated by NMR spectroscopy. In the 31P NMR spectra, the signal derived from the P–Al interaction appears at δ = 13.8 ppm in all three samples. The multiplet of the undefined phosphine species was observed between δ = 3 and 9 ppm. With increasing phosphine content, a shoulder develops in the 31P NMR signal induced by additional iodide anions coming along with the addition of the ionic phosphine. Our results showed that weak P–Al interactions are crucial for phosphine buffered Lewis acidic chloroaluminate ionic liquid cocatalysts. The strong Lewis base tri-tert-butylphosphine coupled with aluminum to give a doublet in the 27Al NMR and a sextet in the 31P NMR spectrum. In contrast, only broad signals were observed for triarylphosphine buffers. A highly dynamic system is obtained from the combination of weakly Lewis basic triarylphosphines in combination with Lewis acidic chloroaluminate melts. The fast P–Al exchange is necessary to maintain a sufficient Lewis acidity to activate nickel catalyst precursor as well as to suppress uncontrolled cationic olefin oligomerization reactions. 5 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids 65 Scheme 7: 31P NMR spectrum of 0.10 equivalents of the cationic phosphine 8 in the ionic liquid [BMIM]+[Al2Cl7]− (C6D6, 25°C). Scheme 8: Excerpts of the 31P NMR spectra of the mixtures containing 0.10 (left), 0.30 (middle) and 0.50 (right) equivalents of the cationic phosphine 8 in the ionic liquid [BMIM]+[Al2Cl7]− (C6D6, 25°C). Conclusions The synthesized cationic triphenylphosphine derivative 8 successfully buffered the highly Lewis acidic [BMIM]+[Al2Cl7]− ionic liquid. Its tetrafluoroborate equivalent 5 was insoluble in the ionic liquid. A nickel complex could be activated with the buffered composition and dimerized propene in a biphasic reaction with a selectivity of 84% to give dimers. Undesired uncontrolled cationic olefin oligomerization reactions were prevented by the addition of 0.30 equivalents of 8 to the ionic liquid. Adding 0.06 equivalents of 8 to a commercially used DIFASOL® ionic liquid composition, the dimer yield could be increased from 80% to 89%. The P–Al interaction was investigated by 31P and 27Al NMR spectroscopy. While no P–Al coupling between buffering triarylphosphines and aluminum centers could be observed, the more basic but sterically hindered tri-tertbutylphosphine coupled with 27Al to give an expected but unusual sextet in the 31P NMR spectrum. The straight-forward synthesis of the cationic phosphine 8 in combination 5 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids 66 with the high yield makes such cationic triphenylphosphine derivatives interesting for biphasic reactions. When triphenylphosphine is replaced by its cationic derivative, leaching effects can be minimized and catalyst lifetimes should increase. Due to the absence of acidic protons, triphenylphosphines with quaternized ammonium substituents can also be applied when highly reactive metal alkyl groups like alkylaluminum bonds are present. Besides dimerization and polymerization reactions, possible applications can be found in biphasic hydroformylation,[27] hydrogenation,[6b] or cross coupling reactions.[4] 5.3 Experimental Section General Remarks All chemical manipulations were carried out using standard Schlenk techniques under argon atmosphere. THF was distilled from Na/K alloy. CH2Cl2 was distilled in two steps from P4O10 and CaH2. For drying methanol, magnesium turnings were dissolved in the solvent before distillation. Demineralized water was degassed with argon prior to use. Acetonitrile (Acros, p.a.) was used as received. 1H, 11B, 13C, 27Al and 31P NMR spectra were recorded on Varian Inova 300 MHz or 400 MHz spectrometers. Chemical shifts are given relative to Me4Si: δ 1H (CHCl3) = 7.24 ppm; δ 1H (CD3(SO)CD2H) = 2.50 ppm; δ 13C (CDCl3) = 77.2 ppm; δ 13C (CD3(SO)CD3) = 39.5 ppm; external 85% H3PO4 [δ 31P = 0 ppm for Ξ (31P) = 40.480747 MHz]; external 1.1 M Al(NO3)3 in D2O [δ 27Al = 0 ppm for Ξ (27Al) = 26.056890 MHz]; external BF3-OEt2 [δ 11B = 0 ppm for Ξ (11B) = 32.083971 MHz]. Chemical shifts are given to ± 0.1 ppm in 13C and 31P NMR spectra and ± 0.4 ppm in 11B and 27Al NMR spectra. The products of the dimerization experiments were characterized by gas chromatography (Agilent 6850). GC-MS spectra were measured on a Thermo Scientific FOCUS DSQ™. Mass spectra were recorded on a Varian MAT CH7 instrument (direct inlet system, electron impact ionization 70 eV). Elemental analyses were performed with a VarioEl III CHN instrument. Acetanilide was used as a standard. Propene (99.3%) was purchased from Riessner Gase, Lichtenfels, and was dried over a column packed with P4O10. EtAlCl2 (0.9 M in heptane) and tri-tert-butylphosphine (95%) were purchased from Acros and used without further purification. Compound 11 was synthesized according to the literature.[4a] Triphenylamine, triphenyl- 5 Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids 67 phosphine, triphenylbismuth, tris(p-chlorophenyl)phosphine, and tris(m-chlorophenyl)- phosphine were purchased from ABCR. 1-Butyl-3-methylimidazolium chloride, dicyclohexylamin, trichlorosilane, p-(dimethylamino)phenyldiphenylphosphine, the boranetetrahydrofurane complex, AlCl3 (ReagentPlus®), and compound 13 were purchased from Sigma-Aldrich and used without further purification. Diphenylphosphinoferrocene was obtained from MCAT. Synthesis of the Ionic Liquids The ionic liquids were synthesized by directly mixing the corresponding amounts of AlCl3 and chloride salt in a cooled Schlenk tube. In the case of EtAlCl2 containing systems, a 0.9 molar solution of EtAlCl2 in heptane was added to the liquid, and the solvent was removed in vacuo. The obtained ionic liquids were stored in Schlenk tubes. Procedure for the Dimerization Reactions using Buffered Ionic Liquids Prior to the experiments, about 2 ml of the ionic liquid were filled into a Schlenk tube and mixed with the buffer. After the buffer had dissolved upon stirring, the nickel complex 15 was added. The obtained homogeneous solution was syringed into the reaction vessel. Its amount was determined by the weight difference of the syringe. In the case of partly soluble buffers, the mixture was stirred until the solid was distributed homogeneously in the liquid. Then, the slurry was syringed into the reaction vessel. For all dimerization experiments, a 300 ml glass autoclave equipped with a stirring bar and a magnetic stirrer with a stirring rate of 1200 min–1 was used. The glass autoclave was kept in a drying oven at 150°C for several hours before an experiment. After addition of the active ionic liquid, propene (40 to 60 ml) was condensed into the glass autoclave by liquid nitrogen cooling. Then, the autoclave was placed in a metal box for safety reasons, and the temperature was regulated by an external water bath. After the experiment, the pressure was slowly released by opening a valve. The product fraction was filtered through a short plug of silica and analyzed by gas chromatography. Synthesis of [4-(Trimethylammonium)phenyldiphenylphosphine Oxide] Iodide 3 Hydrogen peroxide (5.0 g, 30% in H2O) was slowly added to a solution of 1 (11.11 g, 36.38 mmol) in CH2Cl2 (100 ml). The mixture was stirred for 30 minutes and evaporated to dryness. The crude phosphine oxide 2 was dissolved in CH2Cl2 (100 ml) followed by 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 74 the composition. At lower reaction temperatures, the selectivities to give dimers increased significantly. Propene dimers were obtained with selectivities of up to 98%. A cation screening with 100 ammonium and phosphonium halide salts was performed. NMethylpyrrolidine hydrochloride gave the best results in terms of selectivity and lifetime. Propene dimerization reactions in BiPh3 buffered chloroaluminate melts were catalyzed by various soluble nickel compounds with similar performances. The introduction of basic, sterically demanding tricyclohexylphosphine ligands led to higher degrees of branching, however, at the expense of lower dimer selectivities. 6.1 Introduction Selective dimerization reactions of α-olefins attracted a lot of attention in the second half of the 20th century, after Ziegler described the “nickel effect” in 1955.[1] Nickel impurities from previous hydrogenation reactions were found to catalyze selective ethene dimerization reactions to give butenes in the presence of alkylaluminum compounds. Ziegler was originally interested in chain growth and polymerization reactions. However, his discovery also launched an intensive research on selective nickel catalyzed dimerization reactions of short chain olefins in the following decades.[2] The DIMERSOL® process of the Institut Français du Pétrole (IFP) is the largest industrial process based on nickel complexes activated by alkylaluminum cocatalysts. It was developed by Chauvin et al. in the 1970s[3] to dimerize propene (DIMERSOL® G) or 1-butene (DIMERSOL® X) selectively. Currently, 35 licensed DIMERSOL® units with typical capacities between 20000 and 90000 tons per year are in use for selective dimerization reactions of short chain olefins.[4] About 30 years ago, Wilkes et al. and Osteryoung et al. reported the first room temperature ionic liquids based on aluminum chloride.[5] In the 1980s, these chloroaluminate ionic liquids were thoroughly investigated in terms of their physical and electrochemical properties.[6] It took about ten more years until chloroaluminate ionic liquids started to attract attention as new media for biphasic catalysis.[7] Initially, Lewis acidic chloroaluminate melts were used as catalysts for Friedel Crafts alkylation reactions.[8] In the 1990s, Chauvin et al. developed nickel catalyzed biphasic 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 75 dimerization reactions of α-olefins in organochloroaluminate ionic liquids.[9] Ethylaluminum groups were found to suppress uncontrolled cationic olefin oligomerization reactions, which occur in solely AlCl3 based ionic liquids.[10] The IFP brought this so called DIFASOL® process to industrial application by retrofitting it to their existing DIMERSOL® units.[4a, 11] Besides alkylaluminum groups, the addition of weak Lewis bases also reduces the “latent acidity”[12] of chloroaluminate melts. While Wasserscheid et al. used pyrrole, pyridine, and quinoline derivatives to prevent undesired cationic olefin oligomerization reactions,[13] we recently reported that triphenylphosphine and triphenylbismuth efficiently buffer even highly acidic chloroaluminate ionic liquids.[14] These buffered chloroaluminate melts were successfully employed to activate nickel complexes for selective propene dimerization reactions. Buffering with Lewis base additives is advantageous compared to the commercial DIFASOL® system. First, buffering with EtAlCl2 is restricted to only slightly acidic compositions. If higher acidities are chosen, the dimer selectivities decrease. Second, at higher alkylaluminum contents, leaching of neutral chloroalkylaluminum compounds formed by disproportionation reactions[15] becomes a major problem.[11b] Also, alkylaluminum cocatalysts contribute to the deactivation of Ni(II) catalysts by reducing Ni(II) to its zero oxidation state.[16] Usually, the melting points of aluminum chloride based ionic liquids display a minimum for highly acidic melts, while a local maximum is observed around equimolar compositions.[6a, 17] Therefore, DIFASOL® systems are limited to cations intrinsically forming low melting liquids, e.g. N,N’-alkylmethylimidazolium or N-alkylpyridinium salts.[13b] In a recent paper, we described that the melting points of acidic chloroaluminate systems are reduced upon addition of BiPh3. Thus, it was possible to obtain buffered room temperature ionic liquid compositions from aluminum chloride and cations, which would only yield solids at ambient temperature without BiPh3.[14a] Furthermore, the scope of potential cations became even larger since low melting highly acidic systems could be used for nickel catalyzed selective olefin dimerization reactions. Because no reactive alkylaluminum groups were present in BiPh3 buffered chloroaluminate ionic liquids, these systems were not limited to quaternary ammonium cations. Instead, 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 76 cheap halide salts like amine hydrochlorides could be used, which can be recovered easily upon hydrolysis. In this paper, we report nickel catalyzed propene dimerization reactions in BiPh3 buffered chloroaluminate melts with 100 mostly unconventional cation compositions. Only two out of 100 mixtures did not form liquids at ambient temperature. Furthermore, the effect of Lewis acidities, buffer contents and reaction temperatures on the dimer selectivities was investigated. Finally, several nickel complexes were tested for their performances in BiPh3 buffered chloroaluminate melts. 6.2 Results and Discussion Variation of Buffering and Acidity Levels As described in our recent paper, not only slightly acidic chloroaluminate melts could be buffered with BiPh3 but also highly acidic [cation]+[Al2Cl7]− systems.[14a] Therefore, the effect of different acidities and BiPh3 contents in chloroaluminate ionic liquids on nickel catalyzed propene dimerization reactions was investigated in detail (Table 1). For the screening experiments, we used our standard nickel complex A as the catalyst precursor (Figure 1). Figure 1: Catalyst precursor used for biphasic propene dimerization reactions. We chose a system based on the commercially available 1-butyl-3-methylimidazolium (BMIM) tetrachloroaluminate, to which the corresponding amounts of AlCl3 were added. With slightly acidic AlCl3 / BMIMCl = 1.20 systems, which were used by Wasserscheid et al.[13b] and for DIFASOL® systems[9c], dimers were obtained starting with 0.07 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 77 equivalents of BiPh3 (2). With increasing buffer content, selectivities increased. Employing 0.30 equivalents of BiPh3, already 96% dimers were obtained (4), however, at the expense of a reduced productivity. Table 1: Nickel catalyzed propene dimerization reactions in BiPh3 buffered chloroaluminate melts with different compositions.[1] No. [AlCl3] / [BMIMCl] [BiPh3] / [BMIMCl] Productivity[2] C6 [%] 1 1.20 0.05 > 15.8[3] Oil[4] 2 1.20 0.07 > 10.8[3] 89.2 3 1.20 0.12 > 8.0[3] 93.1 4 1.20 0.30 6.7 96.0 5 1.30 0.12 > 8.0[3] 81.8 6 1.50 0.12 > 10.6[3] 74.1 7 1.50 0.18 > 12.0[3] 83.2 8 2.00 0.12 > 11.4[3] 74.8 9 2.00 0.18 > 12.3[3] 79.5 10 2.00 0.24 > 7.4[3] 83.3 11 2.00 0.30 > 9.3[3] 85.5 12 2.00 0.60 > 11.5[3] 90.6 [1] Reaction conditions: 2.5 – 3.5 g buffered ionic liquid; catalyst precursor A; [cat] = 10–5 mol / gionic liquid; T = 25°C; stirring rate = 1200 min–1; t = 60 min; 300 ml glass autoclave; 40 – 70 ml liquid propene. [2] gproduct / gionic liquid x h. [3] Complete conversion. [4] Higher oligomers derived from a cationic oligomerization reaction. Due to mass transfer limitations,[13d] the term “activity” was replaced by “productivity”. Thus, above a catalyst concentration of about 10–6 mol / gionic liquid,[14a] the product yield per hour is proportional to the amount of ionic liquid and independent from the catalyst 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 78 concentration. As expected, increasing acidities with a constant buffer level reduced the dimer selectivities. However, from an AlCl3 / BMIMCl ratio of 1.50 to 2.00, the selectivities dropped only a few percent if the buffer level was held constant. Thus, the advantage of low melting points at highly acidic compositions did not have to be paid by disproportional high amounts of buffer. With only 0.12 equivalents of buffer, the maximum acidic system still yielded about 75% dimers (8). Cation Screening Considering these results, we were curious if our system can be extended to rather unusual and cheap cations. In the literature, only a few chloroaluminate systems were mentioned, which did not rely on standard imidazolium or pyridinium cations. Trimethylammonium heptachlorodialuminate was used for cationic oligomerization reactions of 1-decene.[18] The use of triethylamine hydrochloride, dibutylamine hydrochloride, ethylamine hydrochloride, and dimethylamine hydrochloride was reported for Friedel Crafts alkylation reactions.[7e] Dimethylaniline hydrochloride,[19] and several quaternary asymmetric benzyl substituted melts were described, too.[20] However, none of these were used for nickel catalyzed olefin dimerization reactions due to the above mentioned melting point and acidity restrictions. Friedel Crafts alkylation reactions or cationic olefin oligomerization reactions employ the Lewis acidic ionic liquid itself as catalyst. Thus, in these cases a variation of the cation should only result in minor changes on the results of the reactions. In contrast, the performances of nickel catalyzed olefin dimerization reactions in BiPh3 buffered chloroaluminate melts should strongly depend on the cation. The reactions are catalyzed by a nickel complex, not by the ionic liquid itself. It is necessary that the BiPh3 buffer dissolves well in the ionic liquids to minimize leaching effects. Also, the cation determines the viscosities and the solubilities of the substrate olefin within the ionic liquids. Both the viscosity and the solubility of the substrate are crucial since biphasic nickel catalyzed dimerization reactions in chloroaluminate melts suffer from a mass transport limitation of the olefin.[13d] Therefore, we decided to perform a systematic screening of cations including quaternary ammonium salts, hydrochlorides of primary, secondary and tertiary amines, phosphonium salts, heterocyclic ammonium salts, and cations with more than one heteroatom. One hundred different halide salts and their AlCl3 / halide salt = 2.00 compositions were synthesized. These compositions were buffered with 0.30 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 79 equivalents of BiPh3 and tested for nickel catalyzed propene dimerization reactions. The screening reactions were performed in batch experiments. Furthermore, the lifetime of each system was investigated qualitatively. After each experiment, the product phase was decanted until the dimer selectivity dropped significantly. In Table 2, 33 selected cations and the corresponding C6 selectivities obtained in the first batch experiment are shown. The complete dataset can be found in the Supporting Information. Since the reference complex A had no effect on the branching of the hexene dimers,[14a] all C6 fractions consisted of (± 2%) 25% n-hexenes, 69% 2-methylpentenes, and 6% 2,3-dimethylbutenes. This correlated to the typical product distribution obtained from ligand free nickel salts.[13d] Because the productivities of such buffered systems mainly depended on the buffering level,[14a] the productivities of different cations should not be compared directly. Cations yielding compositions with a good BiPh3 efficiency automatically displayed higher dimer selectivities and subsequently lower productivities. Therefore, the goal was to identify systems, which kept high selectivities over many repetitions. Surprisingly, 95 of 100 compositions produced dimers. Only two cations, formamidine hydrochloride and 1,4-diazabicyclo[2.2.2]octane hydrochloride, yielded solids at ambient temperature (see Supporting Information). Even the pentavalent triphenylphosphine dichloride based system (45) successfully dimerized propene to give 50% dimers (45). However, the systems displayed big differences in their overall performances. The best results were obtained with quaternary ammonium salts and hydrochlorides of tertiary amines. In general, hydrochlorides of primary and secondary amines were less selective. However, even small changes on the cation could have severe effects on the lifetimes and selectivities of the systems. For example, the 1-butyl-3-methylimidazolium chloride cation, which was used for the initial acidity screening experiments, was found among the worst quaternary ammonium cations (22). Introduction of an ethyl (17) instead the butyl chain significantly improved the selectivity as well as the lifetime of the system (see Supporting Information). As already reported, the system based on Nmethylpyrrolidine hydrochloride (26) displayed the best performance in terms of selectivity and qualitative lifetime.[14a] In the case of the BiPh3 buffer, there appears to be an optimum range of carbon atoms for ammonium salts, roughly between five and nine. Ionic liquids derived from cations with more or less carbon atoms dropped much faster in terms of selectivity than those within that range. 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 80 Table 2: Selected results of the cation screening for nickel catalyzed propene dimerization reactions in BiPh3 buffered chloroaluminate melts (left: quaternary ammonium chloride salts; middle: hydrochlorides of tertiary amines; right: hydrochlorides of primary and secondary amines and phosphonium salts).[1] No. Cation C6 [%] No. Cation C6 [%] No. Cation C6 [%] 13 92.0 24 93.1 35 90.4 14 91.2 25 92.5 36 83.0 15 90.9 26 91.6 37 79.0 16 90.1 27 90.7 38 76.1 17 90.0 28 90.0 39 73.3 18 89.7 29 89.9 40 66.1 19 89.3 30 89.6 41 92.5 20 89.3 31 89.6 42 91.0 21 89.2 32 89.6 43 85.4 22 86.2 33 89.4 44 68.6 23 75.9 34 89.1 45 49.5 [1] Reaction conditions: 2.5 – 4.0 g buffered ionic liquid; composition [BiPh3] / [cation]+[Al2Cl7]− = 0.30; catalyst precursor A; [cat] = 10–5 mol / gionic liquid; T = 25°C; stirring rate = 1200 min–1; t = 60 min; 300 ml glass autoclave; 40 – 60 ml liquid propene. 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 81 Temperature Effect on Dimer Selectivities Further, we investigated the effect of the reaction temperature on dimer selectivities with 0.30 and 0.60 equivalents of BiPh3 (Table 3). The reactions were performed either at ambient temperature or at 0°C. Table 3: Nickel catalyzed propene dimerization reactions in BiPh3 buffered chloroaluminate melts at 0°C and 25°C.[1] Cation No. [Buffer] / [Cation] Temperature [°C] Productivity[2] C6 [%] 46 0.30 25 > 10.1[3] 90.7 47 0.30 0 > 6.7[3] 95.5 48 0.60 25 4.6 93.8 49 0.60 0 – [4] – 50 0.30 25 > 10.6[3] 89.4 51 0.30 0 > 12.9[3] 96.0 52 0.60 25 > 7.6[3] 90.8 53 0.60 0 1.8 98.0 [1] Reaction conditions: 3.2 – 5.2 g buffered ionic liquid; composition [BiPh3] / [cation]+[Al2Cl7]− = 0.30; catalyst precursor A; [cat] = 10–5 mol / gionic liquid; stirring rate = 1200 min–1; t = 60 min for T = 25°C; t = 120 min for T = 0°C; 300 ml glass autoclave; 40 – 60 ml liquid propene. [2] gproduct / gionic liquid x h. [3] Complete conversion. [4] Composition solid at 0°C. While the selectivity of an ethyldiisopropylamine (Hunig’s base) hydrochloride based system with 0.30 equivalents of BiPh3 increased from 91% at 25°C (46) to 96% at 0°C (47) the mixture containing 0.60 equivalents of BiPh3 solidified at 0°C (49). Thus, the series was repeated with tributylamine hydrochloride. By lowering the temperature to 0°C, the dimer selectivity could be increased by 7% to 96% (51) and 98% (53), respectively. The experiments also showed that above a certain buffer concentration, the addition of more BiPh3 only resulted in minor improvements of the dimer selectivity. 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 82 Thus, a cation screening was mandatory to identify halide salts, which display high lifetimes combined with high selectivities. Effect of Different Nickel Catalyst Precursors on Dimer Selectivities So far, only catalyst precursor A was used as a reference in all screening experiments. Now, a series of typical nickel(II) halides and complexes was tested for their performances in combination with BiPh3 buffered chloroaluminate melts (Table 4). Considering the results of our previous cation screening, N-methylpyrrolidine hydrochloride mixed with two equivalents of aluminum chloride and buffered with 0.30 equivalents of BiPh3 was chosen as reference system. Since propene dimerization reactions in buffered chloroaluminate melts are limited by mass transfer of the olefin,[13d] the productivities of all nickel compounds should be the same if a sufficiently high catalyst concentration was chosen. In a recent paper, we determined the minimum catalyst concentration necessary to exclude catalyst concentration restrictions to be around 10–6 molNi / gionic liquid.[14a] We applied a concentration of 10–5 molNi / gionic liquid to ensure that the catalyst concentration was not the rate determining factor. First, the four binary nickel halides were tested. All of them yielded very high selectivities of more than 90%. However, the productivities were quite different. Nickel fluoride produced 24.9 g product per gram buffered ionic liquid (54). The other halides displayed lower productivities. This behaviour might be explained by the low solubility of these ligand free halide salts. Although 10–5 molNi / gionic liquid were added to the liquid, the salts did not dissolve completely. Nickel chloride seemed to be almost insoluble in the buffered liquid. The bromide and iodide salts did not dissolve very well, too. Thus, the amount of catalyst, which had actually dissolved in the ionic liquid, might have been below 10–6 molNi / gionic liquid. In this case, not only the mass transfer of the olefin but also the catalyst concentration determined the productivity. To prove this theory, the triphenylphosphine adducts of nickel chloride and nickel bromide were tested, too (58, 59). Indeed, both complexes readily dissolved in the ionic liquid and yielded almost the same productivity of around 28 gproduct / gionic liquid x h with a selectivity around 91% to give C6. The productivities of nickel hexafluoroacetylacetonate (60) and bis-(N-isopropylsalicylaldimine)-nickel(II) (61) were only slightly lower. 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 83 The compositions of all C6 fractions were consistent with ligand free nickel salts. Hydrogenation showed that the hexenes consisted of 25% n-hexenes, 69% 2methylpentenes, and 6% 2,3-dimethylbutenes (± 2%). Either, the triphenylphosphine ligands were abstracted from the Lewis acidic ionic liquids, or they did not influence the degree of branching. Table 4: Catalyst screening for nickel catalyzed propene dimerization reactions in BiPh3 buffered chloroaluminate melts.[1] No. Catalyst Precursor C6 [%] Conversion [%] Productivity[2] 54 NiF2 90.9 69 24.9 55 NiCl2 95.4 7 2.5 56 NiBr2 92.6 43 14.7 57 NiI2 93.0 47 16.1 58 (PPh3)2NiCl2 90.8 80 28.8 59 (PPh3)2NiBr2 91.0 81 28.2 60 Ni(hfacac)2 91.8 75 23.1 61 NiL2[3] 90.6 75 27.0 [1] Reaction conditions: 4.4 g buffered ionic liquid; composition [BiPh3] / [N-methylpyrrolidinium]+[Al2Cl7]− = 0.30; [cat] = 10–5 mol / gionic liquid; T = 40°C; stirring rate = 600 min–1; t = 60 min; stirred 300 ml Parr stainless steel autoclave; 130 – 160 g liquid propene; hfacac = hexafluoroacetylacetonate. [2] gproduct / gionic liquid x h. [3] NiL2 = bis-(N-isopropylsalicylaldimine)-Ni(II)). Effect of Tricyclohexylphosphine Ligands on the Dimer Structure Finally, nickel chloride complexed by two sterically demanding tricyclohexylphosphine ligands was employed as catalyst precursor for olefin dimerization reactions in BiPh3 buffered chloroaluminate ionic liquids (Table 5). Sterically hindered phosphines bound to a catalytically active nickel center in propene dimerization reactions are known to favour the formation of highly branched dimers.[2c, 2e, 2f] Especially 2,3-dimethylbutenes 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 90 (24), methyltributylammonium chloride (27), methyltrioctylammonium chloride (34), tetramethylammonium chloride (35), and N,N-dimethylmethyleneiminium chloride (36) were purchased from Sigma-Aldrich. 1-Ethyl-3-methylimidazolium chloride (17), trimethylphenylammonium chloride (19), hexamethylguanidinium chloride (29), and dichloromethylene-dimethyliminium chloride (32) were purchased from Acros. Chloroethane (Chloraethyl, Dr. Henning) was purchased in a local pharmacy. The free amine precursors and haloalkanes used to synthesize the cations described below were purchased from Sigma-Aldrich, Acros, or ABCR. N-Benzylpyridinium chloride (2), 1-benzyl-3-methylimidazolium chloride (8), N-benzyl-Nmethylpyrrolidinium chloride (10), tributylbenzylammonium chloride (14), dimethylcyclohexylbenzylammonium chloride (20), dimethyldibenzylammonium chloride (25), and dimethylethylbenzylammonium (26) chloride were synthesized following the literature procedure for 1-benzyl-3-methylimidazolium chloride (8).[3] Trimethylethylammonium chloride (15), N,N-dimethylpyrrolidinium chloride (16), trimethylcyclohexylammonium chloride (23), N-methylpyridinium chloride (28), and N,N-dimethylpiperidinium chloride (30) were obtained in a one step reaction by chloromethylation of the corresponding tertiary amines with methylchloroformate in almost quantitative yield applying literature procedures.[4] N-Propyl-N-methylpyrrolidinium chloride (4), N-butyl-N-methylpyrrolidinium chloride (9), N-hexyl-N-methylpyrrolidinium chloride (18), and N-butyl-N-methylpiperidinium chloride (7) were synthesized following the literature procedure for N-butyl-N-methylpyrrolidinium chloride (9).[5] N-Ethyl-N-methylpyrrolidinium (3) chloride and dimethylethylcyclohexylammonium chloride (12) were obtained as white precipitates from reactions of the corresponding amines with three equivalents of chloroethane in toluene at 70°C. The reactions were performed in a pressure Schlenk tube with a reaction time of ten days. 6-Azoniaspiro[5.5]undecane bromide[6] (33), 5-azoniaspiro[4.4]nonane bromide[6] (21), dimethylethylphenylammonium bromide[7] (26), 1,4-bis(3-methylimidazolium-1-yl)butane chloride[8] (13), dimethylbenzylphenylammonium chloride[9] (37), and N-butylpyridinium chloride[10] (22) were synthesized following literature procedures. 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 91 Hydrochlorides of Primary and Secondary Amines Ethylamine hydrochloride (40), dimethylamine hydrochloride (42), acetamidine hydrochloride (44), benzylamine hydrochloride (45), methylamine hydrochloride (48), aniline hydrochloride (49), and formamidine hydrochloride (51) were purchased from Sigma-Aldrich. Guanidine hydrochloride (50) was purchased from ABCR. The free amine precursors used to synthesize the cations described below were purchased from Sigma-Aldrich, Acros, or ABCR. Dicyclohexylamine hydrochloride (38), pyrrolidine hydrochloride (39), cyclohexylamine hydrochloride (41), dibenzylamine hydrochloride (43), methylbenzylamine hydrochloride (46), and N,N,N′,N′-tetramethylguanidine hydrochloride (47) were synthesized by slowly adding concentrated aqueous hydrochloric acid to a cooled flask containing the corresponding amines. Optionally, several hydrochlorides could be precipitated by the addition of isopropanol or acetone to the concentrated aqueous solution. All other compounds were obtained by vacuum drying of the aqueous solution at elevated temperatures. Hydrochlorides of Tertiary Amines 1-(2-Chloroethyl)piperidine hydrochloride (53), triethylamine hydrochloride (57), trimethylamine hydrochloride (71), 1-methylimidazolium chloride (86), imidazole hydrochloride (88), and pyridine hydrochloride (91) were purchased from Sigma-Aldrich. The commercially available amine precursors and haloalkanes used to synthesize the cations described below were purchased from Sigma-Aldrich, Acros, or ABCR. N-Propylpyrrolidine, N-isopropylpyrrolidine, N-nbutylpyrrolidine, N-tert-butylpyrrolidine, N-cyclohexylpyrrolidine, and N-propylpiperidine were obtained from the corresponding primary amines and dibromoalkanes applying the literature procedure described for Ntert-butylpyrrolidine.[11] N,N-Dimethyl-1-cyclopenten-1-amine was synthesized according to the literature.[12] N-Methylpyrrolidine hydrochloride (61), N-ethylpyrrolidine hydrochloride (70), N-propylpyrrolidine hydrochloride (75), N-isopropylpyrrolidine hydrochloride (65), N-nbutylpyrrolidine hydrochloride (78), N-tert-butylpyrrolidine hydrochloride (52), N-cyclohexylpyrrolidine hydrochloride (85), N-methylpiperidine hydrochloride (69), N-ethylpiperidine hydrochloride (77), N-propylpiperidine hydrochloride (60), N-cyclohexylpiperidine hydro- 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 92 chloride (81), dimethylethylamine hydrochloride (68), methyldiethylamine hydrochloride (55), dimethylisopropylamine hydrochloride (56), ethyldiisopropylamine hydrochloride (64), tributylamine hydrochloride (76), triisobutylamine hydrochloride (63), triisopentylamine hydrochloride (79), trioctylamine hydrochloride (59), dimethylcyclohexylamine hydrochloride (83), diethylcyclohexylamine hydrochloride (80), methyldicyclohexylamine hydrochloride (82), ethyldicyclohexylamine hydrochloride (73), dimethylbenzylamine hydrochloride (84), tribenzylamine hydrochloride (89), 1,2,2,6,6-pentamethylpiperidine hydrochloride (87), and 1,4-diazabicyclo[2.2.2]octane hydrochloride (92) were synthesized by slowly adding concentrated aqueous hydrochloric acid to a cooled flask containing the corresponding amines. Optionally, several hydrochlorides could be precipitated by the addition of isopropanol or acetone to the concentrated aqueous solution. All other compounds were obtained by vacuum drying of the aqueous solution at elevated temperatures. 2-Methyl-1-pyrroline hydrochloride (66), 1-pyrrolidino-1-cyclopentene hydrochloride (67), 1-piperidino-1-cyclohexene hydrochloride (74), N,N-dimethyl-1-cyclopenten-1amine hydrochloride (72), and N,N-dimethylaniline hydrochloride (90) could not be obtained as solids from aqueous solution and were prepared from their free amines and HCl gas in diethylether following the literature procedure described for triphenylphosphine hydrochloride[13] (99). N-Benzylpyrrolidine hydrochloride (62) and N-benzylpiperidine hydrochloride (58) were synthesized directly from benzyl chloride and pyrrolidine or piperidine following the literature procedure for 1-benzyl-3-methylimidazolium chloride[3] (8). Phosphonium Salts Trihexyltetradecylphosphonium chloride (93), butyltriphenylphosphonium chloride (95), tetrabutylphosphonium chloride (97), tetraphenylphosphonium chloride (98), and triphenylphosphine dichloride (100) were purchased from ABCR. Benzyltributylphoshonium chloride (94) and benzyltriphenylphosphonium chloride (96) were synthesized from the corresponding tertiary phosphine and benzyl chloride following the literature procedure for 1-benzyl-3-methylimidazolium chloride[3] (8). Triphenylphosphine hydrochloride was synthesized according to the literature[13] (99). 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 93 6.4.3 Cation Screening Quaternary Ammonium Salts Table 1: Cation screening for nickel catalyzed propene dimerization reactions in BiPh3 buffered chloroaluminate melts (sorted in descending order of selectivities in run 1).[1] No. Cation Run 1 C6 [%] Prod.[2] Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8 Run 9 Run 10 Run 11 1 95.1 6.0 93.5 7.3 87.6 > 10.1 70.2 9.5 69.3 8.9 62.5 7.5 2 92.3 8.9 91.9 9.7 91.9 > 10.5 89.9 > 11.1 76.9 > 12.0 61.3 7.8 Oil 7.0 3 92.0 7.8 92.1 9.0 90.2 9.2 88.0 > 11.7 80.5 > 10.5 65.4 7.9 61.9 7.1 4 91.6 > 13.1 89.6 > 12.3 87.2 > 14.0 71.8 > 11.4 61.9 9.0 56.2 7.3 5 91.5 > 12.4 89.7 > 12.1 86.2 > 14.1 76.4 > 11.4 74.5 > 11.5 74.0 11.2 77.6 9.0 77.7 8.7 77.2 6.4 68.3 5.9 6 91.4 9.0 80.3 > 15.2 69.5 > 11.2 67.2 9.1 67.2 7.5 63.8 6.8 7 91.2 > 10.8 89.7 > 11.4 86.5 > 10.9 83.7 > 10.6 75.3 > 10.6 70.8 8.9 69.6 6.9 8 90.9 > 9.4 89.5 > 10.7 88.0 > 12.2 75.9 > 11.2 65.6 9.7 59.5 7.4 Oil 6.8 9 90.9 > 11.7 91.4 > 10.8 91.7 > 11.2 91.6 9.9 91.1 > 10.9 90.3 > 11.3 89.0 > 10.2 86.0 > 11.4 78.5 > 10.7 70.9 8.8 71.1 8.4 10 90.9 > 8.7 92.3 > 10.4 91.9 > 9.3 92.0 > 9.2 91.8 > 9.8 89.4 > 9.4 89.1 > 11.7 80.0 > 11.0 73.5 7.7 11 90.7 > 11.9 87.9 > 10.5 86.7 > 11.9 76.6 > 11.5 60.0 9.3 12 90.7 > 13.4 88.9 > 14.1 82.5 > 14.2 67.4 > 14.0 65.3 10.9 62.8 9.7 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 94 No. Cation Run 1 C6 [%] Prod.[2] Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8 Run 9 Run 10 Run 11 13 90.5 8.9 89.9 8.3 90.1 10.4 89.1 9.4 86.7 12.7 77.4 > 14.5 72.3 9.1 62.6 6.1 14 90.3 > 9.5 90.0 > 11.3 86.0 > 13.6 73.6 > 13.1 72.3 8.8 69.3 6.4 65.2 5.4 15 90.3 7.0 91.9 7.6 91.9 7.1 91.8 9.1 89.9 > 10.1 74.9 > 10.4 71.2 8.8 66.7 5.4 16 90.1 10.3 89.8 10.4 89.1 > 11.5 89.0 > 11.6 71.7 > 11.6 63.2 9.2 17 90.0 > 9.9 89.4 > 9.1 90.9 > 10.8 89.3 > 8.7 89.5 > 9.9 88.8 > 9.6 86.5 > 9.6 84.0 > 9.4 67.8 7.7 64.6 8.8 18 89.9 > 11.6 89.2 > 13.3 83.0 > 14.2 63.0 10.8 60.2 9.9 19 89.7 > 13.0 89.9 > 14.3 88.8 > 14.4 85.9 > 12.7 77.8 > 14.5 75.4 > 14.0 74.0 7.6 66.9 4.6 20 89.6 > 7.8 90.1 > 8.9 86.0 > 10.1 79.7 > 9.2 65.1 > 7.6 72.8 > 7.7 55.5 5.1 21 89.6 8.1 91.5 8.4 92.0 9.4 92.2 8.9 92.4 9.2 91.7 9.6 90.8 9.6 90.2 > 10.7 86.8 > 11.1 78.4 > 10.2 77.9 5.5 22 89.3 > 11.5 86.4 > 11.9 85.1 > 12.6 81.8 > 13.4 58.0 8.5 55.9 8.0 23 89.3 > 9.0 91.3 > 9.5 92.2 > 10.5 91.5 > 10.7 91.0 > 10.1 90.9 > 10.9 89.3 > 10.0 84.3 > 12.9 77.2 > 11.0 74.2 8.9 Oil 7.7 24 89.2 > 11.5 86.6 > 12.6 84.9 > 12.8 75.4 > 12.9 64.5 8.8 25 89.0 8.3 84.1 > 10.6 77.3 8.1 59.1 8.8 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 95 No. Cation Run 1 C6 [%] Prod.[2] Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8 Run 9 Run 10 Run 11 26 88.9 > 10.6 89.1 > 10.9 78.2 > 11.9 63.5 > 10.1 27 88.6 > 9.3 84.3 > 9.0 81.0 > 11.2 64.0 > 9.0 28 88.0 > 9.6 87.9 > 9.5 87.2 > 10.1 85.9 > 10.7 83.5 > 10.8 80.4 > 11.7 74.1 > 10.4 71.4 > 10.6 70.1 > 9.72 69.0 > 8.7 64.2 7.4 29 87.3 8.0 69.3 7.5 73.8 7.2 66.5 9.7 30 86.7 > 10.8 87.6 > 10.4 86.1 > 10.7 86.3 > 10.0 85.3 > 10.7 85.3 > 11.7 82.3 > 11.6 70.2 > 9.3 69.3 8.3 31 86.2 > 10.2 85.8 > 11.2 81.4 > 12.2 68.7 > 12.6 63.5 9.0 32 85.1 6.8 82.6 7.8 73.7 > 10.3 70.2 > 9.5 67.1 8.1 65.7 5.3 33 84.9 > 7.8 87.2 > 9.1 87.2 > 9.4 88.1 > 10.3 88.0 > 9.7 87.2 > 9.2 86.6 > 10.4 83.6 > 10.1 79.7 > 10.4 76.5 > 8.3 75.4 7.1 34 80.3 > 10.0 55.9 4.7 35 75.9 > 9.0 70.1 > 11.9 70.9 > 9.8 71.3 > 8.7 70.6 > 7.9 70.8 8.8 64.4 6.1 36 50.5 9.5 Oil 8.6 37 Oil (decomp.) 7.3 [1] Reaction conditions: 2.5 – 4.0 g buffered ionic liquid; composition [BiPh3] / [cation]+[Al2Cl6X]− = 0.30 (X = Cl, Br); chloride salts if not otherwise mentioned; catalyst precursor A; [cat] = 10–5 mol / gionic liquid; T = 25°C; stirring rate = 1200 min–1; t = 60 min; 300 ml glass autoclave; 40 – 60 ml liquid propene; products decanted after each run; > = complete conversion; Oil = only higher oligomers. [2] Productivity given in gproduct / gionic liquid x h. 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 96 Hydrochlorides of Primary and Secondary Amines Table 2: Cation screening for nickel catalyzed propene dimerization reactions in BiPh3 buffered chloroaluminate melts (sorted in descending order of selectivities in run 1).[1] No. Cation Run 1 C6 [%] Prod.[2] Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8 Run 9 38 90.4 > 9.1 88.5 > 8.2 88.5 9.9 70.4 > 11.5 66.5 > 9.9 63.4 6.7 39 83.0 3.1 84.1 4.1 78.5 5.2 70.0 8.5 69.9 6.5 65.0 5.0 40 79.0 6.0 79.3 5.9 74.1 > 9.5 65.2 > 9.7 67.4 6.8 71.5 6.5 67.5 5.4 67.3 3.3 41 76.1 > 8.8 72.6 > 10.8 70.2 > 10.0 70.1 > 9.3 62.0 8.6 42 73.3 > 7.9 71.5 > 7.8 76.6 > 7.9 69.5 > 7.7 68.1 > 8.1 60.7 > 8.4 65.6 > 8.1 66.5 > 7.1 69.2 5.8 43 73.0 > 9.6 65.0 > 10.6 61.3 > 9.2 44 67.7 > 10.1 62.3 > 8.4 60.4 8.3 57.0 5.9 Oil 5.4 45 66.1 > 9.8 68.1 > 9.9 72.1 7.6 69.5 6.3 69.3 4.6 49.7 4.5 Oil 3.9 46 64.3 > 13.8 63.6 > 12.1 Oil > 9.9 47 60.9 > 9.8 48 59.7 > 12.8 68.0 > 13.1 67.2 > 12.1 72.6 > 13.5 73.5 11.1 73.4 5.9 62.1 4.7 49 Oil 6.8 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 97 No. Cation Run 1 C6 [%] Prod.[2] Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8 Run 9 50 Oil 6.5 51 Buffered IL solid [1] Reaction conditions: 2.5 – 4.0 g buffered ionic liquid; composition [BiPh3] / [cation]+[Al2Cl7]− = 0.30; catalyst precursor A; [cat] = 10–5 mol / gionic liquid; T = 25°C; stirring rate = 1200 min–1; t = 60 min; 300 ml glass autoclave; 40 – 60 ml liquid propene; products decanted after each run; > = complete conversion; Oil = only higher oligomers. [2] Productivity given in gproduct / gionic liquid x h. Hydrochlorides of Tertiary Amines Table 3: Cation screening for nickel catalyzed propene dimerization reactions in BiPh3 buffered chloroaluminate melts (sorted in descending order of selectivities in run 1).[1] No. Cation Run 1 C6 [%] Prod.[2] Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8 Run 9 Run 10 Run 11 52 93.1 7.4 93.6 8.3 94.1 7.5 92.0 8.1 78.0 > 10.5 69.6 8.8 56.7 7.4 53 93.1 7.8 93.3 8.2 92.9 9.5 91.5 8.8 87.4 11.0 73.4 8.6 67.4 6.5 54 93.1 8.7 91.1 7.3 91.1 8.3 80.4 > 10.5 57.8 9.8 Oil 7.3 55 92.8 9.3 93.2 8.2 93.0 7.6 93.1 7.5 90.6 8.5 71.6 10.0 63.6 7.0 Oil 6.7 56 92.7 10.1 92.4 9.4 90.0 9.2 87.6 > 10.6 65.9 10.6 61.9 8.7 Oil 8.0 57 92.5 10.1 92.8 9.6 91.8 8.5 91.5 9.6 89.0 > 9.2 83.0 > 12.2 73.1 10.8 69.2 8.1 67.4 6.6 58 92.4 9.1 91.2 9.0 89.0 9.1 84.0 > 12.8 71.4 10.5 62.6 7.4 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 98 No. Cation Run 1 C6 [%] Prod.[2] Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8 Run 9 Run 10 Run 11 59 92.0 7.9 91.3 7.7 89.5 8.6 86.1 > 9.1 75.2 7.1 66.2 4.5 60 91.7 9.0 90.3 8.6 90.0 9.0 86.0 9.1 74.0 9.2 66.7 10.0 65.7 6.3 Oil 6.1 61 91.6 7.6 93.5 8.2 93.3 7.5 93.1 7.6 93.1 8.1 91.8 8.9 91.1 > 11.2 81.8 > 10.5 70.2 8.8 Oil 6.7 62 91.0 > 11.8 91.7 > 11.4 90.4 > 10.5 89.8 > 12.0 88.0 > 14.0 77.8 > 13.0 74.6 > 12.3 72.0 > 11.3 70.9 > 10.8 66.6 6.8 63 90.8 > 9.8 91.5 > 10.8 90.3 > 10.0 89.8 > 11.4 88.0 > 11.1 79.8 > 13.7 74.0 9.3 68.3 5.2 64 90.7 > 10.1 90.0 > 10.0 90.4 > 10.2 90.0 > 10.1 87.7 > 10.5 85.7 > 11.8 72.2 > 10.3 64.4 > 7.4 65 90.0 > 10.6 89.7 > 9.4 90.0 > 11.5 89.2 > 10.9 89.2 > 12.5 86.2 > 11.6 73.9 > 11.0 68.3 > 9.6 67.5 8.0 64.3 7.0 66 90.0 > 8.2 90.0 > 9.5 87.5 > 8.7 81.3 > 9.6 71.6 > 9.3 65.3 8.1 65.2 5.8 58.4 4.6 67 90.0 > 10.9 90.2 > 10.3 89.6 > 10.2 88.1 > 11.2 85.6 > 12.4 76.8 > 10.6 73.7 > 10.6 71.3 8.0 66.5 6.3 68 90.0 > 8.8 90.6 > 8.6 77.5 > 11.3 60.4 > 8.3 69 89.9 > 8.4 89.6 > 9.3 90.2 > 9.0 88.8 > 8.2 88.8 > 8.3 88.3 > 10.1 85.1 > 9.1 79.7 > 9.2 67.0 7.3 70 89.6 > 10.8 90.0 > 10.6 89.7 > 9.4 89.2 > 10.2 88.5 > 11.5 86.0 > 9.2 76.1 > 9.5 70.8 8.0 68.3 5.8 71 89.6 7.6 86.3 > 8.7 82.6 > 9.7 69.4 > 9.4 Oil > 8.0 6 Nickel Catalyzed Propene Dimerization Reactions in BiPh3 Buffered Chloroaluminate Ionic Liquids: High Performance with Unconventional Cations 99 No. Cation Run 1 C6 [%] Prod.[2] Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8 Run 9 Run 10 Run 11 72 89.6 > 13.2 73.1 > 13.4 68.2 11.8 66.3 9.2 73 89.5 > 10.7 89.9 > 12.7 89.0 > 11.6 84.1 > 10.6 75.8 > 11.6 72.7 8.9 67.4 4.5 74 89.5 > 10.3 89.0 > 11.3 90.0 > 10.6 88.3 > 10.9 80.9 > 12.3 72.0 > 9.3 70.4 8.6 65.6 7.8 60.8 5.8 75 89.4 > 8.0 89.7 > 8.8 87.3 > 7.9 86.8 > 11.2 84.7 > 9.6 78.8 > 8.9 70.4 > 8.1 68.3 8.6 67.4 8.3 67.5 6.9 76 89.4 > 10.6 90.0 > 11.1 89.8 > 10.9 88.3 > 10.6 86.3 > 12.7 77.1 > 11.5 70.7 8.6 64.3 4.1 77 89.1 > 8.7 91.7 > 8.6 91.9 > 8.6 92.5 > 8.6 92.0 > 8.5 90.2 > 8.5 91.6 > 8.3 91.1 > 10.0 89.2 > 10.2 77.4 8.6 73.3 6.1 78 89.0 > 9.4 89.9 > 9.6 88.9 > 11.2 84.8 > 12.1 76.4 > 11.7 74.7 > 11.7 71.7 10.1 66.6 7.3 79 88.7 > 9.9 88.2 > 9.7 86.7 > 10.9 86.1 > 12.2 80.9 > 12.0 74.6 > 10.5 69.1 7.4 Oil 2.5 80 88.3 > 8.1 87.0 > 10.1 85.5 > 10.5 81.5 > 10.5 71.1 > 9.7 67.8 7.1 67.2 6.0 81 88.2 > 10.7 88.6 > 10.6 88.5 > 11.2 87.5 > 11.0 86.5 > 11.9 84.0 > 11.9 79.2 > 13.1 73.4 > 10.2 73.0 > 10.0 73.0 8.8 70.9 7.3 82 88.0 > 8.6 89.7 > 9.5 89.4 > 9.7 88.5 > 10.2 87.8 > 11.5 83.6 > 10.3 76.6 > 10.8 70.5 > 9.2 83 87.9 > 12.5 88.3 > 12.2 87.5 > 12.3 87.4 > 12.3 83.8 > 12.8 76.0 > 12.9 66.8 8.3 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 106 Nickel complexes are not only used for olefin oligomerization and dimerization reactions, they were also found to be efficient for olefin polymerization reactions. In 1995, Brookhart et al. first described ethene polymerization reactions with highly active nickel complexes based on novel bulky diimine ligands activated with methylaluminoxane (MAO).[12] Before, polymerization only occurred occasionally with such Ziegler type nickel oligomerization systems. However, predictions how catalysts and ligands have to be designed to produce polymers selectively were difficult.[3b] In the last ten years, SHOP type catalysts were optimized to polymerize ethene, too.[7, 13] However, their activity cannot compete with MAO activated nickel diimine systems. Due to the intensive research launched by a Brookhart report, many nickel diimine based olefin polymerization catalysts are known today.[14] While the SHOP catalyst consists of a single component, which usually displays low activity, most of the nickel based dimerization or polymerization systems rely on aluminum based activators. Common cocatalysts are MAO and its derivatives, ethylaluminum dichloride, diethylaluminum chloride, triethylaluminum, trimethylaluminum, or mixtures thereof.[1a, 12, 14-15] Unfortunately, all aluminum cocatalysts used so far bear an alkyl group in order to guarantee their solubility in nonpolar organic solvents as well as to reduce their Lewis acidity compared to aluminum trichloride.[16] The latter catalyzes side reactions like isomerization, cracking, disproportionation of alkanes as well as alkylation of aromatics with alkenes and cationic oligomerization of olefins.[17] Furthermore, the proposed mechanism based on the insertion of monomers into Ni–H or Ni–C bonds of nickel(II) complexes requires an alkylating agent. Alkylaluminum compounds are pyrophoric and highly sensitive to water, oxygen, or polar impurities. Thus, they are mostly the cost determining factor for commercial applications. Feed streams have to be purified thoroughly prior to the reactions, and the high reactivity requires a cautious handling. Finally, the cocatalysts are destroyed upon product separation of the homogeneous reactions.[8c] Also, alkylaluminum cocatalysts contribute to the deactivation of active Ni(II) catalysts by reducing Ni(II) to its zero oxidation state.[18] Therefore, the development of insensitive AlCl3 based cocatalyst systems would be highly desirable in terms of process economics. 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 107 Figure 3: Brookhart type nickel diimine complex used for ethene polymerization reactions with buffered AlCl3 activators. Ionic liquids synthesized from AlCl3 are an elegant attempt to overcome the recycling and solubility restrictions of systems based solely on AlCl3.[19] In combination with certain organic cations, AlCl3 forms liquids at ambient temperature over a wide range of compositions, which are immiscible with hydrocarbon phases. The Lewis acidities of such chloroaluminate ionic liquids can be adjusted from basic over neutral to acidic simply by varying the AlCl3 / halide salt ratio.[20] However, basic or neutral liquids do not activate nickel catalysts, while acidic compositions predominantly catalyze side reactions like uncontrolled cationic oligomerization of olefins[17b, 21] as well as isomerization[22] and cracking[23] reactions of saturated alkanes. Chauvin et al. suppressed such cationic olefin oligomerization reactions by the addition of EtAlCl2 to slightly acidic chloroaluminate melts. These melts were able to activate nickel complexes for selective dimerization reactions of α-olefins.[24] The IFP brought this so called DIFASOL® process to industrial application by retrofitting it to their existing homogeneous DIMERSOL® units.[9c, 10-11] However, the use of ethylaluminum dichloride restricts the use of DIFASOL® systems. They can only be operated in combination with DIMERSOL® units, which previously remove all impurities from the feed stream. Otherwise, impurities would accumulate in the ionic liquid, and recycling of the melt after its depletion is not intended.[9a, 9c] In 1998, Wasserscheid reported the first biphasic system for nickel catalyzed olefin dimerization reactions exclusively based on AlCl3.[25] He used weak organic bases like pyridine, quinoline, or pyrrole derivatives to reduce the “latent acidity”[26] of chloroaluminate ionic liquids. The interaction of the nitrogen bases with AlCl3 or the [Al2Cl7]− species present in such liquids[19a, 27] suppressed uncontrolled cationic olefin 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 108 oligomerization reactions. The buffered chloroaluminate melts were used to activate nickel complexes for selective dimerization reactions of various α-olefins.[28] Single component equivalents of the Lewis base interaction with aluminum in ionic liquids are intramolecular donor-stabilized aluminum cocatalysts. Several of such aluminum compounds stabilized by pending intramolecular nitrogen,[29] phosphorus,[30] oxygen,[31] or sulfur[32] groups have been described and used as cocatalysts. However, these intramolecular stabilized aluminum compounds often require large synthetic efforts and still contain reactive Al–C bonds. Therefore, they are no cheap alternative to common alkylaluminum cocatalysts for large commercial applications. Recently, we described the use of triphenylphosphine and triphenylbismuth as highly efficient buffers for Lewis acidic chloroaluminate ionic liquids[33] and investigated the interaction of the basic phosphine center with the Lewis acidic aluminum species by 27Al NMR and 31P NMR spectroscopy.[34] In the course of our search for cheap and insensitive cocatalysts for nickel catalyzed olefin dimerization and polymerization reactions, we extended the concept of buffering highly Lewis acidic aluminum chloride centers to two component homogeneous systems. 7.2 Results and Discussion NMR Investigations of Donor-Acceptor Interactions of Aluminum Chloride with NMethylpyrrole and Triphenylbismuth The most efficient buffers for Lewis acidic chloroaluminate melts were found to be Nmethylpyrrole[28a] (N-MP) and triphenylbismuth.[33] Therefore, we first investigated what happened to AlCl3 in toluene if equimolar amounts of these two buffers were added. Surprisingly, the initially insoluble anhydrous AlCl3 completely dissolved upon buffer addition. Obviously, the donor-acceptor interaction between the buffer and the Al center was sufficient to break up the dimeric Al2Cl6 structure of aluminum chloride. The organic buffer strongly increased the solubility of AlCl3 by forming a weak adduct. It was also possible to dissolve buffer adducts of aluminum chloride in CH2Cl2. However, in the case of N-MP, the system started to flocculate slowly. We assume that the strong Lewis acid AlCl3 catalyzed a ring opening reaction of the nitrogen heterocycle. No precipitation 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 109 occurred with the triphenylbismuth adduct over several hours. We examined these interactions by 27Al NMR spectroscopy (Scheme 1). Scheme 1: 27Al NMR spectra of the donor-acceptor interaction of AlCl3 with different donors at 25°C (A: 1 equivalent BiPh3, 0.05 N AlCl3 in toluene, trace C6D6; B: 10 equivalents N-MP, 0.1 N AlCl3 in toluene, trace C6D6; C: 1 equivalent N-MP, 0.1 N AlCl3 in toluene, trace C6D6; D: 1 equivalent N-MP, 0.1 N AlCl3 in CH2Cl2, trace CDCl3;). The measurements were performed in toluene with a few drops of C6D6 and in CH2Cl2 with additional CDCl3. The aluminum containing probehead of the NMR spectrometer appeared around δ = 80 ppm as broad signal. For BiPh3 adducts, an AlCl3 concentration of 0.05 N was chosen. The solubilities of N-MP adducts were higher, and thus, a 0.1 N system was measured. The compounds were mixed and the spectra recorded within a few minutes due to the decomposition of AlCl3 / N-MP adducts. All 27Al NMR signals appeared around δ = 128 ppm. However, the broadness of the signals differed strongly. The broadest peak was observed for the BiPh3 / AlCl3 interaction in toluene with a line width at half maximum of 1200 Hz (A). With a ten fold excess of NMP in toluene, the line width was 480 Hz (B) while it decreased to 330 Hz with only one equivalent of N-MP (C). The sharpest signal with a line width at half maximum of 190 Hz 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 110 was obtained from an equimolar composition of AlCl3 and N-MP in methylene chloride (D). The shoulders in spectra A and B probably belong to decomposition products, which were not identified. The shoulder increased for system B over time. The broadness of the 27Al NMR signal allowed conclusions about the extent of the donor-acceptor interaction. For example, the 27Al NMR signal was broadened if ten equivalents instead of one equivalent of N-MP were added. In contrast, the interaction between N-MP and AlCl3 in methylene chloride was less intensive indicated by a relatively sharp signal. The more polar solvent methylene chloride interacted more strongly with the adduct increasing the average distance between the donor and the acceptor. Homogeneous Propene Dimerization Reactions Further, we were interested if such adducts still displayed sufficiently high Lewis acidities to activate nickel complexes. In buffered ionic liquids, only small amounts of buffer compared to the AlCl3 content were employed. For AlCl3 / buffer adducts in homogeneous solution at least equimolar amounts of buffer were necessary to dissolve aluminum chloride. In a first series of experiments we tried to activate (PCy3)2NiCl2 with 850 equivalents of the homogeneous AlCl3 adducts in order to selectively dimerize propene. Sterically demanding phosphine ligands like tricyclohexylphosphine were known to favour the formation of valuable highly branched dimers in nickel catalyzed dimerization reactions.[2a, 15e, 16] The formation of 2,3-dimethylbutenes is highly desirable due to their high research octane number (RON). They can be sold as high octane motor fuels or used to increase the RON of low octane fractions.[35] All experiments successfully produced dimers (Table 1). The triphenylbismuth adduct yielded 72% dimers in toluene (1) and 59% dimers in CH2Cl2 (2). Even when the CH2Cl2 solvent was removed after adduct formation, the solvent free system yielded 73% dimers (3). Thus, the adduct was soluble in olefins like propene, too. The best result was obtained with N-MP in toluene with 95% C6 products (4). All AlCl3 compositions displayed enormous activities up to almost 32 ton per mol nickel and hour (2). For systems with higher selectivities to C6, the activities decreased while higher activities were obtained with less selective systems. The highest content of 2,3-dimethylbutenes 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 111 was produced with BiPh3 in toluene (53%). Generally, reactions in toluene were more selective than reactions in CH2Cl2. This was in accordance with our 27Al NMR investigations. A sharper 27Al NMR signal indicated a weaker donor-acceptor interaction, and thus, a higher Lewis acidity compared to toluene based systems. The results indicated that the N-MP buffer was superior to triphenylbismuth in terms of dimer selectivity. In contrast, BiPh3 reduced the rates of ligand abstraction more efficiently. Triphenylbismuth buffered systems always yielded higher 2,3-dimethylbutene contents than similar N-MP buffered compositions. This might also be attributed to additional π-interactions of the three aromatic rings in BiPh3. The addition of aromatic additives was found to maintain the phosphine effect over extended periods of time.[24b] Certainly, systems buffered by a combination of N-MP and BiPh3 should result in high selectivities to both dimers and 2,3-dimethylbutenes. For comparison, three experiments using an ethylaluminum dichloride cocatalyst in toluene were also performed. Without buffer, only 36% dimers were formed under the applied reaction conditions (5). If equimolar amounts of BiPh3 or N-MP were added to the ethylaluminum dichloride solution, the dimer selectivities drastically increased to 75% (6) and 95% (7), respectively. As can be seen from both AlCl3 and EtAlCl2 based systems, higher dimer selectivities came along with higher degrees of branching and reduced activities. Thus, such systems provide a lot of possibilities to optimize the results simply by varying parameters like solvents, buffer excess or type and amount of ligand. Definitely, buffered aluminum based cocatalysts are a promising alternative to alkylaluminum systems used in DIMERSOL® plants. Especially the insensitive N-MP / AlCl3 adducts can easily replace EtAlCl2 or Et2AlCl cocatalysts. Upon hydrolysis, residual N-MP can remain in the product phase if it is used to blend gasoline, while only very cheap aluminum chloride is consumed. Aluminum chloride as well as triphenylbismuth and N-MP can be stored in air and unlike aluminum alkyls, they do not require an extremely cautious handling. The components can be mixed quickly prior to use. Thus, the very slow decomposition of N-MP is negligible under the applied conditions. 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 112 Table 1: Homogeneous nickel catalyzed propene dimerization reactions with buffered AlCl3 cocatalysts.[1] No. Lewis Acid Buffer Solvent C6 [%] HEX [%] MP [%] DMB [%] Activity[2] 1 AlCl3 BiPh3 Toluene 71.6 7 40 53 10.24 2 AlCl3 BiPh3 CH2Cl2 59.1 22 73 5 31.72[3] 3 AlCl3 BiPh3 – 73.0 19 57 24 2.67 4 AlCl3 N-MP Toluene 95.3 17 59 24 5.55 5 EtAlCl2 – Toluene 36.1 22 72 6 8.76 6 EtAlCl2 BiPh3 Toluene 75.2 4 41 55 7.62 7 EtAlCl2 N-MP Toluene 95.3 11 53 36 1.13 [1] Reaction conditions: 3 – 6 mg (PCy3)2NiCl2; [Al] / [Ni] = 850; [Al] / [buffer] = 1; 50 ml solvent; T = 40°C; stirring rate = 600 min–1; t = 60 min; stirred 300 ml Parr stainless steel autoclave; 200 ml liquid propene; HEX = n-hexenes; MP = 2-methylpentenes; DMB = 2,3-dimethylbutenes; N-MP = N-methylpyrrole. [2] tproduct / molNi x h. [3] Extrapolated; reaction was stopped after 30 min due to an extremely high activity. Homogeneous Ethene Polymerization Reactions Since nickel complexes were successfully activated to dimerize propene with our system, we extended the study for ethene polymerization reactions. Therefore, a typical Brookhart type nickel diimine complex was employed as catalyst precursor (Scheme 2). The complex is known to polymerize ethene upon activation with methylaluminoxane or Et2AlCl.[12, 14b, 14c] Table 2 summarizes the results of the ethene polymerization reactions with complex 1. In all experiments employing triphenylbismuth or N-methylpyrrole as buffer, polymers were produced (Table 3). While the toluene based catalysts 8 – 10 yielded polyethylenes with similar molecular weights around 140 kg/mol, the system dissolved in methylene chloride yielded polyethylene with a molecular weight of only 75 kg/mol (11). Certainly, this effect was attributed to the increased Lewis acidity in methylene chloride. Among all toluene based systems, the triphenylbismuth buffered system (8) 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 113 led to the polyethylene with the highest PDI in combination with the lowest degree of crystallinity and the lowest melting point. Remarkably, the PDI of the polyethylene obtained with the N-methylpyrrole buffered system was reduced from 2.55 (9) to 2.25 (10) by the addition of an excess of five equivalents of buffer while molecular weights, melting points, and degrees of crystallinity remained unchanged. Table 2: Homogeneous nickel catalyzed ethene polymerization reactions with buffered AlCl3 cocatalysts.[1] No. Solvent Reaction Time [min] Buffer [Al] / [Ni] PE [wt%] Oligomers [wt%] C4 [%][2] 1-C4 [%][3] Activity[4] 8 Toluene 30 BiPh3 500 38 62 83.3 67.6 1.48 9 Toluene 30 N-MP 500 7 93 87.0 53.7 2.77 10 Toluene 30 N-MP 500[5] 19 81 91.5 79.4 0.69 11 CH2Cl2 10 BiPh3 200 12 88 71.9 20.2 16.54 12 CH2Cl2 60 PPh3 500 0 100 not det. not det. 0.04 [1] Reaction conditions: 8 – 15 mg catalyst precursor 1; AlCl3 activator; [buffer] / [Al] = 1; 50 ml solvent; T = 40°C; stirring rate = 600 min–1; stirred 300 ml Parr stainless steel autoclave; 10 bar ethene; PE = polyethylene; N-MP = N-methylpyrrole. [2] Molar amount of butenes within oligomer fraction. [3] Molar amount of 1-butene within butenes. [4] tproduct / molNi x h. [5] [buffer] / [Al] = 5. However, large amounts of oligomers were formed besides the polyethylenes. The highest polymer content of 38 wt% was obtained with BiPh3 in toluene at an Al to Ni ratio of 500 (8). After the buffer was used in excess to aluminum chloride, the polymer content could be increased from 7 wt% for the equimolar composition (9) to 19 wt% with a five fold excess of N-MP in toluene (10). 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 114 Table 3: Polyethylene properties. No. Mn [kg/mol] Mw [kg/mol] PDI Tm [°C] Xc [%][1] 8 49 141 2.87 114.6 4 9 55 141 2.55 121.1 32 10 60 134 2.25 121.8 30 11 26 75 2.88 105.4 2 [1] Crystallinity calculated with ∆Hm = 293 J/g for 100% crystalline PE. To summarize, the triphenylbismuth buffer led to the formation of higher amounts of polymers compared to N-methylpyrrole. In contrast, N-methylpyrrole buffered systems yielded polyethylenes with higher degrees of crystallinity and smaller PDIs. Thus, these buffered aluminum chloride based cocatalyst systems provide a lot of possibilities to maximize the yield of polymers and to influence their properties. Surprisingly, the oligomers formed as by-products consisted mainly of dimers. The system with excess N-MP produced 81 wt% oligomers (10), of which 92% were butenes. Even the selectivity to give 1-butene among these butenes was outstanding with 79%. Obviously, higher buffering levels came along with lower isomerization rates and, thus, large amounts of α-olefinic dimers were formed. Triphenylphosphine, which was able to buffer Lewis acidic chloroaluminate melts,[34] was also tested in addition to BiPh3 and N-MP for its ability to activate the nickel complex in combination with AlCl3. Although AlCl3 dissolved in CH2Cl2 upon contact with PPh3, the resulting solution was almost inactive and produced no polymer (12). An explanation for the formation of polymers competing with highly selective olefin dimerization reactions is given in Scheme 2. The diimine ligand with its bulky isopropyl substituents is mandatory for the nickel center to polymerize ethene. However, the Lewis acidic aluminum chloride center competes with the nickel center for the Lewis basic diimine ligand. As soon as the ligand is abstracted from the metal center, the nickel complex is no longer able to polymerize ethene. Instead, the remaining nickel bromide is a very efficient dimerization catalyst upon activation.[36] The same behaviour 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 115 was observed for the content of highly branched 2,3-dimethylbutenes in the case of propene dimerization reactions. As soon as the sterically demanding tricyclohexylphosphine ligands are abstracted from the nickel center, the formation of 2,3-dimethylbutenes is stopped. Scheme 2: Abstraction of the diimine ligand of complex 1 by AlCl3. The results of the ethene polymerization reactions also confirmed our theory that BiPh3 suppressed the ligand abstraction more efficiently than N-MP. In triphenylbismuth based systems (8, 11), the polymer yields were higher than in N-MP buffered systems (9, 10). However, after ligand abstraction, the N-MP buffered systems displayed higher selectivities to give butenes. Also, the isomerization of 1-butene to yield 2-butenes was reduced in less acidic systems and with higher buffer contents. These results suggested that such nickel based ethene dimerization systems are capable of producing α-olefins with high selectivities. However, subsequent isomerization reactions disguised the real product distributions. The catalytic results indicated that the mechanism of these homogeneous aluminum chloride based systems was similar to the mechanism we proposed for nickel catalyzed olefin dimerization reactions in buffered chloroaluminate ionic liquids.[33] Upon contact with aluminum chloride, a nickel halide complex forms Ni(AlX4)2 (X = Cl, Br) species[37] with weakly coordinating tetrahaloaluminate anions. While strongly coordinating ligands, e.g. bidentate diimines or monodentate basic phosphines, remain at the active nickel center, the weakly coordinating [AlX4]– anions can easily be replaced by olefins or buffer molecules. Since no alkylating agent was present in these systems, we also assume that the catalytic cycle started with the formation of nickelacyclopentane.[33] 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 122 vessel was cooled to 0°C, and the pressure was slowly released. After reaching ambient temperature, the weight gain of the vessel was determined, and the product phase was analyzed by gas chromatography. To determine the content of n-hexenes, 2methylpentenes, and 2,3-dimethylbutenes, the product fraction was hydrogenated with palladium on activated carbon and analyzed by gas chromatography. Acknowledgements Financial support by ConocoPhillips, USA, is gratefully acknowledged. We also thank Dr. W. Kretschmer for the HT-GPC measurements. M.D. thanks the Elite Network of Bavaria (ENB) within the graduate program “Macromolecular Science”. 7.4 References [1] a) K. Ziegler, E. Holzkamp, H. Breil, H. Martin, Angew. Chem. 1955, 67, 541-547; b) K. Fischer, K. Jonas, P. Misbach, R. Stabba, G. Wilke, Angew. Chem. 1973, 85, 1001-1012. [2] a) B. Bogdanovic, Selectivity Control in Nickel-Catalyzed Olefin Oligomerization, in Advances in Organometallic Chemistry, Vol. 17, Academic Press, New York, 1979, pp. 105-140; b) G. Wilke, Angew. Chem. 1988, 100, 189-211; c) J. Skupinska, Chem. Rev. 1991, 91, 613-648. [3] a) E. F. Lutz, J. Chem. Educ. 1986, 63, 202-203; b) W. Keim, Angew. Chem., Int. Ed. 1990, 29, 235-244. [4] J. C. Mol, J. Mol. Catal. A: Chem. 2004, 213, 39-45. [5] M. J. Muldoon, Dalton Trans. 2010, 39, 337-348. [6] B. Cornils, W. A. Herrmann, I. T. Horvath, W. Leitner, S. Mecking, H. OlivierBourbigou, D. Vogt, Multiphase Homogeneous Catalysis, Wiley-VCH, 2005. [7] P. Kuhn, D. Semeril, D. Matt, M. J. Chetcuti, P. Lutz, Dalton Trans. 2007, 515528. [8] a) Y. Chauvin, J. F. Gaillard, D. V. Quang, J. W. Andrews, Chem. Ind. 1974, 375378; b) D. Commereuc, Y. Chauvin, G. Leger, J. Gaillard, Oil & Gas Science and Technology - Rev. IFP 1982, 37, 639-649; c) Y. Chauvin, A. Hennico, G. Leger, J. L. Nocca, Erdoel, Erdgas, Kohle 1990, 106, 7. [9] a) A. Forestière, H. Olivier-Bourbigou, L. Saussine, Oil & Gas Science and Technology - Rev. IFP 2009, 64, 649-667; b) G. Wilke, B. Bogdanovic, P. Hardt, P. Heimbach, W. Keim, M. Kröner, W. Oberkirch, K. Tanaka, E. Steinrücke, D. Walter, H. Zimmermann, Angew. Chem. 1966, 78, 157-172; c) H. OlivierBourbigou, F. Favre, A. Forestière, F. Hugues, Ionic Liquids and Catalysis: the 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 123 IFP Biphasic Difasol Process, in Handbook of Green Chemistry, Wiley-VCH, 2010, pp. 101-126; d) D. Roy, R. B. Sunoj, Org. Biomol. Chem. 2010, 8, 10401051. [10] B. Gilbert, H. Olivier-Bourbigou, F. Favre, Oil & Gas Science and Technology - Rev. IFP 2007, 62, 745-759. [11] F. Favre, A. Forestiere, F. Hugues, H. Olivier-Bourbigou, J. A. Chodorge, Oil Gas Eur. Mag. 2005, 31, 83 - 91. [12] L. K. Johnson, C. M. Killian, M. Brookhart, J. Am. Chem. Soc. 1995, 117, 64146415. [13] S. Mecking, Angew. Chem. 2001, 113, 550-557. [14] a) C. M. Killian, L. K. Johnson, M. Brookhart, Organometallics 1997, 16, 20052007; b) D. Pappalardo, M. Mazzeo, C. Pellecchia, Macromol. Rapid Commun. 1997, 18, 1017-1023; c) F. AlObaidi, Z. Ye, S. Zhu, Polymer 2004, 45, 68236829; d) T. M. J. Anselment, S. I. Vagin, B. Rieger, Dalton Trans. 2008, 45374548. [15] a) R. J. Maldanis, J. S. Wood, A. Chandrasekaran, M. D. Rausch, J. C. W. Chien, J. Organomet. Chem. 2002, 645, 158-167; b) S. A. Svejda, M. Brookhart, Organometallics 1998, 18, 65-74; c) C. G. de Souza, R. F. de Souza, K. Bernardo-Gusmão, Appl. Catal., A 2007, 325, 87-90; d) F. Benvenuti, C. Carlini, M. Marchionna, R. Patrini, A. M. R. Galletti, G. Sbrana, Appl. Catal., A 2000, 199, 123-132; e) C. Carlini, M. Marchionna, A. M. R. Galletti, G. Sbrana, J. Mol. Catal. A: Chem. 2001, 169, 19-25; f) S. Wu, S. Lu, J. Mol. Catal. A: Chem. 2003, 197, 51-59; g) S. Wu, S. Lu, J. Mol. Catal. A: Chem. 2003, 198, 29-38. [16] G. G. Eberhardt, W. P. Griffin, J. Catal. 1970, 16, 245-253. [17] a) A. K. Roebuck, B. L. Evering, Product R&D 1970, 9, 76-82; b) M. Goledzinowski, V. I. Birss, J. Galuszka, Ind. Eng. Chem. Res. 1993, 32, 17951797; c) H. R. Alul, Ind. Eng. Chem. Prod. Res. Dev. 1968, 7, 7-11; d) H. J. Hepp, L. E. Drehman, US 2968684, 1961; e) V. N. Ipatieff, A. V. Grosse, Ind. Eng. Chem. 1936, 28, 461-464; f) H. J. Hepp, E. O. Box, D. A. Uhrick, Erdoel, Erdgas, Kohle 1968, 21, 102; g) V. I. Komarewsky, S. C. Ulick, J. Am. Chem. Soc. 1947, 69, 492-495. [18] F. Peruch, H. Cramail, A. Deffieux, Macromolecules 1999, 32, 7977-7983. [19] a) J. S. Wilkes, J. A. Levisky, R. A. Wilson, C. L. Hussey, Inorg. Chem. 1982, 21, 1263-1264; b) S. Geetha, D. C. Trivedi, Bull. Electrochem. 2003, 19, 37-48; c) F. Fairbrother, N. Scott, H. Prophet, J. Chem. Soc. 1956, 1164-1167. [20] A. A. Fannin, L. A. King, J. A. Levisky, J. S. Wilkes, J. Phys. Chem. 1984, 88, 2609-2614. [21] O. Stenzel, R. Brüll, U. M. Wahner, R. D. Sanderson, H. G. Raubenheimer, J. Mol. Catal. A: Chem. 2003, 192, 217-222. [22] a) G. Roberts, C. M. Lok, C. J. Adams, K. R. Seddon, M. E. Earle, J. Hamill, WO 1998007680, 1998; b) K. Herbst, J. Houzvicka, B. T. Jespersen, J. Zavilla, US 20030181780, 2003; c) T. V. Vasina, L. M. Kustov, V. A. E. Ksenofontov, Y. E. Zubarev, J. Houzvicka, J. Zavilla, US 20030109767, 2003. 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 124 [23] a) R. W. J. Westerhout, J. Waanders, J. A. M. Kuipers, W. P. M. van Swaaij, Ind. Eng. Chem. Res. 1998, 37, 2293-2300; b) C. J. Adams, M. J. Earle, K. R. Seddon, Green Chem. 2000, 2, 21-24; c) M. J. Earle, K. R. Seddon, Pure Appl. Chem. 2000, 72, 1391-1398; d) L. Xiao, K. E. Johnson, R. G. Treble, J. Mol. Catal. A: Chem. 2004, 214, 121-127. [24] a) Y. Chauvin, B. Gilbert, I. Guibard, J. Chem. Soc.-Chem. Commun. 1990, 1715-1716; b) Y. Chauvin, S. Einloft, H. Olivier, Ind. Eng. Chem. Res. 1995, 34, 1149-1155; c) S. Einloft, F. K. Dietrich, R. F. De Souza, J. Dupont, Polyhedron 1996, 15, 3257-3259; d) Y. Chauvin, H. Olivier, C. N. Wyrvalski, L. C. Simon, R. F. de Souza, J. Catal. 1997, 165, 275-278. [25] P. Wasserscheid, WO 9847616, 1998. [26] a) I. C. Quarmby, R. A. Mantz, L. M. Goldenberg, R. A. Osteryoung, Anal. Chem. 1994, 66, 3558-3561; b) I. C. Quarmby, R. A. Osteryoung, J. Am. Chem. Soc. 1994, 116, 2649-2650; c) D. S. McGuinness, W. Mueller, P. Wasserscheid, K. J. Cavell, B. W. Skelton, A. H. White, U. Englert, Organometallics 2001, 21, 175181. [27] C. L. Hussey, T. B. Scheffler, J. S. Wilkes, J. A. A. Fannin, J. Electrochem. Soc. 1986, 133, 1389-1391. [28] a) B. Ellis, W. Keim, P. Wasserscheid, Chem. Commun. 1999, 337-338; b) P. Wasserscheid, M. Eichmann, Catal. Today 2001, 66, 309-316; c) M. Eichmann, W. Keim, M. Haumann, B. U. Melcher, P. Wasserscheid, J. Mol. Catal. A: Chem. 2009, 314, 42-48. [29] a) G. Bähr, G. E. Müller, Chem. Ber. 1955, 88, 251-264; b) H. Schumann, B. C. Wassermann, S. Schutte, B. Heymer, S. Nickel, T. D. Seuß, S. Wernik, J. Demtschuk, F. Girgsdies, R. Weimann, Z. Anorg. Allg. Chem. 2000, 626, 20812095; c) A. Eisenhardt, B. Heuer, W. Kaminsky, K. Köhler, H. Schumann, Adv. Synth. Catal. 2003, 345, 1299-1304; d) E. Poetsch, L. Pohl, K. Weiss, E. M. Auth, T. Metzner, H. Schumann, C. Wassermann, M. Frick, B. Heymer, S. Schutte, U. Dümichen, E. Hecht, US 6596890, 2003; e) H. Schumann, S. Dechert, F. Girgsdies, B. Heymer, M. Hummert, J. Y. Hyeon, J. Kaufmann, S. Schutte, S. Wernik, B. C. Wassermann, Z. Anorg. Allg. Chem. 2006, 632, 251263. [30] H. Schumann, F. Girgsdies, B. Heymer, J. Kaufmann, C. Marschall, W. Wassermann, Z. Anorg. Allg. Chem. 2007, 633, 2268-2273. [31] H. Schumann, S. Dechert, S. Schutte, J.-Y. Hyeon, M. Hummert, B. C. Wassermann, W. Kaminsky, A. Eisenhardt, K. Köhler, J. Eichhorn, Organometallics 2003, 22, 1391-1401. [32] H. Schumann, B. C. Wassermann, B. Heymer, M. Keitsch, F. Girgsdies, Z. Anorg. Allg. Chem. 2001, 627, 1828-1835. [33] M. Dötterl, H. G. Alt, Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids, to be submitted. [34] M. Dötterl, P. Thoma, H. G. Alt, Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids, to be submitted. 7 Buffered Aluminum Chloride as Highly Efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions 125 [35] H. E. Swift, C.-Y. Wu, US 3622649, 1971. [36] M. Dötterl, H. G. Alt, Nickel catalyzed Propene Dimerization Reactions in BiPh3 buffered Chloroaluminate Ionic Liquids: High Performance with unconventional Cations, to be submitted. [37] J. Brynestad, H. L. Yakel, G. P. Smith, Inorg. Chem. 1970, 9, 686-686. [38] a) R. T. Carlin, J. S. Wilkes, J. Mol. Catal. 1990, 63, 125-129; b) W. OchedzanSiodlak, B. Sacher-Majewska, Eur. Polym. J. 2007, 43, 3688-3694; c) W. Ochedzan-Siodlak, K. Dziubek, D. Siodlak, Eur. Polym. J. 2008, 44, 3608-3614. [39] a) M. Pinheiro, R. Mauler, R. de Souza, Macromol. Rapid Commun. 2001, 22, 425-428; b) V. Lecocq, H. Olivier-Bourbigou, Oil & Gas Science and Technology - Rev. IFP 2007, 62, 761-773. [40] Z. Grubisic, P. Rempp, H. Benoit, Journal of Polymer Science Part B: Polymer Letters 1967, 5, 753-759. 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 126 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes Matthias Dötterl[a] and Helmut G. Alt*[a] [a] Lehrstuhl für Anorganische Chemie II, Universitätsstraße 30, NW I, 95440 Bayreuth, Germany. E-mail: [email protected] Keywords: heterogeneous catalysis, nickel, dimerization, polymerization, ionic liquids Manuscript to be submitted Abstract: We developed a new buffered silica based Lewis acid cocatalyst for olefin dimerization and polymerization reactions with nickel complexes. The Lewis acidities of EtAlCl2 and Et2AlCl modified silicas were adjusted by the addition of BiPh3 or Nmethylpyrrole buffers. The buffered heterogeneous cocatalysts were successfully employed to activate nickel complexes for selective propene dimerization reactions to give 91% dimers. Buffer addition also reduced the rate of ligand abstraction from the catalytically active nickel center. In combination with a nickel diimine complex, ethene was polymerized to give 69 wt% polyethylene and 31 wt% oligomers. The polyethylene displayed a molecular weight of 154 kg/mol and a PDI of 3.26. The oligomers consisted of 93% butenes, 46% thereof being 1-butene. Furthermore, triphenylbismuth buffered 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 127 Lewis acidic chloroaluminate ionic liquids were coated on surface modified silica to yield a supported ionic liquid phase (SILP) system. Dimer selectivities of up to 94% were obtained with the SILP system in nickel catalyzed propene dimerization reactions. The newly developed silica based cocatalysts provide promising new olefin dimerization systems suitable for fixed bed applications. 8.1 Introduction Most of the major metal complex catalyzed olefin conversion processes including hydroformylation, metathesis, dimerization, and polymerization reactions were discovered in homogeneous solution. However, all of these processes were converted to biphasic[1] or heterogeneous[2] systems in order to improve the economics for their application in large scale industrial processes. Olefin polymerization reactions, which yield solid products, have to be performed in slurry, bulk-monomer, or gas-phase processes for large capacities.[2d] Therefore, the catalysts are usually supported on an inorganic carrier to prevent reactor fouling and to guarantee uniform polymer particle sizes.[2c, 2d] For example, the old heterogeneous Union Carbide (chromocene/silica) and Phillips (CrO3/silica) catalysts are still responsible for about one third of the global HDPE production.[3] For olefin dimerization reactions, which yield gaseous or liquid products, heterogeneous catalyst systems are advantageous for fixed bed reactors because the do not require expensive technologies. The olefin stream passes the solid catalyst, which is immobilized in the reactor. There are several different approaches to support metal complexes, which have to be activated by aluminum based cocatalysts for olefin dimerization or polymerization reactions. The most common method is the impregnation of an inorganic support material with alkylaluminum compounds like methylaluminoxane (MAO) or trimethylaluminum (TMA) and a metal complex.[2c, 2d, 4] However, the high prices of MAO and TMA restrict their use for the production of valuable high performance polymers. The well known titanium chloride based Ziegler-Natta catalysts supported on MgCl2 also belong to this class.[3a, 5] 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 128 A second possibility is to connect metal complexes covalently to a heterogeneous support.[6] However, the attachment to support materials requires changes in the ligand structure and increased synthetic efforts. Lewis acidic inorganic oxide materials provide an alternative to the above mentioned systems. McDaniel et al. developed a novel cocatalyst system for metallocene complexes based on alkylaluminum compounds and solid acids.[7] The surface of inorganic oxides was modified by high temperature treatment with chlorinated or fluorinated substrates. In combination with triethylaluminum, these solid Lewis acids are able to activate metallocene complexes for ethene polymerization reactions. This industrial research of ChevronPhillips Chemical[7] showed that it is desirable to develop heterogeneous cocatalysts, which can easily be synthesized, stored and used to activate common metal complexes feasible for olefin dimerization and polymerization reactions. A lot of research has already been done on the surface modification of inorganic oxide materials. In particular, hydroxyl groups located at silica surfaces[8] can be modified with reactive aluminum compounds to yield strong heterogeneous Lewis acids. The strongest Lewis acid sites are obtained by treating silica with either AlCl3[9] to release HCl or with EtAlCl2[10] to release ethane. In the presence of both Al–Cl and Al–C bonds the latter preferentially react.[11] These materials catalyze cracking, disproportionation, or isomerization of alkanes as well as cationic oligomerization reactions of olefins.[10a] The acid strength of such grafted aluminum species usually increases with more halogen atoms being attached to the aluminum center.[12] Since SiO–AlCl2 surface groups initiate the above mentioned cationic olefin oligomerization reactions, they cannot be applied as cocatalysts for metal complex catalyzed olefin dimerization or polymerization reactions. First, the Lewis acidities of such strong solid acids have to be reduced to a level, at which said uncontrolled cationic olefin oligomerization reactions do not take place. We identified new strategies to overcome this problem. In a recent paper, we described novel cocatalysts for nickel catalyzed olefin dimerization and polymerization reactions based on a combination of aluminum chloride and weak Lewis bases such as Nmethylpyrrole or triphenylbismuth.[13] The weak Lewis bases buffered the high Lewis acidity of aluminum chloride and kept it in solution by forming a weak adduct. Such organic buffers were previously used by Wasserscheid et al.[14] and our group[15] to 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 129 reduce the “latent acidity”[16] of Lewis acidic chloroaluminate ionic liquids. The buffered chloroaluminate melts were employed to activate nickel complexes for selective dimerization reactions of α-olefins. Thus, we expected that the addition of these buffering substances would control the Lewis acidities of EtAlCl2 modified solid acids in a similar manner. Nickel based systems are particularly interesting because they can be activated with cheap Et2AlCl or EtAlCl2 cocatalysts. Depending on the ligand structure, they were found to catalyze both selective olefin dimerization[17] or polymerization[18] reactions. The second approach relied on silica treated with excess Et2AlCl. Since the Lewis acidity of the surface aluminum species decreases with a decreasing number of halogens,[12] the substitution of SiO–AlCl2 groups with SiO–AlEtCl should come along with reduced Lewis acidities. Because Et2AlCl can react with both ethyl groups to release two molecules ethane per molecule Et2AlCl,[11] the use of excess Et2AlCl should favour the formation of SiO–AlEtCl species. Upon buffering with N-methylpyrrole and BiPh3, these SiO–AlEtCl species were expected to activate nickel complexes for selective olefin dimerization and polymerization reactions. Furthermore, EtAlCl2 modified silica should provide a suitable support material for buffered chloroaluminate ionic liquids to obtain pseudo heterogeneous supported ionic liquid phase[19] (SILP) dimerization systems. 8.2 Results and Discussion Buffered Silica Supported EtAlCl2 as Cocatalyst for Heterogeneous Nickel Catalyzed Propene Dimerization Reactions For all experiments, Davicat® SI1102 silica with a particle size of 0.4 – 0.8 mm from W.R. Grace & Co. was used. It has a surface area of 311 m2/g, a pore volume of 1.12 cm3/g and a pore diameter of 144 Å. The silica was oven dried at 350°C prior to all manipulations. In a preliminary experiment, treatment of dehydrated silica with excess EtAlCl2 resulted in a weight gain of about 15.7%. Under the approximation that each molecule of EtAlCl2 reacted with only one surface OH group to release ethane,[11] the density of reactive surface OH groups was calculated to be 1.6 mmol/g. 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 130 Figure 1: Catalyst precursors used for propene dimerization (A, B) and ethene polymerization reactions (C). Scheme 1 shows the surface modification of dehydrated silica I with EtAlCl2 and Et2AlCl to give silica II and III, respectively. First, the highly acidic EtAlCl2 modified silica II was buffered with N-methylpyrrole or BiPh3 (Table 1) and applied as cocatalyst for nickel catalyzed propene dimerization reactions with complex A (Figure 1). Both systems were tested in batch experiments with liquid propene. In order not to destroy the silica structure, stirring was avoided. Triphenylbismuth was dispersed from a methylene chloride solution, which was removed afterwards. Its high molecular weight required reduced loadings. In the case of N-methylpyrrole, a cosolvent had to be applied to obtain a homogeneous distribution of the volatile heterocyclic buffer. In both experiments, dimers were produced with high activities. Also, the dimer fraction consisted of remarkable amounts of 60% (1) and 47% (2) of valuable 2,3-dimethylbutenes. Without the buffer, viscous oils consisting of higher oligomers from cationic propene oligomerization reactions were obtained. However, the selectivities to give dimers were poor. Also, the N-methylpyrrole buffered system 2 was inactive in a subsequent experiment. Consequently, the Lewis acidity was still too high. Higher amounts of buffer could further reduce the acidity. However, huge amounts of the heavy BiPh3 would be necessary, while N-methylpyrrole systems deactivated quickly. 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 131 Scheme 1: Surface modification of dehydrated silica (I) with ethylaluminum dichloride (II) and diethylaluminum chloride (III) and buffering of the Lewis acidic surface aluminum species. 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 138 the dimer yield increased to 90% (14), which was similar to the unsupported system. With a reduced loading of 80 wt%, mainly higher oligomers were obtained (17). For this system loaded with 80 wt% buffered [BMIM]+[Al2Cl7]− ionic liquid, a buffer content of one equivalent was necessary to raise the dimer selectivity to 90% again (18). Triphenylbismuth equalled about 37 wt% of the total weight in [BMIM]+[Al2Cl7]− / BiPh3 = 0.60 systems. With one equivalent of buffer, even 50 wt% of the catalytically active melt consisted of BiPh3. Therefore, small changes in terms of loading or composition led to drastic changes in performance. For ionic liquids derived from cations with lower molecular weights, lower loadings would certainly be sufficient. A system with a less acidic composition of [AlCl3] / [BMIMCl] = 1.50 and 0.60 equivalents of BiPh3 yielded 94% dimers in the first experiment (19a). Then, the products were decanted, and the batch experiment was repeated. In four subsequent catalytic runs, the system maintained a dimer selectivity of around 90% (19b – e) before it dropped significantly. All systems displayed low productivities compared to their biphasic analogues.[15c] However, no stirring was applied in order not to destroy the silica structure. With an improved mixing, the productivities could certainly be increased. The reference complex B had no effect on the branching of the hexene dimers,[15a] all C6 fractions consisted of (± 2%) 25% n-hexenes, 69% 2-methylpentenes, and 6% 2,3-dimethylbutenes. This correlated to the typical product distribution obtained from ligand free nickel salts.[14d] Triphenylbismuth buffered chloroaluminate ionic liquids supported on surface modified silica provide novel heterogeneous cocatalyst systems for nickel catalyzed olefin dimerization reactions. If ionic liquids based on amine hydrochlorides are used, both the buffer and the amine can be recovered by a simple acid base extraction.[15a] Especially the application for gas phase ethene dimerization reactions is very promising since no leaching of either the ionic liquid or the buffer can occur. Furthermore, alkylaluminum free ionic liquid compositions are less sensitive to oxygen or polar impurities. Conclusion We developed a new class of buffered solid acid cocatalysts for nickel catalyzed olefin dimerization and polymerization reactions. The Lewis acidity of EtAlCl2 modified silica was reduced by the addition of BiPh3 or N-methylpyrrole buffers. Uncontrolled cationic 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 139 olefin oligomerization reactions, which usually occur in the presence of SiO–AlCl2 species, were suppressed by the donor-acceptor interactions. Silica modification with Et2AlCl instead of EtAlCl2 yielded solid cocatalysts with reduced Lewis acidities. The cocatalysts were successfully employed to activate a nickel complex for selective propene dimerization reactions to give 84% dimers. The addition of small amounts of N-methylpyrrole to the catalyst suspension increased the dimer selectivity up to 91%. Furthermore, the addition of either BiPh3 or N-methylpyrrole hindered the ligand abstraction from the catalytically active metal center. The same Et2AlCl modified silica with additional N-methylpyrrole buffer was used to activate a nickel diimine complex for ethene polymerization reactions to give 69 wt% polyethylene and 31 wt% oligomers. Remarkably, 93% of the oligomers were butenes with 46% thereof being 1-butene. The polyethylene was partially crystalline and displayed a molecular weight of 154 kg/mol with a PDI of 3.26. Buffered, surface modified silicas provide a promising new class of cocatalysts for heterogeneous nickel catalyzed olefin dimerization and polymerization reactions. Surface modification was accomplished in one step with cheap Et2AlCl. The heterogeneous cocatalysts were able to activate common nickel complexes and can be applied in fixed bed tubes for olefin dimerization reactions. Selectivity control could be achieved by the addition of small amounts of an N-methylpyrrole buffer to the olefin feed stream. In combination with nickel based polymerization catalysts, the same systems could be applied in slurry ethene polymerization reactions. Our results indicated that the catalytically active species were immobilized on the silica surface. Finally, BiPh3 buffered Lewis acidic chloroaluminate ionic liquids were coated on surface modified silicas to yield supported ionic liquid phase (SILP) systems for nickel catalyzed propene dimerization reactions. Dimer selectivities up to 94% were obtained with the SILP systems. In contrast to the above mentioned buffered surface modified silica cocatalysts, no external buffer had to be added. The BiPh3 buffer was immobilized in the ionic liquid. Such SILP systems are predestined for gas phase ethene dimerization reactions since no leaching can occur in gas phase reactions. 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 140 8.3 Experimental Section General Information All chemical manipulations were carried out using standard Schlenk techniques under argon atmosphere. n-Heptane and n-pentane were distilled from Na/K alloy under inert atmosphere. Methylene chloride was distilled in two steps from P4O10 and CaH2, toluene from P4O10 and Na/K alloy. The products of the dimerization experiments were characterized by gas chromatography (Agilent 6850) and GC-MS (FOCUS DSQ™ Thermo Scientific). Differential scanning calorimetry experiments were conducted at a heating rate of 10°C/min under N2 atmosphere with a Perkin-Elmer Diamond DSC, calibrated with indium. The endothermic maximum of the second heating cycle was taken as Tm. The molecular weights (Mw/Mn) of the ethylene polymers were determined by gel permeation chromatography on a Polymer Laboratories Ltd. (PL-GPC220) chromatograph at 150°C using 1,2,4-trichlorobenzene as the mobile phase. The sample was prepared by dissolving the polymer (0.1% weight/volume) in the mobile phase solvent in an external oven and was run without filtration. The molecular weight was referenced to polystyrene (Mw = 510 – 3100000 g/mol) standards using the universal calibration method.[25] The reported values are the average of at least two independent determinations. Ethene (99.9%) and propene (99.3%) were purchased from Riessner Gase, Lichtenfels, and were dried over a column packed with P4O10. EtAlCl2 (1.8 N in toluene) and Et2AlCl (0.9 N in toluene) were purchased from Acros. Triphenylphosphine and triphenylbismuth were purchased from ABCR. N-Methylpyrrole, 1-butyl-3methylimidazolium tetrachloroaluminate (BASF), AlCl3 (ReagentPlus®), and (PPh3)2NiCl2 were purchased from Sigma-Aldrich and used as received. Complexes A[15a] and C[18] were synthesized according to the literature. Preparation of the Surface Modified Silicas II and III For all experiments, Davicat® SI1102 silica with a particle size of 0.4 – 0.8 mm from W.R. Grace & Co. was used. It has a surface area of 311 m2/g, a pore volume of 1.12 cm3/g and a pore diameter of 144 Å. Prior to all manipulations, the silica was dehydrated for 4 hours at 350°C to give silica I. The dehydrated silica I was filled into a glass frit under argon counter flow, and half of its volume was covered with toluene. For 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 141 silica II, 2 mmol EtAlCl2 (1.8 N in toluene) per g, for silica III 3 mmol Et2AlCl (0.9 N in toluene) per g were added. The frit was equipped with a pressure valve, and the suspension was stirred with a spatula once in a while. The mixture was left standing over night until no more bubbles could be observed. The modified silica was washed three times with toluene and n-pentane, dried in vacuo, and stored in a Schlenk tube. Preparation of Buffered Surface Modified Silica Cocatalyst Compositions The corresponding amount of surface modified silica was weighed into a glass autoclave. For BiPh3 buffered systems (1, 4), the silica was covered with CH2Cl2. Then, the buffer was added, and the autoclave was gyrated. To the CH2Cl2 suspension the catalyst precursor A was added from a stock solution in CH2Cl2. Subsequently, the methylene chloride was removed in vacuo to yield the heterogeneous BiPh3 buffered propene dimerization system. For the unbuffered experiment (3), the same procedure was applied without the addition of BiPh3. For N-methylpyrrole buffered systems (2, 5 – 9), the silica was covered with n-heptane or toluene. The corresponding amount of N-methylpyrrole was added to the suspension. For the experiments in Table 2, a dilution series of N-methylpyrrole in n-heptane and toluene was performed. The autoclave was gyrated and left standing for a few minutes. Then, the solid catalyst precursor A was added. After it had dissolved, the system was used for heterogeneous N-methylpyrrole buffered propene dimerization reactions. Preparation of Triphenylbismuth Buffered 1-Butyl-3-methylimidazolium Chloroaluminate Ionic Liquid Supported on Silica AlCl3 was dissolved in commercial 1-butyl-3-methylimidazolium tetrachloroaluminate to obtain the desired compositions. The corresponding amounts of ionic liquid and BiPh3 were added to a glass autoclave and mixed until a homogeneous solution was obtained. Subsequently, the buffered ionic liquid was diluted with CH2Cl2 and silica was added followed by the addition of catalyst precursor B. Upon its dissolution, the solvent was removed in vacuo to yield a free flowing granular powder. 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 142 Procedure for Propene Dimerization Reactions with Heterogeneous Catalysts (Experiments 1 – 9, 11 – 19) For all propene dimerization reactions, a 300 ml glass autoclave was used. No stirring was applied in order not to destroy the silica structure and to guarantee identical conditions. The glass vessel was kept in a drying oven at 150°C for several hours before an experiment. An autoclave containing the heterogeneous catalyst composition was cooled, and vacuum was applied. Subsequently, propene (50 ml) was condensed into the glass autoclave by liquid nitrogen cooling. Then, the autoclave was placed in a metal box for safety reasons, and the temperature was regulated by an external water bath. After the experiment, the pressure was slowly released by opening a valve. After reaching ambient temperature, the weight difference of the autoclave was determined, and the product phase was analyzed by gas chromatography. To determine the content of nhexenes, 2-methylpentenes, and 2,3-dimethylbutenes, the product fraction was hydrogenated with palladium on activated carbon, and the products were analyzed by gas chromatography. In the case of experiments 3 and 19, the product phase was decanted, analyzed by gas chromatography, and the experiment was repeated. Procedure for Ethene Polymerization Reactions with an N-Methylpyrrole Buffered Surface Modified Silica Cocatalyst (Experiment 10) For the ethene polymerization reaction, a stirred 300 ml Parr stainless steel autoclave was used. The steel vessel was kept in a drying oven at 150°C for several hours before an experiment, and it was filled with argon. The autoclave was cooled to 0°C with an ice bath. Then, a solution of N-methylpyrrole (0.05 ml) in toluene (50 ml) was added to the reactor followed by the addition of surface modified silica III (12.60 g). Complex C (18.0 mg) was suspended in toluene (20 ml), and the suspension was filled into the reactor. The system was put under vacuum, and the temperature was raised to 40°C. Then, after 5 minutes of stirring, an ethene pressure of 10 bar was applied. Cooling of the exothermic reaction was achieved manually by an ice bath. After the experiment, the vessel was quickly cooled to –20°C. The pressure was released, the weight difference of the vessel was determined (23.24 g), and a cooled sample was directly analyzed by gas chromatography. The silica / polyethylene composite material was separated by filtration, washed several times with acetone and water, and its weight was determined after 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 143 drying (28.62 g silica + polyethylene). A sample of the composite material was stirred for 2 hours in 1,2,4-trichlorobenzene at 120°C to dissolve the polyethylene. The viscous solution was quickly filtered, and the polymer was precipitated by the addition of acetone. The pure polyethylene sample obtained after filtration and drying was analyzed by differential scanning calorimetry and high temperature gel permeation chromatography. Acknowledgements Financial support by ConocoPhillips, USA, is gratefully acknowledged. We also thank Dr. W. Kretschmer for the HT-GPC measurement. M.D. thanks the Elite Network of Bavaria (ENB) within the graduate program “Macromolecular Science”. 8.4 References [1] a) H. Bahrmann, H. Bach, Phosphorus Sulfur Relat. Elem. 1987, 30, 611 - 614; b) W. Keim, Angew. Chem., Int. Ed. 1990, 29, 235-244; c) A. Forestière, H. Olivier-Bourbigou, L. Saussine, Oil & Gas Science and Technology - Rev. IFP 2009, 64, 649-667. [2] a) J. C. Mol, J. Mol. Catal. A: Chem. 2004, 213, 39-45; b) H. Olivier-Bourbigou, F. Favre, A. Forestière, F. Hugues, Ionic Liquids and Catalysis: the IFP Biphasic Difasol Process, in Handbook of Green Chemistry, Wiley-VCH, 2010, pp. 101126; c) J. Chien, Top. Catal. 1999, 7, 23-36; d) G. G. Hlatky, Chem. Rev. 2000, 100, 1347-1376. [3] a) J. C. Chadwick, T. Garoff, J. R. Severn, Traditional Heterogeneous Catalysts, in Tailor-Made Polymers, Wiley-VCH, 2008, pp. 43-78; b) F. J. Karol, G. L. Karapinka, C. Wu, A. W. Dow, R. N. Johnson, W. L. Carrick, J. Polym. Sci., Part A: Polym. Chem. 1972, 10, 2621-2637. [4] a) M. Kristen, Top. Catal. 1999, 7, 89-95; b) J. R. Severn, J. C. Chadwick, V. Van Axel Castelli, Macromolecules 2004, 37, 6258-6259. [5] J. J. A. Dusseault, C. C. Hsu, Polym. Rev. 1993, 33, 103 - 145. [6] a) F. Benvenuti, C. Carlini, M. Marchionna, R. Patrini, A. M. R. Galletti, G. Sbrana, Appl. Catal., A 2000, 204, 7-18; b) C. Carlini, M. Marchionna, R. Patrini, A. M. Raspolli Galletti, G. Sbrana, Appl. Catal., A 2001, 207, 387-395; c) W. Skupinski, S. Malinowski, J. Organomet. Chem. 1975, 99, 465-474; d) W. Skupinski, S. Malinowski, J. Organomet. Chem. 1976, 117, 183-187; e) P. Preishuber-Pflugl, M. Brookhart, Macromolecules 2002, 35, 6074-6076; f) H. G. Alt, Dalton Trans. 2005, 3271-3276; g) H. G. Alt, C. Görl, Self-Immobilizing Catalysts for Olefin Polymerization, in Tailor-Made Polymers, Wiley-VCH, 2008, pp. 305-326; h) J. C. Hicks, C. W. Jones, Tethering Olefin Polymerization 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 144 Catalysts and Cocatalysts to Inorganic Oxides, in Tailor-Made Polymers, WileyVCH, 2008, pp. 239-260. [7] M. P. McDaniel, M. D. Jensen, K. Jayaratne, K. S. Collins, E. A. Benham, N. D. McDaniel, P. K. Das, J. L. Martin, Q. Yang, M. G. Thorn, A. P. Masino, Metallocene Activation by Solid Acids, in Tailor-Made Polymers, Wiley-VCH, 2008, pp. 171-210. [8] a) M. Sato, T. Kanbayashi, N. Kobayashi, Y. Shima, J. Catal. 1967, 7, 342-351; b) R. J. Peglar, F. H. Hambleton, J. A. Hockey, J. Catal. 1971, 20, 309-320; c) R. Anwander, C. Palm, O. Groeger, G. Engelhardt, Organometallics 1998, 17, 20272036; d) J. H. Clark, Acc. Chem. Res. 2002, 35, 791-797. [9] a) R. S. Drago, E. E. Getty, J. Am. Chem. Soc. 1988, 110, 3311-3312; b) R. S. Drago, E. E. Getty, US 4719190, 1988; c) E. E. Getty, R. S. Drago, Inorg. Chem. 1990, 29, 1186-1192; d) S. Sato, G. E. Maciel, J. Mol. Catal. A: Chem. 1995, 101, 153-161; e) S. S. Ray, R. R. Sarin, D. K. Tuli, M. M. Rai, S. Ghosh, A. K. Bhatnagar, S. Sivaram, T. P. Mohandas, D. H. Gholap, M. J. G. Yanjarappa, US 6096678, 2000; f) J. K. Shorrock, J. H. Clark, K. Wilson, J. Chisem, Org. Process Res. Dev. 2001, 5, 249-253; g) V. R. Choudhary, K. Mantri, J. Catal. 2002, 205, 221-225; h) M. Xu, A. Arnold, A. Buchholz, W. Wang, M. Hunger, J. Phys. Chem. B 2002, 106, 12140-12143; i) V. Sage, J. H. Clark, D. J. Macquarrie, J. Mol. Catal. A: Chem. 2003, 198, 349-358. [10] a) S. C. Ho, M. M. Wu, US 5326920, 1994; b) S. C. Ho, M. M. Wu, US 5294578, 1994; c) S. C. Ho, M. M. Wu, US 5451704, 1995. [11] J. P. Blitz, R. E. Diebel, C. A. Deakyne, J. M. Christensen, V. M. Gun'ko, J. Phys. Chem. B 2005, 109, 5667-5677. [12] F. J.-y. Chen, C. Le Deore, R. Spitz, A. Guyot, US 5648580, 1997. [13] M. Dötterl, H. G. Alt, Buffered Aluminum Chloride as highly efficient Cocatalyst for Olefin Dimerization and Polymerization Reactions, to be submitted. [14] a) P. Wasserscheid, WO 9847616, 1998; b) B. Ellis, W. Keim, P. Wasserscheid, Chem. Commun. 1999, 337-338; c) P. Wasserscheid, M. Eichmann, Catal. Today 2001, 66, 309-316; d) M. Eichmann, W. Keim, M. Haumann, B. U. Melcher, P. Wasserscheid, J. Mol. Catal. A: Chem. 2009, 314, 42-48; e) D. S. McGuinness, W. Mueller, P. Wasserscheid, K. J. Cavell, B. W. Skelton, A. H. White, U. Englert, Organometallics 2001, 21, 175-181. [15] a) M. Dötterl, H. G. Alt, Heavy Metal with Heavy Impact: Olefin Dimerization Reactions in Triphenylbismuth Buffered Chloroaluminate Ionic Liquids, to be submitted; b) M. Dötterl, P. Thoma, H. G. Alt, Facile Synthesis of new Cationic Triphenylphosphine Derivatives and their Use for Propene Dimerization Reactions in Buffered Chloroaluminate Ionic Liquids, to be submitted; c) M. Dötterl, H. G. Alt, Nickel catalyzed Propene Dimerization Reactions in BiPh3 buffered Chloroaluminate Ionic Liquids: High Performance with unconventional Cations, to be submitted. [16] a) I. C. Quarmby, R. A. Mantz, L. M. Goldenberg, R. A. Osteryoung, Anal. Chem. 1994, 66, 3558-3561; b) I. C. Quarmby, R. A. Osteryoung, J. Am. Chem. Soc. 1994, 116, 2649-2650. 8 Silica Supported Cocatalysts for Olefin Dimerization and Polymerization Reactions with Nickel Complexes 145 [17] a) G. G. Eberhardt, W. P. Griffin, J. Catal. 1970, 16, 245-253; b) B. Bogdanovic, Selectivity Control in Nickel-Catalyzed Olefin Oligomerization, in Advances in Organometallic Chemistry, Vol. 17, Academic Press, New York, 1979, pp. 105140. [18] L. K. Johnson, C. M. Killian, M. Brookhart, J. Am. Chem. Soc. 1995, 117, 64146415. [19] a) A. Riisager, R. Fehrmann, S. Flicker, R. van Hal, M. Haumann, P. Wasserscheid, Angew. Chem., Int. Ed. 2005, 44, 815-819; b) A. Riisager, R. Fehrmann, M. Haumann, P. Wasserscheid, Eur. J. Inorg. Chem. 2006, 2006, 695-706. [20] C. Carlini, M. Marchionna, A. M. R. Galletti, G. Sbrana, J. Mol. Catal. A: Chem. 2001, 169, 19-25. [21] B. Gilbert, H. Olivier-Bourbigou, F. Favre, Oil & Gas Science and Technology - Rev. IFP 2007, 62, 745-759. [22] a) D. Pappalardo, M. Mazzeo, C. Pellecchia, Macromol. Rapid Commun. 1997, 18, 1017-1023; b) F. AlObaidi, Z. Ye, S. Zhu, Polymer 2004, 45, 6823-6829. [23] a) C. DeCastro, E. Sauvage, M. H. Valkenberg, W. F. Hölderich, J. Catal. 2000, 196, 86-94; b) J. Joni, M. Haumann, P. Wasserscheid, Adv. Synth. Catal. 2009, 351, 423-431; c) J. Joni, M. Haumann, P. Wasserscheid, Appl. Catal., A 2010, 372, 8-15. [24] G. P. Smith, A. S. Dworkin, R. M. Pagni, S. P. Zingg, J. Am. Chem. Soc. 1989, 111, 525-530. [25] Z. Grubisic, P. Rempp, H. Benoit, Journal of Polymer Science Part B: Polymer Letters 1967, 5, 753-759. 9 Solubility Behaviour of TiCl4, ZrCl4, and HfCl4 in Chloroaluminate Ionic Liquids 146 9 Solubility Behaviour of TiCl4, ZrCl4, and HfCl4 in Chloroaluminate Ionic Liquids Matthias Dötterl[a], Isabelle Haas[a] and Helmut G. Alt*[a] [a] Lehrstuhl für Anorganische Chemie II, Universitätsstraße 30, NW I, 95440 Bayreuth, Germany. E-mail: [email protected] Keywords: hafnium, zirconium, titanium, chloroaluminates, ionic liquids Manuscript to be submitted Abstract: The solubility of NbCl5, TaCl5, TiCl4, ZrCl4, and HfCl4 in neutral [BMIM][AlCl4] (BMIM = 1-butyl-3-methylimidazolium) was determined. While TiCl4 was immiscible with the neutral ionic liquid, 0.80 molar equivalents of ZrCl4 and stoichiometric amounts of HfCl4 dissolved in the melt at ambient temperature. The crystal structures and the unit cell parameters of [Ti2Cl10][BMIM]2, [Zr2Cl10][BMIM]2, and [Hf2Cl9][PhNMe3] were determined. [Ti2Cl10][BMIM]2, and [Zr2Cl10][BMIM]2 were crystallized from basic chloroaluminate melts. With a trimethylanilinium cation, [Hf2Cl9][PhNMe3] crystallized from an equimolar composition of PhNMe3Cl, AlCl3, and HfCl4 abstracting a chloride ligand from [AlCl4]– to give highly Lewis acidic [Al2Cl7]– anions. 9.1 Introduction About 30 years ago, Wilkes et al. and Osteryoung et al. reported the first room temperature ionic liquids based on aluminum chloride.[1] In combination with organic cations like N-alkylpyridinium or 1-alkyl-3-methylimidazolium chloride, AlCl3 formed liquids at ambient temperature over a wide range of compositions. For example, a 9 Solubility Behaviour of TiCl4, ZrCl4, and HfCl4 in Chloroaluminate Ionic Liquids 147 mixture of 1-ethyl-3-methylimidazolium chloride with two equivalents of AlCl3 was still found to be a liquid at – 80°C.[2] The Lewis acidities of such chloroaluminate ionic liquids are determined by the ratio of AlCl3 to organic chloride salt.[3] In mixtures containing excess chloride salt, the [AlCl4]– anion is the dominant aluminum species.[3] With increasing AlCl3 content, the equilibrium is shifted towards [Al2Cl7]– species.[1a, 1c, 4] The presence of [Al3Cl10]– was also detected in highly acidic melts.[5] Usually, the melting points display a minimum with a ratio AlCl3 / halide salt = 0.33 or 0.67, while a local maximum is observed for neutral compositions.[2a, 6] Later, chloroaluminate ionic liquids were also employed as solvents for other metal salts.[7] Especially basic melts with free chloride ligands dissolved a variety of metal halides to give anionic chlorometallate complexes.[8] Acidic chloroaluminate melts were used to dissolve small amounts of titanium tetrachloride. The mixtures were used for electrochemical oxidation and metal deposition reactions.[9] Titanium and zirconium tetrachloride additives in acidic chloroaluminate melts were reported to promote alkylation reactions of isoparaffins with olefins.[10] Imidazolium cations were also applied for the formation of ionic liquids with a variety of other metal chlorides other than aluminum chloride.[11] However, the quantitative solubility behaviour of Lewis acidic metal chlorides in neutral tetrachloroaluminate melts has not been investigated yet. 9.2 Results and Discussion Cationic Propene Oligomerization Reactions Initiated by Mixtures of Metal Chlorides and [BMIM][AlCl4] We were interested in new Lewis acidic ionic liquid compositions, which catalyze reactions typical for acidic chloroaluminate ionic liquids like cationic oligomerization of olefins,[12] isomerization[13] and cracking[14] reactions of saturated alkanes, or Friedel Crafts alkylation reactions.[15] Therefore, we chose the commercially available 1-butyl-3-methylimidazolium (BMIM) tetrachloroaluminate as standard solvent for preliminary experiments with different metal chlorides. This neutral ionic liquid itself did not catalyze cationic olefin oligo-