José A.
Fuentes
,
Alexandra M. Z.
Slawin
and
Matthew L.
Clarke
*
School of Chemistry, University of St Andrews, St Andrews, Fife, KY16 9ST, UK. E-mail: mc28@st-andrews.ac.uk; Fax: +44 (0) 1334 463808; Tel: +44 (0) 1334 463850
First published on 8th February 2012
Pre-catalysts of type [PdCl(allyl)(monophosphine)] where the monophosphine is one of the tri-oxo-adamantyl cage phosphines such as 1,3,5,7-tetramethyl-6-phenyl-2,4,8-trioxo-6-phospha-adamantane give very high reactivity in the regioselective hydroxycarbonylation and alkoxycarbonylation (hydro-esterification) of styrene at 60 °C. This high reactivity enables the use of tert-butanol as nucleophile in the alkoxycarbonylation of styrene.
Control of enantioselectivity in alkene hydroxycarbonylation was a particularly intractable problem, but some very selective catalysts (e.e.'s up to 95%) have emerged recently.3k It is likely that part of the reason such high enantioselectivity was observed was the ability to work near room temperature in contrast to the forcing conditions (∼100–150 °C) normally employed in hydroxy and alkoxy carbonylations. A striking feature of the low temperature catalysts of type 1 (Fig. 1) is their unusual dimetallic structure in which a diphosphine bridges two Pd centres, giving one phosphorus per palladium.
We have not been able to locate any other reports in the literature featuring pre-catalysts in which a single monophosphine is the only neutral ligand introduced with the palladium. Thus, while there are many reports of highly regioselective catalysts derived from monophosphines that are also quite reactive, the catalysts are generally formulated with excess phosphine or as bis-phosphine complexes, presumably due to concerns about catalyst stability.
In this paper, we report that complexes of type [Pd(Cl)(L2)(η3–C3H5)], where L2 represents 1,3,5,7-tetramethyl-6-phenyl-2,4,8-trioxo-6-phospha-adamantane (Fig. 1), give more active catalysts for regioselective alkoxycarbonylation of styrene, and that this finding enables the reaction to be extended to a challenging tert-butoxycarbonylation.
A key objective was to establish if a Pd complex with one phosphine would deliver better reactivity than a bis-phosphine-Pd complex. Given that an excess of phosphine is always employed in the literature, catalyst stability may become an issue, and for this reason we selected the phenyl-cage phosphine, L2 (and phosphonite counterparts, L3 and L4 (Fig. 1)),5 since it is thought that L2 readily forms stable mono-ligated Rh catalysts under hydroformylation conditions, even using excess phosphine ligand.6 The bidentate analogues of these ligands are also important ligands in closely related Pd catalysed alkene alkoxycarbonylation processes.4d Thus, in addition to investigating the general class of Pd carbonylation catalysts containing only one mono-phosphine in detail for the first time, an examination of phospha-adamantyl cage phosphines in hydroxycarbonylation and alkoxycarbonylation of styrene was also clearly warranted.
Palladium (II) allyl complexes were chosen as stable catalyst precursors that only contain one phosphine per palladium. Complexes of formula [PdCl(L)(η3–C3H5)] were generated in situ for the initial catalyst screening, but in the case of PPh3, L2 and L3, complexes 5–7 were also isolated in pure form and fully characterised. The new complexes, 6 and 7 are readily prepared by simply mixing commercially available [PdCl(η3–C3H5)]2 with two equivalents of ligand in Et2O or Et2O/THF mixtures at room temperature (see supporting information). Complexes 6 and 7 were isolated as the expected mixture of diastereomers, since the allyl unit has syn/anti coordination modes and there is no symmetry plane in the complexes.7 The complexes were recrystallised by slow evaporation in an NMR tube from CDCl3 solvent to give crystals suitable for X-ray diffraction studies. The X-ray structures of complexes 6 and 7 are shown in Fig. 2 and 3 respectively. Complex 6 contains two independent molecules in the asymmetric unit. The allyl ligand is η3-co-ordinated and shows bond lengths and angles typical for this type of complex. More specifically, the shorter carbon-palladium bond length of Pd(1)–C(3) relative to Pd(1)–C(1) is in accordance with it being trans to the lower field chloride ligand, and has been observed in other complexes of this general type.7 The expansion of the Cl(1)–Pd(1)–P(1) angle to 95.69(7)° to accommodate the bulky phosphine is also an expected feature for such a complex. Complex 7 shares many structural features with complex 6 (although only one isomer is present in the asymmetric unit). However, the Cl(1)–Pd(1)–P(1) angle is significantly larger at 100.39(4)°; this may reflect increased steric demand of this ligand.
![]() | ||
Fig. 2 X-ray structure of complex 6. |
![]() | ||
Fig. 3 X-ray structure of complex 7. |
We arbitrarily chose ethoxycarbonylation of styrene (Scheme 1 and Table 1) as our model reaction to study these new pre-catalysts in comparison to triphenylphosphine based systems (5 and 9) and the known bis-phosphine complex, 8 (Fig. 1).6a As can be seen from Table 1, the allyl complexes of ligands L2–4 are excellent catalysts for this reaction (Table 1, Entries 4 and 5), surpassing the commonly employed triphenylphosphine-based catalysts. The Pd-allyl complex of triphenylphosphine, 5 is not a significant improvement over the more commonly employed bis-phosphine complex, 9 (Table 1, Entries 1 and 2), which is a complete contrast to when complex 6 is compared to the bis-phosphine complex, 8 (Table 1, Entry 4 compared to Entry 6). The activity of the phosphonite-derived catalysts was lower than complex 6, but the performance of L3 and L4 do represent amongst the most active catalysts we have examined in this type of reaction (out of ∼20 distinct catalysts), and phosphinites also represent a broad ligand class that has, to the best of our knowledge, not been examined before in this type of reaction. The catalyst loadings were too low to obtain meaningful 31P{1H} NMR spectra at the end of the catalytic reactions to confirm ligand stability. However, the activity observed, surpassing all monophosphines except L2, and the fact that L4, which would be expected to be the least hydrolytically stable ligand, outperforms L3, combined with the observations on the unusual stability of these phosphonites5 argues that these ligands are robust under the reaction conditions. Further studies were carried out using L2 since it is now commercially available and easier to access than the phosphonites.
Entrya | T/°C | Alcohol | Catalyst (%) | t/h | Conv.b (%) | b:lc | Yieldd (%) |
---|---|---|---|---|---|---|---|
a Reactions carried out at 30 bar of CO, at 80 °C, using 1 mmol styrene in 1.5 ml of EtOH and 5 mol% PTSA.hydrate and 5 mol% LiCl promoters unless otherwise indicated. b Esters/(esters + styrene) × 100 determined by 1H-NMR integration. c Ratio in the isolated products determined by 1H-NMR integration. d Isolated yield of esters. e 20 mol% of LiCl and PTSA. f Reaction carried out in 3.0 ml of degassed 2-butanone in presence of 2.5 equivalents of degassed EtOH (with respect to 2 mmol styrene). g Reaction carried out in 1.5 ml of degassed 2-butanone in presence of 2.5 equivalents of degassed alcohol (with respect to 1 mmol styrene). h Metal precursor and ligand added as stock solution in 2-butanone prior to other reagents. | |||||||
1 | 80 | EtOH | 5 (0.1) | 24 | 28 | n.d. | n.d. |
2e | 80 | EtOH | 9 (1.0) | 24 | 58 | 7![]() ![]() |
40 |
3 | 80 | EtOH | 6 (0.1) | 24 | >99 | 32![]() ![]() |
89 |
4f | 80 | EtOH | 6 (0.05) | 24 | >99 | >100![]() ![]() |
88 |
5f | 80 | EtOH | 7 (0.05) | 24 | >99 | >200![]() ![]() |
98 |
6f | 80 | EtOH | 8 (0.05) | 24 | 7 | 19![]() ![]() |
n.d. |
7f | 60 | EtOH | [PdCl(allyl)]/L2 (0.05)h | 19 | >99 | >200![]() ![]() |
n.d. |
8f | 60 | EtOH | [PdCl(allyl)]/L3 (0.05)h | 19 | 54 | >200![]() ![]() |
n.d. |
9f | 60 | EtOH | [PdCl(allyl)]/L4 (0.05)h | 19 | 93 | >200![]() ![]() |
n.d. |
10f | 60 | EtOH | 6 (0.05) | 22 | >99 | >200![]() ![]() |
83 |
11f | 60 | EtOH | 7 (0.05) | 22 | 38 | n.d | n.d. |
12g | 80 | BnOH | 6 (0.1) | 24 | >99 | >100![]() ![]() |
89 |
13g | 80 | i–PrOH | 6 (0.1) | 24 | >99 | >200![]() ![]() |
96 |
![]() | ||
Scheme 1 Alkene carbonylation reactions investigated in this study. |
Complex 6 was used in several other alkoxy and hydroxycarbonylations. alkoxycarbonylation using benzyl alcohol or iso-propanol are straightforward (Table 1, Entries 9 and 10), while hydroxycarbonylation of styrene is also promoted significantly more effectively using catalysts 6 relative to triphenylphosphine-based systems (Table 2).
Entrya | Cat (%) | Conv.b (%) | b:l | Yieldc (%) |
---|---|---|---|---|
a Reactions carried out at 30 bar of CO, at 80 °C, in 1.5 ml of degassed 2-butanone in presence of 2.5 equivalents of degassed water w.r.t. 1.0 mmol styrene and 5 mol% TsOH·H2O and LiCl for 24 h. b Acids/(acids + styrene) × 100 determined by 1H-NMR integration. c Isolated yield of acids after extractive work-up. d Reaction carried out at 30 bar of CO, at 60 °C, in 3.0 ml of degassed 2-butanone in presence of 2.5 equiv. of water w.r.t. 2.0 mmol styrene and 5 mol% TsOH·H2O and LiCl for 22 h. | ||||
1a | 5 (0.1) | 63 | >200![]() ![]() |
33 |
2a | 6 (0.1) | 90 | >100![]() ![]() |
77 |
3d | 6 (0.05) | 83 | >200![]() ![]() |
76 |
4d | 7 (0.05) | 7 | n.d. | n.d. |
Alkoxycarbonylation using tert-butanol as nucleophile would be a useful process. On one hand, expanding this reaction to include tertiary alcohols seems worthwhile, but more importantly, t-butyl esters are the standard protecting group used for carboxylic acids in organic chemistry, since they deactivate nucleophilic attack on the carboxyl function and remove its acidity but can be readily deprotected by acids such as TFA or HCl under mild conditions. The only example of intermolecular alkoxycarbonylation with a tertiary alcohol reported an 8% conversion after 94 h at 100 °C and high pressures of CO.8a The fact that tert-butanol can be used as a source of iso-butylene in ethoxycarbonylation8b must also be testament to tert-butoxycarbonylation of alkenes being a problematic reaction. Our first attempt using conditions that work well for other alcohols also did not give good results, although catalyst 6 is clearly superior in this reaction (Table 3, Entry 1–3). In addition to lower conversions, problems were encountered with chemoselectivity. The origin of the carboxylic acid by-product was not investigated; it can be formed either by hydroxycarbonylation from adventitious water (PTSA was used as it's hydrate), or by acid catalysed ester cleavage of the tert-butyl ester products. An unexpected side product is the iso-propyl ester detected by GCMS and by the characteristic pattern in the 1H NMR spectra; commercial tert-butanol contains traces of iso-propanol, but this is amplified in the products by a factor of approximately 20, providing semi-quantitive evidence of the relative reactivity of i–PrOH and tert-butanol in these reactions. This side product was accompanied by traces of other unidentified products in the crude reaction mixtures. Since catalyst 6 was so superior to the others in this instance, we optimised conditions using 6 only. Reduced reaction temperatures deliver a significant improvement and the tert-butyl ester could be isolated in pure form by column chromatography in good yield. Other acidic promoters could also be used providing they are acidic enough, but the semi-optimised protocol uses PTSA hydrate at lower concentration delivering tert-butyl ester in 77% yield after chromatography (Table 3, Entry 10).
Entrya | T/°C | Cat (%) | LiCl (%) | Acid (%) | t/h | t-Bu Esterb (%) | Acidc (%) | Otherd (%) | b:le | Yieldf (%) |
---|---|---|---|---|---|---|---|---|---|---|
a Reaction carried out in 1.5 ml of degassed 2-butanone in presence of 2.5 equivalents of degassed t-BuOH w.r.t. 1.0 mmol styrene. b Esters determined by 1H-NMR integration using Et4Si as internal standard. c Acids determined by 1H-NMR integration using Et4Si as internal standard. d Trace unknown side products: The majority is i-Pr ester from i-PrOH impurity (0.4%) in t-BuOH. Remaining mass balance not quantified is styrene. e Ratio in the isolated products determined by 1H-NMR integration. f Isolated yield of pure t-Bu ester after chromatography. g Reaction carried out in 1.5 ml of t-BuOH as solvent. | ||||||||||
1 | 80 | 6 (0.5) | 5 | PTSA, 5 | 23 | 55 | 17 | Traces | 45![]() ![]() |
48 |
2 | 80 | 5 (0.5) | 5 | PTSA, 5 | 23 | 11 | 11 | n.d. | n.d. | n.d. |
3 | 80 | 7 (0.5) | 5 | PTSA, 5 | 23 | 12 | 9 | n.d. | n.d. | n.d. |
4g | 60 | 6 (1.0) | 5 | PTSA, 5 | 23 | 80 | 13 | 3 | 86![]() ![]() |
78 |
5g | 60 | 6 (1.0) | 2.5 | PTSA, 2.5 | 23 | 88 | 11 | 1 | 100![]() ![]() |
73 |
6 | 60 | 6 (1.0) | 2.5 | Acetic, 5 | 23 | <2 | <2 | n.d | n.d | n.d |
7 | 60 | 6 (1.0) | 2.5 | Citric, 2.5 | 23 | 2 | 5 | — | n.d | n.d |
8 | 80 | 6 (0.5) | 2.5 | Al(OTf)3, 1 | 23 | 66 | 12 | 4 | 40![]() ![]() |
n.d |
9g | 60 | 6 (1.0) | 2.5 | Oxalic, 2.5 | 23 | 88 | 8 | 3 | 100![]() ![]() |
71 |
10g | 60 | 6 (1.0) | 2.5 | PTSA, 1 | 23 | 83 | 5 | 3 | 98![]() ![]() |
77 |
11g | 80 | 6 (0.5) | 2.5 | PTSA, 1 | 23 | 79 | 6 | 4 | 64![]() ![]() |
70 |
In summary, we have explored the use of the adamantyl cage phosphines in alkene carbonylation, and discovered that pre-catalysts of type [Pd(Cl)(L2)(η3–C3H5)], in which the phosphine to palladium ratio is one are very active catalysts for this reaction, operating at low temperatures at low catalyst loadings and giving significantly better results than the benchmark triphenylphosphine system. Leading on from this, we have also shown that this catalyst allows alkoxycarbonylation to proceed with tert-butanol as nucleophile.
Footnote |
† Electronic supplementary information (ESI) available: Full experimental details, analytical data and selected NMR spectra. CCDC reference numbers 859192 and 859193. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c2cy00521b |
This journal is © The Royal Society of Chemistry 2012 |