Charge-tagged polar phosphine ligands in Pd-catalysed reactions in aqueous and ionic media

Article abstract of DOI:10.1016/j.tetlet.2018.01.073

A new range of polar imidazolium and phosphate-containing ligands was synthesised from readily available starting materials in high yielding multi-step transformations. These ligands were used to generate Pd catalysts for Suzuki and Heck C–C coupling reac

Full text of DOI:10.1016/j.tetlet.2018.01.073

Tetrahedron Letters xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Charge-tagged polar phosphine ligands in Pd-catalysed reactions in
aqueous and ionic media
a
a
a,b,
⇑
Vivien L. van Zyl , Alfred Muller , D. Bradley G. Williams
a
Department of Chemistry, University of Johannesburg, P.O. Box 524, Auckland Park 2006, South Africa
University of Technology Sydney, School of Mathematical and Physical Sciences, P.O. Box 123, Broadway, Sydney, NSW 2007, Australia
b
a r t i c l e i n f o
a b s t r a c t
Article history:
A new range of polar imidazolium and phosphate-containing ligands was synthesised from readily avail-
able starting materials in high yielding multi-step transformations. These ligands were used to generate
Pd catalysts for Suzuki and Heck C–C coupling reactions in organic and organic/aqueous media. The cat-
alysts performed well in aqueous media in the Suzuki reaction and less well in the Heck reaction, related
to substrate solubility in the aqueous media. When moving to ionic liquids, the Heck reaction dramati-
cally improved, especially in media compatible with the polar catalysts and the non-polar reagents. In
all cases, the catalysts were stable to the formation of Pd black, a form of degradation that frequently
befalls Pd catalysts. The catalysts could be successfully recycled without loss of activity.
Ó 2018 Elsevier Ltd. All rights reserved.
Received 29 November 2017
Revised 18 January 2018
Accepted 23 January 2018
Available online xxxx
Keywords:
Palladium
Homogeneous catalyst
Ionic liquid
Ligand
Introduction
organic reaction product to very low levels. Polar ligands also assist
the move towards more benign solvent systems, even if recovery of
The primary role of ligands in the context of homogeneous
catalysis is to modify the metal in question towards stability, activ-
ity, product selectivity, isolation, and recycling. The role that
ligands play in determining many of the features of metal-based
catalysts has led to a prolific number and wide variety of ligands
the catalyst is not under consideration, because aqueous or ionic
liquid-based solvents can be employed. Thus, it was an objective
7
of the present study to investigate the synthesis, and application
in catalysis, of different P-type ligands with the polarity being
derived from phosphate-containing and imidazolium-containing
moieties. These ligands would be assessed in Pd-catalysed CAC
1
2
being prepared, including those that are phase-switchable, which
1
a
8
9
are highly useful in the mechanistic study of catalysed reactions.
While heterogeneous catalysts are readily recovered and reused,
bond-forming Heck and Suzuki reactions in polar organic,
organic aqueous and ionic liquid systems, including recycling
experiments.
3
homogeneous catalysts suffer a persistent disadvantage because
of the difficulties associated with recovering and recycling them.
Polar phosphine ligands assist the recovery and reuse of catalysts
in several possible ways. Firstly, if the ligands are sufficiently polar,
they may be extracted into an aqueous phase post-reaction, having
Results and discussion
2
been held in the organic phase during the reaction. Secondly, they
The synthesis of these ligands was achieved by multi-step
sequences which converted suitable starting materials into more
complex derivative structures, making use of Ph PCl and Ph PH
may be held in an altogether separate phase, which may be aque-
ous or another non-miscible phase such as, for example, an ionic
2
2
4
liquid. Such biphasic systems facilitate the recovery and reuse of
as sources of the P atom. These two reagents are useful as elec-
5
catalysts, a pertinent example being the Rhône-Poulenc process.
trophiles and nucleophiles, respectively. In the case of Ph PCl as
2
Here, a Rh catalyst is used in a biphasic medium to convert lower
the substrate, a nucleophile may attack the P atom, ejecting chlo-
ride to form a new PAC bond. In the latter, the PAH is deproto-
6
alkenes into higher value aldehydes and alcohols. Triphenylphos-
6
phine trisulphonate is the ligand, and it imparts to the catalyst a
nated thereby converting it into
a nucleophile suitable for
level of water solubility which reduces leaching of Rh into the
reactions with, for example, ,b-unsaturated esters. In all cases,
a
the products of these reactions were transformed into polar
entities.
⇑
Corresponding author at: University of Technology Sydney, School of Mathe-
matical and Physical Sciences, P.O. Box 123, Broadway, Sydney, NSW 2007,
Australia.
4
-Bromobenzyl alcohol (1) was protected as its silyl ether 2
(Scheme 1), and the resulting product subjected to lithium-halogen
exchange with n-butyllithium.
1
0
E-mail address: Bradley.williams@uts.edu.au (D. Bradley G. Williams).
Reaction of the lithiated
https://doi.org/10.1016/j.tetlet.2018.01.073
0
040-4039/Ó 2018 Elsevier Ltd. All rights reserved.
Please cite this article in press as: van Zyl V.L., et al. Tetrahedron Lett. (2018), https://doi.org/10.1016/j.tetlet.2018.01.073

2
V.L. van Zyl et al. / Tetrahedron Letters xxx (2018) xxx–xxx
Scheme 1. Synthesis of mono phosphate-based polar ligands.
tic signals for the imidazolium moiety in the H NMR and 13C NMR
1
2
intermediate with Ph PCl, and P-protection of the reaction product
with BH
3
-THF, afforded the triphenylphosphine derivative 3. If the
spectra thereof, accompanied by a sharp singlet at À5.1 ppm in its
3
1
P-atom was left unprotected during conversion of the benzyl alco-
hol into its corresponding chloride under the action of thionyl chlo-
ride (see below), the P atom was oxidised into its unwanted P(O)
derivative. Unmasking of the OH group under the action of tetra-
n-butylammonium fluoride on 3 furnished benzyl alcohol 4 in
P NMR spectrum, which provided evidence for deprotection of
the P atom. Alternatively, benzyl alcohol 4 could be reacted with
dimethyl chlorophosphate (CAUTION! This substance is highly
toxic and requires special handling techniques) to convert it into
phosphate 7. Deprotection of the P atom and partial hydrolysis of
the phosphate ester of 7 with neat dihexylamine, followed by
cation exchange with NaH, afforded the corresponding sodium
phosphate 8, which was characterised as its phosphine oxide 9.
In this instance, dihexylamine was chosen for the P deprotection
step because the ligand would be soluble in the aqueous layer
and it was desirable for the amine-borane complex to be soluble
6
6% overall yield from 1, after column chromatography on silica
1
gel. A two proton singlet at 4.72 ppm in the H NMR spectrum of
the product was consistent with the anticipated benzyl alcohol
CH group. This product afforded the anticipated broad doublet
2
31
signal at 20.8 ppm (J = 61.3 Hz) in the P NMR spectrum thereof,
accompanied by a range of doublet signals in the aromatic region
1
3
1
of the C NMR spectrum, caused by PAC coupling over a number
of bonds. Uneventful conversion of the benzyl alcohol into its chlo-
ride analogue 5 under the action of thionyl chloride and the reac-
tion of 5 with excess N-methylimidazole at 100 °C, followed by P-
deprotection with neat diethylamine,11 afforded the desired imida-
zolium-containing phosphine 6 in 60% yield from 4. The free phos-
phine and the diethylamine-borane complex were readily
in the organic phase. The H NMR spectrum of the P@O product
9 showed a characteristic two proton doublet at 4.88 ppm (JH,P
=
8.2 Hz) and a three proton doublet at 3.56 ppm (JH,P = 9.5 Hz),
3
1
while the P NMR spectrum thereof showed two sharp singlet sig-
nals at 31.7 ppm (P(O)) and 1.1 ppm (OPO).
In a second approach towards polar phosphines, diphenylphos-
phine-borane 10 was lithiated using n-butyllithium and the
separated by extraction with CH
was soluble in the organic phase while the amine-borane complex
was soluble in the aqueous phase. Compound 6 produced diagnos-
2
Cl
2
and water: the phosphine
lithium phosphide was subsequently reacted with
-butyrolactone in a Michael-type reaction to produce adduct 11
(Scheme 2). The lactone was readily reduced to the corresponding
a-methylene-
c
Scheme 2. Synthesis of bis imidazolium-based polar ligands.
Please cite this article in press as: van Zyl V.L., et al. Tetrahedron Lett. (2018), https://doi.org/10.1016/j.tetlet.2018.01.073

V.L. van Zyl et al. / Tetrahedron Letters xxx (2018) xxx–xxx
3
Scheme 3. Synthesis of bis phosphate-based polar ligands.
diol by reaction thereof with LiAlH
4
to produce the desired product
from the results that the ligands produce catalysts that are consis-
tently good for the Heck and Suzuki reactions performed in DMF.
The catalysts all performed well to mediate the Suzuki reaction
1
2 in 85% yield. Reaction of the diol with thionyl chloride followed
by treatment of the dichloro intermediate 13 with N-methylimida-
zole at elevated temperature produced bisimidazolium chloride
in DMF/H
2
O. In contrast, the Heck reactions performed in DMF/
1
4. The P atom was released immediately prior to the catalysis
2
H O did not fare so well. It is most likely that the difference in out-
by treatment of 14 with dihexylamine to deliver free phosphine
come is due to the polarity of the reaction partners, and it is unli-
kely that the 4-bromobenzonitrile, used as a substrate for the
oxidative addition step that initiates both the Heck and the Suzuki
reaction, causes the difference. Rather, for the Heck reaction the
nonpolar butyl acrylate is used as the substrate while in the Suzuki
reaction the polar phenylboronic acid is used as the substrate. Pre-
sumably, it is the low solubility of the former that hampers the
progress of the Heck reaction in the polar organic/aqueous binary
solvent. These outcomes are consistent with our previous results
using simpler catalyst systems in polar organic and in aqueous
1
1
5, which was not characterised due to its sensitive nature. Diol
2 was also reacted with dimethyl chlorophosphate (CAUTION!
This substance is highly toxic and requires special handling tech-
niques), the product of which was deprotected, partially hydrol-
ysed and subjected to cation exchange to produce the water
soluble bisphosphate 17 (76% from 12), which was characterised
as its P@O derivative 18 (Scheme 3).
The ligands were then employed in Pd-catalysed Suzuki- and
Heck reactions (Scheme 4, Table 1). Reactions were performed in
polar organic medium and in organic/aqueous medium, and the
catalysts were benchmarked against baseline systems which were
devoid of ligands. In all instances, the ligands produced catalysts
that were stable and from which no Pd black was formed. It is clear
8
media.
Additional Heck reactions were therefore performed in ionic
liquids (Table 2) in an attempt to improve the outcome. As a gen-
eral comment on the outcomes of the reactions, the catalysts con-
stituted from the new ligands performed as well as or better than
the triphenylphosphine benchmark reaction. The ionic liquid of
4
choice was [bmim]BF (see Table 2 footnote for definition of the
abbreviations for the ionic liquids), which is the most polar of
the ionic liquids employed. While the [bmim] and [hmim] ionic
liquids could conceivably form carbene structures with Pd, the
[
C8dmim] ionic liquid cannot because the 2-position is blocked.
That there was little difference between the outcomes when using
this ionic liquid and those where the 2-position is free, would point
towards the notion that carbenes potentially generated from the
ionic liquids are not involved in the catalysis. The ligands demon-
strate their real usefulness during recycling experiments. Firstly,
no Pd black was noted for any of the new ligands during four con-
secutive runs, indicating that stable catalysts were formed. Sec-
ondly, NMR spectra of concentrates of the organic extracts (post
reaction in each of the four runs) failed to show the presence of
the ligands, demonstrating minimal leaching thereof, along with
the Pd, into the organic phase. Recharging the substrates to the
Scheme 4. Suzuki and Heck reactions using polar catalysts.
Pd/6/[bmim]BF
4
catalyst medium in the four consecutive runs,
Table 1
a,b
Pd-catalysed Heck- and Suzuki reactions.
Entry
Ligand
Suzuki product 19
Heck product 20
DMF
DMF/H
2
O
DMF
2
DMF/H O
1
2
3
4
5
6
7
–
8
9
6
12
14
17
55%
73%
75%
74%
77%
86%
75%
60%
75%
76%
78%
80%
86%
75%
10%
65%
72%
67%
72%
76%
75%
5%
20%
20%
25%
28%
45%
43%
a
Suzuki reaction: Pd(OAc)
2
(5 mol%), ligand (10 mol%), DMF or 1:1 DMF/H PO
2
O (2 mL), K
3
4
(1.0 equiv.), 4-bromobenzonitrile (1.0 equiv.), phenylboronic acid (1.0 equiv.),
1
00 °C, 24 h.
Heck reaction: Pd(OAc)
b
2
(5 mol%), ligand (10 mol%), DMF or 1:1 DMF/H O (2 mL), Et N (1.0 equiv.), 4-bromobenzonitrile (1.0 equiv.), butyl acrylate (1.0 equiv.), 110 °C,
2
3
2
4 h.
Please cite this article in press as: van Zyl V.L., et al. Tetrahedron Lett. (2018), https://doi.org/10.1016/j.tetlet.2018.01.073

4
V.L. van Zyl et al. / Tetrahedron Letters xxx (2018) xxx–xxx
Table 2
Pd-catalysed Heck reactions in ionic liquids.a,b
Ligand
[bmim]BF
74%
4
[hmim]BF
4
[C8dmim]BF
4
[bmim][NSO
2
CF
]
3 2
[hmim][NSO
2
CF
]
3 2
2 3 2
[C8py][NSO CF ]
TPP
6
61%
63%
59%
58%
68%
56%
58%
65%
59%
48%
48%
45%
51%
55%
52%
88% (83%, 84%, 81%)c
81%
1
5
a
TPP = triphenylphosphine, bmim = 1-butyl-3-methylimidazolium, hmim = 1-hexyl-3-methylimidazolium, C8dmim = 2,3-dimethyl-1-octylimidazolium, C8py = N-
octylpyridinium.
b
2 3
Heck reaction: Pd(dba) (5 mol%), ligand (10 mol%), ionic liquid (1 mL), Et N (1.0 equiv.), 4-bromobenzonitrile (1.0 equiv.), butyl acrylate (1.0 equiv.), 110 °C, 24 h.
Values in parentheses represent recycling experiments in which the catalyst was used in a total of four runs.
c
once the organic phase had been decanted, allowed the recyclabil-
ity of the catalysts to be assessed. These runs afforded the Heck
products in 88%, 83%, 84%, and 81% yield, respectively. The high
level of repeatability of the runs when making use of the recycled
catalyst indicates the retention of high levels of the catalyst in the
ionic liquid phase and that it suffered little or no degradation or
inactivation.
A. Supplementary data
Supplementary data (experimental methods, characterisation
data) associated with this article can be found, in the online ver-
sion, at https://doi.org/10.1016/j.tetlet.2018.01.073.
References
1.
(a) Limberger J, Leal BC, Monteiro BC, Dupont J. Chem Sci. 2015;6:77;
b) Cornils B, Herrmann W, eds. Applied Homogeneous Catalysis with
Organometallic Compounds, Vol. 1. Weinheim: VCH; 1996.
2. Mokhadinyana MS, Williams DBG, Cole-Hamilton DJ. Angew Chem Int Ed.
012;51:1648.
Conclusion
(
Imidazolium and phosphate-containing phosphines, produced
in good overall yield in several steps from readily accessible start-
ing materials, form stable Pd catalysts that are capable of effecting
CAC bond-forming reactions in polar organic and organic/aqueous
mixed solvent systems. All ligands form stable and active catalysts
for Suzuki and Heck reactions in ionic liquids and the outcomes of
the reactions are less linked with the nature of the Pd-ligand sys-
tem than they are with the solubility of the substrates. If the sub-
strates are only slightly soluble in the organic/aqueous medium,
then the reactions proceed in low yields. The presence of the imi-
dazolium or the phosphate groups clearly does not hamper the cat-
alyst activity. Recycling experiments revealed an ability to reuse
the catalysts over several runs without loss of activity.
2
3
4
.
.
De Jong KP, ed. Synthesis of Solid Catalysts. Weinheim: Wiley VCH; 2009.
(a) Škoch K, Císa rˇ ová I, Št eˇ pni cˇ ka P. Organometallics. 2016;35:3378;
(b) Williams DBG, Ajam M, Ranwell A. Organometallics. 2006;25:3088.
(a) Behr A, Kämper A, Kuhlmann R, Vorholt AJ, Franke R. Catal Sci Technol.
5
.
2
016;6:208;
b) Wiebus E, Cornils B. In: Cole-Hamilton DJ, Tooze RP, eds. Catalyst Separation,
Recovery and Recycling: Chemistry and Process Design. Dordrecht: Springer;
006:105–143.
Bhanage BM, Divekar SS, Deshpande RM, Chaudhari RV. Org Process Res Dev.
000;4:342.
7. Williams DBG, Ajam M, Ranwell A. Organometallics. 2007;26:4692.
(
2
6.
2
8
9
.
.
Williams DBG, Lombard H, Holzapfel CW. Synth Commun. 2001;2077.
Williams DBG, Evans SJ, de Bod H, Mokhadinyana MS, Hughes T. Synthesis.
2009;3106;
Williams DBG, Shaw ML. Tetrahedron. 2007;63:1624.
1
1
0. Snieckus V. Chem Rev. 1990;90:879.
1. Williams DBG, Lombard H, van Niekerk M, Coetzee PP, Holzapfel CW.
Phosphorus, Sulfur Silicon Relat Elem. 2002;177:2799.
Acknowledgement
We thank the NRF, THRIP, University of Johannesburg, Sasol and
the University of Technology Sydney for funding.
Please cite this article in press as: van Zyl V.L., et al. Tetrahedron Lett. (2018), https://doi.org/10.1016/j.tetlet.2018.01.073