Highly electron rich alkyl- and dialkyl-N-pyrrolidinyl phosphines: An evaluation of their electronic and structural properties

Article abstract of DOI:10.1039/b108467d

Phosphines containing two N-bound pyrrolidine groups and one alkyl or aryl group have been prepared and shown to be unusually electron rich donor ligands when compared to either tris(N-pyrrolidinyl)phosphine or the trialkyl-or triaryl-phosphines. It is proposed that the electron donating ability of a pyrrolidine group towards phosphorus had been underestimated as only two pyrrolidinyl groups can donate towards the phosphine via the nitrogen lone pair. trans-L²Rh(CO)Cl complexes have been prepared from the phosphines and used to quantify the electron donor characteristics. Two of these complexes have been characterised by X-ray crystallography. The reaction of these phosphines with iron(II) complexes has also been studied. Platinum(II) complexes of the ligands have been prepared and also (in three examples) characterised by X-ray crystallography, which has enabled the steric and bonding properties to be evaluated. tert-Butyl(dipyrrolidinyl)phosphine is one of the most electron rich phosphines known. The dialkyl(pyrrolidinyl)phosphines have been found to contain less potent N→P donation, and are somewhat less good donor ligands.

Full text of DOI:10.1039/b108467d

View Article Online / Journal Homepage / Table of Contents for this issue
Highly electron rich alkyl- and dialkyl-N-pyrrolidinyl phosphines:
an evaluation of their electronic and structural properties
Matthew L. Clarke, Gemma L. Holliday, Alexandra M. Z. Slawin and J. Derek Woollins*
School of Chemistry, University of St. Andrews, St. Andrews, Fife, UK KY16 9ST.
E-mail: jdw3@st-andrews.ac.uk
Received 19th September 2001, Accepted 3rd January 2002
First published as an Advance Article on the web 19th February 2002
Phosphines containing two N-bound pyrrolidine groups and one alkyl or aryl group have been prepared and shown
to be unusually electron rich donor ligands when compared to either tris(N-pyrrolidinyl)phosphine or the trialkyl-
or triaryl-phosphines. It is proposed that the electron donating ability of a pyrrolidine group towards phosphorus
had been underestimated as only two pyrrolidinyl groups can donate towards the phosphine via the nitrogen lone
pair. trans-L2Rh(CO)Cl complexes have been prepared from the phosphines and used to quantify the electron
donor characteristics. Two of these complexes have been characterised by X-ray crystallography. The reaction of
these phosphines with iron() complexes has also been studied. Platinum() complexes of the ligands have been
prepared and also (in three examples) characterised by X-ray crystallography, which has enabled the steric and
bonding properties to be evaluated. tert-Butyl(dipyrrolidinyl)phosphine is one of the most electron rich phosphines
known. The dialkyl(pyrrolidinyl)phosphines have been found to contain less potent N P donation, and are
somewhat less good donor ligands.
phosphorus atom is thought to arise from donation towards
Introduction
the phosphorus from the nitrogen lone pair. X-Ray crystal
Phosphine ligands have many important applications in
structures of tris(alkylamino)phosphines and their metal com-
organometallic chemistry and catalysis. One of the many
plexes3,11 show these ligands to contain two short P–N bonds
attractive features of these ligands is the widespread availability
with planar nitrogens, and one long P–N bond with a non-
of ligands that possess significantly different electronic prop-
erties. Some phosphines have a π-acceptor character that
planar nitrogen atom. This suggested to us that only two of
the nitrogen lone pairs could donate electron density towards
phosphorus, while the third nitrogen substituent acts merely
as an electronegative atom bound to the phosphorus, and there-
fore reduces the overall basicity of the phosphine. If this were
the case, a “hybrid” ligand that contains one electron donating
alkyl group and two electron donating amino groups might
be an extremely electron rich phosphine ligand, and have
approaches that of carbon monoxide, whereas other phos-
phines are very strong σ-donors.1–4
Consequently, a metal complex modified by a strong π-
acceptor ligand will have entirely different properties to a
metal complex of σ-donor ligands. Metal complexes of strongly
electron donating alkylphosphines undergo many reactions
which are not possible with arylphosphines.5 This is often due
numerous applications in catalysis. In this paper we describe the
to oxidative addition being a much easier process for an
electron rich metal centre. In the last few years, there have
been several new catalytic processes developed that rely on
the increased reactivity of an electron rich metal complex.
synthesis of these new bis(N-pyrrolidinyl)alkylphosphines and
metal complexes which shed light on their properties.12
Important recent examples are catalytic C–H bond activation,6
hydroformylation of alkenes to give commercially valuable
alcohols instead of aldehydes,7 and palladium catalysed Heck,
Suzuki, and amination reactions of aryl chlorides, which do not
normally react with most other phosphine based catalysts.8–10
The synthesis of new electron donating ligands that offer differ-
ent properties to those already existing has therefore become an
important challenge. We were interested in using P–N bond
formation to prepare ligands that might have a different
combination of steric and electronic properties to that of
existing ligands (see Fig. 1).
Experimental
General
All manipulations were carried out under an atmosphere of
nitrogen, unless stated otherwise. All solvents were either
freshly distilled from an appropiate drying agent (THF, Et2O,
CH2Cl2) or obtained as anhydrous grade. H, and 31P NMR
1
spectra were recorded using a “Varian 2000” 300 MHz spec-
trometer. IR spectra were recorded as KBr discs (prepared in
air) on a Perkin Elmer PE1720 FTIR/RAMAN spectrometer.
Pt(COD)Cl2, CpFe(CO)2(I), [CpFe(CO)2(MeCN)]BF4 and
(tripyrollidinyl)phosphine were prepared by literature
methods.3,13–15 All other materials used in this paper were
obtained from Aldrich Chemical Company and used as
received. All peaks above 700 cmϪ1 in the IR spectrum are
reported to serve as a fingerprint.
General experimental procedure for synthesis of phosphines
Fig. 1 Proposed electronic properties of (2).
Pyrrolidine (5 equiv. for (dipyrrolidinyl)phoshines, 2.5 equiv.
for (monopyrrolidinyl)phosphines) was added dropwise to a
solution of the appropriate chloro-phosphine in diethyl ether.
Tris(alkylamino)phosphines are known to be electron rich
phosphine ligands. The high basicity (σ-donor strength) of the
DOI: 10.1039/b108467d
J. Chem. Soc., Dalton Trans., 2002, 1093–1103
This journal is © The Royal Society of Chemistry 2002
1093

View Article Online
This was stirred for a few hours (overnight in the case of
tBuPCl2) prior to filtration under nitrogen and removal of
solvent to give a colourless mobile liquid in essentially quanti-
tative yields. The phosphines are air and moisture sensitive, but
can be handled easily using standard Schlenk techniques. The
purity and identity of these phosphines was determined
spectroscopically.
and had solvent reduced to about 0.2–1 mL in vacuo. Hexane or
diethyl ether was then added dropwise to precipitate out/
crystallise the product. This was collected on a frit in air, and
washed with hexane (2 × 3 mL), and then dried in vacuo to yield
pure product. On some occasions, a second crop of micro-
crystals could be obtained from the filtrate by repeating the
above procedure. Typically yields were 60%.
Phenyl(dipyrrolidinyl)phosphine (3). IR(νmax/cmϪ1
)
2971,
trans-Carbonyl-chloro-bis(phenyl(dipyrrolidinyl)phosphine)-
rhodium (7). Anal. Calcd for C29H42N4P2RhClOؒ0.5 CH2Cl2: C,
50.23; H, 6.14; N, 7.94. Found: C, 50.35; H, 6.16; N, 7.99%.
IR(νmax/cmϪ1) 2960, 2861, 1949, 1480,1435, 1191, 1100, 1064,
1010. 31P NMR (121.4 MHz; CDCl3): δP 88.4 (JP–Rh = 136 Hz).
1H NMR (300 MHz; CDCl3): δH 1.80 (8H, s, br,), 3.10 (8H, m),
7.30 (6H, m, ArH), 7.7 (4H, m, ArH).
1638, 1459, 1291, 1163, 1061, 834, 720. 31P NMR (121.4 MHz;
C6D6) δP 64.46 (s). 1H NMR (300 MHz; C6D6) δH 1.51 (8H, qn
app. {app. = apparent}), 3.01 (4H, m, br), 3.16 (4H, m, br),
7.01–7.05 (5H, m, ArH). HRMS: Found: 248.1451 (Mϩ),
C14H21N2P requires: 248.1442.
Methyl(dipyrrolidinyl)phosphine (4). IR(νmax/cmϪ1) 2962, 2841
1478, 1434, 1340, 1190, 1093, 1057, 997, 746, 702. 31P NMR
(121.4 MHz; C6D6) δP 74.07 (s). 1H NMR (300 MHz; C6D6) δH
1.05 (3H, dd, 1.1, 6.6 Hz), 1.30 (8H, qn app.), 2,82 (8H, m).
HRMS: Found: 186.1280 (Mϩ), C9H19N2P requires:186.1286.
trans-Carbonyl-chloro-bis(methyl(dipyrrolidinyl)phosphine)-
rhodium (8). Anal. Calcd for C19H38N4P2RhClOؒ0.5CH2Cl2: C,
40.29; H, 6.76; N, 9.64. Found: C, 40.76; H, 7.20; N, 9.87%.
IR(νmax/cmϪ1) 2965, 2857, 1947, 1457, 1349, 1324, 1281, 1261,
1197, 1080, 1014. 31P NMR (121.4 MHz; CDCl3): δP 95.2 (JP–Rh
tert-Butyl(dipyrrolidinyl)phosphine (5). IR(νmax/cmϪ1) 2965,
1592, 1459, 1395, 1150, 1029, 910. 31P NMR (121.4 MHz;
CDCl3): δP 99.35. 1H NMR (300 MHz; CDCl3): δH 1.46 (9H, d,
J = 12.9 Hz), 1.92 (8H, qn app.), 3.42 (8H, qn app.).
HRMS (E.S.ϩ): Found: 229.1832; C12H26N2P (MHϩ)
requires:229.1833.
1
= 130 Hz). H NMR (300 MHz; CDCl3): δH 1.65 (6H, t, J = 7
Hz), 1.80 (16H, qn app.), 3.10 (16H, m). 13C NMR (75.5 MHz,
CDCl3): δC 13.9 (t app., JC–P = 21.1 Hz, JC–Rh = 21 Hz,), 25.2
(CH2), 47.1 (CH2), CO too weak to observe. MS (E.S.ϩ):
Found: 539.1333; (MHϩ) requires:539.1342.
trans-Carbonyl-chloro-bis(tert-butyl(dipyrrolidinyl)phosphine)-
rhodium (9). Anal. Calcd for C25H50N4P2RhClO: C, 48.20; H,
Methyl(dipyrrolidinyl)phosphine selenide. Anal. Calcd for
8.09; N, 9.00. Found: C, 48.24; H, 8.43; N, 8.96%. IR(νmax/cmϪ1
)
C9H19N2PSe: C, 40.76; H, 7.22; N, 10.6. Found: C, 40.87; H,
7.41; N, 11.17%. 31P NMR (121.4 MHz; CD3) δP 65.95 (JP–Se
=
2861, 1942, 1637, 1456, 1393, 1361, 1184, 1103, 1060. 31P NMR
(121.4 MHz; CDCl3): δP 114.1 (JP–Rh = 133 Hz). 1H NMR (300
MHz; CDCl3) δH 1.31 (9H, . app., J = 11 Hz), 1.82 (8H, m), 3.30
(8H, m). 13C NMR (75.5 MHz, CDCl3) δC 25.2 (CH2), 27.6
(CH3), 37.8 (t app., JC–P = 20.1 Hz, JC–Rh = 20 Hz, R4C), 48.7
(CH2), 187.1 (dt, JC–P = 15.6, JC–Rh = 77 Hz (Rh–P–CO)).
HRMS (E. S.ϩ): Found: 587.2498; C25H50N4P2ORh (M Ϫ Cl)
requires:587.2515.
735 Hz). 1H NMR (300 MHz; CD3) δH 1.7–1.95 (11H, m), 3.05
(8H, m).
Diisopropyl(pyrrolidinyl)phosphine (10). IR(νmax/cmϪ1) 2949,
1463, 1379, 1362, 1346, 1194, 1133, 1063, 1004, 876. 31P NMR
1
(121.4 MHz; CDCl3): δP 64.82. H NMR (300 MHz; CDCl3):
δH 1.05 (12H, dd, J = 14.0, 7.1 Hz), 1.7 (4H, m), 1.93 (2H, m),
3.0 (4H, m). 13C NMR (75.4 MHz; CDCl3) δ 15.85 (d, J = 7.6
Hz), 22.15 (d, J = 11.9 Hz), 23.2 (d, J = 4.33 Hz), 47.2 (d,
J = 10.82). Found: 187.1484: C14H19NP requires: 187.1487.
trans-Carbonyl-chloro-bis(diisopropyl(pyrrolidinyl)phosphine)-
rhodium (12). Anal. Calcd. for C21H44N2P2OClRh: C, 46.63;
H, 8.20; N, 5.18. Found: C, 46.80; H, 8.96; N, 5.18%. IR(νmax
/
cmϪ1) 2967, 2864, 1954, 1455, 1382, 1361, 1347, 1240, 1118,
1066, 1021, 1009, 879. 31P NMR (121.4 MHz; CDCl3): δP 89.4
(d, JP–Rh = 127.6 Hz). 1H NMR (300 MHz; CDCl3): δH 1.3 (24H,
m), 1.80 (4H, m), 2.72 (8H, m), 3.36 (8H, m) 13C NMR (75.4
Di-tert-butyl(pyrrolidinyl)phosphine (11). One equivalent of
n-butyllithium was added to a cold (Ϫ60 ЊC) THF solution
of pyrrolidine. This was allowed to warm to room temperature.
The reaction was left stirring for 30 min prior to the addition of
di-tert-butylchlorophosphine (via syringe). This solution was
stirred overnight. THF was removed in vacuo, and the residue
extracted with toluene, filtered under nitrogen (to remove LiCl)
and pumped dry under vacuum to give the product as a colour-
less liquid. IR(νmax/cmϪ1) 2965, 2868, 1685, 1476, 1389, 1366,
1167, 1072, 1011, 981. 31P NMR (121.4 MHz; CDCl3): δP 83.4.
1H NMR (300 MHz; CDCl3): δH 1.11 (18H, d, J = 11.8 Hz),
1.68 (4H, m), 3.16 (4H, m). HRMS: Found: 215.1810;
C12H26NP requires: 215.1803.
MHz; CDCl3) δC 19.3 (d, JC–P = 19.6 Hz), 26.4, 27.2 (dd, JC–P
=
13, 13 Hz), 51.0, CO too weak to observe. Found: 541.1748;
C21H45N2P2OClRh (MHϩ) requires:541.1750.
trans-Carbonyl-chloro-bis(di-tert-butyl(pyrrolidinyl)phos-
phine)rhodium (13). Anal. Calcd. for C25H52P2N2OClRh: C,
50.30; H, 8.78; N, 4.69; Found: C, 49.97; H, 8.78; N, 4.72%.
IR(νmax
/cmϪ1) 2952, 2870, 1995, 1956, 1480, 1457, 1388, 1363,
1179, 1119, 1066, 1010, 808. 31P NMR (121.4 MHz; CDCl3):
1
δP 103.3 (d, JP–Rh = 134 Hz). H NMR (300 MHz; CDCl3):
General procedure for the synthesis of trans-L2Rh(CO)Cl
complexes
δH 1.55 (36H, dd, J = 6.7, 6.7 Hz), 1.74 (8H, m), 3.50 (8H, m).
Found: 597.2363 C25H53P2N2OClRh requires: 597.2377.
[Rh(CO)2Cl]2 was added in one portion to a stirred solution of
the appropriate phosphine (4.6 equiv.). The rhodium complex
rapidly dissolves with evolution of carbon monoxide to give a
solution of the desired rhodium complex. For L = phenyl-
(dipyrrolidinyl)phosphine size exclusion chromatography was
carried out to remove the excess ligand. For L = diethyl-
(pyrrolidinyl)phosphine, the rhodium complex (described in
ref. 22) proved difficult to isolate due to its sensitivity to air and
extreme solubility. For all other ligands, the reaction solution
was transferred via cannula to another Schlenk flask (this was
a convenient method to remove traces of rhodium metal that
are sometimes observed at the bottom of the reaction vessel),
General procedure for the synthesis of platinum complexes
To a stirred solution of ligands (3)–(5) (two equiv.) in dichloro-
methane was added (COD)PtCl2 (one equiv.) in one portion.
The resultant solution was stirred for two hours, before the
solvent was removed to near dryness. Diethyl ether was then
added via a syringe to obtain a white precipitate that was
collected on a frit and washed with diethyl ether (2 × 5 mL) and
hexane (3 × 5 mL). Drying in vacuo gave the desired platinum
complexes. Recrystallisation by slow diffusion of Et2O into
CH2Cl2 solutions of these compounds gave crystals suitable for
X-ray analysis.
1094
J. Chem. Soc., Dalton Trans., 2002, 1093–1103

View Article Online
cis-Dichloro-bis(methyl(dipyrrolidinyl)phosphine)platinum(II)
(17). Anal. Calcd for C18H38N4P2Cl2Pt: C, 33.86; H, 6.00; N,
8.78. Found: C, 34.23; H, 6.10; N, 8.49%. IR (νmax./cmϪ1) 2965,
2860, 1449, 1352, 1292, 1195, 1115, 1080, 1008, 876. 31P NMR
Carbonyl-cyclopentadienyl-iodo-(tripyrrolidinylphosphine)iron
(19). Yield: 120 mg, 0.232 mmol, 14.3%. Anal. Calcd for
C18H29N3PFeIO: C, 41.80; H, 5.65; N, 8.12. Found: C, 41.33; H,
5.85; N, 7.70%. IR (νCO = 1941 cmϪ1). 31P NMR (121.4 MHz;
1
1
(121.4 MHz; CDCl3): δP 46.4 (JP–Pt = 4232 Hz). H NMR (300
CDCl3): δ 136.7. H NMR (300 MHz; CDCl3): δ 1.90 (12H, s,
MHz; CDCl3): δH 1.73 (6H, d, J = 7 Hz), 1.90 (16H, m), 3.30
(16H, m).
br), 3.1 (12H, m), 4.65 (5H, s, br). HRMS (E. S.ϩ): Found:
518.0520; C28H30N3P1O1Fe1I1 (MHϩ) requires 518.0520.
The unpurified by-product separated from the reaction
mixture can be identified as dicarbonyl-cyclopentadienyl-
(tripyrrolidinylphosphine)iron iodide (20). Yield: 460 mg, 0.844
mmol, 52%. IR (νCO = 1991, 2036 cmϪ1). 31P NMR (121.4 MHz;
cis-Dichloro-bis(phenyl(dipyrrolidinyl)phosphine)platinum(II)
(16). Anal. Calcd for C28H42P2N4PtCl2: C, 44.10; H, 5.55; N,
7.35. Found: C, 44.22; H, 5.68; N, 6.79%. IR (νmax./cmϪ1) 2965,
2870, 1483, 1434, 1349, 1331, 1246, 1190, 1105, 1070, 1009, 918,
866. 31P NMR (121.4 MHz; CDCl3): δP 46.1 (JP–Pt = 4358 Hz).
1H NMR (300 MHz; CDCl3): δH 1.90 (16H, m), 3.30 (16H, m),
7.0–7.5 (10H, m).
1
CDCl3): δ 117.6. H NMR (300 MHz; CDCl3): δ 1.96 (12H, s,
br), 3.2 (12H, m), 5.53 (5H, s, br). HRMS (E. S.ϩ): Found:
418.1353; C19H29N3P1O2Fe1 (M Ϫ Iϩ) requires 418.1347.
Carbonyl-cyclopentadienyl-iodo-(phenyl(dipyrrolidinyl)phos-
phine)iron (21). Yield: 250 mg, 0.477 mmol, 38.5%. Anal. Calcd
for C20H26N2PFeIO: C, 45.83; H, 5.00; N, 5.34. Found: C, 45.78;
H, 4.96; N, 4.97%. IR (νCO = 1942 cmϪ1). 31P NMR (121.4 MHz;
cis-Dichloro-bis(tripyrrolidinylphosphine)platinum(II) (15). 31P
NMR monitoring of a mixture of (COD)PtCl2 and tri-
pyrrolidinephosphine showed a major platinum containing
species [δp: 40.7 (1JP–Pt = 4950 Hz)] which we assign as the cis
compound, (15). This formulation is further supported by mass
spectrometry of the crude (isolated) product which shows peaks
due to MHϩ, M Ϫ Cl, M Ϫ 2Clϩ, (M Ϫ Cl Ϫ pyrrolidine), (M
Ϫ 2Cl Ϫ pyrrolidine), (M Ϫ Cl Ϫ 2pyrrolidine), (M Ϫ 2Cl Ϫ
2pyrrolidine) etc. IR (ν max./cmϪ1) 2959, 2864, 1459, 1340,
1240, 1193, 1071, 1009, 916, 871. Recrystallisation only gave a
few crystals that were analysed as being pure (15) by X-ray
crystallography. There was not sufficient pure material for
complete analytical characterisation. Fortunately, our primary
interest in the synthesis of (15) was to determine its crystal
structure.
1
CDCl3): δ 125.1. H NMR (300 MHz; CDCl3): δ 1.80 (8H, s,
br), 2.8–3.4 (8H, m), 4.15 (5H, s, br), 7.3–7.8 (5H, m, ArH).
HRMS (E. S.ϩ): Found: 525.0246; C20H27N2POFeI (MHϩ)
requires 525.0255.
The unpurified by-product separated from the reaction
mixture can be identified as dicarbonyl-cyclopentadienyl-
(phenyl(dipyrrolidinyl)phosphine)iron iodide (22). Yield: 70
mg, 0.127 mmol, 10.2%. IR (νCO = 1992, 2037 cmϪ1). 31P NMR
1
(121.4 MHz; CDCl3): δ 111.1. H NMR (300 MHz; CDCl3):
δ 2.0 (8H, s, br), 2.9–3.3 (8H, m), 5.2 (5H, s, br), 7.5–7.7 (5H, m,
ArH). HRMS (E. S.ϩ): Found: 425.1080; C21H27N2P1O2Fe1
(M Ϫ Iϩ) requires 425.1081.
Iron complexes derived from [CpFe(CO)2(MeCN)]BF4
trans-Dichloro-bis(tert-butyl(dipyrrolidinyl)phosphine)-
platinum(II) (18).
A solution of tert-butyl(dipyrrolidinyl)-
[CpFe(CO)2(MeCN)]BF4, (70.4 mg, 0.231 mmol, 1 equiv.) and
phenyl(dipyrrolidinyl)phosphine (0.149 g, 0.600 mmol, 2.6
equiv.) were refluxed in dichloromethane for 2 hours. Solvent
was then reduced to ca. 0.5 mL prior to the addition of diethyl
ether (5 mL) which gave a beige precipitate. This was collected
(in air) by filtration and washed with Et2O (3 × 5 mL) and
hexane (3 × 5 mL) which removed any traces of MeCN and free
ligand. Yields are essentially quantitative.
phosphine in CH2Cl2 was added dropwise to a solution of
K[PtCl3(C2H4)] in acetone. After 30 min, the solvents were
removed in vacuo, and the residue extracted with CH2Cl2 and
centrifuged to remove KCl. Removal of solvent from this
solution, followed by washing with diethyl ether gave the
desired complex in quantitative yield. Anal. Calcd for
C24H50N4P2PtCl2: C, 39.89; H, 6.97; N, 7.75. Found: C, 40.15;
H, 6.82; N, 7.65%. IR (νmax./cmϪ1) 2964, 2859, 1457, 1394, 1362,
1243, 1185, 1100, 1070, 1011, 994, νPt–Cl = 341. 31P NMR (121.4
Dicarbonyl-cyclopentadienyl-(tripyrrolidinylphosphine)iron
tetrafluoroborate (23). Anal. Calcd for C19H29N3PFeO2BF4 C,
45.18; H, 5.79; N, 8.32. Found: C, 45.26; H, 5.99; N, 8.82%. IR
(νCO = 2071, 2027 cmϪ1). 31P NMR (121.4 MHz; CDCl3):
1
MHz; CDCl3): δP 83.6 (JP–Pt = 2702 Hz). H NMR (300 MHz;
CDCl3): δH 1.47 (18H, dd., J = 7.6, 7.6 Hz), 1.84 (16H, m), 3.40
(16H, m). HRMS (E. S.ϩ): Found: 722.2608; C24H50N4P2Pt1Cl2
(MHϩ) requires 722.2614.
1
δ 118.0. H NMR (300 MHz; CDCl3): δ 1.8 (12H, s, br), 3.0
(12H, s, br), 5.4 (5H, s, br).
General procedure for the formation of monodentate iron
complexes
Dicarbonyl-cyclopentadienyl-(phenyl(dipyrrolidinyl)phos-
phine)iron tetrafluoroborate (24). Anal. Calcd for C21H26-
N2PFeO2BF4ؒ0.5CH2Cl2 C, 49.42; H, 5.21; N, 5.36. Found: C,
49.19; H, 5.34; N, 5.54%. IR (νCO = 2042, 2011 cmϪ1). 31P NMR
(121.4 MHz; CDCl3): δ 111.5. H NMR (300 MHz; CDCl3):
δ 2.0 (8H, s, br), 3.0 (8H, s, br), 5.0 (5H, s,), 7.2–7.6 (5H, m).
A Schlenk flask containing a preweighed amount of the desired
ligand (1.08 equiv.) was briefly opened under a stream of
nitrogen to allow the introduction of [CpFe(CO)2]2 (3 mol%)
and CpFe(CO)2I (1 equiv.). This flask was then sealed with a
rubber septum, evacuated and flushed with nitrogen. Toluene
(15 mL) was then added and the flask heated at 90 ЊC. The
reactions were monitored for completion by removal of small
aliquots by syringe and analysing by solution IR spectroscopy.
The reactions are cooled and filtered through a sinter (briefly
in air) to collect the ionic product (which is pure by NMR
spectroscopy), and a dark green/brown solution. Solvent is
reduced to dryness and the dark powder is dissolved in the
minimum volume of dichloromethane and added to an alumina
column that was then eluted with toluene until a green band
containing the pure product is collected. We have found that it
is acceptable to carry out this chromatography in air using
normal solvents, despite the fact that the iron complexes do
slowly decompose in solution when left exposed to the air for
several hours. Yields are unoptimised.
1
X-Ray crystallography
The experimental details for (15) and (17) have already been
reported (CCDC reference numbers 152757 and 152758).12
Crystal structures were obtained using a Bruker SMART
diffractometer with graphite-monochromated Mo-Kα radi-
ation (λ = 0.71073 Å). Intensity data were collected using 0.3
or 0.15Њ width ω steps accumulating area detector frames
spanning a hemisphere of reciprocal space for all structures. All
data were corrected for Lorentz, polarisation and long term
intensity fluctuations. Absorption effects were corrected on the
basis of multiple equivalent reflections. Structures were solved
by direct methods and refined by full-matrix least squares
against F 2 (SHELXTL).16
J. Chem. Soc., Dalton Trans., 2002, 1093–1103
1095

View Article Online
Table 1 Crystal data for compounds (13), (14), (16) and (24)
(13)
(14)
(16)
(24)
Empirical formula
C25H52ClN2OP2Rh
C37H66ClOP2Rh
C28H42Cl2N4P2Pt
C21H26BF4FeN2O2P
M
596.99
Monoclinic
C2/c
727.20
Triclinic
P1
762.59
512.07
Monoclinic
P21/n
Crystal system
Space group
Triclinic
¯
P1
a/Å
b/Å
c/Å
α/Њ
β/Њ
γ/Њ
U/Å
26.7010(8)
15.1617(5)
24.5230(3)
90
116.033(1)
90
8920.5(4)
12
0.791
9.9574(1)
10.2575(1)
10.8125(1)
113.782(1)
109.177(1)
90.819(1)
940.65(2)
1
9.8077(1)
10.4935(3)
16.7513(5)
94.161(2)
101.821(2)
116.604(1)
1481.95(6)
2
11.6620(2)
14.6306(1)
13.7467(3)
90
97.238(1)
90
2326.80(7)
4
0.769
Z
µ/mmϪ1
0.636
5.048
Reflections measured
Independent reflections
Final R1, wR2 [I > 2σ(I )]
22067
6403
0.0366 0.0752
5922
5012
0.0250 0.0582
6547
4236
0.0406 0.0842
10002
3349
0.0408 0.0969
The data for (14) were examined in some detail because of
are all liquids and stable for considerable periods of time if kept
under an atmosphere of dry nitrogen. In an attempt to prepare
a crystalline derivative of methyl(dipyrrolidinyl)phosphine, (4)
was also characterised by its quantitative conversion to its
selenide. This compound which is also a liquid, shows the
expected selenium satellites in its phosphorus NMR spectrum.
the possibility of space group ambiguity. There are no system-
¯
atic absences to help distinguish between P1 and P1 but the
intensity statistics clearly indicate a non-centrosymmetric space
group [mean |E × E Ϫ1| = 0.741, for centrosymmetric space
groups 0.968 is expected, for non-centrosymmetric 0.736 is
expected], refinements were also attempted in both P1 and P1
The magnitude of 1JP᎐Se (735 Hz) is significantly smaller than in
¯
᎐
to ascertain if a disordered centrosymmetric model could give a
better fit. The refinement as reported gave the most satisfactory
results. Details of data collection and crystal refinement are
summarised in Table 1.
tris(dialkylamino)phosphines. This is in line with previous
studies which have shown that this coupling constant is smaller
1
for alkylphosphines (Me P᎐Se, J
= 648 Hz; (Me N) P᎐Se,
᎐
2 3
᎐
3
P᎐Se
᎐
1JP᎐Se = 795 Hz) than in nitrogen substituted phosphines.18
᎐
CCDC reference numbers 171245–171248.
See http://www.rsc.org/suppdata/dt/b1/b108467d/ for crystal-
lographic data in CIF or other electronic format.
b: Electronic properties of new phosphino-amines
Molloy and Petersen3 prepared tris(N-pyrrolyl)phosphine
which has exceptional π-acceptor character. This is due to the
stabilisation of any negative charge accepted from the metal on
nitrogen by an aromatic resonance structure in this ligand.
They also prepared tris(N-pyrrolidinyl)phosphine as a com-
parison and showed this ligand to be a very electron rich
phosphine. The method they used to gauge the electronic
characteristics of their ligands was measurement of νCO in the
infrared spectrum of the trans-(R3P)2Rh(CO)Cl complexes.
These are readily formed in high purity from [Rh(CO)2Cl]2 and
an excess of ligand (Scheme 2).
Molecular modelling/cone angle measurements
PC Spartan Pro (Wavefunction Inc., CA; www.wavefun.com)
was used to calculate the cone angles using molecular mechan-
ics minimisation of a phosphine. Then the half angles were
measured using the program’s measure angle function and
selecting the three relevant angles (making sure the metal atom
is set 2.28 Å from the P atom).
The Chemical Database Service search engine ConQuest was
used to determine values of real angles. It was attempted, wher-
ever possible, to measure the cone angle from a simple nickel
complex or the sulfide of the phosphine, so as to avoid counter-
ligand steric effects.
Results
a: Phosphine synthesis
The amino-phosphines described in this study were prepared by
the addition of an excess of pyrrolidine to the appropriate
phosphine dichloride (see Scheme 1, pyrrolidine was chosen as it
is a very basic secondary amine).
Scheme 2
Vastag and co-workers have shown that a good correlation
exists between the value of ν(CO) in L2Rh(CO)Cl and
Ni(CO)3(L).19 The position of νCO in the easily prepared
rhodium complexes is known for a huge variety of phosphines,
and is always in agreement with the expected donor strength of
the phosphine. By way of ensuring the accuracy and validity of
our values, we have reprepared trans-(R3P)2Rh(CO)Cl com-
plexes of tris(N-pyrrolidinyl)phosphine, Ph3P, and Cy3P. The
values quoted in Table 2 are those found on our spectrometer
(literature values are in parentheses). The rhodium complexes
formed in the in situ IR experiments can be isolated by size
exclusion chromatography (to remove the excess ligand) and/or
by precipitation of very concentrated CH2Cl2 solutions with
hexane. However, it was not possible to grow good quality crys-
tals of (6) to (9) for X-ray analysis. The complexes all show
Scheme 1
P–N bond formation, which has been studied extensively in
these laboratories17 is generally facile, high yielding and gives
products of high purity, providing air and moisture are absent
1
during the reaction. 31P and H NMR of the air and moisture
sensitive phosphine products showed them to be essentially
pure, and they were therefore purified no further. The ligands
1096
J. Chem. Soc., Dalton Trans., 2002, 1093–1103

View Article Online
Table 2 Comparison of νCO of trans-L2Rh(CO)Cl complexes for pyrrolidine based ligands with other phosphines
Entry
L
νCO a [L2Rh(CO)Cl]
Ref.
Tolman cone angleb/Њ
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
(PhO)3P
(p-CF3C6H4)3P
Ph3P
PhMe2P
Me3P
Et3P
(2016)
(1990)
1965 (1966)
(1965)
(1960)
3
3
128
145
145
122
118
132
170
182
—
145
145
136
157
155
170
3c
21
3
(1956)
19
22c
—
3
Cy3P
1943 (1942)
tBu3P
P(N(CH3)2)3
(2)
—
e
(1959)
1951 (1952)
1949
1947
1942
1954
1955
3c,
d
c,
c,
c,
c,
c,
d
d
d
d
d
(3)
(4)
(5)
(12)
(13)
a To ensure the accuracy and validity of our values, we have reprepared trans-(R3P)2Rh(CO)Cl complexes of tris(N-pyrrolidinyl)phosphine, Ph3P, and
Cy3P. The values quoted are those found on our spectrometer (literature values are in parentheses). b Cone angle taken from ref. 1 and 2, and in the
case of the new ligands estimated assuming similar steric influence for phenyl and pyrollidinyl. c IR spectra recorded as KBr discs. d This work.
e Tetrahedral rhodium complex, so cannot be compared
the expected doublets in their 31P NMR spectra (1JP–Rh between
128 and 136 Hz). The 13C NMR spectra of these complexes
generally did not show the (very weak) carbonyl resonance.
However, in the case of complex (9), the spectrum was acquired
overnight to reveal the expected doublet of triplets that arises
by coupling of the carbonyl ligand to phosphorus and
rhodium. As can be seen from Table 2, the position of νCO for
PhMe2P is, as expected, in between electron rich Me3P and the
less basic Ph3P. Phenylbis(N-pyrrolidinyl)phosphine, however,
has νCO at lower wavenumber than tris(N-pyrrolidinyl)phos-
phine, and therefore can be assumed to be a more electron
rich ligand. This is evidence (along with the properties of
DMAP (4-dimethylaminopyridine) to promote the reaction.
Since pyrrolidine is a particularly nucleophilic amine, it seems
likely that it will prove impossible to prepare tBu2P substituted
amines under the standard mild conditions that are generally
used throughout this work. The desired phosphine [(11)] could
be prepared if pyrrolidine is first deprotonated with nBuLi prior
to addition of tBu2PCl (Scheme 3).
methylbis(N-pyrrolidinyl)phosphine
and
tert-butylbis(N-
pyrrolidinyl)phosphine) that only two pyrrolidinyl groups
can contribute towards the strong donor strength of (N-
pyrrolidinyl)phosphines. The values of νCO for (4) and (5) are
significantly lower than those of most other highly electron rich
alkylphosphines (compare entries 12 and 13 to entries 5 and 6)
which are often used in catalysis. Tri-tert-butylphosphine, which
is generally thought of as the most electron donating phosphine
actually forms a tetrahedral (R3P)2Rh(CO)Cl complex and
cannot be directly compared.20
The new ligands may be of particular use as they deliver a
donor strength that is normally reserved for very bulky ligands
(defined by their large cone angle). The cone angles of the three
new ligands can readily be estimated as Molloy and Petersen
have already shown that a pyrrolidinyl group has a similar steric
effect to a phenyl group.3 It follows that the cone angle of
methyl(dipyrrolidinyl)phosphine is the same as (or similar to)
MePh2P (136Њ). The cone angles of (2) and (3) are 145Њ. It can
be seen from Table 2 that triethylphosphine and (4) have
approximately similar steric properties, yet the new ligand
is a considerably stronger donor ligand. It is known that a
phosphine with a small cone angle and strong donating power
will bind very tightly to a metal centre and stabilise organo-
metallic compounds with respect to reductive elimination.23
Ligand (4) can therefore be expected to be applied with success
in organometallic chemistry and catalysis. Ligand (5) tert-
butyl(dipyrrolidinyl)phosphine is actually a fairly bulky ligand,
but given that is perhaps even more strongly electron donating
than Cy3P and somewhat smaller, it again delivers a different
combination of steric and electronic effects to existing ligands,
and promises to find application in transition metal chemistry
and catalysis.
Scheme 3
Complexes of the type trans-L2Rh(CO)Cl could be readily
prepared by standard methods. The electronic properties of
these dialkyl(pyrrolidinyl)phosphines are also somewhat
surprising. We expected ν(CO) for the rhodium complexes of
diisopropyl(pyrrolidinyl)phosphine (10) and di-tert-butyl-
pyrrolidinylphosphine(11)tobeataround1940–1947cmϪ1. This
would represent a composite of the known electronic properties
of each of the two groups. In actual fact, they represent signifi-
cantly less strong donor ligands than could be expected (Table
2, entries 14 and 15). We propose that this may be due to
the steric bulk of these phosphines preventing the nitrogen
substituents from adopting an orientation that allows them
to interact strongly with the phosphorus atom.24 These results
provide us with a useful lead in our ongoing work on catalysis
of the Suzuki reaction of aryl chlorides.25 We have at this stage
prepared hemi-labile, weakly chelating ligands derived from
iPr2PCl and Cy2PCl. The ligands do make effective catalysts,
but were found to be less electron rich than we might have
hoped. Given that a strongly electron donating phosphine is
likely to be desirable in this reaction, our future work will utilise
tert-butyldiaminophosphines as a hemilabile ligand. These
investigations will be reported in due course.
Using the reported correlation between ν(CO) of L2Rh-
(CO)Cl and Ni(CO)3(L) complexes, a rough estimate of the
values of ν(CO) in the extremely toxic nickel complexes can be
made. These values clearly place tert-butyl(dipyrrolidine)phos-
phine as one of the most strongly electron donating phosphines
known.26 Tolman used the values of ν(CO) in Ni(CO)3(L) com-
pounds to estimate a single substituent parameter χ, to describe
the contribution of a substituent, R to the overall basicity
of the phosphine R1R2R3P. 1 It is obvious that a substituent
We have also prepared two dialkyl(pyrrolidinyl)phosphines.
Ligand (10) is readily prepared by the addition of an excess of
pyrrolidine to iPr2PCl. Pyrrolidine does not react with tBu2PCl
even at elevated temperatures or using stronger bases such as
J. Chem. Soc., Dalton Trans., 2002, 1093–1103
1097

View Article Online
Table 3 Selected bond lengths (Å) and angles (Њ) for (13), the values
for the second independent molecule are not significantly different
parameter cannot be assigned to a pyrrolidine group, as it will
vary depending on the other substituents on phosphorus. We
suspect that the variability in donor strength of pyrrolidine
phosphines carries through to other phosphino-amines as other
phosphino-amines show similar structural features to those
reported and analysed here.
Rh(1)–P(1)
Rh(1)–P(2)
Rh(1)–Cl(1)
Rh(1)–C(34)
Rh(2)–C(54)
2.370(1) P(1)–N(1)
2.375(1) P(2)–N(41)
2.378(1) P(2)–N(21)
1.793(5) C(34)–O(34)
1.789(7)
1.681(4)
1.691(4)
1.684(3)
1.111(5)
c: X-Ray crystal structures of rhodium complexes
P(1)–Rh(1)–Cl(1)
P(2)–Rh(1)–Cl(1)
P(1)–Rh(1)–P(2)
C(34)–Rh(1)–P(1)
C(34)–Rh(1)–P(2)
C(25)–N(21)–P(2)
C(22)–N(21)–P(2)
92.62(4)
92.43(4)
173.43(4)
87.9(1)
87.3(1)
121.4(3)
129.9(3)
P(2)–Rh(1)–Cl(1)
92.43(4)
92.62(4)
106.8(4)
128.6(3)
121.6(3)
106.6(3)
P(1)–Rh(1)–Cl(1)
C(2)–N(1)–C(5)
C(5)–N(1)–P(1)
C(2)–N(1)–P(1)
C(25)–N(21)–C(22)
Recrystallisation of complex (13), by cooling a hexane/CH2Cl2
solution overnight gave crystals suitable for an X-ray crystal
structure determination. The molecular structure of (13) is
shown in Fig. 2 with selected bond lengths and angles in Table
Table 4 Selected bond lengths (Å) and angles (Њ) for (14)
Rh(1)–P(1)
Rh(1)–P(2)
2.352(3)
2.358(3)
Rh(1)–Cl(1)
Rh(1)–C(37)
2.422(4)
1.93(1)
P(1)–Rh(1)–Cl(1)
P(2)–Rh(1)–Cl(1)
91.0(1)
88.6(1)
C(37)–Rh(1)–P(1)
C(37)–Rh(1)–P(2)
89.1(3)
91.2(3)
Table 5 Selected bond lengths (Å) and angles (Њ) for (15)
Pt(1)–P(1)
Pt(1)–P(2)
Pt(1)–Cl(1)
Pt(1)–Cl(2)
P(1)–N(1)
2.270(3)
2.246(3)
2.398(3)
2.371(3)
1.63(1)
P(1)–N(6)
P(1)–N(11)
P(2)–N(16)
P(2)–N(21)
P(2)–N(26)
1.72(1)
1.64(1)
1.66(1)
1.68(1)
1.678(9)
P(1)–Pt(1)–Cl(1)
P(2)–Pt(1)–Cl(2)
Cl(2)–Pt(1)–Cl(1)
P(1)–Pt(1)–P(2)
P(1)–Pt(1)–Cl(2)
P(2)–Pt(1)–Cl(1)
C(2)–N(1)–C(5)
C(5)–N(1)–P(1)
C(2)–N(1)–P(1)
C(7)–N(6)–P(1)
C(7)–N(6)–C(10)
C(10)–N(6)–P(1)
86.3(1)
90.4(1)
85.1(1)
C(15)–N(11)–C(12)
C(12)–N(11)–P(1)
C(15)–N(11)–P(1)
C(20)–N(16)–C(17)
C(17)–N(16)–P(2)
C(20)–N(16)–P(2)
C(25)–N(21)–C(22)
C(25)–N(21)–P(2)
C(22)–N(21)–P(2)
C(27)–N(26)–P(2)
C(27)–N(26)–C(30)
C(30)–N(26)–P(2)
110(1)
124.4(9)
123(1)
110(1)
123(1)
123.9(8)
110(1)
126(1)
Fig. 2 Molecular structure of (13).
3. There are one and a half independent molecules in the unit
cell, one of which lies about a crystallographic two-fold axis.
The rhodium centre is square planar with little deviation
from idealised geometry. A comparison of our data with that
collated by Molloy and Petersen [Rh–C bond lengths in the
range 1.77(1)–1.845(15) Å, Rh–P bonds in the range 2.282(4)–
2.428(1) Å, and Rh–Cl bond lengths in the range 2.350(4)–
2.479(1) Å] is perhaps of interest. They found that the very
electron poor phosphine, tripyrollylphosphine gave very short
Rh–P and Rh–Cl distances, with an elongated Rh–C bond.
This is readily rationalised by considering back bonding within
the π-acceptor and π-donor ligands. Ligand (11) behaves as a
bulky, moderately electron rich ligand. The Rh–P bond lengths
are amongst the longest reported for these types of complex (av.
2.374(1) Å), whereas the Rh–C bonds are relatively short (av.
1.791(7) Å). It is to be expected that a weaker Rh–P interaction
should strengthen the metal–carbonyl bond. The Rh–Cl bond
lengths are fairly typical for these types of complex. The P–N
bonds within the structure are planar, but are completely
asymmetrical with respect to the geometry about the N atom.
In particular, the carbon nearest the bulky tert-butyl groups is
severely bent away to minimise repulsive interaction between
these groups (mean C–N–P angle = 129.2(3)Њ). This effect is not
seen in the other three structures reported in this paper, and we
tentatively suggest that it is the inability of the nitrogen atom to
98.2(1)
170.9(1)
175.4(1)
107(1)
126.4(8)
125.5(9)
124.7(10)
111.9(10)
121.6(10)
119(1)
124.4(8)
110.7(7)
122.6(7)
Crystals were grown easily by slow diffusion of Et2O into a
CH2Cl2 solution of trans-(Cy3P)2Rh(CO)Cl. The molecular
structure (shown in Fig. 3) shows the rhodium centre to be
adopt a more symmetrical geometry that weakens the N
P
donor interaction and therefore reduces the donor strength of
this ligand. The P–N bond lengths are not exceptional, but are,
nevertheless longer than those found on planar nitrogens in the
other three structures reported here (Tables 4, 5, and 6).
Fig. 3 Molecular structure of Cy3P2Rh(CO)Cl (14).
It was mentioned in the discussion of the IR data that
tri-tert-butylphosphine forms a tetrahedral rhodium complex
of type Rh(L)2(CO)Cl, as verified by X-ray crystallography.20
Tricyclohexylphosphine is thought to be nearly as bulky as
tri-tert-butylphosphine, so it was not clear whether the complex
trans-(Cy3P)2Rh(CO)Cl (14) contained some distortion away
from the square planar geometry expected. It was therefore of
interest to determine its structure by X-ray crystallography.
square planar with little deviation from square planar geom-
etry. The bond lengths around the metal centre are somewhat
anomalous, with very long Rh–P, Rh–Cl and Rh–C bond
lengths (Table 4). The Rh–C bond (1.93(1) Å) is especially
surprising, as it would be expected that an electron rich ligand
that does not form a particularly strong Rh–P bond would lead
to strong π-back bonding to the carbonyl ligand. We interpret
1098
J. Chem. Soc., Dalton Trans., 2002, 1093–1103

View Article Online
Table 6 Selected bond lengths (Å) and angles (Њ) for (17). There are two independent molecules. The related values for the second independent
molecule are given in square brackets
Pt(1)–P(1)
Pt(1)–P(2)
Pt(1)–Cl(1)
Pt(1)–Cl(2)
2.226(2) [2.246(2)]
2.255(2) [2.229(2)]
2.372(2) [2.375(2)]
2.391(2) [2.370(2)]
P(1)–N(1)
P(1)–N(6)
P(2)–N(11)
P(2)–N(16)
1.652(7) [1.690(7)]
1.675(7) [1.627(7)]
1.643(6) [1.692(7)]
1.676(6) [1.651(7)]
P(1)–Pt(1)–Cl(1)
91.32(8)8 [88.49(8)]
89.32(8) [91.00(9)]
86.14(8) [91.00(9)]
93.37(7) [93.90(8)]
109.8(6) [110.3(7)]
124.4 (5) [120.3(6)]
110.7(7) [109.7(6)]
122.3 (6) [123.9(6)]
C(15)–N(11)–C(12)
C(12)–N(11)–P(2)
C(15)–N(11)–P(2)
C(20)–N(16)–C(17)
C(17)–N(16)–P(2)
C(20)–N(16)–P(2)
C(2)–N(1)–P(1)
109.9(7) [110.3(7)]
124.3(5) [120.6(6)]
125.3(6) [120.9(6)]
110.9(6) [110.0(8)]
118.3(5) [121.2(6)]
124.3(6) [125.2(7)]
122.4(5) [123.4(6)]
120.3(5) [125.9(5)]
P(2)–Pt(1)–Cl(2)
Cl(2)–Pt(1)–Cl(1)
P(1)–Pt(1)–P(2)
C(2)–N(1)–C(5)
C(5)–N(1)–P(1)
C(7)–N(6)–C(10)
C(10)–N(6)–P(1)
C(7)–N(6)–P(1)
this data with caution as chloride and carbonyl ligands have
been found to be disordered in these types of complexes.27
d: Platinum complexes of pyrrolidinylphosphines
We have also characterised the dichloroplatinum complexes,
(R3P)2PtCl2, of the four N-pyrrolidinylphosphines (15) to (18).
These are formed quantitatively from (COD)PtCl2 and two
equivalents of phosphines (3) and (4) (Scheme 4). These com-
Scheme 4
Fig. 4 Molecular structure of (15).
pounds all show the expected singlets with platinum satellites
typical of these compounds. The sizes of JP–Pt reflect the
1
smaller coupling constants observed when phenyl groups are
compared to alkyl groups, and the larger constants typical of
compounds that contain either P–N or P–O bonds. Tris-
(pyrrolidinyl)phosphine (2), does not react so cleanly with
(COD)PtCl2. However, (15) is the major Pt complex formed
(31P NMR monitoring), and a few good quality crystals of
this compound could be obtained from CH2Cl2/Et2O (slow
diffusion). Bulky electron rich ligands do not react with
(COD)PtCl2 in such a straightforward way as smaller or less
basic ligands. Ligand (2) appears to be an intermediate case.
tert-Butyl(dipyrrolidine)phosphine does not react with
(COD)PtCl2 to give the desired product. (An unidentified,
unstable mixture of products is formed.) A complex, (18) of
trans geometry can be readily obtained from Zeises salt
(Scheme 5).
Fig. 5 Molecular structure of (17).
Scheme 5
We have been successful in obtaining the crystal structure of
three of the complexes by X-ray diffraction. The molecular
structures of tris(N-pyrrolidinyl)phosphine- (15), methylbis(N-
pyrrolidinyl)phosphine- (17) and phenylbis(N-pyrrolidinyl)-
phosphine-dichloroplatinum() (16) are shown in Figs 4, 5 and
6 respectively. Selected bond lengths and angles can be found in
Tables 5, 6 and 7. The crystal structures of (15) and (17) have
been described in our preliminary communication,12 so are only
briefly discussed here.
A feature observed in all three solid state structures is an
asymmetry regarding the Pt–P and Pt–Cl bond lengths. Thus
Fig. 6 Molecular structure of (16).
J. Chem. Soc., Dalton Trans., 2002, 1093–1103
1099

View Article Online
Table 7 Selected bond lengths (Å) and angles (Њ) for (16)
Table 8 Data used for Fig. 7
Pt(1)–P(1)
Pt(1)–P(2)
Pt(1)–Cl(1)
Pt(1)–Cl(2)
2.252(2)
2.268(2)
2.351(2)
2.393(2)
P(1)–N(1)
P(1)–N(6)
P(2)–N(16)
P(2)–N(21)
1.667(6)
1.678(7)
1.652(7)
1.673(7)
Spartan
calculated
angle/Њ
Actual
crystallographic
angle/Њ
Tolman
angle/Њ
Phosphine
PH3
PH2Ph
PF3
P(OMe)3
P(OEt)3
PMe3
PMe2Ph
PCl3
PHPh2
P(OMe)Ph2
PEt3
P(OEt)Ph2
PEt2Ph
P(CF3)3
PEtPh2
P(O-o-Tol)3
87
101
104
107
109
118
122
124
128
132
132
133
136
137
140
141
142
145
154
51.76
69.55
55.74
115.20
112.11
84.16
45.43
61.88
49.73
98.27
96.57
88.12
109.38
52.73
P(1)–Pt(1)–Cl(1)
P(2)–Pt(1)–Cl(2)
Cl(2)–Pt(1)–Cl(1)
P(1)–Pt(1)–P(2)
P(1)–Pt(1)–Cl(2)
P(2)–Pt(1)–Cl(1)
C(2)–N(1)–C(5)
C(5)–N(1)–P(1)
C(2)–N(1)–P(1)
90.40(8)
87.44(8)
85.77(8)
96.39(8)
175.41(8)
173.21(8)
109.7(6)
119.8(5)
125.5(5)
C(20)–N(16)–C(17)
C(17)–N(16)–P(2)
C(20)–N(16)–P(2)
C(25)–N(21)–C(22)
C(25)–N(21)–P(2)
C(22)–N(21)–P(2)
C(7)–N(6)–P(1)
106.6(7)
126.6(6)
126.3(6)
110.9(6)
125.9(6)
123.2(6)
120.9(5)
109.6(6)
122.0(5)
96.75
61.70
C(7)–N(6)–C(10)
C(10)–N(6)–P(1)
103.39
126.17
124.30
135.12
124.19
87.86
113.34
164.00
130.89
130.25
131.30
87.03
120.73
104.10
114.75
132.67
83.49
124.30
135.20
127.62
130.87
129.17
one of the phosphorus–platinum bonds is ca.0.02 Å longer than
the other [e.g. Pt(1)–P(1) = 2.270(3)Å, Pt(1)–P(2) = 2.246(3) Å
in (15)]. The elongation of one of the Pt–P bonds also results
(in all three structures) in a shortening of the Pt–Cl bond trans
to it, undoubtedly due to the trans influence [e.g. Pt(1)–Cl(1) =
2.398(3) Å, Pt(1)–Cl(2) = 2.371(3) Å in (15)]. This type of solid
state effect has been observed in the crystal structures of
some other (R3P)2PtCl2 complexes. A study by Nelson and
co-workers showed that this difference in bond length is
reflected in the solid state 31P NMR spectra of such molecules.28
A particularly interesting part of structure (15) concerns the
P–N bond lengths and angles. The sum of angles around each
nitrogen was similar (356–360Њ) and planar. The P–N bond
lengths do not show any particular pattern. This was
unexpected as the X-ray structure of trans-carbonyl-chloro-
bis(tripyrrolidinylphosphine)rhodium shows two of the pyrrol-
idine rings to have planar N atoms (sum of angles around
nitrogen = 354 to 360Њ), and one tetragronally distorted N atom
(sum of angles around nitrogen = 347, 350Њ) with a 0.02 Å
longer P–N bond length (av. 1.667(3) Å vs. 1.688(3) Å). There
have been several other crystal structure studies of tris(di-
alkylamino)phosphines such as tris(N-piperidinyl)phosphine,
tris(N-morpholino)phosphine and the selenides, sulfides, and
transition metal complexes of related phosphines.11 Each of
these structures shows two planar nitrogens and one more
tetrahedral nitrogen. We were therefore surprised to find
the crystal structure of (15) not showing this phenomenon.
Husebye and co-workers have alerted us to a similar structural
effect observed in their tellurium complexes derived from
tris(dimethylamino)phosphine selenides.29
The crystal structure of (17) shows a similar co-ordination
environment to complex (15), but with some differences that
may reflect the smaller size (and stronger donor power) that we
anticipated for ligand (4). The Pt–P bond lengths are ca. 0.02 Å
shorter in complex (17) (average 2.239(2) Å vs. 2.258(3) Å). This
is an indication that this phosphine may be binding to the
platinum more strongly than tripyrrolidinylphosphine. In the
crystal structures of other cis-bisphosphine platinum dichloride
complexes, it appears that the Pt–P bond length is dominated
by the steric bulk of the phosphine ligands. Hence, in
(Cy3P)2PtCl2, particularly long (av. 2.294(4) Å) Pt–P bonds are
observed,30 whereas the complexes of the small cone angle
ligands, Me3P and PF3 show particularly short bond lengths
(2.243(10) and 2.142(3) Å respectively).31 It is therefore likely
that the smaller size of methyl(dipyrrolidinyl)phosphine with
respect to tripyrrolidinylphosphine is responsible for the shorter
bond lengths observed. The angle P(1)–Pt(1)–P(2) between the
phosphines is considerably smaller than in complex (15)
(93.90(8)Њ, 93.37(7)Њ vs. 98.2(1)Њ), and is also consistent with
ligand (4) being less sterically demanding than (2).
t
PPh2 Bu
PPh3
P(NMe2)3
confirmation in which repulsive interactions are kept to a
minimum. (It is probably easier to pack two rigid hexagons in
an eclipsed face/edge conformation than a non planar ring with
sp3 hybridised carbon atoms.) The nitrogen atoms within this
structure are all approximately planar (355, 352.5, 359.5, 360Њ).
The crystal structure of (15) reveals different structural
features to the other crystal structure studies of tris(dialkyl-
amino)phosphines. This may suggest that the bonding observed
in these compounds is slightly subtler than we originally sup-
posed, and could also be related to the exact co-ordination
environment of the ligand. The co-ordination environment
provided by the platinum complex is also likely to have an effect
on the structural features observed within the phosphines of
complex (17), especially considering the genuine but relatively
small distortions observed. In the crystal structure of complex
(16), there is a variation between P–N bond lengths, but with
no particular trend with regard to the bond angles. There is
considerable X-ray (and photoelectronic spectral) data on tris-
(dialkylamino)phosphines which supports the idea that one of
the P–N bonds is different in character (less N–P π-donation)
than the other two, our structural data suggest that conclusions
can only be drawn when certain structural phenomena reappear
time after time.
e: Crystallographic cone angles versus Tolman cone angles
The Tolman cone angle of a phosphine R1R2R3P is defined as
2/3 of the sum of half-angles, where half-angles can be measured
from simple molecular models. M–P bond lengths are set at
2.28 Å and the phosphine substituent bond lengths and orient-
ation are set using intuition and perhaps known crystallo-
graphic data. The substituents are arranged to give a minimum
cone angle. The half-angle is that between metal, phosphorus,
and the outer edge of the outermost atom on a phosphorus
substituent. It is a useful estimate of the steric bulk of a ligand.1
It is perhaps unwise to use this method to evaluate very small
differences between ligands. It is an attractive idea to suggest
crystallographic cone angles (calculated by measuring M–P–X,
where X is the middle of the outermost atom on each sub-
stituent) as a replacement for the Tolman value. This has the
advantage of being a real value that is not subject to the
accuracy and/or bias of a model. However, Muller and Mingos
showed that while the mean average of a large number of
crystal structures has a good correlation with the Tolman value,
there is a considerable range throughout the series.32 We have
found that the correlation between Tolman, crystallographic
and molecular modelled (PC Spartan) cone angles (see Table 8)
The crystal structure of (16) is broadly similar to (15). The
angle between the phosphine ligands (P(1)–Pt(1)–P(2) =
96.39(8) Å, see Table 7) is slightly smaller than that found in the
tripyrrolidinylphosphine structure. This may be a result of
the phenyl groups within the structure adopting an eclipsed
1100
J. Chem. Soc., Dalton Trans., 2002, 1093–1103

View Article Online
give a greater proportion of the neutral products.34 (The ionic
compounds are most likely intermediates in the synthesis of
the neutral compounds.) It has been noted that reaction of
CpFe(CO)2I with more electron rich phosphines such as Cy3P
and Bz3P often gives some ionic products in which iodide is
substituted from the co-ordination sphere of the iron centre. In
contrast, the reaction of triphenylphosphine with CpFe(CO)2I
gives an 80% yield of the neutral compound within 1 hour in
refluxing benzene (using ≈3 mol% [CpFe(CO)2]2, see Scheme 6).
Fig. 7 Differences (in Њ) between the Tolman, PC Spartan Pro and
crystallographic (CDS) cone angles.
Table
9
Comparison of νCO of (Cp)Fe(CO)(I)L complexes for
pyrrolidine based ligands with iron complexes of other phosphines
Scheme 6
Entry
L
νCO a/cmϪ1
Ref.
Cone angleb/Њ
We have prepared iron() complexes of type CpFe(CO)-
I(PR3) from two of the N-pyrrolidinylphosphines by the
method of Colville et. al.34 The reaction of these phosphines
with CpFe(CO)2I yields a mixture of insoluble ionic com-
pounds, [CpFe(CO)2(L)]I, which can be filtered off, and the
soluble CpFeCO(L)I complexes which are readily purified by
chromatography using an alumina column.
1
2
3
3
4
5
6
7
(PhO)3P
(EtO)3P
(iPrO)3P
(C6H5)3P
PhMe2P
Bz3P
(1984)
(1961)
(1960)
(1957)
(1950)
(1949)
1941
34
34
34
34
34
34
128
109
130
145
122
165
145
145
(2)
(3)
Tripyrrolidinylphosphine reacts with CpFe(CO)2I under
[CpFe(CO)2]2 catalysis to give, after 90 min at 90 ЊC in toluene, a
3.6 : 1 mixture of ionic to neutral compounds (19) and (20). In
order to obtain a larger yield of neutral compound, phenyl-
(dipyrrolidinyl)phosphine was reacted for three hours at 90 ЊC
in toluene. This gave a 1 : 3.8 mixture of ionic to neutral com-
pounds. In both cases, it is more difficult to prepare the neutral
species than if less electron rich ligands were used. The FAB
mass spectrum of the iron complexes is especially informative
in confirming the formulation. The CpFeCO(L)I compounds
show similar fragmentation patterns in which MHϩ, (M Ϫ
CO)ϩ, (M Ϫ I)ϩ, (M Ϫ I Ϫ CO)ϩ and (M Ϫ CpFe(CO)I) are
detected. These compounds were additionally characterised
by showing good agreement with calculated values in their high
resolution electrospray mass spectra. The ionic complexes,
[CpFe(CO)2(L)]I, show (M Ϫ I)ϩ and (M Ϫ CO Ϫ I)ϩ peaks in
their FAB mass spectrum, and were also characterised by giving
acceptable high resolution electrospray mass spectra. All
1942
a Chloroform or toluene solution. b Cone angle data are taken from
ref. 1. c This work.
is reasonable for trialkylphosphines but less good for those
containing P–F, P–Cl or P–H groups (Fig. 7).
In spite of all this, it is perhaps still useful to glance at crystal-
lographic cone angles of similar metal complexes as these do
represent what a given phosphine will ‘look like’ when com-
plexed to that metal fragment (see Table 9). This may be a more
useful measure when a series of metal complexes (e.g. trans-
L2Rh(CO)Cl) have been used to catalyse or undergo a reaction.
The crystallographic data in this case are likely to be valid sets
of information that may shed light on the reactivity patterns
involved. The crystallographic cone angles in complexes of type
trans-L2Rh(CO)Cl are: L = PPh3, 111.9Њ; L = p-Tol3P, 106.4Њ; L
= Me3P, 80.3Њ; L = Ph2MeP, 97.6Њ. This can be compared with
that found in the two structures reported here, which give
crystallographic cone angles of 125Њ and 124Њ for (13) L =
di-tert-butylpyrrolidine and (14), L = Cy3P respectively. The
Tolman cone angle predicted for these two phosphines is 170Њ.
1
the iron complexes were additionally characterised by H and
31P NMR, and IR spectroscopy, and the purity of the neutral
complexes was further confirmed by acceptable chemical
analyses. Although the position of νCO in the IR spectrum is not
generally used as a quantitative measure of phosphine basicity,
both neutral complexes show ν(CO) at significantly lower wave-
number than in the iron complexes of Ph3P, PhMe2P or Bz3P.
The co-ordination chemical shift is greater in the neutral
complexes, [for L = (2), ∆δ = 31.8 in (19), 12.7 in (20); L = 3,
∆δ = 60.5 in (21), 46.6 in (22)].
f: Iron carbonyl complexes of pyrrolidinyl phosphines
It is well known that iron complexes of type CpFe(CO)(L)I and
[CpFe(CO)2(L)]I can be prepared from phosphines and
[CpFe(CO)2]2.33 The proportion of ionic and neutral complexes
varies depending on the nature of the phosphines and reaction
conditions. In 1987, Coville and co-workers demonstrated that
[CpFe(CO)2]2 can efficiently catalyse this reaction, and hence
Another method to prepare cationic iron complexes of type
[CpFe(CO)2(L)]X is the reaction of a phosphine with the labile
precursor [CpFe(CO)2(MeCN)]BF4 (Scheme 7).15
J. Chem. Soc., Dalton Trans., 2002, 1093–1103
1101

View Article Online
Table 10 Selected bond lengths (Å) and angles (Њ) for (24)
Fe(1)–P(1)
Fe(1)–C(17)
Fe(1)–C(18)
Fe(1)–C(19)
P(1)–N(1)
2.226(1) Fe(1)–C(40)
2.085(4) Fe(1)–C(30)
2.086(4) Fe(1)–C(21)
2.101(4) Fe(1)–C(20)
1.650(3) P(1)–N(6)
1.130(5) C(40)–O(40)
1.773(4)
1.772(4)
2.095(4)
2.108(4)
1.655(3)
1.135(5)
C(30)–O(30)
C(30)–Fe(1)–C(40)
C(40)–Fe(1)–P(1)
O(30)–C(30)–Fe(1)
C(10)–N(6)–P(1)
C(2)–N(1)–C(5)
C(2)–N(1)–P(1)
96.7(2)
89.9(1)
178.2(4)
126.4(3)
109.6(3)
128.4(3)
C(30)–Fe(1)–P(1)
93.0(1)
177.6(4)
120.8(3)
109.4(3)
122.0(3)
O(40)–C(40)–Fe(1)
C(7)–N(6)–P(1)
C(10)–N(6)–C(7)
C(5)–N(1)–P(1)
Scheme 7
This type of synthesis is more straightforward for cationic
compounds as there is no possibility for neutral iron dicarbonyl
complexes being formed. Refluxing either tripyrrolidinylphos-
phine or phenyl(dipyrrolidinyl)phosphine with (CpFe(CO)2-
(MeCN)]BF4 in dichloromethane rapidly gave the desired
complexes [(23) and (24) respectively] in high purity.
A few brown crystals of (24) could be grown by layering
a CH2Cl2 solution with diethyl ether. The molecular structure
of (24) is shown in Fig. 8 and consists of discrete cations and
tri-alkylphosphine) is often reluctant to behave in a similar
fashion to ‘normal’ phosphines, the somewhat smaller cone
angles of (4) and (5) should broaden their applications further.
In any case these studies have led us to a greater understanding
of the donor strength of phosphino-amines which we are
now applying to functionalised phosphines which are showing
considerable promise in catalysis.25
Acknowledgements
We are grateful to the EPRSC for financial support and the use
of the EPRSC National Mass Spectrometry Service (Swansea).
We would also like to thank the JREI for equipment grants.
References
1 C. A. Tolman, Chem Rev., 1977, 77, 313.
2 Md. M. Rahman, H.-Y. Lie, K. Eriks, A. Prock and W. P. Giering,
Organometallics, 1989, 8, 1 and refs. therein.
3 K. G. Molloy and J. L. Petersen, J. Am. Chem. Soc., 1995, 117, 7696.
4 M. F. Ernst and D. M. Roddick, Inorg. Chem., 1989, 28, 1624.
5 M. C. Simpson and D. J. Cole-Hamilton, Coord. Chem. Rev., 1996,
155, 163 and refs. therein.
6 T. Sakakura, T. Sodeyama, K. Sasaki, K. Wada and M. Tanaka,
J. Am. Chem. Soc., 1990, 112, 7221 and refs. therein.
7 J. K. MacDougall, M. C. Simpson, M. J. Green and
D. J. Cole-Hamilton, J. Chem. Soc., Dalton Trans., 1996, 1161 and
refs. therein.
Fig. 8 Molecular structure of (24).
8 A. F. Littke and G. C. Fu, J. Org. Chem., 1999, 64, 10; A. F. Littke
and G. C. Fu, Angew. Chem,. Int. Ed., 1999, 38, 2411.
9 J. P. Wolfe, R. A. Singer, B. H. Yang and S. L. Buchwald, J. Am.
Chem. Soc., 1999, 121, 9550.
10 B. C. Hamann and J. F. Hartwig, J. Am. Chem. Soc., 1998, 120,
7370.
11 A. H. Cowley, M. Lattman, P. M. Stricklen and J. G. Verkade, Inorg.
Chem., 1982, 21, 453; C. Romming and Jon Songstad, Acta Chem.
Scand., Ser. A, 1978, 32, 689; C. Romming and Jon Songstad, Acta
Chem. Scand., Ser. A, 1979, 33, 187; C. Romming and Jon Songstad,
Acta Chem. Scand, Ser. A., 1982, 36, 665.
12 A portion of this work has appeared in preliminary form:
M. L. Clarke, D. J. Cole-Hamilton, A. M. Z. Slawin and
J. D. Woollins, Chem. Commun., 2000, 2065.
13 H. C. Clark and L. E. Manzer, J. Organomet. Chem., 1973, 59, 411.
14 T. s Piper and G. Wilkinson, J. Inorg. Nucl. Chem., 1956, 2, 38.
15 H. Schumann, M. Speis, W. P. Bosmann, J. M. M. Smits and
P. T. Beurkens, J. Organomet. Chem., 1991, 403, 165.
16 SHELXTL structure solution and determination package, Bruker
AXS, Madison, WI, 1999.
anions. The cationic iron centre displays octahedral piano stool
geometry, with the Cp ligand occupying three faces of this
octahedron. Two of the carbons within the Cp ligand show
slightly elongated bonds lengths, which reflects their positions
trans to carbon monoxide ligands which have a higher trans
effect relative to the phosphine ligand. The carbon monoxide
ligands are linear and show similar Fe–C bond lengths to that
found in the crystal structure of [CpFe(CO)2PPh3]ϩ (1.773(4) Å
in (24), vs. 1.776 Å).35 The iron–phosphorus bond length
[2.226(1) Å, see Table 10] is somewhat shorter than that found
in the analogous triphenylphosphine complex [2.242(1) Å].
Since PPh3 and phenyl(dipyrrolidinyl)phoshpine are thought
to have a similar cone angle, it is likely that the shorter P–Fe
bond arises from stronger bonding from the more electron rich
ligand. The P–N bonds within the structure are fairly typical
for this type of ligand, and have planar nitrogen atoms (sum of
angles at N = 360(3) and 356.6(3)Њ).
17 J. D. Woollins, J. Chem. Soc., Dalton Trans, 1996, 2893;
A. M. Z. Slawin, M. B. Smith and J. D. Woollins, Chem. Commun.,
1996, 2095; P. Bhattacharyya, A. M. Z. Slawin, M. B. Smith,
D. J. Williams and J. D. Woollins, J. Chem. Soc., Dalton Trans, 1996,
3647.
18 (a) R. D. Kroshefsky, R. Weiss and J. G. Verkade, Inorg. Chem.,
1979, 18, 469; (b) P. A. W. Dean, Can. J. Chem., 1979, 18, 469.
19 S. Vastag, B. Heil and L. Marko, J. Mol. Catal., 1979, 5, 189.
20 R. L. Harlow, S. A. Westcott, D. L. Thorn and R. T. Baker, Inorg.
Chem., 1992, 31, 323.
21 D. de Montauzon and R. Poilblanc, J. Organomet. Chem., 1975, 93,
397.
22 W. Strohmeier, W. Rehder-Strirnweiss and G. Reischig,
J. Organomet. Chem., 1971, 17, 393.
Summary and conclusions
In this work, we have shown the consequences of an esoteric
structural phenomenon on the donor strength of phosphino-
amines. It is hoped that two of the phosphines prepared in the
course of this study [(4) and (5)], will find applications in
organometallic chemistry and catalysis. These phosphines have
a different combination of electronic and steric properties from
those reported before, with tert-butyl(dipyrrolidinyl)phosphine
being amongst the most electron donating phosphines
known. As tri-tert-butylphosphine (the most electron rich
1102
J. Chem. Soc., Dalton Trans., 2002, 1093–1103

View Article Online
23 F. Ozawa, T. Ito, Y. Nakamura and A. Yamamoto, Bull. Chem. Soc.
Jpn., 1981, 54, 1868.
28 J. A. Rahn, L. Baltuis and J. H. Nelson, Inorg. Chem., 1990, 29,
750.
24 Further support for this proposal comes from the rhodium complex
of the less bulky diethyl(pyrrolidinyl)phosphine which we have
prepared and analysed by IR and NMR spectroscopy but failed to
isolate. The position of ν(CO) (1952 cmϪ1) in this complex is at a
lower position than the di-tert-butyl substituted ligand which would
be very unexpected unless the degree of N P donation differed
markedly between the two complexes.
25 M. L. Clarke, D. J. Cole-Hamilton and J. D. Woollins, J. Chem. Soc.,
Dalton Trans., 2001, 2721.
26 A special example of a phosphine that is very basic (towards
protons) is the commercially available Verkade superbase.
M. A. Laramay and J. G. Verkade, J. Am. Chem. Soc., 1990, 112,
9421.
29 M. D. Rudd, S. V. Lindeman and S. Husebye, Acta Chem. Scand.,
1996, 50, 759.
30 T. S. Cameron, H. C. Clark, A. Linden, A. M. Nicholas and
M. J. Hampden-Smith, Inorg. Chim. Acta, 1989, 162, 9.
31 W. M. Attia, G. Balducci and M. Calligaris, Acta Crystallogr., Sect.
C, 1987, 43, 1053.
32 T. E. Muller and D. W. P. Mingos, Transiion. Met. Chem., 1995, 20,
533.
33 Gmelin Handbook of Inorganic Chemistry; Organic compounds
B11/B12, Springer Verlag, Berlin, 1983/1984.
34 N. J. Colville, E. A. Darling, A. W. Hearn and P. Johnson,
J. Organomet. Chem., 1987, 328, 375.
35 G. A. Sim, D. I. Woodhouse and G. R. Know, J. Chem. Soc., Dalton
Trans., 1979, 629.
27 K. R. Dunbar and S. C. Haefner, Inorg. Chem., 1992, 31, 2676.
J. Chem. Soc., Dalton Trans., 2002, 1093–1103
1103