A new class of rhodium(I) κ.1-P and κ2-P, N complexes with rigid PTN(R) ligands (PTN = 7-phospha-3-methyl-1,3,5- triazabicyclo[3.3.1]nonane)

Article abstract of DOI:10.1021/om051099r

The open cage ligands PTN(R) (PTN = 7-R-phospha-3-methyl-1,3,5- triazabicyclo[3.3.1]nonane; R = Me, Ph), derived from cage cleavage of the water-soluble phosphine PTA (PTA = 1,3,5-triaza-7-phosphaadamantane), were used to synthesize neutral [RhCl(cod)(PTN(R)] (10) and [RhI(CO){PTN-(Me)}] (12) and cationic [Rh(cod){PTN(R)}][BAr₄^F] (11) rhodium(I) complexes, which were comprehensively characterized both in solution and in the solid state (cod = 1,5-cyclooctadiene; BAr₄^F = B[3,5-(CF₃)₂C₆H₃]₄). Two modes of coordination to the metal were identified, κ¹-P (for 10) and κ2-P,N (11, 12). For κ¹-P coordination, the free N donor site rests in close proximity to the Rh center imparted by the rigid PTN framework. The complexes were tested as catalysts for (biphasic) olefin hydroformylation and regioselective C=C bond reduction under transfer hydrogenation conditions, together with acetophenone hydrogenation. DFT calculations on simplified models helped to assess the bonding properties within the complexes and to determine the amount of cage strain that accompanies the structural modifications of the ligand when transformed from a κ¹-P to κ²-P,N coordination mode.

Full text of DOI:10.1021/om051099r

Organometallics 2006, 25, 2189-2200
2189
A New Class of Rhodium(I) K1-P and K2-P,N Complexes with Rigid
PTN(R) Ligands (PTN )
7-Phospha-3-methyl-1,3,5-triazabicyclo[3.3.1]nonane)
Andrew D. Phillips,† Sandra Bolan˜o,† Sylvain S. Bosquain,† Jean-Claude Daran,‡
Raluca Malacea,‡ Maurizio Peruzzini,*,† Rinaldo Poli,*,‡ and Luca Gonsalvi*,†
Consiglio Nazionale delle Ricerche, Istituto di Chimica dei Composti Organometallici (ICCOM-CNR),
Via Madonna del Piano 10, 50019 Sesto Fiorentino (Firenze), Italy, and Laboratoire de Chimie de
Coordination (UPR CNRS 8241), 205 Route de Narbonne, 31077 Toulouse Cedex 4, France
ReceiVed December 29, 2005
The open cage ligands PTN(R) (PTN ) 7-R-phospha-3-methyl-1,3,5-triazabicyclo[3.3.1]nonane; R )
Me, Ph), derived from cage cleavage of the water-soluble phosphine PTA (PTA ) 1,3,5-triaza-7-
phosphaadamantane), were used to synthesize neutral [RhCl(cod)(PTN(R)] (10) and [RhI(CO){PTN-
(Me)}] (12) and cationic [Rh(cod){PTN(R)}][BArF ] (11) rhodium(I) complexes, which were compre-
4
hensively characterized both in solution and in the solid state (cod ) 1,5-cyclooctadiene; BArF4 ) B[3,5-
(CF3)2C6H3]4). Two modes of coordination to the metal were identified, κ1-P (for 10) and κ2-P,N (11,
12). For κ1-P coordination, the free N donor site rests in close proximity to the Rh center imparted by
the rigid PTN framework. The complexes were tested as catalysts for (biphasic) olefin hydroformylation
and regioselective CdC bond reduction under transfer hydrogenation conditions, together with
acetophenone hydrogenation. DFT calculations on simplified models helped to assess the bonding properties
within the complexes and to determine the amount of cage strain that accompanies the structural
modifications of the ligand when transformed from a κ1-P to κ2-P,N coordination mode.
(e.g., four atom) chain chelate8 or usually by inserting a phenyl,9
pyridyl,10 ferrocenyl,11 or other functionalities containing π-bond-
ing into the ligand backbone.12
Introduction
During the past decade, ligand design for homogeneous
catalysts has advanced from simple bidentate diphosphines and
diamines to mixed donor atom moieties, such as phosphino-
ethers, phosphino-thioethers, and aminophosphines.1 These
catalytic systems have been applied to hydroformylation,
hydrogenation, and hydrosilylation reactions.2-4 Numerous
configurations of phosphorus-nitrogen (P,N) ligands have been
reported and can be generically divided into two categories,
namely, those based on structurally rigid sp2 nitrogen centers,
e.g., unsaturated pyridine5 and imine6 groups, or sp3 nitrogen
centers, which constitute saturated amine functionalities.7,8 For
sp3 N-based ligands, rigidity is enforced either using a short
Among aminophosphine ligands, the neutral water-soluble
phosphine 1-phospha-3,5,7-triazaadamantane (PTA, 1)13 was
successfully used to prepare κ1-P-coordinated rhodium catalysts
capable of regioselective hydrogenation14 and hydroformyla-
tion15 of olefins in aqueous media. Long-term catalyst stability
and low activity are two of the most problematic aspects for
systems based on PTA;16 thus a shift from a monodentate to a
bidentate hemilabile ligand system featuring similar stereoelec-
tronic properties and water solubility could be of interest. On
this basis we have considered the use of the bowl-shaped alkyl
(8) (a) For example: (a) Simonsen, K. P.; Suzuki, N.; Hamada, M.;
Kojima, M.; Ohba, S.; Saito, Y.; Fujita, J. Bull. Chem. Soc. Jpn. 1989, 62,
3790. (b) Dahlenburg, L.; Gotz, R. J. Organomet. Chem. 2001, 619, 88.
(c) Kashiwabara, K.; Morikawa, A.; Suzuki, T.; Isobe, K.; Tatsumi, K. J.
Chem. Soc., Dalton Trans. 1997, 1075. (d) Berger, H.; Nesper, R.; Pregosin,
P. S.; Ruegger, H.; Worle, M. HelV. Chim. Acta 1993, 76, 1520.
(9) (a) Kostas, I. D.; Screttas, C. G. J. Organomet. Chem. 1999, 585, 1.
(b) McKay, I. D.; Payne, N. C. Can. J. Chem. 1986, 64, 1930.
(10) Clarke, M. L.; Slawin, A. M. Z.; Wheatley, M. V.; Woollins, J. D.
J. Chem. Soc., Dalton Trans. 2001, 3421.
(11) (a) Butler, I. R.; Cullen, W. R.; Rettig, S. J. Organometallics 1986,
5, 1320. (b) Nishibayashi, Y.; Segawa, K.; Arikawa, Y.; Ohe, K.; Hidai,
M.; Uemura, S. J. Organomet. Chem. 1997, 545, 381.
(12) For example, an interesting zwitterionic indenide-Rh complex has
been recently reported. Stradiotto M.; Cipot, J.; McDonald, R. J. Am. Chem.
Soc. 2003, 125, 5619.
(13) Phillips, A. D.; Gonsalvi, L.; Romerosa, A.; Vizza, F.; Peruzzini,
M. Coord. Chem. ReV. 2004, 248, 955.
(14) Smolenski, P.; Wajda-Hermanowicz, K. J. Organomet. Chem. 1998,
570, 1395.
(15) Pruchnik, F.; Smolenski, P.; Galdecka, E.; Galdecka, Z. New J.
Chem. 1998, 22, 1395.
(16) (a) Cermak, J.; Kvicalova, M.; Blechta, V. Collect. Czech. Chem.
Commun. 1997, 62, 355. (b) Cermak, J.; Kvicalova, M.; Blechta, V.
Proceedings ISHC-10; Princeton, NJ, 1996 (abstract B54).
* To whom correspondence should be addressed. E-mail: l.gonsalvi@
iccom.cnr.it, mperuzzini@iccom.cnr.it, poli@lcc-toulouse.fr.
† Consiglio Nazionale delle Ricerche.
‡ Laboratoire de Chimie de Coordination (UPR CNRS 8241).
(1) Slone, C. S.; Weinberger, D. A.; Mirkin, C. A. Prog. Inorg. Chem.
1999, 48, 233-350.
(2) van Leeuwen, P. W. N. M.; Casey, C. P.; Whiteker, G. T. Rhodium
Catalysed Hydroformylation; van Leeuwen, P. W. N. M., Claver, C., Eds.;
Kluwer Academic Publishers: Dordrecht, The Netherlands, 2000.
(3) Agbossou, F.; Carpentier, J.-F.; Mortreux, A. Chem. ReV. 1995, 95,
2485.
(4) Fache, F.; Schulz, E.; Tommasino, L. M.; Lemaire, M. Chem. ReV.
2000, 100, 2159.
(5) For example: (a) Newkome, G. R. Chem. ReV. 1993, 93, 2067. (b)
Anderson, M. P.; Casalnuovo, A. L.; Johnson, B. J.; Mattson, B. M.;
Mueting, A. M.; Pignolet, L. H. Inorg. Chem. 1988, 27, 1649. (c)
Jaaskelainen, S.; Haukka, M.; Riihimaki, H.; Pursiainen, J. T.; Pakkanen,
T. A. J. Organomet. Chem. 2004, 689, 1064.
(6) For example: (a) Ghilardi, C. A.; Midollini, S.; Moneti, S.; Orlandini,
A.; Scapacci, G. J. Chem. Soc., Dalton Trans. 1992, 3371. (b) Marxen, T.
L.; Johnson, B. J.; Nilsson, P. V.; Pignolet, L. H. Inorg. Chem. 1984, 23,
4663.
(7) Lee, H.-S.; Bae, J.-Y.; Ko, J.; Kang, Y. S.; Kim, H. S.; Kim, S.-J.;
Chung, J.-H., Kang, S. O. J. Organomet. Chem. 2000, 614, 83.
10.1021/om051099r CCC: $33.50 © 2006 American Chemical Society
Publication on Web 03/30/2006

2190 Organometallics, Vol. 25, No. 9, 2006
Phillips et al.
Table 1. Solution 31P NMR Data for 2, 10, 11, and 12a
Chart 1
compound
δ (ppm)
1JRhP (Hz)
solvent
2a
2b
-91.8b
-77.1b
-48.2c
-43.9c
-40.2c
-37.4c
-26.5c
C6D6
CDCl3
C6D6
10a
10b
11a
11b
12
151.2c
152.3c
143.5c
142.3c
154.4c
C6D6
CD2Cl2
CD2Cl2
CD2Cl2
Chart 2
a All signals are referenced with respect to 85% aqueous H3PO4.
b Reference 23. c This work.
Results and Discussion
Syntheses and Characterizations of Rh(I) Complexes with
PTN(R) Ligands. PTN(R) ligands were obtained by reductive
cleavage of P-functionalized PTA salts (see Scheme 1).22,23 The
choice of the substituent on P was limited to R ) Me and Ph,
as negligible yields are obtained for R ) benzyl, cyclohexyl,
and tert-butyl due to nonselective and inefficient cleavage.22
We synthesized and fully characterized two new classes of
Rh(I) 1,5-cyclo-octadiene (COD) complexes, featuring 2 in a
monodentate κ1-P or chelate κ2-P,N coordination mode. The
preparation of the monodentate PTN(R) complexes from [Rh-
(cod)Cl]2 is straightforward (Scheme 2) and affords the products
[RhCl(cod){PTN(R)}] (R ) Me, 10a; R ) Ph, 10b) as air- and
moisture-stable yellow solids in high yields.
or aryl P-substituted 7-phospha-3-methyl-1,3,5-triazabicyclo-
[3.3.1]nonanes abbreviated as PTN(R) (2), synthesized originally
by Schmidbaur et al.17
The bridgehead position of 2 anchors the ligand conformation,
with the constrained curvature creating a preset bidentate
capability. This structural feature is also advantageous in a
nonchelating situation where the inflexible framework ensures
that the amine functionality remains always in close proximity
to the transition metal center, a feature that has been suggested
to facilitate the intramolecular base-assisted heterolytic cleavage
of H2.18 Other examples of chelated (κ2-P,N)-Rh complexes with
aminophosphine (3),19 aminophosphinite (4),20 and aminophos-
phinite (5)21 have been reported.
The coordination chemistry of 2 has so far received scarce
attention, including Au and Mo complexes (Chart 3). Depending
on the nature of the ancillary substituent attached to Au, 2 adopts
either a mono- or bidentate coordination mode,22 and examples
of both were obtained as [κ1-P-AuCl{PTN(R)}] (6 R ) Me,
Ph), [κ1-P-AuCl{PTN(Me)}2] (7), and [κ2-P,N-AuMe2{PTN-
(Me)}][AuCl4] (8), the latter exhibiting high water solubility.
A second group of PTN(R) complexes containing molybdenum
have also been reported by Schmidbaur and co-workers.
Although no crystals suitable for X-ray diffraction studies were
obtained, in all cases, the κ2-P,N coordination mode of 2 in
[Mo(CO)4{PTN(R)}] (9, R ) Me, Ph) is inferred from the
number and intensity of IR-active CO stretching frequencies.22
Complex 10a is readily soluble in chlorinated solvents,
toluene, and benzene and moderately soluble in water with
S(25 °C) ) 3.80 mg mL-1, at a value comparable to [RhCl-
(PTA)3]24 and [Rh(acac)(CO)(PTA)].14 In contrast, the presence
of the phenyl group on P significantly reduces the aqueous
solubility for 10b, S(25 °C) ) 0.74 mg mL-1. Toluene solutions
of 10 instantly produce black precipitates when heated beyond
100 °C, whereas of a sample of 10a in D2O heated to 80 °C
showed no change in the 31P{1H} spectrum after 5 h. Routine
EI-MS characterization is reported in the Experimental Section.
Complexes 10 were reacted with 1 equiv of NaB(C6H3-
(CF3)2)4 (abbreviated as NaBArF )25 in dichloromethane to afford
4
κ2-P,N complexes [Rh(cod){PTN(R)}](BArF ) (R ) Me, 11a;
4
R ) Ph, 11b, Scheme 3), which were isolated as air-sensitive
bright yellow compounds.
The neutral, air- and moisture-sensitive rhodium complex κ2-
P,N-[RhI(CO){PTN(Me)}] (12) was obtained directly by react-
ing the dimer Rh2(CO4)I2 with 2 equiv of 2a in dichloromethane
or toluene and stirring for 1 h at room temperature, as shown
in Scheme 4. The water solubility of 12 [S(25 °C) ) 0.2 mg
L-1] is remarkably decreased from that of 11a, as observed for
other neutral carbonyl complexes.
In this article we report the synthesis and spectroscopic and
structural data together with a computational analysis of 2 ligated
to rhodium(I) centers, in both a mono- and bidentate coordina-
tion mode. Finally, results are given for catalytic tests using
10-12 in olefin hydroformylation and transfer hydrogenation
of benzylidene acetone and acetophenone in organic and
biphasic organic/water solvent systems.
The values for 31P{1H} NMR chemical shifts for 10-12,
summarized in Table 1, are in the normal range of other four-
coordinate Rh complexes of the type RhX2L2.26 For 12, 31P
NMR shows a chemical shift of -26.0 ppm in CD2Cl2, which
is the highest field signal for all Rh-PTN(R) compounds
reported.
1H NMR chemical shifts are summarized in Table 2. 1H-1H
COSY and 1H-13C HETCOR experiments were used to provide
definitive assignments of the CH2 groups of 2 in 11a and 11b.
(17) Assmann, B.; Angermaier, K.; Schmidbaur, H. J. Chem. Soc., Chem.
Commun. 1994, 941.
(18) Andrieu, J.; Richard, P.; Camus, J.-M.; Poli, R. Inorg. Chem. 2002,
41, 3876.
(19) Faissner, R.; Huttner, G.; Kaifer, E.; Rutsch, P. Eur. J. Inorg. Chem.
2003, 1681.
(20) Kuznetsov, V. F.; Facey, G. A.; Yap, G. P. A.; Alper, H.
Organometallics 1999, 18, 4706.
(21) (a) Bondoux, D.; Mentzen, B. F.; Tkatchenko, I. Inorg. Chem. 1981,
20, 839. (b) Pradat, C.; Riess, J. G.; Bondoux, D.; Mentzen, B. F.;
Tkatchenko, I.; Houalla, D. J. Am. Chem. Soc. 1979, 101, 2234. (c) Wachter
J.; Jeanneaux, F.; Le Borgne, G.; Reiss, J. G. Organometallics 1984, 3,
1034.
(23) Fluck, E.; Weissgraber, H. J. Chem. Ztg. 1977, 101, 204.
(24) Darensbourg, D. J.; Joo´, F.; Kannisto, M.; Katho, A.; Reibenspies,
J. H.; Daigle, D. J. Inorg. Chem. 1994, 33, 200.
(25) Brookhart, M.; Grant., B.; Volpe, A. F. Organometallics 1992, 11,
3920.
(26) Verkade, J. G.; Mosbo, J. A. In Phosphorus-31 NMR Spectroscopy
in Stereochemical Analysis; Verkade, J. G., Quin, L. D., Eds.; VCH: New
York, 1987; Chapter 13.
(22) Assmann, B.; Angermaier, K.; Paul, M.; Riede, H.; Schmidnaur,
H. Chem. Ber. 1995, 128, 891.

Rhodium(I) κ1-P and κ2-P,N Complexes
Organometallics, Vol. 25, No. 9, 2006 2191
1
Table 2. Solution H NMR Data (δ, ppm, TMS referenced) for 2, 10, 11, and 12
compound
CH2(COD)
CH(COD)
CH3-N
CH2 (PTN(R))
R-P (2JPH, Hz)
2aa
n/a
n/a
1.95-2.34
2.25-2.57
n/a
1.89-2.32
2.09-2.56
n/a
n/a
1.97 s
2.23 s
2.22 s
2.39 s
2.86 s
2.42 s
2.45 s
3.36-4.02
3.65-4.05
3.02-4.26
3.77-4.53
3.65-4.40
3.43-4.550
3.56-4.65
0.93 d (5.40)
2ba
7.33-7.67
10ab
11aa
12a
5.73 br s
5.26 br s
n/a
5.77 br s
5.30 br s
3.56 br s
1.14 dd (6.60, 2.40c)
1.16 d (9.90)
1.50 dd (10.46, 2.55c)
7.31-8.19
10bb
11ba
7.28-7.60
c 3
a In CD2Cl2 solution. b In C6D6 solution.
J
.
RhH
Chart 3
Scheme 1
Scheme 2
In 11b, the CH2 signals of COD (protons HX and HY, Figure
1) show a complicated splitting pattern. Both the H and 13C
NMR show that one set of the CH groups belonging to COD
(HY, Figure 1) is shifted extremely upfield, δ(1HY) ) 3.559
ppm (cf. δ(1HX) ) 5.303 ppm) and δ(13C) ) 70.95 (cf. 109.12)
ppm. These effects are attributed to the ring current generated
by the π-system of the phenyl substituent and its direct face-on
positioning with respect to the (CH)Y group of COD, which is
evident from the crystal structure of 11b (vide infra). The
chemical shift of the P-CH2 groups of 2 (HA and HB) are less
affected (and HC to an even lesser extent), as observed by a
different orientation from the phenyl group.
1
Scheme 3
For 12, the CO stretching frequency was observed at 1989
cm-1 in CH2Cl2 and at 1974 cm-1 in KBr, identical to that found
in (P,sp2-N)-chelates of Rh(I) including [RhCl(CO){P(Ph)2-
CPPh2dN(p-CNC6F4)}].27
X-ray Crystal Structure Determinations. Suitable crystals
of 10, 11, and 12 were grown, and the corresponding X-ray
crystal structure data were collected. Selected bond lengths and
angles for all structures are provided in Table 3, while Table 8
(Experimental Section) summarizes the main crystallographic
data.
Scheme 4
All compounds feature a square planar geometry, two
coordination positions being defined either by the centers of
the C-C bonds of COD or, in the case of 12, by iodide and
CO ligands (Figure 3). Compounds 10 contain a coordinated P
donor and a dangling N(Me) group, the fourth coordination
position being occupied by the chloride ligand, whereas 2 adopts
the κ2-P,N coordination mode in compounds 11 and 12 in
keeping with the spectroscopic solution data.
n
From the magnitudes of the JPC it was possible to assign the
CH2 groups adjacent to phosphorus (Figure 1). For the
resonances related to the CH and CH2 groups of COD and 2,
there are significant differences between 11a and 11b, reflecting
the rigidity of the PTN(R) framework, as apparent from Figure
2.
(27) Katti, K. V.; Santarsiero, B. D.; Pinkerton, A. A.; Cavell, R. G.
Inorg. Chem. 1993, 32, 5919-5925.

2192 Organometallics, Vol. 25, No. 9, 2006
Table 3. Selected Bond Lengths (Å) and Angles (deg) for Compounds 10, 11, and 12
Phillips et al.
parameter
10a
10ba
11a
11b
12
Rh-P
2.3073(5)
2.3739(5)
2.323(1)
2.394(1)
2.334(1)
2.368(1)
2.229(2)b
2.239(1)
2.206(1)
Rh-Cl
Rh-I
2.711(5)
2.219(3)
Rh-N
3.401(2)
2.206(2)
2.199(2)
2.108(2)
2.107(2)
1.374(3)
1.409(3)
3.055(2)
2.189(4)
2.179(4)
2.107(4)
2.106(4)
1.385(6)
1.407(6)
3.226(2)
2.191(4)
2.179(4)
2.116(4)
2.098(4)
1.368(6)
1.407(6)
2.204(3)b
2.229(5)
2.230(5)
2.146(5)
2.132(5)
1.353(8)
1.376(7)
2.213(3)
2.226(4)
2.280(4)
2.155(3)
2.130(3)
1.352(6)
1.406(6)
Rh-C
(trans to P)
Rh-C
(trans to Cl or N)
CdC (trans to P)
CdC (trans to Cl or N)
Rh-C(O)
Rh(C)-O
P-Cexo
1.803(5)
1.138(5)
1.829(2)
1.875(2)
1.863(2)
1.455(2)
1.474(2)
1.465(2)
87.58(2)
1.829(4)
1.870(4)
1.864(4)
1.465(4)
1.475(4)
1.472(4)
90.22(4)
1.828(4)
1.873(4)
1.864(4)
1.452(4)
1.466(4)
1.452(4)
89.07(4)
1.775(5)b
1.800(5)b
1.805(5)b
1.525(6)b
1.570(6)b
1.577(6)b
1.813(4)
1.844(4)
1.853(4)
1.498(5)
1.512(5)
1.517(5)
P-Cendo
1.844(4)
1.840(5)
1.481(5)
1.510(5)
1.521(5)
N-Cexo
N-Cendo
Cl-Rh-P
I-Rh-N
P-Rh-C(O)
N-Rh-P
Cl-Rh-N
C-Rh-C
98.7(1)
90.6(2)
84.1(1)
80.99(2)
95.05(2)
81.7(2)
62.65(4)
81.69(4)
81.1(2)
60.35(4)
60.33(4)
81.9(2)
83.3(1)b
84.1(1)
80.0(2)
79.9(2)
80.5(1)
80.5(1)
81.3(2)
81.8(2)
80.8(2)
Cl-Rh-P-N
Cl-Rh-C-Pexo
-97.6(2)
82.2(2)
-80.5(1)
93.1(1)
-81.7(1)
98.8(1)
a The two sets of parameters refer to the crystallographically independent molecules in the unit cell. b Due to a statistical disorder between the terminal
P and N atoms in this structure, the corresponding bond distances and angles in the P,N-ligand are less reliable.
Table 4. Comparison of Mayer Bond Indices for the
PTN(Me) Ligand, Free (2a) and Ligated through
P-Coordination (13) and P,N-Chelation (14, 15)a
parameter
2a
13
14
15
P-CH3
P-CH2
0.698
0.721
0.997
0.880
0.868
0.945
0.926
0.922
0.924
0.909
0.889
1.016
1.024
0.879
0.825
0.015
1.075
0.634
0.693
0.908
0.911
0.906
0.915
0.981
0.996
0.909
0.908
0.940
1.201
0.222
1.009
0.738
P-CH2-Ni
Ni-CH2-Ni
Ni-CH2-Nt
Ni-CH2
0.828
0.922
0.872
1.041
0.881
0.908
0.971
0.940
Figure 1. Labeling scheme of protons in the [Rh(COD)(PTN(R)]+
complexes 11a and 11b. The ring current effects of the phenyl group
Nt-CH3
Rh-P
Rh-Nt
0.801
n/a
n/a
0.939
1.379
0.234
1
in 11b significant alter the H chemical shifts of HA, HB, HC, and
HY.
a For nitrogen labeling see Figure 5.
PTN(Me) moiety, the hydrogen assuming an axial position with
respect to the Rh center. The Rh‚‚‚N distances (see Table 3)
are approximately 0.4-0.8 Å shorter than the sum of van der
Waals radii for Rh and N (3.800 Å), suggesting the presence
of a weak interactions between the Rh and N centers, due to
the rigid ligand conformation and possibly shortened through
crystal packing.
In 10b, the majority of substituent orientations are identical
among the two isomers present. The primary difference is in
the rotation of the phenyl group, with the Rh-P-Cipso-Cortho
torsion angle at -12.09(5)° versus -29.40(5)°; in the former
the phenyl group is almost coplanar with the bridgehead C and
P and terminal N atoms, while for the second conformer it is
almost eclipsed with the P-C bond of 2b. An analysis of the
crystal structure apparently showed no indications of disorder.
The presence of the two different conformers is likely due to a
small energy rotation barrier for the C(Ph)-P bond. The crystal
packing forms a pseudo coordination polymer (see Figure 4)
where the units are sequentially linked through chloride-PTN
interactions ranging from 2.734 to 2.910 Å. The different
1
Figure 2. Comparison of the H NMR spectra of 11a and 11b
demonstrating the effects of the phenyl group in 11b.
Complex 10a shows intermolecular hydrogen bonding (3.045
Å) between the rhodium center and the N-methyl group of the

Rhodium(I) κ1-P and κ2-P,N Complexes
Organometallics, Vol. 25, No. 9, 2006 2193
Figure 3. ORTEP diagrams of complexes RhCl(COD)(PTN(R)) (R ) Me, 10a; Ph, 10b), the cationic fragments [Rh(COD)(PTN(R))]+
(R ) Me, 11a; Ph, 11b), and the neutral complex [RhI(CO)(PTN(Me))] (12). All diagrams are drawn with 50% probability ellipsoids.
Hydrogen labels have been omitted for clarity. For 10b, only one isomer is shown.
number of Cl‚‚‚H (of the PTN fragment) interactions between
units results in a variation in Rh-Cl bond length and phenyl
group orientation within the individual conformers.
The Rh‚‚‚N distances in 10a and 10b are significantly shorter
than the distance between Rh and the dangling NEt2 group in
the [RhCl(CO){Ph2PCH(o-C6H4NMe2)CH(Ph)NHPh}] complex
(3.66 Å).18 The ligand is oriented such that it lies with the P,N
vector orthogonal to the plane defined by the Rh, Cl, and P
centers. The Cl-Rh-P-N torsion angle is larger for 10a
(97.62(8)°) than for 10b (86.52(15)°, 81.71(16)°). The Rh-P
distance is slightly shorter for 10a, 2.3073(5) Å, than for 10b,
2.334(1), 2.323(1) Å. For 10a, the shorter Rh-P bond results
in a longer Rh‚‚‚N interaction (3.401(2) Å) and a shorter Rh-
Cl bond length (2.374(1) Å). Both the longer P-Rh and shorter

2194 Organometallics, Vol. 25, No. 9, 2006
Phillips et al.
Figure 4. Left: Packing diagram (section) of 10b highlighting the intermolecular interactions between hydrogen atoms of PTN(Ph) and
chloride atoms. Right: Packing diagram (section) of 12 demonstrating the axial intermolecular interactions between the PTN(Me) and
rhodium center.
Cl-Rh bond lengths may suggest that 2b is a slightly weaker
electron donor than 2a.
11b is orthogonal to the plane defined by P, Rh, and N, and it
is faced toward one side of COD. This observation is consistent
with the NMR analysis showing downfield chemical shifts and
distributed resonances associated with COD and PTN(R)
substituents (see above). This demonstrates that the phenyl group
is maintained in an identical fixed position for the solution and
in the solid state. Another interesting aspect of the structures
of 11 is the significant increase in the P-Cendo and N-Cendo
bonds (see Table 3), suggesting that 2 has been stretched in
order to accommodate a chelating coordination mode. Analo-
gous for compounds 10a and 10b, the Rh-C bonds are longer
and the CdC bond is longer for the atoms of the COD ligand
located trans to phosphorus.
A survey of the Cambridge structural database28 shows to
date 31 structures containing rhodium, COD, a nonbridging
chloride, and a monocoordinating phosphine, the simplest being
[RhCl(cod)(PPh3)],29 having an identical Rh-P bond distance
(2.308(2) Å) and a slightly longer Rh-Cl bond (2.361(2) Å)
compared to 10a and 10b. However, comparison of the P-trans
Rh-CH(cod) distances shows that 10a and 10b have shorter
distances than in [RhCl(cod)(PPh3)] (2.205(7) and 2.238(7) Å)
and correspondingly a longer P-trans C(H)dC(H) bond (cf.
[RhCl(cod)(PPh3)] 1.340(10) Å).29 The stronger π-back-donation
to COD indicates that the PTN(R) ligand is a stronger donor
than PPh3.
In contrast to 11a, 12 does not have disordering of the P,N
donor sites due to the specific influences of the trans-effects
caused by CO and I. Thus, the structure of 12 is significantly
more reliable in terms of metric parameters associated with the
chelated PTN ligand than 11a. However, as a general observa-
tion, 11a features shorter Rh-P and Rh-N bond lengths than
12, which is consistent with stronger bonding associated with
a cationic Rh center in the former. In comparison to RhCl(CO)
complexes with 5, 12 has longer Rh-P and Rh-N bond lengths,
combined with a slightly wider P-Rh-N bond angle; moreover,
the Rh-C(O) distance is shorter in 12 than in RhCl(CO)
complexes with 5. This suggests that 2 is a weaker ligand than
5 and, probably, more hemilabile. The internal PTN framework
of 12 demonstrates some shortening of the P-Cendo and N-Cendo
bonds, which is unresolved in the COD analogue. The packing
of the unit cell is similar to the other Rh-PTN complexes,
containing weak intermolecular contacts between hydrogens of
2. The most interesting property is an axial interaction between
Rh and H3A (2.920 Å), which is continuously repeated as an
extended polymeric structure (Figure 4). A related example was
reported in the cis-[PdCl2(PTA)2] structure, where two strong
Pd‚‚‚H(CH) axial interactions (2.86 Å) form a dimeric unit.30
In 11 the cation and anion fragments are well separated from
each other, the closest interactions being between H(PTN(R))‚‚‚
F(BArF ) at 2.449 Å for both R ) Me and Ph. The formation
4
of the Rh-N bond is accompanied by a shorter Rh-P bond
(see Table 3), i.e., approximately 0.1 Å shorter than that of the
parent compounds 10. This result is consistent with the stronger
ligand coordination characteristic of cationic rhodium(I) com-
plexes.
The P-Rh-N bite angle for 11a and 11b (ca. 84°) is midway
between the six- (86.7°) and five-membered systems (82.8°).
A comparison of 10a and 11a with [Au{PTN(Me)}Me2][AuMe2-
Cl2] reveals the P-M-N bond angles are identical22 within
experimental error at a value of 82.7(5)°.
Due to the statistical disorder of the chelating P and N atoms
within 11a (described in the X-ray Structure Determination
section), it is inconclusive whether the P-Rh and N-Rh
distances are different from those of 11b. The phenyl plane in
(28) (a) Cambridge Structural Database version 5.26 (February 2005).
(b) Allen, F. H. Acta Crystallogr. B 2002, 58, 380. (c) Orpen, A. G. Acta
Crystallogr. B 2002, 58, 398.
(29) Horn, Q. L.; Jones, D. S.; Evans, R. N.; Ogle, C. A.; Masterman,
T. C. Acta Crystallogr. E 2002, 58, m51.

Rhodium(I) κ1-P and κ2-P,N Complexes
Organometallics, Vol. 25, No. 9, 2006 2195
Figure 5. DFT-calculated geometries for 13, 14 (representing simpler models of 10a and 11b), and 15 (unsimplified model of 12). Unlabeled
atoms refer to carbon. Nt and Ni denote terminal and internal nitrogen centers, respectively. Blue-colored bonds show those that are strengthened
compared to the reference model, the uncoordinated PTN(Me) 2a. Likewise, red bonds indicate those that have weakened compared to 2a.
Figure 6. Left: HOMO of PTN(Me) 2a demonstrating the antibonding repulsions between the phosphorus and nitrogen atoms. Right:
SHOMO of RhPTN(Me), 13, showing the orbital interaction between Rh dxy and the σ*(P-C) of the PTN ligand, which is partially responsible
for weakening the P-C bonds. The other metal-ligand bonds have been omitted for clarity.
Computational Analysis. A density functional theory in-
vestigation on 2a and model complexes of 10a and 11a was
performed in order to elucidate the electronic effects associated
with P-coordination to Rh including the changes induced
through κ2-P,N chelation. The complete 2a ligand was optimized
in Cs symmetry in the lowest energy configuration with exo-
substituents on both P and N, which is the most suitable
conformer for coordination. The energy-minimized model
exactly matches the exo,exo-geometry observed in the disordered
crystal structure reported by Schmidbaur and co-workers.22
Complexes 10a and 11a were simplified by replacing COD by
two ethylene groups, yielding the model complexes [RhCl-
(C2H4)2{PTN(Me)}], 13, and [Rh(C2H4)2{PTN(Me)}]+, 14.
Complex 12 was modeled (15) without any simplifications and
is shown in Figure 5. In general, the optimized geometric
parameters of 13, 14, and 15 accurately represent the corre-
sponding values experimentally observed for 10a, 11a, and 12
within experimental error, especially those pertaining to the
ethylene groups, 2a, the P-Rh-Cl angle (for 13), the P-Rh-N
bite angle (for 14), and the torsion angle N-P-Rh-Cl. Minor
overestimations were observed for the Rh-P and Rh-Cl bonds
(+0.02 and +0.01 Å, respectively). A large variation, up to 7°,
is noted with the C-Rh-C bond angles, probably caused by
the replacement of the cyclic COD by two less constrained
ethylene ligands.
Reoptimization in C1 resulted in a skewing of the ethylene
groups within the square planar coordination geometry by 11.2°,
as observed in the crystal structure of 11a, yielding a stable
minimum. Hence the distortion has an electronic origin related
to rhodium coordination and does not result from the COD.
The skewing energy is trivial and was calculated to be only 0.3
kcal mol-1, which is in agreement with the fluxional behavior
1
observed in solution. Indeed, the H NMR spectra of 11 show
only a single resonance for HX and HY. Model 15 was optimized
with Cs symmetry analogous to the crystal structure 12, and no
imaginary frequencies were observed.
The HOMO of free 2a consists of an out-of-phase combina-
tion of lone pairs associated with the three N atoms and a major
contribution from the phosphorus lone pair (Figure 6). As in
the case of the tightly constrained PTA, the intra-atomic
repulsions between the two internal N centers and the terminal
nitrogen and phosphorus centers are caused by the rigidly pinned
structure of 2a, thus resulting in a unusually higher energy
expected for the HOMO.31 This is in line with the experimental
observation that PTN is methylated at phosphorus rather than
at nitrogen, as expected on the basis of pure hard-soft rules.22
A single-point calculation consisting of only the PTN(Me)
fragment with the geometry fixed in the coordinating position
observed in complexes 13 and 14 shows a 1.74 and 11.27 kcal
mol-1 difference in total energy from that of 2a. Moreover, the
energy of the HOMO increased by 7.31 and 10.92 kcal mol-1
,
The optimization of 14 in Cs symmetry, with the ethylene
groups bisected by the N-Rh-P plane, yielded a saddle point
respectively. A large portion of this energy change can be
with a corresponding single imaginary frequency of -60.2 cm-1
.
(31) Phillips, A. D.; Peruzzini, M.; Mealli, C. EURESCO Conference,
Sant Feliu de Guixols (Spain), September 4-9, 2004, Book of Abstracts,
Poster 58, and unpublished results.
(30) Desiraju, G. R.; Steiner, T. The Weak Hydrogen Bond; Oxford
University Press: New York, 1999.

2196 Organometallics, Vol. 25, No. 9, 2006
Phillips et al.
associated with an increase in bond strain when the internal
framework of the ligand is stretched for metal coordination,
especially high in the κ2-P,N chelating mode.
Mayer bond indices32 are particularly useful for measuring
bond strain energy rather than bond lengths, which in the case
of the PTN(Me)-chelated complexes are unreliable due to site
disorder between the coordinating P and N centers. Using the
free 2a as a reference point, continuing to P-coordination and
finally to the P,N-chelate mode, the differences between 2a and
13, 14, and 15 are presented in Table 4. It can be noticed that
the P-coordination of PTN(Me) in 13 induces an overall bond
strengthening within the ligand, which can be attributed mainly
to a substantial reduction in internal interatomic repulsion
between P and N centers. In contrast, upon P,N-chelation in
the cationic Rh complex 14, a bond strength reduction occurs
particularly for P-CH2 and Nt-CH2 and is counterbalanced
by a strengthening of the Nt-CH3 and Pt-CH3 bonds. The
bridgehead core component of the PTN ligand in all model
complexes remains invariant to the mode of coordination and
the oxidation state of the metal. Hence, when the PTN ligand
is forced into a chelate mode to satisfy the electrophilic demand
of the metal in 14 and 15, both ends of the ligand are stretched
to wrap around the Rh center. In comparison, the interaction
between Rh and Nt is considerable weaker than that of the Rh-
P. The increase of the latter from 13 to 14 does not affect the
π-back-bonding to the P-trans ethylene ligand, as indicated by
an almost identical CdC Mayer bond index for both species.
Also expected is the increase of the Rh-P Mayer index on going
from 14 to 15, paralleling the bond shortening. The interaction
between the Nt and Rh centers in 13 is extremely weak and
practically nonexistent, as suggested by the Mayer bond index
value.
An examination of the MO structure of 14 reveals a
significant σ*(P-CH2) contribution to the occupied SHOMO
(second highest occupied MO), which is characterized by an
interaction with the Rh dxy orbital; see Figure 6. This type of
occupation contributes significantly to the weakening of the
P-CH2 bonds, as evidenced by the Mayer bond index. Certainly,
the bond stretching associated with the chelation mode combines
with the σ-electron-withdrawing effect caused by the presence
of the nitrogen atoms to lower the energy of the σ*(P-CH2)-
containing MO, which in turns promotes increased π-back-
donation from Rh.
In summary, the DFT studies reveal significant changes to
the PTN ligand on passing from the P-mono- to the P,N-
chelating-coordination mode. Overall, it is evident that the PTN
ligand has a preference for maintaining P-coordination to Rh.
In contrast, a higher energy cost, in terms of bond strain, is
associated with P,N-chelation, which suggests that the ligand
may show hemilabile behavior in the presence of polar solvents
or other types of nucleophiles. Comparison of the Mayer bond
indices for Rh-P and Rh-N bonds in cationic (14) and neutral
(15) Rh(I) complexes indicates a stronger interaction with PTN
in the latter complex, implying a higher tendency for hemilabile
behavior for 14.
Figure 7. Reaction profiles for styrene hydroformylation in CHCl3
at 55 °C. Run 1 (0), [Rh(COD)Cl]2/2a ) 1:2; run 2 (O), [Rh-
(acac)(COD)]/2a ) 1:1; run 3 (4), [Rh(acac)(COD)]/2b ) 1:1.
Table 5. Styrene Hydroformylation under Homogeneousa
and Biphasic Conditionsb
compJLlex conv (%) branched (%) linear (%)
l/b
EtPh (%)
10aa
10ab
10ba
10bb
11aa
11bb
12b
99.3
57.7
88.4
71.0
80.8
99.9
40.6
95.4
52.9
84.5
66.9
74.7
95.3
39.5
3.7
2.3
3.7
3.2
5.5
4.7
0.1
0.039
0.043
0.044
0.047
0.074
0.049
0.003
0.2
1.9
0.2
0.9
0.5
0.0
1.0
a Styrene: 2.5 mmol; cat.: 0.01 mmol; H2:CO (1:1), 600 psi; CH3Cl:
40 mL; 55 °C; 6 h. b As above, except H2O/n-octane (2:1), 60 °C.
Cl-containing phosphine adducts are less suitable precatalysts
than other Rh salts, because of the generation of strong acidity
(HCl) under catalytic conditions.2 However, the presence of
external bases such as amines is often an efficient remedy, and
the use of hemilabile aminophosphine ligands has recently
received considerable attention.18,34 The positive effect of the
nitrogen donor basicity and proximity to the metal center has
been demonstrated.34b Therefore, a comparison of the activity
for the Cl- and acac- precursors in the presence of ligands 2
was considered of interest.
Under mild hydroformylation conditions (CO/H2 600 psi;
CHCl3, 100 mL; 55 °C), using the complexes made in situ, the
COD ligand will likely be replaced by CO. The runs were
monitored by sampling the reaction mixtures, and the conversion
profiles are shown in Figure 7. The systems tested showed
comparable efficiency (complete conversion after 8 h), chemose-
lectivity (ca. 90% total aldehyde yield), and regioselectivity (b/l
ratio ca. 91:9), without detectable traces of ethylbenzene among
the (unidentified) byproducts. No induction period was observed
for the conversion of styrene. The conversion profiles are very
similar, suggesting that the active species must be very close
in nature for all the systems tested (Figure 7). Conversely, a
blank test carried out with [RhCl(cod)]2 without 2 under identical
experimental conditions showed that an induction period is
required for aldehyde formation. These results are significant
when compared with those recently reported in the presence of
similar P,N ligands under essentially identical conditions, which
showed no activity (TOF < 1 h-1) with PPh3 from the chloride
Olefin Hydroformylation Tests. Complexes 10, 11, and 12
were tested as catalysts for styrene and 1-hexene hydroformyl-
ation and transfer hydrogenation of benzylideneacetone (BZA)
and acetophenone using HCO2Na/H2O or KOH/PriOH. The
catalytic tests were carried out either using complexes formed
in situ by addition of the ligands to either [RhCl(cod)]2 or [Rh-
(acac)(cod)]33 or by introducing the preformed complexes in
the reaction mixture. It is well known that [RhCl(cod)]2 and
(33) (a) Fornika, R.; Six, C.; Go¨rls, H.; Kessler, M.; Kru¨ger, C.; Leitner,
W. Can. J. Chem. 2001, 79, 642. (b) Yamamoto, Y.; Fujita, M.; Miyaura,
N. Synlett. 2002, 767. (c) Tanaka, M.; Hua, R. Pure Appl. Chem. 2002, 74,
181.
(34) (a) Camus, J. M.; Andrieu, J.; Richard, P.; Poli, R.; Baldoli, C.;
Maiorana, S. Inorg. Chem. 2003, 42, 2384. (b) Andrieu, J.; Camus, J.-M.;
Richard, P.; Poli, R.; Gonsalvi, L.; Vizza, F.; Peruzzini, M. Eur. J. Inorg.
Chem. 2006, 51. (c) Andrieu, J.; Camus, J. M.; Balan, C.; Poli, R. Eur. J.
Inorg. Chem. 2006, 62.
(32) Mayer, I. Int. J. Quantum Chem. 1986, 29, 477.

Rhodium(I) κ1-P and κ2-P,N Complexes
Organometallics, Vol. 25, No. 9, 2006 2197
Table 6. 1-Hexene Hydroformylation under Homogeneous
Conditionsa
Styrene hydroformylation tests were then repeated in the
presence of the preformed aminophosphine complexes, and the
final conversion and yields (no sampling) are reported in Table
5, both under strictly homogeneous (chloroform) and biphasic
(n-octane/water) conditions. High conversions and chemose-
lectivity to the branched aldehydes (>94%) were achieved as
expected for rhodium complexes, and only traces of hydroge-
nated product (ethylbenzene, <0.5%) were determined. Control
runs with [Rh(cod)Cl2]2 showed that in CHCl3 at 55 °C a slightly
higher l/b ratio is obtained at complete conversion (0.11,
corresponding to 88.9% branched and 9.9% linear aldehyde),
whereas in water/octane at 60 °C a higher amount of ethylben-
zene is produced (5.9% against 0.1 in CHCl3) accompanied by
a small increase in regioselectivity (l/b 0.08).
Preliminary mechanistic 31P HPNMR studies in sapphire tube
conditions showed that all phosphorus species are present in
solution as doublets due to coupling with Rh during and after
exposing CHCl3 solutions of the complexes to CO/H2 pressure
(600 psi) and prolonged heating to 55 °C, regardless of the
presence of styrene.35
Further hydroformylation studies were carried out using
1-hexene as substrate, and the results are summarized in Table
6. As expected, the linear aldehyde is the favored product, with
l/b ratios ranging from 1.35 (10b in CHCl3, 4 h) to 3.13 (10a
in THF, 1 h). The best conversion was achieved with 10b (99%),
although a larger amount of isomerized alkenes was found by
GC analysis of the reaction mixture after the run. The composi-
tion of the syngas mixture has only a slight effect on the final
conversion and yields, both being higher in the presence of a
1:1 CO/H2 mixture. The yield of isomerization is not largely
affected by this variable. Remarkably, our systems afford very
little amounts of hydrogenation products, except in the case
when CHCl3 is used as solvent at 55 °C instead of THF at 60
°C, where alcohols were also accounted for among the products.
The previously reported hydroformylation of 1-hexene with
the [Rh(acac)(CO)(PTA)] complex,36 carried out in water,
resulted in an l/b ratio and TON in the same range as our set of
catalysts. The catalytic styrene hydroformylation results are
comparable with those reported by Andrieu et al.18,34b using Ph2-
PCH(o-C6H4CH2NHPh) (γ-P,N) as ligand, for which HPNMR
studies showed that under catalytic conditions it behaves as a
conv %
(h)
linear
%
branched
%
isom
%
hydrog
%
catalyst
l/b
10a
10a
10ab
10ac
10b
10b
10bc
11a
11a
11b
20.2(1)
61.5(4)
83.0(4)
92.5(4)
63.4(1)
99.3(4)
82.5(4)
27.4(1)
70.0(4)
52.6(1)
12.2
31.9
43.5
50.7
33.6
48.3
33.5
14.8
35.8
29.8
3.9
13.4
20.9
26.2
12.3
26.1
24.9
5.5
3.13
2.38
2.08
1.94
2.73
1.85
1.35
2.69
2.27
2.64
4.0
16.1
18.4
15.5
17.4
20.9
19.9
7.0
0.1
0.1
0.2
0.1
0.1
4.0
4.2d
0.1
0.8
0.2
15.8
11.3
17.6
11.3
a 1-Hexene: 2.5 mmol; cat.: 0.01 mmol; H2:CO (2:1), 600 psi; THF:
30 mL; 60 °C; 4 h. b CO/H2 (1:1), 600 psi. c CHCl3: 30 mL, 55 °C, CO/
H2 (1:1), 600 psi. d Including alcohols.
Table 7. Hydrogen Transfer Tests for the Reduction of
BZAa and Acetophenoneb
selectivity
average
TON
catalyst
substrate
BZA
acetophenone
BZA
acetophenone
BZA
acetophenone
BZA
conv (%) % Ac % Bc % Cc
10a
10a
10b
10b
11a
11a
11b
11b
59.9
69.3
96.4
57.7
59.3
22.3
98.6
99.8
1.4
0.2
1.2
0.0
60
347
97
288
60
112
100
87
100
100
100
100
98.8
100
17.4
>99.9
acetophenone
a BZA:catalyst:NaHCO2 (100:1:1000); MeOH (3 mL); H2O (3 mL);
80 °C, 5 h. b Conditions: catalyst:acetophenone:KOH (1:500:5); PriOH
(10 mL); 80 °C, 6 h. c GC values based on pure samples. A ) 4-phenyl-
2-butanone; B ) 4-phenyl-3-buten-2-ol; C ) 1-phenyethanol.
precursor.34b That study has demonstrated that the TOF is
sensitive to the distance between the N and P functions for
amines of comparable basicity, proving that the proximity of
the basic amine to the catalytic center has a beneficial role in
the catalytic rate-determining step. Only the species with ligand
Ph2PCH2NMe2, having the most basic amine in closest proxim-
ity to the phosphine donor, was a similar activity observed for
the chloride and acac catalytic precursors,34b as observed in the
present case for ligands 2.
Table 8. Selected Crystallographic Data for Rh(PTN(R))Cl(cod) (R ) Me, 10a; R ) Ph, 10b),
Rh(cod)(PTN(R))][B(3,5-(CF3)2C6H3)4] (R ) Me, 11a; R ) Ph, 11b), and RhI(PTN(Me))CO, 12
10a
10ba
11a
11b
12
formula
fw, g mol-1
color, habit
cryst syst
space group
a, Å
C15H28ClN3PRh
419.73
yellow, needle
Pbca
orthorhombic
15.3268(7)
18.0089(9)
12.3321(6)
90
90
90
3403.9(3)
8
C40H60Cl2N6P2Rh2
963.60
yellow, prism
P21/c
monoclinic
9.9872(6)
12.8682(8)
32.2080(19)
90
99.792(5)
90
4079.0(4)
4
C48H42BF24N3PRh
1332.44
yellow, prism
P1h
triclinic
13.3739(16)
14.1424(15)
16.9699(19)
105.601(9)
98.540(10)
114.454(11)
2687.7(5)
2
C52H42BF24N3PRh
1309.58
yellow cube
P1h
C8H16N3OPRhI
431.02
orange prism
Pbca
orthorhombic
12.5944(12)
12.1571(9)
17.0443(14)
90
90
90
2609.7(4)
8
triclinic
12.2076(13)
14.7591(14)
16.6244(18)
73.349(12)
70.373(12)
88.792(12)
2693.6(5)
2
b, Å
c, Å
R, deg
â, deg
γ, deg
V, Å3
Z
cryst dimens, mm
0.14 × 0.0 × 0.02
0.23 × 0.21 × 0.06
0.4 × 0.26 × 0.11
0.2 × 0.16 × 0.12
0.2 × 0.16 × 0.12
d
calc, g cm-1
1.638
1.569
1.646
1.615
2.194
µ, mm-1
1.252
34 414
5753
0.0272
1.056
28 917
7997
0.0403
0.568
19 023
9688
0.0500
0.470
26 840
9859
0.0402
3.778
6852
2952
0.0344
no. of measd reflns
no. of unique reflns
R1 [I > 2σ(I)]
wR2, all
0.0579
0.0723
0.1120
0.1036
0.0840
S (GOF)
0.915
0.929
0.898
1.021
1.082
no. of params
192
471
861
851
138
a Two crystallographically independent molecules are present in the unit cell.

2198 Organometallics, Vol. 25, No. 9, 2006
Phillips et al.
purified by washing three times with a 10% aqueous solution of
NaOH followed by three washings with distilled water. After drying
and filtration with MgSO4, the styrene was twice trap-to-trap
distilled under static vacuum and degassed for 30 min with a stream
monodentate phosphine, and this may imply a hemilabile nature
of the PTN ligands in these reactions.
Transfer Hydrogenation Tests. The catalytic activity of Rh
complexes in transfer hydrogenation reactions is well docu-
mented.37 Complexes 10 and 11 were tested in the transfer
hydrogenation of benzylidene acetone (BZA) and acetophenone.
Very high chemoselectivity to CdC bond hydrogenation was
obtained with all complexes in the transfer reduction of BZA
using NaHCO2 as hydrogen source in water (Table 7). Interest-
ingly, a ligand effect was observed on the conversion of BZA
using 2a instead of 2b, regardless of the coordination mode of
the ligand. The highest conversions (>95%) were observed in
the presence of 2b. Hydrogen transfer reduction of acetophenone
was achieved using the KOH/PriOH protocol (Table 7). Interest-
ingly, good conversions (57-69%) were achieved using the
monocoordinated Rh(I) neutral complexes 10, while moderate
conversions were obtained with the cationic κ2-P,N-chelated
complexes 11. At higher substrate/catalyst ratio (1000), hydro-
gen transfer to acetophenone still occurred albeit with lower
conversion (20%, TON 198). Cationic complexes 11 generally
displayed TON values similar to the cationic aminophosphine
rhodium precatalyst described by Gao et al.38 Control experi-
ments using [Rh(cod)Cl2]2 showed that BZA conversion was
negligible under reaction conditions described in Table 7.
of argon. [RhCl(cod)]2 and [Rh2I2(CO)4]40 were prepared in
39
accordance with standard literature procedures. The aminophos-
phines 2a,b were synthesized with a slightly modified procedure
from that reported in the literature.17,22 NaBArF was synthesized
4
as described by Brookhart et al.25 All other reagents (technical
grade) were purchased from commercial sources and used as
received. The NMR spectra were recorded using a Bruker AC-200
operating at 200 MHz, a Varian VXR instrument operating at 300
MHz, or a Bruker Avance DRX operating at 400 MHz. Deuterated
dichloromethane was dried with activated 4 Å molecular sieves
and C6D6 with stirring over Na-K alloy. Both solvents were
degassed prior to sample preparation. Chemicals shifts for 1H and
13C are referenced with respect to external TMS, and 31P was
referenced to external aqueous 85% H3PO4. FTIR spectra were
measured on a Perkin-Elmer Spectrum BX instrument using KBr
pellets or solution cells. Elemental combustion microanalyses (C,
H, N) were obtained using a Perkin-Elmer 2400 series II elemental
analyzer. Mass spectra were obtained with a Nermag R10-10 mass
spectrometer with the electron-impact technique. GC analyses were
performed on a Shimadzu GC-14 A gas chromatograph equipped
with a flame ionization detector and a 30 m (0.25 mm i.d., 0.25
µm film thickness) SPB-1 Supelco fused silica capillary column.
GC/MS analyses were performed on a Shimadzu QP 5000 apparatus
equipped with a column identical with that used for GC analysis.
Synthesis of [RhCl(cod){PTN(R)}] (10). In a dry and N2-purged
50 mL Schlenk flask were introduced 0.493 g (1.00 mmol) of
[RhCl(cod)]2 and 5 mL of toluene. After 5 min of stirring, 2.00
mmol (2a, 0.346 g; 2b, 0.471 g) dissolved in 10 mL of toluene
was added dropwise over a period of 30 min. After 1 h of stirring
at room temperature, the volume of the solution was reduced to 5
mL under reduced pressure. Addition of 8 mL of n-pentane resulted
in the formation of a yellow powder, which was collected by
decanting off the mother liquor, washed with three 2 mL portions
of n-pentane, and dried overnight under vacuum.
Conclusions
In summary, a new class of Rh(I) neutral and cationic
complexes bearing the water-soluble aminophosphine cage
ligand PTN have been synthesized, showing both κ1-P and κ2-
P,N coordination modes. All the new PTN complexes were fully
characterized by conventional spectroscopic methods in solution
and by X-ray diffractometry in the solid state. The geometry of
the ligand was correlated to the stability of chelation mode
versus monocoordination by DFT calculations, suggesting that
the straining constraints imposed by the rigid PTN ligand favors
κ1-P coordination to Rh. The PTN ligand may behave as
hemilabile under forcing olefin hydroformylation conditions and
keep their κ1-P or κ2-P,N coordination mode during milder
transfer hydrogenation of activated olefins and ketones. Mecha-
nistic studies are currently in progress to verify this hypothesis.
R ) Me (10a). Yield: 0.748 g (93%). Anal. Calcd (found) for
C15H28N3PRh: C, 42.92 (43.02); H, 6.72 (6.32); N, 10.01 (9.70).
Melting point (dec): 172 °C. 1H NMR (299.96 MHz, 25 °C,
2
3
C6D6): δ(ppm) 1.14 (dd, 3H, JPH ) 6.6 Hz, JRhH ) 2.4 Hz,
P-CH3), 1.94 (m, 4H, CH2(COD)), 2.22 (s, 3H, N-CH3), 2.34 (br
s, 4H, CH2(COD)), 3.04 (m, 2H, P-CHH-N), 3.08 (m, 2H, N-CHH-
N), 3.56 (m, 1H, N-CHH-N), 3.85 (m, 1H, N-CHH-N), 3.92 (m,
2H, P-CHH-N), 4.23 (m, 2H, N-CHH-NMe), 5.72 (br s, 4H, CH2-
(COD)). 13C NMR (75.43 MHz, 25 °C, C6D6): δ(ppm) 28.95 (s,
CH2(COD)), 33.43 (s, CH2(COD)), 38.86 (s, N-CH3), 50.45 (d,
1JPC ) 14.6 Hz, P-CH3), 66.96 (d, 1JPC ) 13.6 Hz, P-CH2-N), 70.71
Experimental Section
General Procedures. Manipulations of all products and reagents
were carried under a purified N2 or Ar atmosphere with standard
Schlenk techniques. All glassware was predried overnight, and
subsequently several purge/refill cycles were performed before the
introduction of solvents or reagents. Toluene and n-pentane were
dried by reflux over sodium, while dichloromethane was refluxed
with CaH2. All solvents were degassed prior to use. Styrene was
3
(d, JPC ) 10.8 Hz, N-CH2-N), 76.47 (s, N-CH2-NMe), 98.98 (s,
CH(COD)). 31P NMR (121.42 MHz, 25 °C, C6D6): δ(ppm) -48.21
1
(d, JRhP ) 151.2 Hz). MS (EI): m/z 419 (M+, 14.3%), 311
(M+ - COD, 28.8%), 268 (M+ - COD - CH2NCH3, 11.6%),
232 (M+ - COD - CH2NCH3 - HCl, 10.3%).
(35) Bosquain, S. S.; Phillips, A. D.; Gonsalvi, L.; Peruzzini, M.
Unpublished results.
(36) Pruchnik, F. P.; Smolenski, P.; Wajda-Hermanowicz, K. J. Orga-
nomet. Chem. 1998, 570, 63.
R ) Ph (10b). Yield: 0.863 g (89.6%). Anal. Calcd (found) for
C20H30N3PRh: C, 49.86 (49.82); H, 6.28 (6.12); N, 8.72 (8.31).
Melting point (dec): 173 °C. 1H NMR (299.96 MHz, 25 °C,
C6D6): δ(ppm) 1.90 (m, 4H, CH2(COD)), 2.34 (br s, 4H, CH2-
(COD)), 2.42 (s, 3H, N-CH3), 3.44 (m, 2H, P-CHH-N), 3.92 (m,
1H, N-CHH-N), 4.13 (m, 2H, P-CHH-N), 4.53 (m, 2H, N-CHH-
(37) (a) Toros, S.; Heil, B.; Kollar, L. J. Organomet. Chem. 1980, 197,
85. (b) Kollar, L.; Toros, S.; Heil, B. J. Organomet. Chem. 1980, 192, 253.
(c) Spogliarich, R.; Kaspar, J.; Graziani, M. J. Organomet. Chem. 1986,
306, 407. (d) Kvintovics, P.; Bakos, J.; Heil, B. J. Mol. Catal. 1985, 32,
111. (e) Mestroni, G.; Zassinovich, G.; Alessio, E. J. Mol. Catal. 1989, 49,
175. (f) Kvintovics, P.; James, B. R.; Heil, B. J. Chem. Soc., Chem.
Commun. 1986, 1810. (g) Leitner, W.; Brown, J. M.; Brunner H. J. Am.
Chem. Soc. 1993, 115, 152. (h) Thorpe, T.; Blacker, J.; Brown, S. M.
Tetrahedron Lett. 2001, 42, 4041. (i) Pamies O.; Backvall J. E. Chem. Eur.
J. 2001, 7, 5052. (j) Nindakova, L. O.; Shainyan, B. A.; Belonogova, L. N.
Russ. J. Org. Chem. 2003, 39, 1484.
N
Me), 5.77 (br s, 4H, CH2(COD)), 7.34 (m, 4H, o-,m-C6H5), 8.16
3
(t, JHH ) 7.8 Hz, 1H, p-C6H5). 13C NMR (75.43 MHz, 25 °C,
C6D6): δ(ppm) 29.01 (s, CH2(COD)), 33.17 (s, CH2(COD)), 39.39
(s, N-CH3), 49.32 (d, JPC ) 10.71 Hz, P-CH2-N), 68.05 (s, CH-
1
(39) Giordano, G.; Crabtree, R. H. Inorg. Synth. 1979, 19, 218.
(40) Fulford, A.; Hickey, C. E.; Maitlis, P. M. J. Organomet. Chem.
1990, 398, 311.
(38) Gao, J.-X.; Yi, X.-D.; Xu, P.-P.; Tang, C.-L.; Wan, H.-L.; Ikariya,
T. J. Organomet. Chem. 1999, 592, 290.

Rhodium(I) κ1-P and κ2-P,N Complexes
Organometallics, Vol. 25, No. 9, 2006 2199
3
(COD)), 71.08 (d, JPC ) 9.60 Hz, N-CH2-N), 76.10 (s, N-CH2-
dissolved with 5 mL of dry and degassed toluene. To this solution,
36.7 g (0.21 mmol) of PTN(Me) in 5 mL of toluene was slowly
added, and after 1 h stirring at ambient temperature, an orange
precipitate was obtained. After removal of the mother liquor, the
remaining solid was washed with three 2 mL portions of pentane
and dried under reduced vacuum overnight. Yield: 0.055 g (64%).
Anal. Calcd (found) for C8H16N3OIPRh: C, 22.29 (22.45); H, 3.74
1
NMe), 97.77 (s, CH(COD)), 129.53 (s, p-C6H5), 131.42 (d, JPC
)
85.01 Hz, i-C6H5), 132.75 (d, 3JPC ) 9.81 Hz, m-C6H5), 139.56 (d,
2JPC ) 15.61 Hz, o-C6H5). 31P NMR (121.42 MHz, 25 °C, C6D6):
δ(ppm) -43.90 (d, JRhP ) 152.3 Hz). MS (EI): m/z 481 (M+,
1
0.85%), 373 (M+ - COD, 10.1%), 330 (M+ - COD - CH2NCH3,
9.8%), 294 (M+ - COD - CH2NCH3 - HCl, 12.2%).
1
Synthesis of [Rh(cod){PTN(R)}][BArF ] (11). In a dry and N2-
(3.89); N, 9.75 (9.70). H NMR (400.13 MHz, 25 °C, CD2Cl2):
4
2
3
δ(ppm) 1.56 (dd, 3H, JPH ) 10.5 Hz, JRhH ) 2. 6 Hz, P-CH3),
purged 50 mL Schlenk flask were introduced 0.50 mmol of chloride
precursor (0.210 g, 10a; 0.241 g, 10b) 0.432 g (0.50 mmol) of
NaBArF , and 10 mL of dichloromethane. After overnight stirring
at ambient temperature, the resulting cloudy solution was filtered
to give a clear yellow solution. The volume of the solution was
reduced to 4 mL under reduced pressure. The addition of 8 mL of
n-pentane yielded a yellow powder, which was collected by
decanting off the mother liquor, washed with three 2 mL portions
of pentane, and dried overnight under vacuum.
2
2.86 (s, 3H, N-CH3), 3.67 (br d, 2H, JHH ) 12.8 Hz, N-CHH-
2
NMe), 3.74 (m, 1H, JHH ) 8.6 Hz, N-CHH-N), 3.78 (m, 1H,
4
2JHH ) 8.6 Hz, N-CHH-N), 4.04 (m, 2Ha + 2Hb, P-CHaHb-N), 4.39
(d, 2H, JHH ) 13.0 Hz, N-CHH-NMe). 13C NMR (100.63 MHz,
3
1
25 °C, CD2Cl2): δ(ppm) 29.68 (s, N-CH3), (d, JPC ) 22.5 Hz,
P-CH3), 56.25 (d, 1JPC ) 30.2 Hz, P-CH2-N), 70.79 (d, 3JPC ) 9.1
Hz, N-CH2-N), 79.42 (s, N-CH2-NMe) 189.91 (CO). 31P NMR
1
(121.42 MHz, 25 °C, CD2Cl2): δ(ppm) -25.98 (d, JRhP ) 154.0
Hz). IR: νCO (cm-1) ) 1974 (KBr), 1989 (CH2Cl2).
R ) Me (11a). Yield: 0.487 g (78.2%). Correct analytical data
could not be obtained (low C, H, N), probably because of
contamination by NaCl (the presence of sodium was confirmed by
X-ray Structure Determinations. Large transparent yellow
crystals suitable for single-crystal X-ray diffraction were obtained
under air- and moisture-free conditions by slow diffusion of a layer
of dry and degassed pentane into a saturated solution of dry and
degassed solvents, toluene for 10a and 10b and dichloromethane
for 11a, 11b, and 12. A single crystal of each compound was
mounted under inert perfluoropolyether on the tip of a glass fiber
and cooled in the cryostream of an Oxford-Diffraction XCALIBUR
CCD diffractometer for 10a, 10b, 11a, and 12 or a Stoe IPDS
diffractometer for 11b. Data were collected using monochromatic
Mo KR radiation (λ ) 0.71073). The structures were solved by
direct methods (SIR9741 or SHELXS-9742) and refined by least-
squares procedures on F2 using SHELXL-97.42 In 11a and 11b some
CF3 groups displayed very large anisotropic displacement param-
eters for the F atoms, and they were then treated as disordered over
two positions using the tools available in SHELXL-97. It is
interesting to note that in compound 11a the isotropic thermal
parameters for the P and N atoms coordinated to the rhodium
displayed unrealistic differences, thus indicating that the coordina-
tion sites are partially occupied by either P or N atoms. The P and
N atoms were considered as sharing the same site, and free
refinement of the occupancy factors led to a P/N ratio of 90:10 on
one site and 10:90 on the other site. All H atoms attached to carbon
were introduced in idealized positions and treated as riding models.
The drawing of the molecules was realized with the help of
ORTEP3.43 Crystal data and refinement parameters are shown in
Table 8. Crystallographic data are also given as CIF (see Supporting
Information).
1
flame test). Melting point (dec): 97 °C. H NMR (299.96 MHz,
25 °C, C6D6): δ(ppm) 1.16 (d, 3H, 3JPH ) 9.9 Hz, P-CH3), 2.292
(m, 4H, CH2(C8H12)), 2.39 (s, 3H, N-CH3), 2.534 (m, 4H, CH2-
(COD)), 3.04 (m, 2H, P-CHH-N), 3.56 (m, 1H, N-CHH-N), 3.85
(m, 1H, N-CHH-N), 3.92 (m, 2H, P-CHH-N), 4.23 (br d, 2H,
3JHH ) 14.4 Hz, N-CHH-NMe), 5.263 (br s, 4H, CH2(COD)), 7.69
(s, 4H, p-C6H3(CF3)2), 7.85 (s, 8H, o-C6H3(CF3)2)). 13C NMR (75.43
MHz, 25 °C, CD2Cl2): δ(ppm) 29.59 (s, CH2(COD)), 32.75 (s,
CH2(COD)), 46.97 (s, N-CH3), 52.32 (d, 1JPC ) 18.5 Hz, P-CH3),
1
70.92 (s, N-CH2-N), 74.28 (d, JPC ) 11.7 Hz, P-CH2-N), 80.79
(s, N-CH2-NMe), 109.29 (s, CH(COD)), 118.34 (s, p-C6H3(CF3)2),
1
2
125.49 (q, JFC ) 272.5 Hz, CF3), 129.75 (q, JFC ) 31.3 Hz,
m-C6H3(CF3)2), 135.69 (s, o-C6H3(CF3)2), 162.64 (non-binomial q,
1JCB ) 49.8 Hz, i-C6H3(CF3)2). 31P NMR (121.42 MHz, 25 °C,
1
CD2Cl2): δ(ppm) -40.15 (d, JRhP ) 143.5 Hz). The MS (EI)
spectrum shows fragments deriving from the BArF4 anion [m/z 796
(BArF - CF3 + 2H, 2.3%), 777 (796 - F, 1.5%), 727 (796 -
4
CF3, 1.5%), 650 (BArF +, 5.4%), 631 (650 - F, 7.5%)] but no
3
fragments deriving from the cationic rhodium complex.
R ) Ph (11b). Yield: 0.475 g (72.6%). Correct analytical data
could not be obtained (low C, H, N), probably because of
contamination by NaCl (the presence of sodium was confirmed by
1
flame test). Melting point (dec): 128 °C. H NMR (299.96 MHz,
25 °C, CD2Cl2): δ(ppm) 2.12 (m, 2H, CH2(COD)), 2.28 (m, 2H,
CH2(COD)), 2.45 (s, 3H, N-CH3), 2.53 (m, 4H, CH2(COD)), 3.56
(br s, 2H, CH(COD)), 4.01 (m, 2H, P-CHH-N), 3.92 (m, 1H,
N-CHH-N), 4.14 (m, 2H, P-CHH-N), 4.53 (m, 2H, N-CHH-NMe),
DFT Calculations. Geometry optimizations, Mayer bond in-
dexes, and energies were calculated with density functional theory
using the Gaussian 03 suite of programs.44 The three-parameter
hybrid gradient-corrected functional (B3P86) as developed by
Becke45 with the nonlocal correlation of Perdew46 was used. For
each step of the geometry optimizations, self-consistent iterations
were performed until a convergence criterion of 10-8 was achieved.
A double-ú Gaussian-type basis set47 was augmented with polariza-
3
5.30 (br s, 2H, CH(C8H12)), 7.31 (t, JHH ) 7.8 Hz, 1H, p-C6H5),
7.57 (m, 4H, o-,m-C6H5), 7.65 (s, 4H, p-C6H3(CF3)2), 7.85 (s, 8H,
o-C6H3(CF3)2)). 13C NMR (75.43 MHz, 25 °C, C6D6): δ(ppm)
29.74 (s, CH2(C8H12)), 32.40 (s, CH2(COD)), 46.97 (s, N-CH3),
1
51.29 (d, JPC ) 16.6 Hz, P-CH2-N), 70.95 (s, CH(COD)), 75.60
3
(d, JPC ) 11.8 Hz, N-CH2-N), 80.80 (s, N-CH2-NMe), 109.12 (s,
CH(C8H12)), 118.38 (s, p-C6H3(CF3)2), 132.71 (s, p-C6H5), 124.96
1
1
(d, JPC ) 37.1 Hz, i-C6H5), 125.49 (q, JFC ) 272.5 Hz, CF3),
2
3
(41) Altomare, A.; Burla, M. C.; Camalli, M.; Cascarano, G. L.;
Giacovazzo, C.; Guagliardi, A.; Moliterni, A. G. G.; Polidori, G.; Spagna,
R. SIR97 a program for automatic solution of crystal structures by direct
methods. J. Appl. Crystallogr. 1999, 32, 115.
(42) Sheldrick, G. M. SHELX97, Programs for Crystal Structure Analysis
(Release 97-2); Institu¨t fu¨r Anorganische Chemie der Universita¨t: Tam-
manstrasse 4, D-3400 Go¨ttingen, Germany, 1998.
(43) Farrugia, L. J. ORTEP3 for Windows. J. Appl. Crystallogr. 1997,
30, 565.
(44) Frisch, M. J.; et al. Gaussian 03, Revision C.02; Gaussian Inc.:
Wallingford CT, 2004.
129.75 (q, JFC ) 32.2 Hz, m-C6H3(CF3)2), 130.46 (d, JPC ) 9.8
Hz, o-C6H5), 130.71 (d, 2JPC ) 7.8 Hz, m-C6H5), 135.69 (s, o-C6H3-
(CF3)2), 162.64 (non-binomial q, 1JCB ) 49.79 Hz, i-C6H3(CF3)2).
31P NMR (121.42 MHz, 25 °C, CD2Cl2): δ(ppm) -37.36 (d,
1JRhP ) 142.3 Hz). The MS (EI) spectrum shows fragments deriving
from the BArF4 anion [m/z 796 (BArF4 - CF3 + 2H, 0.058%), 777
(796 - F, 0.033%), 727 (796 - CF3, 0.037%), 650 (BArF +
,
3
1.01%), 631 (650 - F, 0.92%)], plus two fragments deriving from
the cationic rhodium complex [1055 (M+ - PTN(Ph) - F, 0.023%);
861 (1055 - ArF - F, 0.048%)].
Synthesis of [RhI(CO){PTN(Me)}] (12). In a dry and N2-purged
50 mL Schlenk flask, 57 mg (0.1 mmol) of [Rh2I2(CO)4] was
(45) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(46) Perdew, J. P. Phys. ReV B 1986, 33, 8822.
(47) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.; DeFrees,
D. J.; Pople, J. A.; Gordon, M. S. J. Chem. Phys. 1982, 77, 3654.

2200 Organometallics, Vol. 25, No. 9, 2006
Phillips et al.
tion functions48 and with an additional set of d-orbitals for C, N,
and P and an additional p-orbital for hydrogen (6-31+G(d,p)).
Rhodium was modeled with the LAN2DZ basis set of double-ú
quality, substituting the core electron wave functions with an
effective potential.49 Optimized structures were verified to be
stationary minima, as indicated by the absence of imaginary
frequencies, expect in the case of the transition state (one imaginary
was observed).
(contract HPRN-CT-2002-00176). Drs P. Barbaro, C. Mealli,
(ICCOM-CNR), and Dr. E. Manoury (LCC) are gratefully
acknowledged for extra NMR experiments and useful discus-
sions regarding computational studies and catalysis. Dr. C.
Giannelli (University of Florence) is kindly acknowledged for
supplementary GC analysis. Thanks are expressed to “FIRENZE
HYDROLAB”, a project sponsored by Ente Cassa di Risparmio
di Firenze, for supporting this research activity.
Supporting Information Available: Crystallographic informa-
tion files (CIF) for 10a, 10b, 11a, 11b, and 12; coordinates for
optimized models 2a, 13, 14, and 15; full author list for ref 47;
experimental details for catalytic tests. This material is available
free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. We thank the EC for promoting this
scientific activity with financial support including grants through
the Marie Curie Research Training Network “Hydrochem”
(48) Frisch, M. J.; Pople, J. A.; Binkley, J. S. J. Chem. Phys. 1984, 80,
3265.
(49) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
OM051099R