103Rh chemical shifts in complexes bearing chelating bidentate phosphine ligands

Article abstract of DOI:10.1021/om980980k

Experimental and theoretical methods have been employed to investigate the influence of the chelating phosphine ligand on the ¹⁰³Rh chemical shift in complexes containing the [(P₂)Rh] fragment (P₂ = chelating bidentate pho

Full text of DOI:10.1021/om980980k

1196
Organometallics 1999, 18, 1196-1206
103Rh Ch em ica l Sh ifts in Com p lexes Bea r in g Ch ela tin g
Bid en ta te P h osp h in e Liga n d s
Walter Leitner,*,† Michael Bu¨hl,*,‡ Roland Fornika,| Christian Six,†
Wolfgang Baumann,§ Eckhard Dinjus,| Magnus Kessler,† Carl Kru¨ger,† and
Anna Rufin´ska†
Max-Planck-Institut fu¨r Kohlenforschung, Kaiser-Wilhelm-Platz 1,
45470 Mu¨lheim/ Ruhr, Germany, Organisch-chemisches Institut, Universita¨t Zu¨rich,
Winterthurestrasse 190, 8057 Zu¨rich, Switzerland, Institut fu¨r Organische Katalyseforschung
an der Universita¨t Rostock, Arbeitsgruppe “Komplexkatalyse”, Rostock, Germany, and
Forschungszentrum Karlsruhe, Technik und Umwelt, Institut fu¨r Technische Chemie,
Karlsruhe, Germany
Received December 3, 1998
Experimental and theoretical methods have been employed to investigate the influence
of the chelating phosphine ligand on the 103Rh chemical shift in complexes containing the
[(P2)Rh] fragment (P2 ) chelating bidentate phosphine). The δ(103Rh) values obtained by
2D(31P,103Rh{1H}) NMR spectroscopy for a series of neutral rhodium complexes [{R2P(CH2)n-
PR2}Rh(hfacac)] (R ) Ar, Ph, Cy, Me, n ) 1-4, hfacac ) hexafluoroacetylacetonate) have
been compared. Systematic variation of the phosphine ligand has allowed separation of
electronic and geometrical effects. The purely electronic influence of para substituents in
complexes [{(p-XC6H4)2P(CH2)4P(p-XC6H4)2}Rh(hfacac)] correlates directly with the Hammett
σP constants of X, but leads to variations in the chemical shift of less than 80 ppm between
X ) CF3 and X ) OMe. In contrast, geometrical changes in complexes [(P2)Rh(hfacac)] lead
to variations in the chemical shift over a range of approximately 800 ppm. The individual
contributions of various structural parameters on the δ(103Rh) values have been assessed
by density-functional-based calculations for suitable model compounds. The same approach
has been extended to the rationalization of the trends in 103Rh chemical shifts of cationic
complexes with four P donor ligands around a Rh(+I) center and selected anionic Rh(-I)
complexes [(P2)2Rh]-. This analysis allows for the first time a direct corroboration of
geometrical variations and their effect on 103Rh chemical shifts, demonstrating that
correlations of reactivity with 103Rh chemical shifts can give valuable information on
structure/reactivity relationships.
In tr od u ction
ligand effects3 to either structural (e.g., steric bulk, bite
angle) or electronic (e.g., basicity) properties. Despite
the practical value of such distinctions in catalyst
development,4,5 both structural and electronic param-
The fragment [(P2)Rh] (P2 ) chelating bidentate phos-
phine of general formula R2P-(X)-PR2) plays a crucial
role in a plethora of catalytic reactions. The controlling
influence of the chelating phosphine ligand on the
activity and the selectivity of such transformations is
well documented.1 A number of attempts have been
made to develop “ligand parameters”2 that relate these
(3) We restrict our discussion to intrinsic ligand effects, i.e., direct
effects of ligands on metal centers of catalytically active intermediates.
In a practical catalytic system, however, other possibilities must also
be taken into account, e.g., formation of different active intermediates,
partial dissociation, etc. For example see: (a) Graf, E.; Leitner, W.
Chem. Ber. 1996, 129, 91. (b) Portnoy, M.; Milstein, D. Organometallics
1993, 12, 1655. (c) Shaw, B. L.; Perera, S. D. Chem. Commun. 1998,
1863.
(4) For mainly geometric variations see: (a) Casey, C. P.; Whiteker,
G. T.; Melville, M. G.; Petrovich, L. M.; Gavney, J . A., J r.; Powell, D.
R. J . Am. Chem. Soc. 1992, 114, 5535. (b) Kranenburg, M.; van der
Burgt, Y. E. M.; Kramer, P. C. J .; van Leeuwen, P. W. N. M.; Goubike,
K.; Fraanje, J . Organometallics 1995, 14, 3081. (c) Hofmann, P.; Meier,
C.; Hiller, W.; Hickel, M.; Riede, J .; Schmidt, M. U. J . Organomet.
Chem. 1995, 490, 51.
(5) For mainly electronic variations see: (a) J acobsen, E. J .; Zhang,
W.; Gu¨ler, M. L. J . Am. Chem. Soc. 1991, 113, 6703. (b) RajanBabu,
T. V.; Ayers, T. A.; Casalnuovo, A. L. J . Am. Chem. Soc. 1994, 116,
4101. (c) Casalnuovo, A. L.; RajanBabu, T. V.; Ayers, T. A.; Warren,
T. H. J . Am. Chem. Soc. 9869. (d) Chan, A. S. C.; Pai, C.-C.; Yang,
T.-K.; Chen, S.-M. J . Chem. Soc., Chem. Commun. 1995, 2031. (e) Park,
S.-B.; Murata, K.; Matsumoto, H.; Nishiyama, H. Tetrahedron: Asym-
metry 1995, 6, 2487. (f) Schnyder, A.; Togni, A.; Wiesli, U. Organo-
metallics 1997, 16, 255. (g) Palucki, M.; Finney, N. S.; Pospisil, P. J .;
Gu¨ler, M. L.; Ishida, T.; J acobsen, E. N. J . Am. Chem. Soc. 1998, 120,
948.
† Max-Planck-Institut fu¨r Kohlenforschung.
‡ Organisch-chemisches Institut.
§ Institut fu¨r Organische Katalyseforschung an der Universita¨t
Rostock.
| Forschungszentrum Karlsruhe.
(1) (a) Comprehensive Organometallic Chemistry; Abel, E. W., Stone,
F. G. A., Wilkinson, G., Eds.; Pergamon Press: Oxford, 1995; Vol. 12.
(b) Crabtree, R. H. The Organometallic Chemistry of the Transition
Metals; Wiley: New York, 1994. (c) Homogeneous Transition Metal
Catalyzed Reactions; Moser, W. R., Slocum, D. W., Eds.; Adv. Chem.
Ser. 230; 1992. (d) Applied Homogeneous Catalysis with Organometallic
Compounds; Cornils, B., Herrmann, W. A., Eds.; VCH: Weinheim,
1996.
(2) For a classical textbook example: Tolman, C. A. Chem. Rev.
1977, 77, 312. For more recent discussions: Brown, T. L. Inorg. Chem.
1995, 34, 2718. Drago, R. S.; J oerg, S. J . Am. Chem. Soc. 1996, 118,
2654. Fernandez, A.; Reyes, C.; Wilson, M. R.; Woska, D. C.; Prock,
A.; Giering, W. P. Organometallics 1997, 16, 342. J oerg, S.; Drago, R.
S.; Sales, J . Organometallics 1998, 17, 589.
10.1021/om980980k CCC: $18.00 © 1999 American Chemical Society
Publication on Web 03/05/1999

103Rh Chemical Shift
Organometallics, Vol. 18, No. 7, 1999 1197
eters arguably influence the properties of a catalytically
active metal center simultaneously.6
acetonate) as model compounds for the investigation of
the intrinsic influence of bidentate chelating ligands on
structure, reactivity, and spectroscopic properties of the
[(P2)Rh] fragment.13 Experimental data for structurally
closely related compounds suggested a potentially useful
empirical correlation between the 103Rh shifts of com-
plexes [(P2)Rh(hfacac)] and their catalytic activity in the
hydrogenation of CO2 to formic acid.13a During extension
of these studies to a broader range of ligands, however,
the connections between structural features, 103Rh
shifts, and chemical reactivity were found to be very
subtle.13b
The chemical shift of a transition metal is a very
sensitive probe for electronic and geometric changes in
the coordination sphere of the nucleus under scrutiny.7-10
Although there is no a priori reason for an intrinsic
relation between NMR shifts and chemical reactivity,
a number of empirical correlations of this type have
been established.9b-e,11-13 In a seminal study, Bo¨nne-
mann and von Philipsborn correlated 59Co shifts and
catalytic properties of substituted cyclopentadienyl
Co(I) complexes.12 Similarly, direct relationships be-
tween 103Rh shifts and rate constants have been de-
scribed for stoichiometric exchange processes involving
cyclopentadienyl Rh(I)9b and Rh(III)9c complexes. The
chemical shift difference between the diastereomeric
intermediates in catalytic hydrogenation of prochiral
olefins using chiral phosphine rhodium complexes has
been related to the enantioselective discrimination.9d
More recently, A° kermark et al. have correlated stability
constants of complexes [(acac)Rh(alkene)2] with 103Rh
shifts.9e
To proceed from purely empirical correlations in
practical systems to a deeper understanding of the
underlying principles, the intrinsic influence of the
controlling ligand on the NMR shift of the metal center
in the [(P2)Rh] fragment has to be elucidated. Despite
an increasing body of data,10,13 in particular from the
seminal work by the group of Elsevier,10a,b the factors
governing 103Rh shifts in Rh(I) phosphine complexes are
far from being fully understood. A detailed analysis of
the solid state NMR data of related complexes contain-
ing [(P2)M] (M ) Mo, W, Pt) fragments by Lindner et
al. has revealed a remarkable sensitivity of the principal
components of the 31P chemical shift tensor on small
structural changes, suggesting that changes in the
coordination geometry might play a major role in
determining the chemical shift of the corresponding
metal centers as well.14 Recently, density-functional
theory based calculations have been shown to reliably
reproduce trends in δ(103Rh) shifts for various rhodium
complexes15 and may thus help in the interpretation of
experimental results obtained from systematic studies.
Herein we present a detailed experimental and com-
putational study on the influence of the geometrical and
electronic properties of chelating phosphine ligands on
the 103Rh chemical shift in complexes containing [(P2)-
Rh] fragments. For neutral 1,3-diketonato complexes,
electronic variation was achieved by introducing sub-
stituents with different inductive effects in the para
position of a chelating bis(aryl)phosphine. Geometric
parameters, for instance P-Rh-P angles ranging from
70° to 100°, were adjusted by variations in the backbone
and the groups R at phosphorus. In addition, selected
ionic complexes containing the [(P2)Rh] fragment were
investigated for comparison.
We have recently introduced the four-coordinate 16e
complexes [(P2)Rh(hfacac)] (hfacac ) hexafluoroacetyl-
(6) For interesting investigations on the relationship between bite
angle and electronic structure see: Hofmann, P.; Meier, C.; Englert,
U.; Schmidt, M. U. Chem. Ber. 1992, 125, 353.
(7) (a) Benn, R.; Rufin´ska, A. Angew. Chem. 1986, 98, 851; Angew.
Chem., Int. Ed. Engl. 1986, 25, 861. (b) von Philipsborn, W. Pure Appl.
Chem. 1986, 58, 513. (c) Mason, J . Chem. Rev. 1987, 87, 1299. (d)
Pregosin, P. S., Ed. Transition Metal Nuclear Magnetic Resonance;
Elsevier: Amsterdam, 1991. (e) von Philipsborn, W. Chem. Soc. Rev.,
in press.
(8) For recent examples with transition metals other than rhodi-
um: 55Mn: Dowler, M. E.; Le, T. X.; DeShong, P.; von Philipsborn,
W.; Vo¨hler, M.; Rentsch, D. Tetrahedron 1993, 49, 5673. 187Os: Bell,
A. G.; Koz´min´ski, W.; Linden, A.; von Philipsborn, W. Organometallics
1996, 15, 3124. 91Zr: Bu¨hl, M.; Hopp, G.; Beck, S.; Prosenc, M.-H.;
Rief, U.; Brintzinger, H.-H. Organometallics 1996, 15, 778. 57Fe: Benn,
R.; Brenneke, H.; Frings, A.; Lehmkuhl, H.; Mehler, G.; Rufin´ska, A.;
Wildt, T. J . Am. Chem. Soc. 1988, 110, 5561. Meier, E. J . M.;
Koz´min´ski, W.; Linden, A.; Lustenberger, P.; von Philipsborn, W.
Organometallics 1996, 15, 2469.
(9) For recent examples with rhodium and variation in ligands other
than phosphines: (a) Graham, P. B.; Rausch, M. D.; Ta¨schler, K.; von
Philipsborn, W. Organometallics 1991, 10, 3049. (b) Koller, M.; von
Philipsborn, W. Organometallics 1992, 11, 467. (c) Tedesco, V.; von
Philipsborn, W. Organometallics 1995, 14, 3600. (d) Bender, B. R.;
Koller, M.; Nanz, D.; von Philipsborn, W. J . Am. Chem. Soc. 1993, 115,
5889. (e) A° kermark, B.; Blomberg, M. R. A.; Glaser, J .; O¨ hrstro¨m, L.;
Wahlberg, S.; Wa¨rnmark, K.; Zetterberg, K. J . Am. Chem. Soc. 1994,
116, 3405. (f) Brunet, J .-J .; Commenges, G.; Neibecker, D.; Philippot,
K.; Rosenberg, L. Inorg. Chem. 1994, 33, 6373. (g) Law, D. J .; Bigam,
G.; Cavell, R. G. Can. J . Chem. 1995, 73, 635. (h) Asaro, F.; Costa, G.;
Dreos, R.; Pellizer, G.; von Philipsborn, W. J . Organomet. Chem. 1996,
513, 193. (i) Donkervoort, J . G.; Bu¨hl, M.; Ernsting, J . M.; Elsevier, C.
J . Eur. J . Inorg. Chem. 1999, 27.
(10) For recent examples of rhodium complexes with structural
variations in phosphorus ligands: (a) Ernsting, J . M.; Elsevier, C. J .;
de Lange, W. G. J .; Timmer, K. Magn. Reson. Chem. 1991, 29, S118.
(b) Elsevier, C. J .; Kowall, B.; Kragten, H. Inorg. Chem. 1995, 34, 4836.
(c) Carlton, L. J . Organomet. Chem. 1992, 43, 103. (d) Cherkasova, T.
G.; Varshavsky, Yu. S.; Podkorytov, I. S.; Osetrova, L. V. Rhodium
Express 1993, 0, 21. (e) Gavrilov, K. N.; Ignatenko, A. V.; Teleshev, A.
T. Russ. J . Coord. Chem. 1995, 21, 461. (f) Bo¨rner, A.; Kless, A.; Holz,
J .; Baumann, W.; Tillack, A.; Kadyrov, R. J . Organomet. Chem. 1995,
490, 213. (g) Bo¨rner, A.; Kless, A.; Kempe, R.; Heller, D.; Holz, J .;
Baumann, W. Chem. Ber. 1995, 128, 767. (h) J olly, P. W.; Hopp, G.;
Passelaigne, E.; von Philipsborn, W. Personal communications.
Exp er im en ta l Section
Gen er a l Rem a r k s. All reactions involving air- or moisture-
sensitive materials were performed under argon using stan-
dard Schlenk techniques in dried and deoxygenated solvents.
The complexes [(cod)Rh(hfacac)] (cod ) 1,5-cyclooctadiene, 4)16
and [(P2)Rh(hfacac)]13b (7, 9b-d , 11) were synthesized as
described previously. The ligands 6, 8, 10 were commercial
products or prepared following known procedures.17 Cl2P(CH2)4-
(13) (a) Fornika, R.; Go¨rls, H.; Seemann, B.; Leitner, W. J . Chem.
Soc., Chem. Commun. 1995, 1479. (b) Angermund, K.; Baumann, W.;
Dinjus, E.; Fornika, R.; Go¨rls, H.; Kessler, M.; Kru¨ger, C.; Leitner,
W.; Lutz, F. Chem. Eur. J . 1997, 3, 755.
(11) For
a
computational prediction of such
a
correlation for
(14) Lindner, E.; Fawzi, R.; Mayer, H. A.; Eichele, K.; Hiller, W.
Organometallics 1992, 11, 1033.
(15) (a) Bu¨hl, M. Organometallics 1997, 16, 261. (b) Bu¨hl, M. Chem.
Phys. Lett. 1997, 267, 251.
(16) Leipoldt, J . G.; Grobler, E. C. Trans. Met. Chem. 1986, 11, 110.
(17) (a) Issleib, K.; Mu¨ller, D.-W. Chem. Ber. 1959, 92, 1397. (b) Burt,
R. J .; Chatt, J .; Hussein, W.; Leigh, G. J . J . Organomet. Chem. 1979,
182, 203. (c) Brunner, H.; Lautenschlager, H.-J . Synthesis 1989, 706.
(d) Dinjus, E.; Leitner, W. Appl. Organomet. Chem. 1995, 9, 43.
vanadium-catalyzed ethylene polymerization see: Bu¨hl, M. Angew.
Chem., Int. Ed. Engl. 1998, 37, 142.
(12) (a) Bo¨nnemann, H.; Brijoux, W.; Brinkmann, R.; Meurers, W.;
Mynott, R.; von Philippsborn, W.; Egolf, T. J . Organomet. Chem. 1984,
272, 231. (b) Bo¨nnemann, H. Angew. Chem. 1985, 97, 264; Angew.
Chem., Int. Ed. Engl. 1985, 24, 248. (c) For other examples for relations
between structure and 59Co shifts see: Benn, R.; Cibura, K.; Hoffmann,
P.; J onas, K.; Rufin´ska, A. Organometallics 1985, 4, 2214.

1198 Organometallics, Vol. 18, No. 7, 1999
Leitner et al.
PCl2 (1) was synthesized according to the method of Kuchen.18
Mass spectra were measured on a Finnigan MAT SSQ 710
with EI-MS technique. Standard NMR spectra were recorded
in 5 mm tubes on a Bruker AC 200 spectrometer operating at
from tBuOMe afforded analytically pure 9a as deeply red,
almost black crystals suitable for X-ray diffraction.
9a : 582 mg (0.81 mmol, 82%), mp 115 °C. Anal. (C30H47F6-
O2P2Rh) C; H; F; P: calcd, 50.15, 6.59, 15.86, 8.62; found,
50.08, 6.58, 16.08, 8.59. MS (EI): m/ z 718 (M•+, 100), 636 (M•+
- C6H10, 31), 545 (M•+ - 2C6H10, 23).
1
200.13, 50.29, 188.15, and 80.15 MHz for H, 13C, 19F, and 31P.
Chemical shifts δ are reported in ppm relative to external
CFCl3 for 19F, H3PO4 for 31P, and TMS for 1H and 13C, using
the solvent resonance as a secondary standard if possible.
Coupling constants J are given in hertz.
Solid Sta te NMR. Solid state CP/MAS NMR spectra were
recorded as described previously19 on a Bruker MSL-300
spectrometer operating at 121.5 MHz for 31P and 75.47 MHz
for 13C using NaH2PO4 (δ 0.314 relative to 85% H3PO4) and
adamantane (δ 38.40 relative to TMS) as an external standard,
respectively. Coupling constants J (PP′) were determined from
J resolved 2D spectra.19,20
IR (KBr): 2925 s (ν(CH2), Cy), 2852 s (ν(CH2), Cy), 2665 w,
2654 w, 1672 w, 1621 w, 1593 m (ν(CdO) hfacac), 1565 m,
1536 m, 1504 m, 1479 m (ν(CdO) hfacac), 1447 m (δ(CH2)),
1335 m, 1254 s (CF3), 1193 s, 1143 s (CF3), 1095 m, 1076 m,
1068 m, 1024 w, 1000 m, 942 m, 914 w, 888 m, 851 m, 812 w,
784 m, 759 m, 741 w, 721 m, 677 m, 588 m, 531 m, 510 m
cm-1
.
1H NMR (THF-d8): δ 5.95 (1H, s, CHC(CF3)), 2.50 (2H, dt,
2J (PH) ) 9 Hz, 3J (RhH) ) 3 Hz, CH2(PCy2)2) 2.12-1.14 (44H,
2
m, Cy-H) ppm. 13C{1H} NMR (THF-d8): δ 171.0 (q, J (FC) )
1
Deter m in a tion of δ(103Rh ). 2D(31P,103Rh{1H}) spectra were
recorded on a Bruker ARX 400 spectrometer (B0 ) 9.4 T) at
298 K. The standard four-pulse HMQC sequence21 was per-
formed twice with variation of the 103Rh frequency offset and
the t1 increment to ensure that signals in the F1 dimension
33.0 Hz, CO), 120.1 (q, J (FC) ) 284 Hz, CF3), 90.1 (s, CHC-
(CF3)), 27.7 (dt, 1J (PC) ) 9.5 Hz, 2J (RhC) ) 6.5 Hz, CH2(PCy2)2)
37.0(m)/30.4(s)/30.3(s)/27.94(m)/27.90(m)/27.1(s) (Cy-C) ppm.
19F{1H} NMR (THF-d8): δ -76.7 ppm. 31P{1H} NMR (THF-
d8): δ -4.2 (d, 1J (RhP) ) 164 Hz) ppm. 2D(31P,103Rh) {1H}
NMR (THF-d8): δ 1062 ppm. 13C{1H} CP/MAS NMR (297 K):
(
103Rh) are not folded. Typical conditions: 128 experiments in
1
δ 170.2/169.6 (CO), 119.9 (q, J (FC) ) 290 Hz, CF3), 89.8 (s,
t1, 4 or 8 scans each, 4 s relaxation delay, total experimental
time ca. 40-80 min. The 103Rh shifts are given relative to
¥(103Rh) ) 3.16 MHz7,22 and were reproducible within 1 ppm.
Syn th esis of P h osp h in e Liga n d s 3a -e. A solution of
Cl2P(CH2)4PCl2 (1.14 mL, 6.00 mmol) in 10 mL of THF was
added dropwise to a THF solution containing 23.9 mmol of
p-XC6H4MgBr at -30 °C. The resulting mixture was allowed
to warm to room temperature and stirred for 1 h. Dioxane (20
mL) was added, and the solution was filtered over Celites. The
filtrate was washed with 20 mL of aqueous NH4Cl solution,
and the aqueous phase was extracted with THF (2 × 10 mL).
The combined organic layers were dried over MgSO4 and
filtered, and the solvents were removed in vacuo. The resulting
off-white or slightly yellow products were recrystallized from
EtOH to give the colorless ligands (40-80% yield) in >95%
purity (1H, 31P NMR). Selected spectroscopic data are sum-
marized in Table 1; full details are available as Supporting
Information.
Syn th esis of Com p lexes 5a -e. One equivalent of the
appropriate chelating phosphine 3a -e in 15 mL of THF was
added dropwise to a solution of 4 (300 mg, 0.72 mmol) in 15
mL of THF at -78 °C. The reaction mixture was allowed to
warm to room temperature, and the volatiles were removed
in vacuo. The glassy residues were dried under vacuum for
24 h to yield quantitatively 5a -e as red-brown powders
containing small amounts of THF and cyclooctadiene, but less
than 5% phosphorus-containing impurities according to NMR
analyses. The 103Rh chemical shifts were determined from
these samples without further purification. Characteristic
spectroscopic data are given in Table 1; full details including
MS and 13C NMR data are available as Supporting Informa-
tion.
CHC(CF3)), 43.1 - 27.2 (CH2(PCy2)2, Cy-C). 31P{1H} CP/MAS
2
1
NMR (310 K): δ -5.6 (dd, J (PP′) ) 117 ( 5 Hz, J (RhP) )
164 ( 5 Hz, P), -6.7 (dd, 2J (P′P) ) 117 ( 5 Hz, 1J (RhP′) )
170 ( 5 Hz, P′).
Cr ysta l Da ta of 9a : C30H47F6O2P2Rh, M ) 718.53, trigonal,
space group R3hc (No. 167), a ) b ) c ) 17.6900(9) Å, R ) â )
γ ) 96.738(4)°, V ) 5411.2(5) Å3 (λ ) 0.710 69 Å), Z ) 6, Fc )
1.323 g cm-3, F(000) ) 2232. Crystal dimensions 0.32 × 0.46
× 0.46 mm3, µ ) 6.16 cm-1
.
Da ta Collection a n d P r ocessin g. Data were collected on
an Enraf-Nonius CAD4 diffractometer in the ω/2θ scan mode
using graphite-monochromated Mo KR radiation. Unit cell
parameters were determined from least-squares refinement
of setting angles for 25 reflections. Data were corrected for
Lorentz and polarization effects. There was no correction for
crystal decay. A total of 5368 (904 independent) reflections
were measured (1.76 e θ e 27.45, 1 e h e 22, 0 e k e 15, 0
e l e 11) at T ) 293(1) K.
Str u ctu r e An a lysis a n d Refin em en t. The structure was
solved by the heavy atom method using SHELX-93.23 Refine-
ment was by full-matrix least-squares on F2 with 413 posi-
tional and anisotropic thermal parameters for non-H atoms.
H atoms were fixed at calculated positions with C-H ) 0.96
Å and an overall temperature factor. No absorption correction
was applied. Weighting scheme w ) q/[σ2(Fo2) + (0.0641P)2 +
2
12.2717P] where P ) (Fo + 2Fc2)/3. Final R indices (I > 2σ-
(I)), R ) 0.0653 and wR2 ) 0.2068; residual electron density
0.478 e Å-3. Essential geometric molecular data are depicted
in Figure 2.
Syn th esis of Com p lex 9c. Complex 9c was obtained as
brown powder according to the method described for 5a -e,
using solutions of 4 (265 mg, 0.63 mmol) and 8c (278 mg, 0.64
mmol) in 20 mL of THF each. Crystallization from tBuOMe
gave 425 mg (0.57 mmol, 90%) of 9c in the form of a brown
microcrystalline powder.
Syn th esis a n d X-Ra y Cr ysta l Str u ctu r e An a lysis of
Com p lex 9a . The procedure was essentially the same as
described earlier for 9b,13b using solutions of 4 (414 mg, 0.99
mmol) and 8a (405 mg, 0.99 mmol) in 40 mL of THF each.
Addition of the phosphine solution was carried out extremely
slowly in order to avoid formation of the doubly substituted
complex [(8a )2Rh][hfacac]13b and/or bridged binuclear com-
plexes (see text for details). Crystallization of the crude product
1H NMR (THF-d8): δ 6.02 (1H, s, CHC(CF3)), 2.30 (2H, d,
2J (PH) ) 12 Hz, CH2(PCy2)2) 1.95-1.10 (46H, m, Cy-H) ppm.
13C{1H} NMR (THF-d8): δ 170.9 (q, 2J (FC) ) 33.4 Hz, CO),
1
120.2 (q, J (FC) ) 285 Hz, CF3), 90.0 (s, CHC(CF3)), 19.2 (dt,
1
2
| J (PC) + 3J (P′C)| ) 32.4 Hz, J (RhC) ) 1.9 Hz, CH2(PCy2)2),
24.1 (d, 2J (RhC) ) 2.9 Hz, CH2CH2(PCy2)2) 36.9(m)/30.9(s)/
29.9(s)/28.4(m)/28.2(m)/27.4(s) (Cy-C) ppm. 19F{1H} NMR
(THF-d8): δ -76.8 ppm. 31P{1H} NMR (THF-d8): δ 42.7 (d,
1J (RhP) ) 186 Hz) ppm. 2D(31P,103Rh) {1H} NMR (THF-d8):
δ 608 ppm.
(18) Diemert, K.; Kuchen, W.; Kutter, J . Phosphorus Sulfur 1983,
15, 155.
(19) Bach, I.; Po¨schke, K.-R.; Goddard, R.; Kopiske, C.; Kru¨ger, C.;
Rufin´ska, A.; Seevogel, K. Organometallics 1996, 15, 4959.
(20) Wu, G.; Wasylishen, R. E. Inorg. Chem. 1992, 31, 145.
(21) Bax, A.; Griffey, R. H.; Hawkins, B. L. J . Magn. Reson. 1983,
55, 301.
(22) Mann, B. E. In Transition Metal Nuclear Magnetic Resonance;
Pregosin, P. S., Ed.; Elsevier: Amsterdam, 1991; p 177.
(23) Sheldrick, G. M. SHELXL-93, Program for Crystal Structure
Refinement; University Go¨ttingen: Germany, 1993.

103Rh Chemical Shift
Organometallics, Vol. 18, No. 7, 1999 1199
Sch em e 1. Syn th esis of pa r a -Su bstitu ted Liga n d s Ar 2P (CH2)4P Ar 2 F or m in g Seven -Mem ber ed Ch ela te
Rin gs u p on Coor d in a tion
Sch em e 2. Syn th esis of Neu tr a l Com p lexes [{R2P (CH2)n P R2}Rh (h fa ca c)]
Com p u ta tion a l Deta ils. Geometries have been fully op-
timized with the Gaussian 94 package24 employing the gradi-
ent-corrected density functionals of Becke (1988)25 and Perdew
(1986)26 for exchange and correlation, respectively, together
performed using the MORPHY program34 and the B3LYP/II′
density, i.e., employing basis II only on Rh, on the P atoms,
and on the 1,3-propanedionato moiety, and IGLO-basis DZ33
on the rest of the ligands.
with
a quasi-relativistic effective core potential and the
corresponding [6s5p3d] valence basis set for Rh,27 and standard
6-31G* basis28 on the ligands (designated BP86/ECP1).
Magnetic shielding tensors have been computed with the
GIAO (gauge-including atomic orbitals-DFT method as imple-
mented29 in the Gaussian 94 program, employing Becke’s
three-parameter exchange DFT/Hartree-Fock hybrid func-
tional30 and the correlation functional of Lee, Yang, and Parr,31
a fine integration grid (finegrid option), together with basis
II, i.e., a well-tempered [16s10p9d] basis32 on Rh (contracted
from the 22s14p12d set and augmented with two d-shells of
the well-tempered series) and the recommended IGLO-basis
II33 for the ligands (which is essentially of polarized triple-ú
quality), designated GIAO-B3LYP/II. Experimental 103Rh chemi-
cal shifts of a number of rhodium complexes have been well
reproduced using this particular combination of theoretical
methods.15b Theoretical relative 103Rh chemical shifts (refer-
enced to the usual standard frequency of ¥ ) 3.16 MHz) have
been calculated as δ ) σ(standard) - σ(calc), employing the
σ(standard) value from ref 15b, -878 ppm. This value has been
obtained from σ(calc) vs δ(expt) linear regressions employing
BP86/ECP1 geometries.
Resu lts a n d Discu ssion
1. Neu tr a l Com p lexes of th e Typ e [(P 2)Rh (1,3-
d ik eton a te)]. In vestiga tion of Electr on ic Effects.
Chelating phosphines 3a -e were synthesized bearing
electron-withdrawing or -donating substituents in the
para position of the phenyl rings of the well-known dppb
ligand 6d , in order to probe the influence of electronic
properties of the phosphine ligand on the 103Rh shift. It
is well established that para substituents in aryl
phosphorous compounds do not affect the steric proper-
ties of a given ligand.2a,b The phosphines 3a -e were
obtained in reasonable to good yields by reaction of bis-
(dichlorophosphino)butane18 1 with the appropriate aryl
Grignard reagents 2a -e according to Scheme 1. The
neutral Rh(I) complexes [(3a -e)Rh(hfacac)] (5a -e) were
isolated as dark red-brown glassy solids from the
reaction of [(cod)Rh(hfacac)] (4) with 1 equiv of 3a -e
in THF in analogy with the synthesis of the unsubsti-
tuted parent complex 7d (Scheme 2).13b Selected spec-
troscopic data of the free ligands 3a -e, 6d and the
complexes 5a -e, 7d are summarized in Table 1 together
with the Hammett σP constants35 for the substituents
X.
Topological analysis of the total electron density has been
(24) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
J ohnson, B. G.; Robb, M. A.; Cheeseman, J . R.; Keith, T.; Petersson,
G. A.; Montgomery, J . A.; Raghavachari, K.; Al-Laham, M. A.;
Zakrzewski, V. G.; Ortiz, J . V.; Foresman, J . B.; Peng, C. Y.; Ayala, P.
Y.; Chen, W.; Wong, M. W.; Andres, J . L.; Replogle, E. S.; Gomperts,
R.; Martin, R. L.; Fox, D. J .; Binkley, J . S.; DeFrees, D. J .; Baker, J .;
Stewart, J . J . P.; Head-Gordon, M.; Gonzales, C.; Pople, J . A. Gaussian
94; Pittsburgh, PA, 1995.
The variations in the 13C chemical shifts of the ipso-
carbon atoms in 3a -e, 6d and 5a -e, 7d upon substitu-
tion are noticeable, whereas changes in δ(31P) are
relatively small. In both cases, no direct relationship
between Hammett σP constants and chemical shifts are
revealed. In contrast, δ(103Rh) of 5a -e, 7d can be
directly correlated with σP, but variations are also
(25) Becke, A. D. Phys. Rev. A 1988, 38, 3098.
(26) Perdew, J . P. Phys. Rev. B 1986, 33, 8822; 1986, 34, 7406.
(27) Andrae, D.; Ha¨ussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Theor. Chim. Acta 1990, 77, 123.
(28) Hehre, W. J .; Radom, L.; Schleyer, P. v. R.; Pople, J . A. Ab Initio
Molecular Orbital Theory; Wiley: New York, 1986.
(29) Cheeseman, J . R.; Trucks, G. W.; Keith, T. A.; Frisch, M. J . J .
Chem. Phys. 1996, 104, 5497.
(30) Becke, A. D. J . Chem. Phys. 1993, 98, 5648.
(34) Popelier, P. L. A. Comput. Phys. Commun. 1996, 93, 212.
(35) Note that Hammett constants and absolute magnetic shieldings
have the same notations and should not be confused. The values for
σP used here are taken from: March, J . Advanced Organic Chemistry,
3rd ed.; Wiley: New York, 1985; p 244.
(31) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(32) Huzinaga, S.; Klobukowski, M. J . Mol. Struct. 1988, 167, 1.
(33) Kutzelnigg, W.; Fleischer, U.; Schindler, M. In NMR Basic
Principles and Progress Vol. 23; Springer Verlag: Berlin, 1990; p 165.

1200 Organometallics, Vol. 18, No. 7, 1999
Leitner et al.
a
Ta ble 1. Selected An a lytica l Da ta of F r ee Liga n d s (p-XC6H4)2P (CH2)4P (p-XC6H4)2 6d , 3a -e a n d
b
Cor r esp on d in g Com p lexes [(P 2)Rh (h fa ca c)] 7d , 5a -e
δ (13C)c
(ligand)
[ppm]
δ (13C)c
(complex)
[ppm]
δ (13P)
(ligand)
[ppm]
δ (13P)
(complex)
[ppm]
J RhP
(complex)
[ppm]
δ (103Rh)
(complex)
[ppm]
e
σP
X
(Hammett)
6d /7d
3a /5a
3b/5b
3c/5c
3d /5d
3e/5e
H
138.7
129.9
135.4
134.2
136.8
143.1
137.3
128.6
134.3
132.8
136.6
141.2
-14.3d
-19.1
-17.5
-18.0
-17.3
-14.2
48.6
45.5
46.8
47.8
48.4
50.3
191
192
191
191
191
190
623
646
639
599
589
570
0
OMe
Me
F
Cl
CF3
-0.28
-0.14
0.15
0.24
0.53
a
b
d
In CDCl3. In THF-d8. c Aromatic ipso-carbon directly bound to phosphorus. In benzene-d6. e See ref 35.
Ta ble 2. Exp er im en ta l a n d Ca lcu la ted 103Rh
Ch em ica l Sh ifts (δ in p p m ) for Selected Com p lexes
[{R2P (CH2)n P R2}Rh (1,3-d ik eton a te)]
R
n ) 1
n ) 2
n ) 3
n ) 4
Ph
Cy
Me
7a : 1130a
9a : 1062b
12a : 1466c
7b: 438a
9b: 368a
12b: 514c
11: 370a
7c: 567a
9c: 608b
12c: 659c
7d : 623b
9d : 845a
12d : 724c
a
b
Experimental value, from ref 13. Experimental value, this
work. c From DFT calculations on the acac model compounds
12a -d without additional constraints.
ed 103Rh nuclei with δ(103Rh) ) 1130 and 1062, respec-
tively. The experimentally determined 103Rh chemical
shifts for the systematic series of neutral complexes
[{R2P(CH2)nPR2}Rh(hfacac)] with R ) Ph (7a -d ) and
R ) Cy (9a -d ) are compiled in Table 2. Independent of
the substituent at phosphorus, the dependence of
δ(103Rh) on the length n of the chelate bridge shows the
typical U-shaped behavior for 1 e n e 4: With respect
to the five-membered chelate rings 7b and 9b, both
smaller and larger rings in the respective series are
characterized by strong deshielding of the 103Rh chemi-
cal shift.10a,b,14 However, an increased shielding rather
than a deshielding of the 103Rh nuclei would be antici-
pated for the small P-Rh-P angles in 7a , 9a if the
above described empirical correlation for n g 2 would
be generally valid.
F igu r e 1. Correlation between δ(103Rh) and Hammett σP
constants35 for complexes [{(p-XC6H4)2P(CH2)4P(p-XC6H4)2}-
Rh(hfacac)] (5a -e, 7d ).
moderate, within less than 80 ppm (Figure 1). Correla-
tions with the Hammett σP constants have been estab-
lished also for the basicity of related aryl phosphines36
and are generally interpreted in terms of inductive
effects. Thus, the same inductive effects will affect the
HOMO-LUMO gap and lead to variations in the
paramagnetic contribution of the magnetic shielding.
Apparently, such inductive effects have very little
impact on the 103Rh chemical shift in complexes con-
taining the [(P2)Rh] fragment.10a,b
In vestiga tion of Geom etr ica l Effects. Coordina-
tion geometries ofcomplexes [{R2P(CH2)nPR2}Rh(hfacac)]
can be controlled by the length n of the chelate backbone
and the steric bulk of the substituents R.13 The size of
the P-Rh-P angle together with the bulkiness of the
groups R and the flexibility of the chelate ring can play
an important role for the behavior of the [(P2)Rh] frag-
ment in catalytic reactions, as has recently been ratio-
nalized in the so-called accessible molecular surface
(AMS) model.13b In the series of compounds 7b-d and
9b-d with n ) 2-4, larger chelate rings were shown
to be associated with a deshielding of the 103Rh nucleus
and with an enhanced catalytic activity for CO2 hydro-
genation, independent of the substituent R. At the same
time, the P-Rh-P angle as determined by X-ray
crystallographic analyses showed an apparent linear
correlation with δ(103Rh).13
Elucidation of the molecular structure of 9a by single-
crystal X-ray diffraction analysis (Figure 2) unambigu-
ously proves that it exists as a mononuclear 16e complex
bearing a four-membered chelate ring in the solid
state.37 The rhodium atom and the two central carbon
atoms of the phosphine and the hfacac chelate rings
reside on a C2 symmetry axis. The CF3 groups are in
statistical disorder as observed also in other solid state
structures of Rh-hfacac complexes.13b Bond lengths and
angles within the four-membered chelate38,39 and within
the hfacac ligand13b are in the range expected from other
Rh(I) complexes containing similar moieties. The coor-
dination geometry about Rh is slightly distorted from
an ideal square planar arrangement with an angle θ
between the O-Rh-O and P-Rh-P planes of 12.6°.
(37) The ability of bis(phosphino)methane ligands to bridge two
rhodium metals is well established. See for example: (a) Sanger, A.
R. J . Chem. Soc., Chem. Commun. 1975, 893. (b) Sanger, A. R. J . Chem.
Soc., Dalton Trans. 1977, 120. (c) Shafiq, F.; Kramarz, K. W.;
Eisenberg, R. Inorg. Chim. Acta 1993, 213, 111. (d) Anderson, D. J .;
Kramarz, K. W.; Eisenberg, R. Inorg. Chem. 1996, 35, 2688.
Introduction of the methylene-bridged ligands R2-
PCH2PR2 (R ) Ph, 6a ; R ) Cy, 8a ) using the same
methodology as for n g 2 affords complexes 7a and 9a ,
which are, however, characterized by strongly deshield-
(38) R ) Ph: Pignolet, L. H.; Doughty, D. H.; Novicki, S. C.;
Casalnuovo, A. L. Inorg. Chem. 1980, 19, 2172. Chebi, D. E.; Fanwick,
P. E.; Rothwell, I. P. Organometallics 1990, 9, 2948. Ge, Y.-W.; Sharp,
t
P. R. Inorg. Chem. 1991, 30, 1671; R ) Bu, see ref 6.
(39) Fornika, R.; Go¨rls, H.; Leitner, W.; Six, C. Manuscript in
preparation.
(36) Bush, R. C.; Angelici, R. J . Inorg. Chem. 1988, 27, 681.

103Rh Chemical Shift
Organometallics, Vol. 18, No. 7, 1999 1201
-9.8/-10.9, solid state) and the 103Rh,31P coupling
constant (J (Rh,P) ) 170 Hz) are very similar to those
of 9a , suggesting that this species also contains a four-
membered [(P2)Rh] fragment. Therefore, we tentatively
assign this species as the hfacac-bridged dimer [(8a )-
Rh(µ-κ2O,O′-hfacac)2Rh(8a )], 9a ′. In agreement with the
opening of the hfacac chelate, the 103Rh chemical shift
of 9a ′, 326 ppm, is remarkably shielded compared to its
monomeric congener 9a . A second phosphorus-contain-
ing complex was formed from 9a in benzene-d6 solution
upon standing (ca. 8% after 78 h) and showed an
apparent doublet (∆v ) 101 Hz) at δ(31P) ) 0.6. This
splitting is well in the range of the higher order coupling
1
constants | J (RhP) + 3J (RhP)| (ca. 114 Hz) reported for
binuclear complexes with 6a as a bridging ligand
(apparent doublets, AA′A′′A′′′XX′ spin system).37a,b There-
fore, this second minor component can reasonably be
assigned as the bridged binuclear complex [(hfacac)Rh-
(µ-κ2P,P′-8a )2Rh(hfacac)], 9a ′′, although the correspond-
ing 103Rh shift was not determined. In benzene-d6, the
hfacac methine protons of 9a , 9a ′, and 9a ′′ resonated
at 6.24, 6.03, and 6.34 ppm, respectively, and their
intensities were fully consistent with the above assign-
ments.
F igu r e 2. Molecular structure of [(Cy2PCH2PCy2)Rh-
(hfacac)] (9a ) as determined by single-crystal X-ray dif-
fraction analysis. Heavy atoms are represented as spheres,
and hydrogen atoms have been omitted for clarity. Selected
bond distances (Å) and angles (deg): Rh(1)-P(1), 2.195-
(4); P(1)-C(1), 1.840(5); P(1)-C(5), 1.79(2); P(1)-C(11),
1.90(2); Rh(1)-O(1), 2.10(1); O(1)-C(3), 1.20(2); C(3)-C(4),
1.40(2); C(2)-C(3), 1.54(3); P(1)*-Rh(1)-P(1), 73.7(2);
C(1)-P(1)-Rh(1), 97.4(2); P(1)*-C(1)-P(1), 91.4(2); C(5)-
P(1)-Rh(1), 124.8(7); C(11)-P(1)-Rh(1), 113.2(7); C(11)-
P(1)-C(5), 107.6(9); C(5)-P(1)-C(1), 107.3(6); C(11)-P(1)-
C(1), 104.0(7); O(1)*-Rh(1)-P(1); 100.0(4); O(1)-Rh(1)-
P(1), 169.4(4); O(1)*-Rh(1)-O(1), 87.5(5); C(3)-O(1)-
Rh(1), 125.(1); C(4)-C(3)-O(1), 129(2); C(3)*-C(4)-C(3),
123(1); C(2)-C(3)-O(1), 115(2).
The above analysis proves unambiguously that the
strong deshielding of the rhodium nucleus in 7a and
9a is indeed an intrinsic property of the four-membered
[(P2)Rh] chelate ring of these complexes and not related
to possible isomerizations in solution.37 The deshielding
effect of the small bite angle is also consistent with
observations made earlier using the same chelating
phosphines together with other co-ligands,10a but no
satisfactory explanation is currently available for this
fact or for the relationship between coordination geom-
etry and δ(103Rh) in general. Discussion of trends in
chemical shifts of transition metal compounds hitherto
usually invoked the Ramsey equation, which describes
the paramagnetic contribution to the chemical shift in
terms of three key contributions: electron density,
radial term, and mean excitation energy.43 However,
geometric variations as imposed by structural properties
of chelating ligands may affect all three terms simul-
taneously, and quantification or even rationalization is
therefore often a matter of ambiguity.10a Hence, we
decided to adopt an alternative approach by separating
the individual geometric effects on δ(103Rh) using the
established tools of DFT calculations for chemical shifts
of transition metals.15,44 The acetylacetonate model
compounds 12a -d and 13 were used in these calcula-
tions (Chart 1).
The geometries of model complexes [{R2P(CH2)nPR2}-
Rh(acac)] (acac ) acetylacetonate, R ) Me, n ) 1-4;
12a -d ) have been optimized at the BP86/ECP1 level,
and the chemical shifts were calculated at the GIAO-
B3LYP/II level, Table 2). For the complex with n ) 2
(12b, C2 symmetry), a δ(103Rh) value of 514 ppm is
computed. Taking into account the generally observed
shift difference of 100-150 ppm between acac and
hfacac complexes,9e,22 this value is in very good accord
with the experimental value, 370 ppm for [{Me2P(CH2)2-
PMe2}Rh(hfacac)] (11).13 Direct comparison of the chemi-
In the 31P CP/MAS spectra of 9a , the isolated 31P-
31P spin pairs of the chelating ligand are composed of
two spectroscopically inequivalent nuclei with chemical
shifts of -5.6 and -6.7 ppm, respectively.14,40 A coupling
constant J (P,P) ) 117 ( 5 Hz was determined for 9a
by 2D-J -resolved experiments, similar to the large
couplings detected in Ni(0) complexes bearing four-
membered chelates of the related ligand tBu2PCH2-
PtBu2.41 Two distinct resonances at 170.2 and 169.6 ppm
are also observed for the quaternary carbonyl carbons
of the hfacac ligand in the 13C CP/MAS spectra of 9a .
Allowing for the nonidentical orientations of chemical
shift tensors in the solid state,40 the similarity of the
NMR data in the solid state and in solution clearly
indicates that the mononuclear structure of 9a with the
four-membered chelate is retained in solution. Ad-
ditional evidence is provided by the ring contribution
to the 31P chemical shift ∆δR ) 64 ppm, which is in the
typical range for strained four-membered [(P2)Rh] frag-
ments.42
In addition to 9a a minor byproduct was sometimes
observed in the NMR spectra of the crude product both
in solution and in the solid state. This compound was
formed in variable amounts (ca. 10%) depending on the
rate of addition and the reaction temperature. The
values of the 31P chemical shift (δ -7.2, THF-d8; δ
(40) (a) Kubo, A.; McDowell, C. A. J . Chem. Phys. 1990, 92, 7156.
(b) Davies, J . A.; Dutremez, S. Coord. Chem. Rev. 1992, 114, 61.
(41) Rufin´ska, A.; Nickel, T.; Po¨rschke, K.-R. Unpublished results.
(42) Garrou, P. E. Chem. Rev. 1981, 81, 229.
(43) Ramsey, N. F. Phys. Rev. 1950, 78, 699.
(44) Malkin, V. G.; Malkina, O. L.; Casida, M. E.; Salahub, D. R. J .
Am. Chem. Soc. 1994, 116, 5898.

1202 Organometallics, Vol. 18, No. 7, 1999
Leitner et al.
F igu r e 3. Definition of key geometric parameters for
structural description of neutral four-coordinated com-
plexes of the type [(P2)Rh(1,3-diketonate)]: (left) top view;
(right) projection along the pseudo-C2 axis of the O-Rh-O
chelate. See text for details.
Ch a r t 1. Mod el Com p lexes Used for
Com p u ta tion a l Stu d ies on Com p lexes
[{R2P (CH2)n P R2}Rh (1,3-d ik eton a te)]
cal shifts computed for the other complexes 12a ,c,d to
experimental data is difficult because of the different
substituents R employed in the actual compounds.
However, the aforementioned U-shaped dependence of
the δ(103Rh) values and their relative ordering as a
function of the chelate-ring size is qualitatively repro-
duced by these model calculations. These results em-
phasize again the reliability of DFT calculations to
reproduce at least qualitative trends and encouraged
us to use this methodology for a detailed analysis of the
relationship between coordination geometry and 103Rh
chemical shift.
Detailed analysis of the experimental and computed
molecular structures of complexes [(P2)Rh(1,3-diketo-
nate)] reveals that variations in the chelate ring size
are inevitably associated with changes of several geo-
metrical parameters, notably the “bite angle” â, the
P-Rh distance r, the distortion from planarity as
measured by the twist angle θ, and the deviation from
ideal alignment of the phosphorous lone pair with the
Rh-P bond axis as measured by the “tilt angle” τ
(Figure 3). These parameters cannot be varied indepen-
dently in real systems,14 as illustrated for example by
a linear correlation between bite angles â and P-Rh
distance r in complexes of the type [(P2)Rh(hfacac)] (for
â > 82°).13b To probe the individual contributions of each
parameter to δ(103Rh) separately, model calculations
have been performed for [(H3P)2Rh(acac)] (13) (C2v
symmetry45). Starting from the fully optimized geometry
of 13, 103Rh chemical shifts have been calculated for
suitably distorted structures, i.e., by successively vary-
ing one parameter with all others fixed at constant
values.
F igu r e 4. Variation of δ(103Rh) with structural parameters
(+ computed for model compound 13, ] experimental data
from Table 2 and ref 13); the chosen parameters (see Figure
3 for definition) and the resulting fitting functions for the
theoretical data are as follows: (a) bite angle â, δ ) 0.498-
(â - 82.4)2 + 709; (b) Rh-P bond length r, δ ) 10376r -
22241; (c) twist angle θ, δ ) 0.344θ2 + 780; (d) tilt angle τ,
δ ) 0.735(τ + 4.5)2 + 708 [the form of the fitting function
for τ depends somewhat on the value of â (i.e., both
structural effects on δ are not simply additive); since the
“increment system” of Table 4 is based on the five-
membered chelate ring as standard, the â value of 12d has
been adopted during the variation of τ].
The results for the variation in â are displayed
graphically in Figure 4a, together with experimental
data points from the present work and ref 13. The
GIAO-B3LYP data indeed show a minimum near 82°,
close to the angles of 82-87° typically found46 in five-
membered P-Rh-P chelate rings. It is obvious, how-
(45) Two possible orientations of the PH3 ligands are possible for
C2v symmetric structures of 13, with the hydrogen atoms in the
O-Rh-P plane pointing either to the outside or to the inside of the
chelate ring. Although the first one is computed to be slightly more
stable, we employed the second one because of its closer resemblance
to the experimental structures of the chelating bisphosphine complexes.

103Rh Chemical Shift
Organometallics, Vol. 18, No. 7, 1999 1203
ever, that the computed deshieldings in going to smaller
or larger bite angles are far too small to account for the
experimentally observed trends.
The explicit dependence of the δ(103Rh) values on the
Rh-P bond length r computed for 13 is depicted in
Figure 4b. The empirical correlation between X-ray
derived Rh-P distances and experimental 103Rh chemi-
cal shifts for complexes with â > 82° noted earlier47 is
also evident from Figure 4b. With the notable exception
of the four-membered chelate ring 9a (see below), both
computed (dotted line) and empirical (dashed line)
correlations now appear qualitatively similar. The large
slopes [10 400 ppm Å-1 (GIAO-B3LYP) and 12 800 ppm
Å-1 (empirical)] indicate a large sensitivity of δ(103Rh)
toward r. It thus appears that the 103Rh chemical shifts
are to a large extent governed by this bond distance,
which is in turn related to the bite angle for â > 82°.13b,47
Similar findings can be noted in the case of complexes
bearing chelating amine ligands.9i
However, the exceptional deshielding of the 103Rh
nucleus in 7a , 9a cannot be explained solely in terms
of the values of â and r; this is immediately apparent
from the offset of the corresponding data points in
Figure 4a,b. As noted above, the coordination geometry
about the Rh atom is not planar in 9a , with a P-Rh-
P/O-Rh-O twist of θ ) 12.6°. Within this range of
distortion, however, the twist angle θ is found to affect
the computed δ(103Rh) value of 13 only marginally
(Figure 4c).
Another geometrical feature differentiating 9a from
the other chelate complexes 9b-d is the alignment of
the phosphorus “lone pairs” with the respective Rh-P
bond axes. Owing to the short and rigid methylene
bridge in 9a , it appears that these lone pairs are
considerably tilted away from the bond axes. This less
optimal alignment of the lone pairs is clearly evident
2
in the Laplacian ∇ F of the total electronic density of
model compound 12a (Figure 5a). Local charge concen-
2
trations as revealed by maxima of -∇ F reflect the shell
structure of a given atom and indicate the positions of
bonds and lone pairs in the valence shell.48
In a geometrical description, the tilt τ of the lone pair
can be expressed as the angle between the direct Rh-P
connection line and the pseudo-C3 axis at the phospho-
rus atom, which forms identical angles to the three
nearest neighbors in the phosphine moiety (Figure 3).
This pseudo-C3 axis presumably coincides with the
direction of the lone pair and the proposed direction of
one of the principal components of the 31P shielding
tensor.14 Using this definition, the lone pairs at phos-
phorus are tilted by ca. 5° and 16° in the solid state
structures of 9b and 9a , respectively. The unusually
large tilt in the latter is indicative of a less efficient
overlap of the orbitals involved and may thus contribute
appreciably to the 103Rh magnetic shielding. This notion
is corroborated by 103Rh chemical shift calculations for
F igu r e 5. Negative Laplacian of the electron density,
2
-∇ F, in the P-Rh-P plane of compounds 12a -d (B3LYP/
(46) The P-Rh-P angles of nearly all rhodium complexes with
ligands R2P(CH2)2PR2 located in the Cambridge Structure Database
are within this range.
(47) Fornika, R. Ph.D. Thesis, Friedrich-Schiller-Universita¨t J ena,
Germany, 1994.
(48) (a) Bader, R. F. W. Atoms in Molecules. A Quantum Theory;
Clarendon Press: Oxford, 1990. (b) Bader, R. F. W. Chem. Rev. 1991,
91, 893. (c) Bader, R. F. W.; Popelier, P. L. A.; Keith, T. A. Angew.
Chem. 1994, 106, 647; Angew. Chem., Int. Ed. Engl. 1994, 33, 620.
II′ level). Unfilled circles and outlined labels are projections
of nuclei into the plot plane. Solid lines are in regions where
electronic charge is concentrated, dashed lines are in
regions where charge is depleted, with contour values of
(2i × 10j atomic units (i ) 1, 2, 3; j ) -1, 0, 1). The arrow
points to the charge concentration associated with the lone
pair on a phosphorus atom of 12a .

1204 Organometallics, Vol. 18, No. 7, 1999
Leitner et al.
Ta ble 3. BP 86/ECP 1 Op tim ized a a n d
Exp er im en ta lb Geom etr ica l P a r a m eter sc of
Selected Com p lexes Con ta in in g th e
[{R2P (CH2)n P R2}Rh ] F r a gm en t
n
R
r
â
θ
τ
source
12a
9a
12b
11
9b
12c
7c
1
1
2
2
2
3
3
4
4
Me
Cy
Me
Me
Cy
Me
Ph
2.229
2.195
2.215
2.176
2.193d
2.223
2.194
2.230d
2.224d
74.5
73.7
86.2
85.1
85.0
93.1
90.9
98.8
98.9
0.0
12.6
0.2
0.0
7.6
0.0
3.8
2.0
5.1
17.1
16.1
5.0
DFT
X-ray
DFT
3.7
X-ray13
X-ray13
DFT
4.2d
-3.0
-3.9d
-7.6d
-7.5d
X-ray13
DFT
12d
9d
Me
Cy
X-ray13
acac complexes. hfacac complexes. c For definition, see text
a
b
d
and Figure 3. Mean value.
F igu r e 6. Comparison of calculated and experimental
103Rh chemical shift changes ∆δ resulting from variation
of the chain length n in complexes [{R2P(CH2)nPR2}Rh(1,3-
diketonate)]. aSum of the DFT-based ∆δ increments (Table
Ta ble 4. In cr em en ta l 103Rh Ch em ica l Sh ift
Ch a n ges ∆δ Ba sed on Str u ctu r a l P a r a m eter s
Resu ltin g fr om Va r ia tion of th e Ch a in Len gth n in
Com p lexes [{R2P (CH2)n P R2}Rh (1,3-d ik eton a te)]a
b
4). Computed change in δ(103Rh) for the model complexes
c
12a -d (from Table 2). Observed changes in δ(103Rh) for
source of
change
parameter
R
of n
∆δ (r) ∆δ (â) ∆δ (θ) ∆δ (τ) ∑∆δ
complexes 7a -d and 9a -d in solution (from Table 2).
DFTb
DFTb
DFTb
X-rayc
X-rayc
X-rayc
Me
Me
Me
Cy
Ph
2f1
2f3
2f4
2f1
2f3
2f4
145
83
156
21
10
321
24
50
127
34
33
132
0
0
1
277
-65
-59
256
-54
-49
446
68
chelate rings in the solid state structures can be
influenced by packing forces and do not necessarily
reflect the intrinsic structural parameters of the [(P2)-
Rh] fragment in solution.13b On the other hand, this may
also reflect the fact that the structural effects on
δ(103Rh) are not simply additive. This is already appar-
ent from the fact that the increments based on the DFT
structures reproduce the actual, DFT-computed chemi-
cal shift trends only qualitatively, not quantitatively
(Figure 6).
225
346
-8
35
3
-29
Cy
375
a
The ∆δ values are obtained by insertion of each parameter
value in the respective fitting equations of Figure 4 as evaluated
for model compound 13 and forming the difference according to
the pair of structures indicated. acac complexes. c hfacac com-
b
plexes.
13 as a function of τ (Figure 4d). Indeed, a steep increase
of δ(103Rh) is computed for larger values of τ.
Although the complexity of the structural influences
does not allow the derivation of a general and quantita-
tive increment system at this point, the individual
contributions ∆δ in Table 4 can help to identify the
structural parameter(s) most important for δ(103Rh).
Thus, the deshielding of the 103Rh nucleus upon increas-
ing the chelate ring size by elongating the CH2 backbone
is governed simultaneously by the increase of the bite
angle â andseven more pronouncedsthe concomitant
lengthening of the Rh-P bond (Table 4, entries 2, 3).
The decrease of â on going from a five- to a four-
membered chelate ring is not associated with a further
decrease of the Rh-P distance, as this would lead to
unacceptably short Rh-P bonds.49 The deshielding
in the four-membered chelate ring is indicated to arise
primarily from the tilt angle τ as apparent from the
large ∆δ(τ) value in entry 1, Table 4. Although the
phosphorus lone pair deviates from the Rh-P axis also
in the seven-membered chelate (Figure 5d), the result-
ing ∆δ(τ) value is much smaller because of the particular
shape of the fitting function in Figure 4d.
Explicit values for the four structural parameters
discussed above are given in Table 3 for experimental
and theoretical structures of selected neutral complexes
[(P2)Rh(hfacac)] and 12, respectively. Data are taken
from X-ray crystallography for 7, 9 (with R ) Ph, Cy)
and DFT optimizations for the model complexes 12 (with
R ) Me). These data can be inserted into the fitting
equations of Figure 4 derived from model complex 13
in order to provide chemical shift differences resulting
from individual geometric variations between two given
structures. The resulting shift differences are sum-
marized in Table 4 for selected pairs of complexes.
The experimental and computational data for the
changes in δ(103Rh) of complexes [(P2)Rh(1,3-diketo-
nate)] resulting from structural variations in the [(P2)-
Rh] fragment are compared in Figure 6. As expected,
the absolute values of these differences depend on the
substituents at phosphorus and on the applied model
approach. Nevertheless, both sets of computational data
reproduce the overall trends observed experimentally
reasonably well, in view of the inevitable simplifications
of the model approach.
When the individual increments are evaluated using
structural parameters from single-crystal X-ray diffrac-
tion studies, the qualitative agreement between trends
in calculated and experimental 103Rh chemical shifts is
less satisfactory (Table 4). One possible reason for this
discrepancy could be that the conformation of the
(50) J ames, B. R.; Mahajan, D. J . Organomet. Chem. 1985, 279, 31.
(51) (a) Marder, T. B.; Williams, I. D. J . Chem. Soc., Chem. Commun.
1987, 1478. (b) Fornika, R.; Go¨rls, H.; Leitner, W. Unpublished results.
(c) Hall, M. C.; Kilbourn, B. T.; Taylor, K. A. J . Chem. Soc. A 1970,
2539. (d) Burch, R. R.; Calabrese, J . C. J . Am. Chem. Soc. 1986, 108,
5359. (e) Anderson, M. P.; Pignolet, L. H. Inorg. Chem. 1981, 20, 4101.
(f) J ones, R. A.; Real, F. M.; Wilkinson, G.; Galas, A. M. R.; Hursthouse,
M. B.; Malik, K. M. A. J . Chem. Soc., Dalton Trans. 1980, 511.
(52) The four-membered rings are puckered in an anti-fashion with
respect to each other, as found in the solid state for 14a ; a completely
planar arrangement (in D2h symmetry), as essentially observed for solid
14b, is higher in energy by 0.3 kcal/mol at the BP86/ECP1 level.
(49) The Rh-P bond lengths of more than 90% of rhodium complexes
with ligands R2P(CH2)nPR2 located in the Cambridge Structure
Database exceed 2.200 Å.

103Rh Chemical Shift
Organometallics, Vol. 18, No. 7, 1999 1205
Ta ble 5. 103Rh Ch em ica l Sh ifts a n d Key Str u ctu r a l Da ta for Selected Ca tion s [{R2P (CH2)n P R2}2Rh ]+
a n d [(R3P )4Rh ]+
cation
n
R
δ(103Rh) (ref)
anion (ref)
P-Rh-P [deg]
P-Rh [Å]a
twist θ [deg]b
tilt τ [deg]a,c
14a
14b
14c
1
1
2
Ph
Cy
Me
-354d (10a)
hfacac (39)
70.80(2)
70.80(2)
85.2(4)
84.0(4)
84.5
82.1(2)/
83.4(2)
83.5
84.03(6)/
82.71(6)
90.97(7)/
89.50(6)
90.0
2.287(1)
2.287(1)
2.282(1)
2.285(1)
2.316
0
0
0
0
4.8
3.6
17.9
17.1
8.2
8.7
7.7
-383d (10a)
BF4- (39)
-1293e (this work)
C9H7- (51a)
[(cod)RhCl2]- (51b)
none (DFT, this work)
ClO4- (51c)
-1345 (DFT, this work)
-1171e (this work)
14d
2
Ph
2.306(7)
7.6
[Ag(C3F7)]- (51d)
hfacac (39)
2.306
2.334(2)
7.9
18.6
7.8
9.0
14e
14f
2
4
Cy
Ph
-1129e (this work)
BF4- (51e)
2.329(2)
38
9.2
15a
15b
H
Me
-1223 (DFT, this work)
-516 (DFT, this work)
none (DFT, this work)
Cl- (51f)
none (DFT, this work)
2.304
2.297(1)
2.337
12.4
43
41.1
4.3
9.9
8.2
93.86(5)
93.8
a
b
d
Mean value. The angle between two P2Rh planes in analogy with Figure 3. c See Figure 3 for definition. In CD2Cl2. e In acetone-
d6. The data reported for cation 14c in ref 10a, δ(103Rh) ) -901, J (Rh,P) ) 98 Hz, correspond most likely to a Rh(III) species resulting
from oxidative addition of CD2Cl2 (cf. Marder, T. B.; Fultz, W. C.; Calabrese, J . C.; Harlow, R. L.; Milstein, D. J . Chem. Soc., Chem.
Commun. 1997, 1543). f One of two independent molecules in the unit cell.
rangement of the [P4Rh]+ coordination sphere on the
Ch a r t 2. Id ea lized Coor d in a tion Geom etr ies for
Ca tion ic a n d An ion ic Com p lexes of F or m u la
chemical shift of the transition metal has been assessed
by exploratory calculations for the model compound [Rh-
(PH3)4]+ (15a ), as well as for [Rh(PMe3)4]+ (15b) and
[Rh{Me2P(CH2)2PMe2}2]+ (14c). For D2-symmetric 15a ,
a slightly twisted structure (12.4° angle between the
P-Rh-P planes containing adjacent pairs of phos-
phines) and a δ(103Rh) value of -1223 ppm are obtained.
Variations of δ(103Rh) with the structural parameters
illustrated in Figure 3 are qualitatively similar to those
discussed above for 13. However, the sensitivity of δ-
[Rh (P 2)2]+/-
(
103Rh) toward these parameters appears somewhat
2. Ion ic Com p lexes Con t a in in g t h e [(P 2)2R h ]
F r a gm en t. Our analysis has so far been restricted to
neutral four-coordinate complexes of Rh(+I) bearing 1,3-
diketonate ligands, and the question arises to what
extent the qualitative structure/shift relationships might
be affected by the electronic configuration or the co-
ligand. We have therefore extended our studies to
coordination compounds of the composition Rh(P2)2
lacking additional ligands in order to test the generality
of our approach. Most of these complexes are d8 cations,
but some can also be reduced to the corresponding d10
anions (Chart 2),53,54 allowing the investigation of the
role of the electronic configuration at the metal center
without concomitant change of the stabilizing ligand.
Ca tion ic Com p lexes [(P 2)2Rh ]+. Cationic rhodium
complexes with two chelating phosphines represent a
large class of coordination compounds with various
applications in homogeneous catalysis.1,50,51e A large
body of crystallographic data for such species is avail-
able,51 and 103Rh chemical shifts have been determined
in a number of cases.10a,h We have recently synthesized
complexes [(P2)2Rh][hfacac], which are rare examples
of transition metal complexes with noncoordinating
hfacac anions.39 Key structural parameters and
δ(103Rh) values for selected cations [(P2)2Rh]+ (14) are
collected in Table 5.
lower for cationic 15a than for neutral 13: for instance,
the computed chemical shift derivative with respect to
the Rh-P bond length is only ca. 6800 for 15a compared
to more than 10 000 ppm Å-1 for 13.
Exchanging the PH3 ligands in 15a with the bulkier
PMe3 (15b, D2d symmetry) induces a much larger twist
(41.1°), in accord with the experimental value (43°).51f
This remarkable deviation from the square planar
arrangement expected for a d8 transition metal center
results from steric repulsion of the P donor groups.35,51
Accordingly, the distortion of cations containing two
chelating phosphines R2P(CH2)nPR2 increases with the
size of R and the chain length n as illustrated nicely in
the series of compounds 14c,d ,e and 14a ,d ,f, respec-
tively. The concomitant deshielding represents an im-
portant contribution to the 103Rh chemical shift in these
cations. For model compound 15a , the explicit variation
of δ with the twist angle θ can be expressed by the fit
equation δ ) 0.300θ2 - 1268, leading to an extreme
deshielding of more than 2000 ppm for a hypothetical
Rh(+I) cation with two perpendicular [P2Rh] units.
For 14c (D2 symmetry) the optimized [(P2)2Rh] moiety
is essentially planar (twist angle θ ) 4.8°), as found in
the solid state,51b,c and δ(103Rh) ) -1345 is computed
in excellent accord with the experimental value of
-1293 ppm. In addition, [Rh(Me2P(CH2)PMe2)2]+ (14g)
with two four-membered chelate rings has been opti-
mized in C2h symmetry.52 The computed 103Rh chemical
shift of -296 ppm is substantially deshielded with
respect to that of 14c, and the difference ∆δ of 1049 ppm
arising from the geometrical variation is again in
reasonable agreement with the experimental data ob-
In analogy with the approach for neutral 1,3-diketo-
nate complexes, the influence of the geometrical ar-
(53) Kunin, A. J .; Nanni, E. J .; Eisenberg, R. Inorg. Chem. 1985,
24, 1852.
(54) Bogdanovic´, B.; Leitner, W.; Six, C.; Wilczok, U.; Wittmann,
K. Angew. Chem. 1997, 109, 518; Angew. Chem., Int. Ed. Engl. 1997,
36, 502.

1206 Organometallics, Vol. 18, No. 7, 1999
Leitner et al.
served with the bulkier substituents (14a /14d ∆δ ) 817,
14b/14e ∆δ ) 746). Like in the case of the neutral 1,3-
diketonate complexes, this large effect can be again
attributed to the deviation of the Rh-P bond axes from
the ideal orientation of the phosphorus lone pair induced
by the strained four-membered ring.
An ion ic Com p lexes [(P 2)2Rh ]-. The bis-chelated
anions [{Ph2P(CH2)nPPh2}2Rh]- (n ) 2, 16a ; n ) 3, 16b)
with rhodium in the formal oxidation state (-I) are
formed from the corresponding Rh(+I) complexes by
electrochemical53 or chemical reduction.53,54
Despite the change in the formal oxidation state, the
observed 103Rh chemical shifts of the anions are quite
similar to those of the respective cations, e.g., -1018
ppm (16a ) vs -1171 ppm (14d ), or -645 ppm (16b) vs
-593 ppm ([{Ph2P(CH2)3PPh2}2Rh]+). A geometry op-
timization starting from the almost planar structure of
the d8 cation 14c affords an essentially perpendicular
arrangement (θ ) 86.5°) of the two P-Rh-P planes in
the corresponding anion [{Me2P(CH2)2PMe2}2Rh]- (16c),
in accord with the tetrahedral coordination expected for
a d10 center both on electronic and steric grounds. A
103Rh chemical shift of -1000 ppm is computed for 16c,
consistent with the experimental results for 16a (-1018
ppm), and this value is indeed in a similar range as
theoretical and experimental δ(103Rh) data of 14c.
Together with the results for the cations discussed
above, these data provide a rationalization for the fact
that different coordination geometries about Rh are of
paramount importance for the 103Rh chemical shifts in
tetracoordinate phosphorus containing Rh(+I) cations
and Rh(-I) anions and often surpass electronic influ-
ences resulting either from inductive effects or from
arguments based on formal oxidation states.
The same qualitative relationships hold for cationic
complexes with four P donor ligands around a Rh(+I)
center. In addition, steric crowding in cationic complexes
[{R2P(CH2)nPR2}2Rh]+ and [Rh(PR3)4]+ can induce large
deviations from an ideal square planar arrangement
expected for a d8 transition metal center, and the
resulting variation in the twist angle of the two [(P2)-
Rh] fragments contributes considerably to the relative
deshielding of the metal nucleus in complexes with large
bite angles and/or bulky substituents at phosphorus. In
anionic complexes [{R2P(CH2)nPR2}2Rh]- a twist of the
two P-Rh-P planes close to the maximum value of 90°
is realized, consistent with the expected tetrahedral
coordination geometry of a d10 Rh(-I) center. This is one
of the major reasons for the similarity of the chemical
shifts of corresponding [(P2)2Rh] cations and anions,
despite the different formal oxidation states.
In general terms, we have developed a novel approach
to rationalize trends in transition metal chemical shifts
based on contributions from discrete variations of geo-
metrical parameters in the coordination sphere. Not
surprisingly, the complex interplay of the parameters
governing the metal shifts precludes the derivation of
a simple quantitative increment system. Nevertheless,
our approach provides the basis for a more rational
analysis and, thus, for a deeper understanding of
structure/shift relationships than possible interpreta-
tions in terms of the Ramsey equation. In this respect,
our results have important implications for correlations
of 103Rh chemical shifts with reactivities and catalytic
activities: identification of the electronic and geometric
factors governing the metal shifts can help to establish
the intrinsic connection between chemical shifts and
reactivities. However, the interdependence of the rel-
evant parameters can be quite complex even for appar-
ently closely related compounds, as demonstrated here
for example for [{R2P(CH2)nPR2}2Rh(hfacac)] complexes.
Hence, experimentally observed shift/reactivity correla-
tions can generally not be expected to be universally
valid and require structural variations to be intro-
duced very systematically, avoiding large simultaneous
changes in parameters which have opposite effects on
the δ(103Rh) values.
Con clu sion
Depending on chelate ring size and substitution
pattern at phosphorus, the 103Rh chemical shifts of
complexes containing the fragment [(P2)Rh] span a
remarkable wide range. Shift differences of over 800
ppm are encountered even for apparently closely related
system like [{R2P(CH2)nPR2}2Rh(hfacac)]. Based on a
combined experimental and theoretical effort, we have
assessed the relative importance of electronic and
structural variations for δ(103Rh). The electronic effects
could be separated in isostructural compounds [{(p-
XC6H4)2P(CH2)4P(p-XC6H4)2}Rh(hfacac)] as revealed by
a linear correlation of δ(103Rh) with the Hammett σP
constant of X, indicating a direct inductive influence.
This electronic influence leads to variations of the
103Rh chemical shifts not exceeding 80 ppm. In contrast,
small variations in the coordination geometry of neutral
complexes [{R2P(CH2)nPR2}Rh(hfacac)] were found to
induce much larger shift differences. Careful analysis
of the individual contributions of changes in key struc-
tural parameters of these essentially square planar
complexes allowed the rationalization of the observed
trends: The relative deshielding of δ(103Rh) on going to
larger chelate-ring sizes (n > 2) is mainly induced by
the parallel increase of the Rh-P distance and the
P-Rh-P angle. In contrast, the large deshielding of the
103Rh nucleus in four-membered chelate rings (n ) 1)
is a consequence of the substantial tilt of the lone pairs
at phosphorus away from the Rh-P axis.
Ack n ow led gm en t. Financial support by the Max-
Planck-Society, the Fonds der Chemischen Industrie,
and the BMBF is gratefully acknowledged. W.L. wishes
to thank U. Beneke, J ena, for preparative assistance,
and the Degussa AG for a generous loan of RhCl3
×
3H2O. M.B. thanks W. Thiel for his continuous support.
We also thank W. von Philipsborn for a preprint of ref
7e. Calculations were performed on Silicon Graphics
PowerChallenges (Organisch-chemisches Institut, Uni-
versita¨t Zu¨rich) and on IBM RS6000 workstations (C4
cluster, ETH Zu¨rich), as well as on a NEC-SX4 super-
computer (CSCS, Manno, Switzerland).
Su p p or tin g In for m a tion Ava ila ble: A listing of analyti-
cal data (1H, 13C, 31P, 19F NMR, EI-MS) for ligands 3a -e and
complexes 5a -e. This material is available free of charge via
the Internet at http://pubs.acs.org. The crystallographic data
of complex 9a have been deposited at the Cambridge Structure
Database under the deposition number 101723.
OM980980K