Selective anion effects in chiral complexes of iridium via diffusion and HOESY data: Relevance to catalysis

Article abstract of DOI:10.1039/b110066c

¹H and ¹⁹F Pulsed Gradient Spin Echo (PGSE) diffusion data, together with ¹H, ¹⁹F-HOESY results are shown to distinguish between different types of anion/cation interactions in chiral dihydrido-P,N-complexes of Ir(III); in CD₂Cl₂ the diffusion coefficients, D, for the BArF and the Ir-cation suggest ion-pairing whereas the D-values for PF₆- reveal independent motion; the PF₆- approaches the cation via a specific pathway; the combined PGSE/HOESY approach offers a unique opportunity for exploring anion effects in organometallic/catalytic chemistry.

Full text of DOI:10.1039/b110066c

Selective anion effects in chiral complexes of iridium via diffusion and
HOESY data: relevance to catalysis†
Daniela Drago,a Paul S. Pregosin*a and Andreas Pfaltzb
a Laboratory of Inorganic Chemistry, ETH HCI, Ho¨nggerberg CH-8093, Zu¨rich, Switzerland.
E-mail: PREGOSIN@INORG.CHEM.ETHZ.CH
b Laboratory of Organic Chemistry, University of Basel, CH-4056 Basel, Switzerland
Received (in Cambridge, UK) 5th November 2001, Accepted 29th November 2001
First published as an Advance Article on the web 23rd January 2002
1H and 19F Pulsed Gradient Spin Echo (PGSE) diffusion
dinate Ir(
) models [Ir(1,5-COD)(3)](anion), 4 and [Ir(CNt-
I
1
data, together with H, 19F-HOESY results are shown to
Bu)(3)](anion), 5. The diffusion results confirm that the cations
2
distinguish between different types of anion/cation inter-
actions in chiral dihydrido-P,N-complexes of Ir(III); in
CD2Cl2 the diffusion coefficients, D, for the BArF and the
move slower with increasing molecular size, but that the PF6
anion moves much faster and thus independently of the cation
(see Fig. 1). In a direct comparison, e.g. [Ir(CNtBu)(3)]PF6, 5a
vs. [Ir(CNtBu)(3)]BArF, 5b or the P(m-tolyl)3 compounds 8a
vs. 8b, the two cations show almost identical diffusion
constants, but the two anions move at very different rates. The
data for the BArF anion in 6b and 9b suggest ion-pairs;10
however, in 4b, 5b, 7b and 8b it is likely that the BArF anion is
moving independently.10 In the two four-co-ordinate BArF
complexes, 4b and 5b, we observe HOESY contacts from the
BArF to the cation. Therefore, even though 4b and 5b are not
ion-paired, the BArF is close-by. There are no close contacts
2
Ir-cation suggest ion-pairing whereas the D-values for PF6
reveal independent motion; the PF62 approaches the cation
via a specific pathway; the combined PGSE/HOESY ap-
proach offers a unique opportunity for exploring anion
effects in organometallic/catalytic chemistry.
It is not unusual to find that both the structure and reactivity of
a transition metal cationic salt depend upon the anion, since
anions may be either weakly or strongly co-ordinating. In
several homogeneously catalysed reactions1,2 one finds that the
anion plays a significant role, e.g. reaction rates may be
excellent or only modest.
Table 1 Diffusion data for 4–9
We have recently suggested3 that Pulsed Gradient Spin Echo
(PGSE4,5) diffusion data can be useful in determining molecular
volume. If the complex possesses a fluorine containing salt
(BF42 or PF6 etc.) one can use PGSE data from both 1H (for the
cation) and 19F (for the anion) to obtain a measure of the ion-
pairing,6,7 e.g. for 1 and 2 the 19F diffusion data reveal that these
anions move at ca. the same rate as their respective larger
cations in CDCl3 solution.5 Consequently, these salts exist as
ion-pairs.
10210 D/m2
s21
Recently, a dependence of the reactivity on the anion in Ir-
catalysed hydrogenation, shown in eqn. (1), has been observed.7
Compound
Nucleus Fragment
Ir(COD)(3)PF6 4a
1H
19F
1H
1H
1H
19F
1H
1H
1H
19F
1H
1H
1H
19F
1H
1H
1H
19F
1H
1H
1H
19F
1H
1H
Cation
PF6
Cation
BArF
9.92(6)
2
13.78(6)
9.73(6)
8.90(6)
9.14(6)
13.84(6)
9.09 (6)
8.92(6)
9.05 (6)
13.65(6)
8.09(6)
7.93(6)
9.67(6)
13.18(6)
9.43(6)
8.55(6)
7.70(6)
13.43(6)
7.87(6)
8.13(6)
8.54(6)
14.27(6)
7.96(6)
8.24(6)
Ir(COD)(3)BArF 4b
Ir(CNtBu)(3)PF6 5a
Cation
2
PF6
Ir(CNtBu)(3)BArF 5b
IrH2(Mebipy)(3)PF6 6a
IrH2(Mebipy)(3)BArF 6b
IrH2(TMEDA)(3)PF6 7a
IrH2(TMEDA)(3)BArF 7b
IrH2(PAr3)2(3)PF6 8a
IrH2(PAr3)2(3)BArF 8b
IrH2(TROPPh2)(3)PF6 9a
IrH2(TROPPh2)(3)BArF 9b
Cation
BArF
Cation
(1)
2
PF6
Cation
BArF
cation
Using the P,N ‘PHOX’ ligand, 3,7 shown complexed to Ir, Aryl
= o-tolyl, the rate of hydrogenation is faster with the anion
BArF than with PF62. We show here that 1H and 19F PGSE and
HOESY measurements on the model compounds 4–9, anion =
2
PF6
cation
BArF
cation
2
PF6 or BArF, reveal marked differences in how these two
anions interact with the corresponding cations.
2
PF6
The model complexes IrH2(chelate)(3), 6–9, were prepared
by standard procedures.8 Table 1 gives PGSE data from 48
different measurements,9 including those for the four-coor-
Cation
BArF
cation
2
PF6
cation
BArF
† Electronic supplementary information (ESI) available: spectroscopic data
for 6b. See http://www.rsc.org/suppdata/cc/b1/b110066c/
286
CHEM. COMMUN., 2002, 286–287
This journal is © The Royal Society of Chemistry 2002

close enough to slow the intermolecular attack leading to the
bridging hydride. We believe this combined diffusion/HOESY
approach to anion effects in catalysis to be unique. The results
do not directly explain the observed differences in catalytic
activity since we employed model compounds; however, the
data from this approach offer a new and more direct way of
studying this type of problem.
We thank Dr Massimiliano Valentini both for the measure-
ments on the 1,5-COD complexes and for his leadership. P. S. P.
thanks the Swiss National Science Foundation and the ETH
Zurich for support. We also thank Johnson Matthey for precious
metal salts.
Notes and references
1 B. M. Trost and R. C. Bunt, J. Am. Chem. Soc., 1998, 120, 70.
2 E. P. Kündig, C. M. Saudan and G. Bernardinelli, Angew. Chem., 1999,
111, 1298; E. P. Kündig, C. M. Saudan and F. Viton, Adv. Synth. Catal.,
2001, 343, 51; D. A. Evans, J. A. Murray, P. von Matt, R. D. Norcross
and S. J. Miller, Angew. Chem., 1995, 107, 864.
3 A. Pichota, P. S. Pregosin, M. Valentini, M. Worle and D. Seebach,
Angew. Chem., Int. Ed., 2000, 39, 153; M. Valentini, P. S. Pregosin and
H. Rœegger, Organometallics, 2000, 19, 2551; M. Valentini, P. S.
Pregosin and H. Ru¨egger, J. Chem. Soc., Dalton Trans., 2000, 4507.
4 W. S. Price, Annu. Rep. Nucl. Magn. Reson. Spectrosc., 1996, 32, 51; P.
Stilbs, Prog. Nucl. Magn. Reson. Spectrosc., 1987, 19, 1.
Fig. 1 PGSE diffusion data for 6a showing that the PF62 anion (8) moves
much faster than the slower Ir-cation (2).
between the BArF and the dihydro Ir(III)-cations in any of these
1
complexes. The H, 19F-HOESY data, e.g. in Fig. 2, for the
PF62 anion indicate a specific interaction with the cation, i.e. in
5 R. M. Stoop, S. Bachmann, M. Valentini and A. Mezzetti, Organome-
tallics, 2000, 19, 4117; S. Beck, A. Geyer and H. H. Brintzinger, Chem.
Commun., 1999, 2477; B. Olenyuk, M. D. Lovin, J. A. Whiteford and P.
J. J. Stang, J. Am. Chem. Soc., 1999, 121, 10434; C. B. Gorman, J. C. J.
Smith, M. W. Hager, B. L. Parkhurst, H. Sierzputowska-Gracz and C. A.
Haney, J. Am. Chem. Soc., 1999, 121, 9958; C. Zuccaccia, G.
Bellachioma, G. Cardaci and A. Macchioni, Organometallics, 2000, 19,
4663.
6 P. Pregosin and M. Valentini, Helv. Chim. Acta, 2001, 84, 2833; Y.
Chen, M. Valentini, P. S. Pregosin and A. Albinati, Inorg. Chim. Acta,
2001, in press.
7 A. Lightfoot, P. Schnider and A. Pfaltz, Angew. Chem., Int. Ed., 1998,
37, 2897.
2
4a–7a the PF6 anion does not make a random approach (see
10) but rather moves toward the cation via the oxazoline ring.11
For 8 and 9 the large phosphine ligands12 block this PF6
2
8 Synthesis of 6b: To [Ir(1,5-COD)(3)](BArF) (30 mg, 0.0190 mmol) in
distilled acetone (0.8 ml) was added 4,4A-dimethyl-2,2A-bipyridine (3.5
mg, 0.0190 mmol). Hydrogen gas was then bubbled gently through the
solution at RT for 3.5 h. The solvent was removed in vacuo and the
yellow residue washed with pentane. Recrystallization from ether/
pentane gave an analytically pure product. Yield: 25 mg (79%).
9 The anion (19F for PF6, 1H for BArF) and the cation (1H) are measured
separately, twice, with 100 ms and 150 ms delays. The diffusion
constant given represents an average of the two runs. The D-values stem
from 22–25 points per measurement to characterize the line and our R-
values are always > 98%. The complexes were dissolved in distilled
CD2Cl2 and measured without spinning. The shape of the gradient was
rectangular, its length was 5 ms, and the strength was varied in the
course of the experiment. The diffusion coefficients reported in Table 1
were estimated using the diffusion coefficient of HDO in D2O as a
reference. Measuring the HDO and the complex under the same
conditions allows one to use the following simple relationship shown to
obtain the D-value of the complex.
approach completely. This regiospecific approach is interesting
but need not be of catalytic relevance.
Ê
ˆ
È
Í
˘
˙
mHDO
mHDO /(D - d / 3)
DHDO
=
=
Á
˜
m
m
/(D - d / 3)
D
Ë
¯
Í
Î
˙
˚
compl
compl
compl
10 We suggest ion-pairing when the BArF anion diffusion constant is close
to that of the cation and simultaneously when the diffusion constant for
the cation decreases on going from the PF62 to the BArF anion.
11 During the preparation of this communication a report on the use of
HOESY in connection with ion-pairing in [IrH2(bipy)(PPh3)2](anion)
appeared: A. Macchioni, C. Zuccaccia, E. Clot, K. Gruet and R. H.
Crabtree, Organometallics, 2001, 20, 2367. This paper emphasizes the
placement of the anion near the bipy and away from the hydrides. There
are no contacts between the hydrides and either PF62 or BArF in 6–9 so
that the anion is remote from the hydride ligands.
12 For the P–olefin bidentate ligand, TROPPh2 see: S. Boulmaaz, M.
Mlakar, S. Loss, H. Schönberg, S. Deblon, M. Wörle, R. J. Nesper and
H. J. Grutzmacher, Chem. Commun., 1998, 2623; J. Liedtke, S. Loss, G.
Alcaraz, V. Gramlich and H. J. Grutzmacher, Angew. Chem., 1999, 111,
1724.
Fig. 2 HOESY spectrum of 6a. There are three types of contacts, one weak
and two strong, to: (a) oxazoline methine H-4, trans to the tBu group; (b) the
bipyridyl methyls H-29 and H-35 and (c) the NCH H-30 and H-33 from the
bipyridyl moiety. The PF6 has no contacts to the hydride ligands (400
MHz, CD2Cl2).
2
We assume that the catalytically active Ir-species are
deactivated via trimerization to bridging hydrides.13 For the
2
cation/PF6 combination, the cation is more available for
intermolecular reaction in that the anion is often further away
from the metal. For the cation/BArF pair, ion-pairing is
possible, although it may be sufficient for the large BArF to be
13 D. F. Chodosh, R. H. Crabtree, H. Felkin and G. E. Morris, J.
Organomet. Chem., 1978, 161, 67 suggest deactivation in a related
system.
CHEM. COMMUN., 2002, 286–287
287