Cyanide 13C NMR hyperfine shifts in paramagnetic cyanide-bridged mixed-valence complexes

Article abstract of DOI:10.1039/b717400d

Paramagnetic (hyperfine) NMR shifts in the ¹³C cyanide bridge and ³¹P resonances in a set of mixed valence complexes [(η⁵-C₅R₅)Ru(PPh₃)L( ¹³CN)Ru(NH₃)₅]ⁿ⁺ (R = H; L = PPh₃, CO, NO⁺; R = Me; L = PPh₃) are sensitive to the extent of intermetallic charge-transfer, and are strongly solvent dependent. The Royal Society of Chemistry.

Full text of DOI:10.1039/b717400d

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
13
Cyanide C NMR hyperfine shifts in paramagnetic cyanide-bridged
mixed-valence complexes
W. Michael Laidlaw* and Robert G. Denning
Received (in Cambridge, UK) 12th November 2007, Accepted 18th January 2008
First published as an Advance Article on the web 8th February 2008
DOI: 10.1039/b717400d
Paramagnetic (hyperfine) NMR shifts in the 13C cyanide bridge
and 31P resonances in a set of mixed valence complexes
[(g5-C5R5)Ru(PPh3)L(13CN)Ru(NH3)5]n+ (R
= H; L =
PPh3, CO, NO+; R = Me; L = PPh3) are sensitive to the
extent of intermetallic charge-transfer, and are strongly solvent
dependent.
Fig. 1 Structures of ruthenꢀium mixed valence complexes: n =
3 (1, 2, 4); n = 4 (3). CF3SO3 anion.
In bimetallic cyanide-bridged mixed valence complexes the
cyanide ion mediates intermetallic electronic communication.
In unsymmetrical systems the extent of donor–acceptor charge
transfer between the metals has, for the most part, been
studied by optical and electrochemical methods.1 With the
exception of infra-red studies2 concerning the CN stretching
frequency, little attention has been given to the influence of the
charge transfer on the bridging group, whose role is usually
described by an empirical transfer integral. In this report we
show that the distribution of unpaired electron spin across the
cyanide bridge can be tracked using NMR paramagnetic
shifts.
Me; L = PPh3) with the labile complex [Ru(NH3)5(OSO2CF3)]-
(CF3SO3)2 in acetone under nitrogen at ambient temperature
in the dark for a few days. Analytically pure materials were
isolated through recrystallization.
Apart from quite sharp peaks in the aromatic carbon region
(90–150 ppm), the 13C NMRz solution spectra (298 K) of 1–4
exhibit a very broad unstructured peak (halfwidth B300–
4500 Hz) far downfield (+400–1300 ppm), corresponding to
a paramagnetic shift between B+320 and +1120 ppm,
attributed to the bridging cyanide. Representative spectra for
1–3 in acetonitrile are shown in Fig. 2.
The 13C cyanide chemical shift primarily depends upon the
identity of ligand L and to a lesser extent upon the solvent. We
find generally that d(13CN) for L = PPh3 4 CO 4 NO+ and,
for a given ligand L, d(13CN) in nitromethane (DN = 2.7) 4
Cyanide-bridged species in which the metals have low-spin
d6–d5 configurations, belong in the weakly coupled
Robin–Day3 Class II. If the d5 acceptor is a group VIII metal
pentaammine moiety, the unpaired electron is primarily loca-
lized on this centre, and its reduction potential varies with the
Gutmann4 donor number (DN) of the solvent. Ligand-
mediated coupling redistributes spin density from the para-
magnetic (d5) centre towards the nominally diamagnetic (d6)
metal and its local coordination environment. As a result
pseudo-contact (dipolar) and contact contributions5 to se-
lected NMR chemical shifts can be a sensitive indicator of
the consequences of the donor–acceptor interactions.
Here we identify paramagnetic (hyperfine) shifts in the 13C
NMR signal of the linking cyanide as well as in the 31P
resonances of phosphines coordinated to the d6 metal in a
set of related compounds based on the cyclopentadienyl
triphenylphosphineruthenium(II) and pentaammine rutheniu-
m(III) moieties (Fig. 1). These shifts respond to variations in
the ligand, L, and to changes in the solvent donicity.
The isotopically labelled cyanide (13C; 100%) diruthenium
compounds 1–4 (Fig. 1) were synthesizedw by reacting a slight
excess of the diamagnetic organometallic species, [(Z5-C5R5)-
Ru(PPh3)L(13CN)] (R = H; L = PPh3, CO, NO+ and R =
Inorganic Chemistry Laboratory, South Parks Road, Oxford, UK
OX1 3QR. E-mail: michael.laidlaw@chem.ox.ac.uk. E-mail:
bob.denning@chem.ox.ac.uk; Fax: +44 (0)1865 272690; Tel: +44
(0)1865 272600
Fig. 2 13C NMR spectra of the cyanide-bridge in 1 (top), 2 and 3
(bottom) in d3-acetonitrile at 298 K. Solvent reference, 1.39 ppm.
ꢁc
This journal is The Royal Society of Chemistry 2008
1590 | Chem. Commun., 2008, 1590–1592

View Article Online
acetonitrile (DN = 14.1) 4 acetone (DN = 17) 4 dimethyl-
formamide (DN = 26.6) 4 dimethylsulfoxide (DN = 29.8).
The 31P resonancesz of 1–4 follow similar trends to those of
13CN, showing positive paramagnetic shifts and large line-
widths (B70–400 Hz). These trends can be rationalized in
terms of the degree of redox asymmetry between the two metal
centres. Optical (and electrochemical) measurements show
that as the p-accepting ability of L increases, in the order
PPh3 o CO o NO+, the energy of the donor metal orbitals
decrease, and MMCT excited states are raised in energy. For
example, in d3-acetonitrile, EMMCT E 15 130 (1), E 18 315 (2)
and 4 25 700 (3) cmꢀ1. Similarly, as the donor number of
the solvent increases the pentaammineruthenium(III) centre
is stabilized (through hydrogen bonding interactions)
with respect to reduction, and the MMCT energy rises line-
arly.6–8 According to Hush theory,6,9 as the degree of redox
asymmetry (and MMCT energy) decreases, as in the sequence
3 4 2 4 1 and in the solvent order dimethylsulfoxide 4
dimethylformamide 4 acetone 4 acetonitrile 4 nitro-
methane, delocalization of the dp electron spin increases,
resulting in enhanced hyperfine interactions for nuclei directly
bonded or close to the formally diamagnetic centre.
observations. Within the closely related set of complexes
(1, 2, 3), the range of 31P shifts induced by varying the solvent
is B185, 100 and 30 ppm, respectively. In contrast, the
corresponding variations in 13CN shifts are B135, 200 and
285 ppm. Apparently, as the degree of redox asymmetry
decreases, in the sequence 3 4 2 4 1, the phosphorus centre
becomes more sensitive whereas the 13CN centre becomes less
sensitive to the increase in spin delocalization. There appears
to be a limit beyond which an increase in spin density at the d6
metal, as sensed by the 31P shift, does not cause a matching
increase in spin density on the bridging carbon, even though
they are bonded to the same metal. Moreover, for compound 4
(Fig. 3), the most strongly coupled system, the 13CN chemical
shift reaches a maximum and begins to fall in the least basic
media, whereas the 31P shift continues to increase. This
behaviour is consistent with the cyanide bridge gating spin
density to the metal–phosphorus centre from the paramag-
netic ammine metal in a way that appears to restrict the
population of spin within the bridge orbitals.
A preliminary analysis suggests that Fermi contact contri-
butions dominate the 13C paramagnetic shifts in the cyanide
bridge (and the 31P shifts). Based on the Hush model, less than
3% of electron spin is delocalized in these compounds,6 so an
estimate of the dipolar electron–nuclear interaction between
the 13C nucleus and the d5 metal centre can be obtained from
experimental EPR g-values.10 On this basis pseudo-contact
contributions are either far too small (for compounds 1 and 2),
or have the wrong sign (for compound 3) to account for the
observed shifts. Local ligand-centred pseudo-contact contri-
butions arising from spin density in the p-system are difficult
to quantify, but may be significant. Nonetheless, the 13C NMR
cyanide resonances appear to be sensitive probes of the local
spin-density in the bridge.
We are not aware of any data on 13C paramagnetic shifts of
cyanide ions bound through carbon to ruthenium(III). Deter-
mining the 13C paramagnetic shifts in the monoruthenium(III)
compounds [(Z5-C5R5)Ru(PPh3)L(13CN)]+ (R = H; L =
PPh3, CO, NO+ and R = Me; L = PPh3) would enable a
limiting reference point for the 13C shifts observed in 1–4.
Unfortunately, electrochemical studies reveal one-electron
oxidation of the organoruthenium(II) centres to be irreversible
when R = H; L = PPh3, CO, and beyond +2 V when L =
NO+. When R = Me and L = PPh3, the Ru(II/III) couple is
reversible but NMR experiments on solutions generated by the
stoichiometric addition of a one-electron oxidant in an NMR
tube have so far been unsuccessful. The limit of 13C and 15N
shift data have been reported for low-spin Fe(III) in hexa-
cyanoferrate(III),11 porphyrins and heme proteins.5,12 The 13C
resonances are far upfield (Bꢀ2000 to ꢀ4200 ppm) whereas
the 15N peaks are found far downfield (B+600 to
+1000 ppm). The alternation in sign is indicative of a domi-
nant contact spin-polarization mechanism. In compounds 1–4,
the downfield 13C shifts are consistent with a similar mechan-
ism since the cyanide nitrogen is bonded to the paramagnetic
metal centre. The peaking of the 13C shifts in Fig. 3 may
therefore reflect the onset of competing spin delocalization and
core polarization effects in mixed-valence systems, and de-
serves a more comprehensive theoretical treatment and experi-
mental investigation. The sign of the 13C bridging cyanide
Counteranion interactions also influence the redox asym-
metry.8 In the weak donor solvent nitromethane, the 13C
resonance of 1 shifts by ꢀ3 ppm when the trifluoromethane-
sulfonate anion is replaced by the more weakly interacting
hexafluorophosphate anion, whereas the 31P peak shifts
by +52 ppm.
Optical measurements reveal there are strong linear correla-
tions between dhyp(31P) and EMMCT. Thus, dhyp(31P) (1)/ppm
= ꢀ0.0467EMMCT/cmꢀ1 + 973.8; R2 = 0.994 and dhyp(31P)
(2)/ppm = ꢀ0.0283EMMCT/cmꢀ1 + 696.7; R2 = 0.995.
(Charge transfer energies for 3 are obscured in the ultraviolet.)
In contrast, the relationships between dhyp(13C) and EMMCT
are unambiguously nonlinear.
Fig. 3 shows a monotonic but nonlinear plot of the hyper-
fine shifts dhyp(31P) against dhyp(13CN) for the complete set of
Fig. 3 Correlation between 13C cyanide and 31P NMR paramagne-
tically induced (hyperfine) chemical shifts in 1–4 in various aprotic
solventsz.
ꢁc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 1590–1592 | 1591

View Article Online
z
13C NMR spectra (125.70 MHz) of the trifluoromethanesulfonate
resonance might tell apart linkage isomers of Class II mixed-
valence complexes13 comprising cyanide-bridged diamagnetic
and paramagnetic centres.
salts of the bimetallic paramagnets were recorded on concentrated
samples (B15 mg in 0.5 mL) at 298 K on a Bruker AVII 500
spectrometer with a 13C optimised cryoprobe in the following solvents
and referenced accordingly (d3-nitromethane (except 4) 62.54 ppm; d3-
acetonitrile 1.39 ppm; d6-acetone 29.9 ppm; d7-dimethylformamide
(except 4) 163.15 ppm; d6-dimethylsulfoxide 39.5 ppm). In addition, 1
and 4 were recorded as hexafluorophosphate salts in nitromethane.
Paramagnetically induced (hyperfine) shifts were calculated from
the 13CN resonance of the appropriate diamagnetic monometallic
precursorw.
We thank Dr T. D. W. Claridge for his time recording the
13C NMR spectra and Jesus College, Oxford for help with
research expenses.
Notes and references
31P NMR spectra (202.38 MHz) were recorded at B298 K on a Varian
UNITY 500 spectrometer on the same samples. Chemical shifts were
referenced to a capillary of D3PO4 (0 ppm) inserted into each solution.
Hyperfine shifts were calculated from 31P resonances of the appro-
priate diamagnetic organometallic fragment. The 13CN and 31P reso-
nances of the organometallic precursors were almost independent of
solvent, varying at most by 2.5 ppm.
w The 13C cyanide labelled organometallic precursors, [(Z5-C5R5)-
Ru(PPh3)213CN] (R = H, Me) were synthesised from Na13CN (CSK
Gas Products) and [(Z5-C5R5)Ru(PPh3)2Cl] following published pro-
cedures:14 (R = H), IR/KCl n(13CN) = 2024 cmꢀ1; NMRz/d2-
dichloromethane d(13CN) = 140.4 ppm, d(31P) = 50.3 ppm; (R =
Me) IR/KBr n(13CN) = 2022 cmꢀ1; NMRz/d6-acetone d(13CN) =
148.4 ppm, d(31P)
=
53.8 ppm. The nitrosyl derivative, [(Z5-
C5H5)Ru(PPh3)(NO)(13CN)](CF3SO3), was synthesised from [(Z5-
C5H5)Ru(PPh3)213CN] according to the literature method,14 but aqu-
eous trifluoromethanesulfonic acid was employed instead of NH4PF6
to isolate the desired salt. Analysis: found (calc.) C, 47.5 (47.6); H, 3.0
(3.2); N, 4.4 (4.4)%; IR/KBr n(13CN) = 2090, n(NO) = 1866 cmꢀ1
(broad); NMRz/d6-acetone d(13CN) = 110.9 ppm, d(31P) = 39.4 ppm.
Yield 85%.
1 For example, (a) M. A. Watzky, A. V. Macatangay, R. A. Van
Camp, S. E. Mazzetto, X. Song, J. F. Endicott and T. Buranda, J.
Phys. Chem. A, 1997, 101, 8441; (b) Y. Dong and J. T. Hupp,
Inorg. Chem., 1992, 31, 3170; (c) C. A. Bignozzi, R. Argazzi, J. R.
Schoonover, K. C. Gordon, R. B. Dyer and F. Scandola, Inorg.
Chem., 1992, 31, 5260; (d) G. U. Bublitz, W. M. Laidlaw, R. G.
Denning and S. G. Boxer, J. Am. Chem. Soc., 1998, 120, 6068.
2 (a) A. V. Macatangay, S. E. Mazzetto and J. F. Endicott, Inorg.
Chem., 1999, 38, 5091; (b) A. V. Macatangay and J. F. Endicott,
Inorg. Chem., 2000, 39, 437; (c) M. A. Watzky, J. F. Endicott, X.
Song, Y. Lei and A. Macatangay, Inorg. Chem., 1996, 35, 3463.
3 M. B. Robin and P. Day, Adv. Inorg. Chem. Radiochem., 1967, 10, 247.
4 V. Gutmann, Electrochim. Acta, 1976, 21, 661.
The carbonyl, [(Z5-C5H5)Ru(PPh3)(CO)(13CN)], was prepared by
heating [(Z5-C5H5)Ru(PPh3)213CN] (0.5 g, 0.7 mmol) in deaerated
decalin (B80 mL) and 1,2-dichlorobenzene (B10 mL) at B180 1C for
2 hours whilst passing carbon monoxide through the solution. When
the reaction was complete [monitored by 1H NMR (CDCl3) d(C5H5)
= 5.0 ppm (product), 4.4 ppm (reactant)] the solution was mixed with
toluene and chromatographed on alumina. After washing the column
with pentane the product was eluted with tetrahydrofuran, isolated by
evaporation and dried under vacuum. Yield 0.3 g (89%). Analysis:
found (calc.) C, 62.4 (62.2); H, 4.1 (4.2); N, 3.1 (2.9)%; IR/KCl
n(13CN) = 2058, n(CO) = 1968 cmꢀ1 (broad); NMRz/d-chloroform
d(13CN) = 131.4, d(31P) = 53.3 ppm.
5 I. Bertini and C. Luchinat, Coord. Chem. Rev., 1996, 150, 29.
6 W. M. Laidlaw and R. G. Denning, J. Chem. Soc., Dalton Trans.,
1994, 1987.
7 C. J. Adams, N. G. Connelly, N. J. Goodwin, O. D. Hayward, A.
G. Orpen and A. J. Wood, Dalton Trans., 2006, 3584.
8 W. M. Laidlaw and R. G. Denning, Inorg. Chim. Acta, 1996, 248, 51.
9 N. S. Hush, Prog. Inorg. Chem., 1967, 8, 391.
The known bimetallic complexes 1 and 4 were synthesised by a slight
modification of the published method.6 The syntheses of the new
compounds 2 and 3 were achieved following a similar method,
specified herewith for 2. [(Z5-C5H5)(PPh3)(CO)Ru(13CN)] (0.1 g,
0.21 mmol) was pre-dissolved in dichloromethane (1–2 mL) and
acetone (B20 mL) and placed under nitrogen (or argon). (Dichloro-
methane is unnecessary for preparing 3). [Ru(NH3)5-
(OSO2CF3)](CF3SO3)2 (0.12 g, 0.19 mmol) was added to the stirred
solution and the mixture was warmed in the dark at 30 1C for a few
days. The solution was rotor-evaporated to dryness (o35 1C). The
residue was dissolved in a few drops of acetone and ethanol was added
dropwise to the swirled solution to produce a crystalline solid.
Diethylether (B2 mL) was added dropwise to the swirled mixture
which was then refrigerated (ꢀ25 1C; 2 hours). The supernatant
solution was removed and the product was isolated by filtration,
washed with diethylether and dried under vacuum. Typical yield
10 W. M. Laidlaw, D. M. Murphy and R. G. Denning, manuscript in
preparation. X-Band EPR studies of 1 (B5 K; acetonitrile) yield g1
E 2.79, g2 E 1.52 and g3 E 1.22. Taking g8 = g1 along the
intermetallic axis and g>
=
12(g2 + g3), the point dipole pseudo-
contact contribution to the 13C cyanide chemical shift at 300 K is
estimated, using dpc E 173(g8 ꢀ g>2)(3cos2y ꢀ 1)/r3, to be
2
B+64 ppm where y = 0 and r = 3.17 A. For 3 (B5 K; acetone)
with g8 = 1.03 and g> = 2.27 a pseudo-contact shift of Bꢀ44 ppm
is calculated at 300 K.
11 (a) B. R. McGarvey and J. Pearlman, J. Magn. Reson., 1969, 1,
178; (b) D. G. Davis and R. J. Kurland, J. Chem. Phys., 1967, 46,
388.
12 (a) I. Morishima and T. Inubushi, J. Am. Chem. Soc., 1978, 100,
3568; (b) H. Fujii, J. Am. Chem. Soc., 2002, 124, 5936; (c) H. Fujii
and T. Yoshida, Inorg. Chem., 2006, 45, 6816.
ꢀ
0.17 g (80%). Portions of 1 and 4 were converted to PF6 salts using
13 For example, (a) A. Geiss and H. Vahrenkamp, Inorg. Chem.,
2000, 39, 4029; (b) C. J. Adams, K. M. Anderson, M. Bardaji,
N. G. Connelly, N. J. Goodwin, E. Llamas-May, A. G. Orpen and
P. H. Rieger, Dalton Trans., 2004, 683.
14 W. M. Laidlaw and R. G. Denning, J. Organomet. Chem., 1993,
463, 199.
ammonium hexafluorophosphate in aqueous methanol.
Analytical data: 1 (blue); IR/KCl n(13CN) = 1970 cmꢀ1: 2 (magenta);
found (calc.) C, 30.1 (30.1); H, 3.1 (3.2); N, 7.2 (7.5)%; IR/KCl
n(13CN) = 2064, n(CO) = 1990 cmꢀ1: 3 (yellow); found (calc.) C,
26.6 (26.5); H, 3.0 (2.8); N, 7.4 (7.7)%; IR/KCl n(13CN) = 2122, n(NO)
= 1892 cmꢀ1: 4 (turquoise green); IR/KBr n(13CN) = 1942 cmꢀ1
.
ꢁc
This journal is The Royal Society of Chemistry 2008
1592 | Chem. Commun., 2008, 1590–1592