A 13C and 15N Solid-State NMR Study of Structural Disorder and Aurophilic Bonding in AuI and AuIII Cyanide Complexes

Article abstract of DOI:10.1021/ic8022198

Solid-state nuclear magnetic resonance has been used to study several cyanoaurates. Carbon-13 and nitrogen-15 NMR spectra of samples enriched with isotopically labeled ¹³C, ¹⁵N cyanide ligands were recorded for stationary samples and

Full text of DOI:10.1021/ic8022198

Inorg. Chem. 2009, 48, 2316-2332
13
15
A C and N Solid-State NMR Study of Structural Disorder and
Aurophilic Bonding in AuI and AuIII Cyanide Complexes
Kristopher J. Harris and Roderick E. Wasylishen*
Department of Chemistry, UniVersity of Alberta, Edmonton, Alberta, T6G 2G2, Canada
Received November 19, 2008
Solid-state nuclear magnetic resonance has been used to study several cyanoaurates. Carbon-13 and nitrogen-15
NMR spectra of samples enriched with isotopically labeled 13C,15N cyanide ligands were recorded for stationary
samples and samples spinning at the magic angle. Several salts of the dicyanoaurate(I) anion, M[Au(CN)2], where
M ) n-butylammonium, potassium, and thallium, were studied via solid-state NMR. A gold(III) cyanide, K[Au(CN)4],
was also investigated. Carbon-13 and nitrogen-15 chemical shift tensors are reported for each salt, as are the
measured 13C,15N direct dipolar coupling constants together with the related derived cyanide bond lengths, r(C,N).
The value for r(C,N) in [(n-C4H9)4N][Au(CN)2], 1.17(5) Å, was determined to be more realistic than a previously
reported X-ray diffraction value of 1.03(4) Å. Large 13C NMR line widths from Tl[Au(CN)2], 250-315 Hz, are attributed
to coupling with 197Au (I ) 3/2) and/or 203/205Tl (I ) 1/2), as confirmed by measurements of the transverse relaxation
constant, T2. Investigation of the carbon-13 chemical shifts for cyanide ligands bound to gold involved in a variety
of metallophilic bonding environments demonstrates that the chemical shift is sensitive to metallophilic bonding.
Differences in Au-Tl metallophilic bonding are shown to cause a difference in the isotropic carbon-13 chemical
shift of up to 15.7 ppm, while differences in Au-Au aurophilic bonding are found to be responsible for a change
of up to 5.9 ppm. The disordered polymeric material gold(I) monocyanide, AuCN, was also investigated using 13
C
and 15N SSNMR. Two-dimensional 13C,13C double-quantum dipolar-recoupling spectroscopy was used to probe
connectivity in this material. The 13C NMR site multiplicity in AuCN is explained on the basis of sensitivity of the
carbon-13 chemical shift to aurophilic bonding of the directly bonded gold atom. This assignment allows estimation
of the position of the linear [-M-CN-]∞ chain’s position with respect to the neighboring polymer chain. For the
samples studied, a range of 7 ( 2% to 25 ( 5% of the AuCN chains are found to be “slipped” instead of aligned
with the neighboring chains at the metal position.
1. Introduction
now understood to occur through a relativistic strengthening
of the van der Waals interaction, an increase which can be
enough to rival the energy of a hydrogen bond.6 Much of
the current research on gold compounds focuses on using
aurophilic bonds for controlling supramolecular struc-
tures,5,7-9 or studying the unusual photophysics often
associated with the presence of aurophilic bonds.6,10-12 For
example, recent studies have found that the photoluminesence
Molecules containing gold have been recognized for some
time as having the potential to form an “extra” bond between
closed-shell gold cations, despite the expected Coulombic
repulsion.1-5 While this interaction between heavy, closed-
shell atoms has been assigned a variety of labels, for
convenience we here apply the terms aurophilic, to describe
gold-gold interactions, and metallophilic, to describe the
analogous heteronuclear bonds. Metallophilic interactions are
(6) Pyykko¨, P. Chem. ReV. 1997, 97, 597–636.
(7) Braga, D.; Grepioni, F.; Desiraju, G. R. Chem. ReV. 1998, 98, 1375–
1405.
(8) Leznoff, D. B.; Lefebvre, J. Gold Bull. 2005, 38, 47–54.
(9) Katz, M. J.; Sakai, K.; Leznoff, D. B. Chem. Soc. ReV. 2008, 37,
1884–1895.
(10) Gade, L. H. Angew. Chem., Int. Ed. 1997, 36, 1171–1173.
(11) Yam, V. W.-W.; Lo, K. K.-W. Chem. Soc. ReV. 1999, 28, 323–334.
(12) Yam, V. W.-W.; Cheng, E. C.-C. Top. Curr. Chem. 2007, 281, 269–
309.
* To whom correspondence should be addressed. Phone: (780)492-4336.
Fax: (780)492-8231. E-mail: Roderick.Wasylishen@ualberta.ca.
(1) Jones, P G. Gold Bull. 1981, 14, 102–118.
(2) Jones, P. G. Gold Bull. 1983, 16, 114–124.
(3) Jones, P. G. Gold Bull. 1986, 19, 46–56.
(4) Scherbaum, F.; Grohmann, A.; Huber, B.; Kru¨ger, C.; Schmidbaur,
H. Angew. Chem., Int. Ed. 1988, 27, 1544–1546.
(5) Schmidbaur, H. Chem. Soc. ReV. 1995, 24, 391–400.
2316 Inorganic Chemistry, Vol. 48, No. 5, 2009
10.1021/ic8022198 CCC: $40.75  2009 American Chemical Society
Published on Web 01/23/2009

Structural Disorder and Aurophilic Bonding
of gold-containing materials is influenced by factors such
as temperature alteration,13 freezing of a solution containing
a gold(I) complex,14 or introduction of an organic solvent
to a solid gold(I) material, either as a vapor15-19 or as a
liquid.20 Rational design of functional materials based on
these properties is already underway in several labs: detection
of alkali metal cations has been shown to be possible through
the design of a suitable gold(I) compound with crown ether
pendants,21,22 and a sensor for volatile organic compounds
based on the vapochromism of a gold(I) material has already
been built.15
The cyano complexes of gold are of great commercial
importance, through their useage in gold extraction,23 and
are also of medicinal interest.24 Aurophilic interactions
between gold atoms in approximately linear dicyanoaurate
anions, [NCAuCN]-, are strong enough that the anions
associate into oligomers in solution25,26 and also affect the
structure of materials containing this anion.5,7-9 Combining
the abilities of both cyanide ligands and gold atoms to bridge
metals, through bidentate or metallophilic bonding, has led
to the development of new heterometallic network materials,
with a plethora of novel features. Research into coordination
polymers composed of cyanide and gold has been recently
reviewed8,9 and has continued to be an area of active
research.13,27-37 Coordination polymers of this type have
demonstrated unusually high optical birefringence;38,39 tun-
able electron lone-pair stereochemistry of Pb(II);37
and the
ability to reversibly absorb water,40,41 organic vapors,18 or
ammonia,19 the latter two causing an observable difference
in the photophysics. The heterometallic nature of these
polymers has also allowed the incorporation of metals with
accessible high-spin electronic states, providing a means to
study magnetic properties.18,28,29,31,32,34-36,40-59 Magnetic
exchange between electron spins on adjacent metal centers
has been shown to occur along diamagnetic gold(I) or
gold(III) cyanide chains.18,35,40,43,45,54,57,58 As well, materials
containing dicyanoaurate bridges and aurophilic bonds have
been investigated in regard to promising behaviors with
respect to transitions between high- and low-spin states of
neighboring metals.47,53
(33) Chen, J.-X.; Zhang, W.-H.; Tang, X.-Y.; Ren, Z.-G.; Li, H.-X.; Zhang,
Y.; Lang, J.-P. Inorg. Chem. 2006, 45, 7671–7680.
(34) Madalan, A. M.; Avarvari, N.; Andruh, M. Cryst. Growth Des. 2006,
6, 1671–1675.
(35) Katz, M. J.; Shorrock, C. J.; Batchelor, R. J.; Leznoff, D. B. Inorg.
Chem. 2006, 45, 1757–1765.
(36) Qu, J.; Gu, W.; Liu, X. J. Coord. Chem. 2008, 61, 618–626.
(37) Katz, M. J.; Michaelis, V. K.; Aguiar, P. M.; Yson, R.; Lu, H.;
Kaluarachchi, H.; Batchelor, R. J.; Schreckenbach, G.; Kroeker, S.;
Patterson, H. H.; Leznoff, D. B. Inorg. Chem. 2008, 47, 6353–6363.
(38) Katz, M. J.; Aguiar, P. M.; Batchelor, R. J.; Bokov, A. A.; Ye, Z.-
G.; Kroeker, S.; Leznoff, D. B. J. Am. Chem. Soc. 2006, 128, 3669–
3676.
(39) Katz, M. J.; Kaluarachchi, H.; Batchelor, R. J.; Bokov, A. A.; Ye,
Z.-G.; Leznoff, D. B. Angew. Chem., Int. Ed. 2007, 46, 8804–8807.
(40) Dong, W.; Zhu, L.-N.; Sun, Y.-Q.; Liang, M.; Liu, Z.-Q.; Liao, D.-
Z.; Jiang, Z.-H.; Yan, S.-P.; Cheng, P. Chem. Commun. 2003, 2544–
2545.
(41) Colacio, E.; Lloret, F.; Kiveka¨s, R.; Ruiz, J.; Sua´rez-Varela, J.;
Sundberg, M. R. Chem. Commun. 2002, 592–593.
(42) Chu, I. K.; Shek, I. P. Y.; Siu, K. W. M.; Wong, W.-T.; Zuo, J.-L.;
Lau, T.-C. New J. Chem. 2000, 24, 765–769.
(13) Assefa, Z.; Omary, M. A.; McBurnett, B. G.; Mohamed, A. A.;
Patterson, H. H.; Staples, R. J.; Fackler, J. P., Jr. Inorg. Chem. 2002,
41, 6274–6280.
(14) White-Morris, R. L.; Olmstead, M. M.; Jiang, F.; Tinti, D. S.; Balch,
A. L. J. Am. Chem. Soc. 2002, 124, 2327–2336.
(15) Baria´in, C.; Mat´ıas, I. R.; Romeo, I.; Garrido, J.; Laguna, M. Appl.
Phys. Lett. 2000, 77, 2274–2276.
(16) Rawashdeh-Omary, M. A.; Omary, M. A.; Fackler, J. P., Jr.; Galassi,
R.; Pietroni, B. R.; Burini, A. J. Am. Chem. Soc. 2001, 123, 9689–
9691.
(43) Leznoff, D. B.; Xue, B.-Y.; Patrick, B. O.; Sanchez, V.; Thompson,
R. C. Chem. Commun. 2001, 259–260.
(17) Ferna´ndez, E. J.; López-de-Luzuriaga, J. M.; Monge, M.; Olmos,
M. E.; Pe´rez, J.; Laguna, A.; Mohamed, A. A.; Fackler, J. P., Jr.
J. Am. Chem. Soc. 2003, 125, 2022–2023.
(18) Lefebvre, J.; Batchelor, R. J.; Leznoff, D. B. J. Am. Chem. Soc. 2004,
126, 16117–16125.
(19) Katz, M. J.; Ramnial, T.; Yu, H.-Z.; Leznoff, D. B. J. Am. Chem.
Soc. 2008, 130, 10662–10673.
(20) Vickery, J. C.; Olmstead, M. M.; Fung, E. Y.; Balch, A. L. Angew.
Chem., Int. Ed. 1997, 36, 1179–1181.
(21) Yam, V. W.-W.; Li, C.-K.; Chan, C.-L. Angew. Chem., Int. Ed. 1998,
37, 2857–2859.
(22) Li, C.-K.; Lu, X.-X.; Wong, K. M.-C.; Chan, C.-L.; Zhu, N.; Yam,
V. W.-W. Inorg. Chem. 2004, 43, 7421–7430.
(23) La Brooy, S. R.; Linge, H. G.; Walker, G. S. Miner. Eng. 1994, 7,
1213–1241.
(24) Ahmad, S. Coord. Chem. ReV. 2004, 248, 231–243.
(25) Rawashdeh-Omary, M. A.; Omary, M. A.; Patterson, H. H. J. Am.
Chem. Soc. 2000, 122, 10371–10380.
(44) Leznoff, D. B.; Xue, B.-Y.; Stevens, C. L.; Storr, A.; Thompson,
R. C.; Patrick, B. O. Polyhedron 2001, 20, 1247–1254.
(45) Leznoff, D. B.; Xue, B.-Y.; Batchelor, R. J.; Einstein, F. W. B.;
Patrick, B. O. Inorg. Chem. 2001, 40, 6026–6034.
(46) Shorrock, C. J.; Xue, B.-Y.; Kim, P. B.; Batchelor, R. J.; Patrick,
B. O.; Leznoff, D. B. Inorg. Chem. 2002, 41, 6743–6753.
(47) Niel, V.; Thompson, A. L.; Muñoz, M. C.; Galet, A.; Goeta, A. E.;
Real, J. A. Angew. Chem., Int. Ed. 2003, 42, 3760–3763.
(48) Colacio, E.; Lloret, F.; Kiveka¨s, R.; Sua´rez-Varela, J.; Sundberg,
M. R.; Uggla, R. Inorg. Chem. 2003, 42, 560–565.
(49) Shorrock, C. J.; Jong, H.; Batchelor, R. J.; Leznoff, D. B. Inorg.
Chem. 2003, 42, 3917–3924.
(50) Stender, M.; White-Morris, R. L.; Olmstead, M. M.; Balch, A. L.
Inorg. Chem. 2003, 42, 4504–4506.
(51) Vitoria, P.; Muga, I.; Gutie´rrez-Zorrilla, J. M.; Luque, A.; Roma´n,
P.; Lezama, L.; Zu´niga, F. J.; Beitia, J. I. Inorg. Chem. 2003, 42,
960–969.
(52) Zhou, H.-B.; Wang, S.-P.; Dong, W.; Liu, Z.-Q.; Wang, Q.-L.; Liao,
D.-Z.; Jiang, Z.-H.; Yan, S.-P.; Cheng, P. Inorg. Chem. 2004, 43,
4552–4554.
(53) Galet, A.; Muñoz, M. C.; Mart´ınez, V.; Real, J. A. Chem. Commun.
2004, 2268–2269.
(26) Rawashdeh-Omary, M. A.; Omary, M. A.; Patterson, H. H.; Fackler,
J. P., Jr. J. Am. Chem. Soc. 2001, 123, 11237–11247.
(27) Che, C.-M.; Wong, W.-T.; Lai, T.-F.; Kwong, H.-L. J. Chem. Soc.,
Chem. Commun. 1989, 243–244.
(28) Dong, W.; Sun, Y.-Q.; Yu, B.; Zhou, H.-B.; Song, H.-B.; Liu, Z.-
Q.; Wang, Q.-M.; Liao, D.-Z.; Jiang, Z.-H.; Yan, S.-P.; Cheng, P.
New J. Chem. 2004, 28, 1347–1351.
(29) Paraschiv, C.; Andruh, M.; Ferlay, S.; Hosseini, M. W.; Kyritsakas,
N.; Planeix, J.-M.; Stanica, N. Dalton Trans. 2005, 1195–1202.
(30) Siemeling, U.; Rother, D.; Bruhn, C.; Fink, H.; Weidner, T.; Tra¨ger,
F.; Rothenberger, A.; Fenske, D.; Priebe, A.; Maurer, J.; Winter, R.
J. Am. Chem. Soc. 2005, 127, 1102–1103.
(54) Leznoff, D. B.; Shorrock, C. J.; Batchelor, R. J. Gold Bull. 2007,
40, 36–39.
(55) Lefebvre, J.; Callaghan, F.; Katz, M. J.; Sonier, J. E.; Leznoff, D. B.
Chem.sEur. J. 2006, 12, 6748–6761.
(56) Geisheimer, A. R.; Katz, M. J.; Batchelor, R. J.; Leznoff, D. B.
CrystEngComm 2007, 9, 1078–1083.
(57) Katz, M. J.; Kaluarachchi, H.; Batchelor, R. J.; Schatte, G.; Leznoff,
D. B. Cryst. Growth Des. 2007, 7, 1946–1948.
(31) Xu, G.-F.; Liu, Z.-Q.; Zhou, H.-B.; Guo, Y.; Liao, D.-Z. Aust.
J. Chem. 2006, 59, 640–646.
(32) Zhou, H.-B.; Wang, S.-P.; Liu, Z.-Q.; Liao, D.-Z.; Jiang, Z.-H.; Yan,
S.-P.; Cheng, P. Inorg. Chim. Acta 2006, 359, 533–540.
(58) Lefebvre, J.; Chartrand, D.; Leznoff, D. B. Polyhedron 2007, 26,
2189–2199.
(59) Lefebvre, J.; Trudel, S.; Hill, R. H.; Leznoff, D. B. Chem.sEur. J.
2008, 14, 7156–7167.
Inorganic Chemistry, Vol. 48, No. 5, 2009 2317

Harris and Wasylishen
Presented in the following is a 13C and 15N solid-state
NMR, SSNMR, study of some prototype gold cyano
complexes, in an effort to both investigate the interesting
fundamental physics of these systems as well as to provide
benchmark data and techniques useful for investigation of
the cyanoaurate coordination polymers mentioned above.
SSNMR studies of cyano 13C and 15N nuclei are well-known
to provide useful information on cyanide-containing materi-
als.60-74 To date, 13C and 15N NMR parameters have only
been reported for one cyano gold complex, Pb[Au-
(CN)2](H2O)x (x ) 0, 1),38 in the solid state. One aim of the
present research is to determine the effect of structural
changes, including changes in metallophilic bonding, on the
NMR properties of the cyanide ligand. Samples investigated
include the dicyanoaurate anion complexed with a series of
counterions, M+[Au(CN)2]-, where M+ ) n-butylammonium
(no aurophilic interactions between anions),75 potassium
(anions associated via aurophilic bonds),76 and thallium (both
Au-Au and Au-Tl metallophilic bonding present).77 In
addition to data on these gold(I) dicyanides, we also present
data for potassium tetracyanoaurate(III), in an effort to
determine the effect of gold’s formal oxidation state on the
cyanide ligand NMR properties. NMR spectra of samples
undergoing magic angle spinning (MAS) as well as stationary
example of the information that can be determined using
SSNMR to investigate a cyano gold coordination polymer.
The monocyanides of the group 11 transition metals resisted
structural characterization until recently, mainly due to
disorder and poor sample crystallinity. A consistent picture
of the structures of these materials has recently been built
up from a combination of diffraction78-85 and SSNMR/
NQR86,87 experiments. Two polymorphs of CuCN are
known, and structures of both the high-temperature, HT-
CuCN, and low-temperature, LT-CuCN, polymorphs of
CuCN have been determined.79,80,83-85 In HT-CuCN, AgCN,
and AuCN, the materials are composed of infinite, linear
[-M-CN-]∞ chains,79,81,82 with neighboring chains aligned
at the metal for AuCN, but staggered for HT-CuCN and
AgCN. Chain alignment in AuCN presumably arises from
the favorable energetics of forming aurophilic bonds.78 LT-
CuCN is also comprised of infinite MCN chains, but these
are packed together in a more complicated way, with bond
angles along the chain deviating slightly from 180°, ranging
from 174.0 to 179.3°.80 Interestingly, neighboring MCN
chains in LT-CuCN are also lined up at the metal, for
example, in two neighboring chains, d(Cu3-Cu3) ) 2.997
Å. In the group 11 monocyanides, it is not possible to
differentiate the carbon from the nitrogen atoms using
diffraction experiments, making the orientation of the cyanide
ligands in these chains unknown.78-85 The relative orienta-
tion of cyanide ligands on opposing sides of the metal in
CuCN and AgCN has been investigated using SSNMR and
NQR spectroscopy.86,87 In these studies, it was determined
that there is significant “head-tail” disorder of the cyanide
ligands. In both materials, 30 ( 10% of the metal sites have
the two coordinating cyanides in opposite orientation,
[-NC-Ag-CN-] or [-CN-Ag-NC-], while 70 ( 10%
have the two cyanides in the same orientation,
[-CN-Ag-CN-].86,87 Diffraction studies have shown that
another type of disorder in these materials is also likely: static
disorder of the long chains with respect to each other.79,81,82
Presented below is a 13C and 15N SSNMR study of AuCN
to investigate the possibility of cyanide orientational disorder,
and to probe for the presence of stacking faults in the chain
alignments.
samples (where possible) were recorded to determine the 13
C
and 15N chemical shift tensors and 13C,15N dipolar coupling
constants.
A SSNMR study of gold monocyanide, a material likely
to possess as-yet unquantified modes of disorder, is also
presented. Additionally, this study of AuCN provides an
(60) Avalle, P.; Harris, R. K.; Hanika-Heidl, H.; Fischer, R. D. Solid State
Sci. 2004, 6, 1069–1076.
(61) Poll, E.-M.; Schu¨tze, J.-U.; Fischer, R. D.; Davies, N. A.; Apperley,
D. C.; Harris, R. K. J. Organomet. Chem. 2001, 621, 254–260.
(62) Poll, E.-M.; Olbrich, F.; Samba, S.; Fischer, R. D.; Avalle, P.;
Apperley, D. C.; Harris, R. K. J. Solid State Chem. 2001, 157, 324–
338.
(63) Poll, E.-M.; Samba, S.; Fischer, R. D.; Olbrich, F.; Davies, N. A.;
Avalle, P.; Apperley, D. C.; Harris, R. K. J. Solid State Chem. 2000,
152, 286–301.
(64) Siebel, E.; Fischer, R. D.; Davies, N. A.; Apperley, D. C.; Harris,
R. K. J. Organomet. Chem. 2000, 604, 34–42.
(65) Bowmaker, G. A.; Churakov, A. V.; Harris, R. K.; Oh, S.-W. J.
Organomet. Chem. 1998, 550, 89–99.
(66) Brimah, A. K.; Schwarz, P.; Fischer, R. D.; Davies, N. A.; Harris,
R. K. J. Organomet. Chem. 1998, 568, 1–12.
(67) Schwarz, P.; Siebel, E.; Fischer, R. D.; Davies, N. A.; Apperley,
D. C.; Harris, R. K. Chem.sEur. J. 1998, 4, 919–926.
(68) Siebel, E.; Fischer, R. D.; Kopf, J.; Davies, N. A.; Apperley, D. C.;
Harris, R. K. Inorg. Chem. Commun. 1998, 1, 346–349.
(69) Schu¨tze, J.-U.; Eckhardt, R.; Fischer, R. D.; Apperley, D. C.; Davies,
N. A.; Harris, R. K. J. Organomet. Chem. 1997, 534, 187–194.
(70) Davies, N. A.; Harris, R. K.; Olivieri, A. C. Mol. Phys. 1996, 87,
669–677.
(71) Eller, S.; Schwarz, P.; Brimah, A. K.; Fischer, R. D.; Apperley, D. C.;
Davies, N. A.; Harris, R. K. Organometallics 1993, 12, 3232–3240.
(72) Behrens, U.; Brimah, A. K.; Soliman, T. M.; Fischer, R. D.; Apperley,
D. C.; Davies, N. A.; Harris, R. K. Organometallics 1992, 11, 1718–
1726.
(78) Bowmaker, G. A.; Kennedy, B. J.; Reid, J. C. Inorg. Chem. 1998,
37, 3968–3974.
(79) Hibble, S. J.; Cheyne, S. M.; Hannon, A. C.; Eversfield, S. G. Inorg.
Chem. 2002, 41, 4990–4992.
(80) Hibble, S. J.; Eversfield, S. G.; Cowley, A. R.; Chippindale, A. M.
Angew. Chem., Int. Ed. 2004, 43, 628–630.
(81) Hibble, S. J.; Cheyne, S. M.; Hannon, A. C.; Eversfield, S. G. Inorg.
Chem. 2002, 41, 1042–1044.
(82) Hibble, S. J.; Hannon, A. C.; Cheyne, S. M. Inorg. Chem. 2003, 42,
4724–4730.
(73) Apperley, D. C.; Davies, N. A.; Harris, R. K.; Brimah, A. K.; Eller,
S.; Fischer, R. D. Organometallics 1990, 9, 2672–2676.
(74) Wu, G.; Wasylishen, R. E. J. Phys. Chem. 1993, 97, 7863–7869.
(75) Schubert, R. J.; Range, K.-J. Z. Naturforsch., B: Chem. Sci. 1990,
45, 1118–1122.
(76) Rosenzweig, A.; Cromer, D. T. Acta Crystallogr. 1959, 12, 709–
712.
(83) Reckeweg, O.; Simon, A. Z. Naturforsch., B: Chem. Sci. 2002, 57,
895–900.
(84) Reckeweg, O.; Lind, C.; Simon, A.; DiSalvo, F. J. Z. Naturforsch.,
B: Chem. Sci. 2003, 58, 155–158.
(85) Wang, J.; Collins, M. F.; Johari, G. P. Phys. ReV. B: Condens. Matter
Mater. Phys. 2002, 65, 180103.
(86) Kroeker, S.; Wasylishen, R. E.; Hanna, J. V. J. Am. Chem. Soc. 1999,
121, 1582–1590.
(77) Blom, N.; Ludi, A.; Bu¨rgi, H.-B.; Tichy´, K. Acta Crystallogr., Sect.
C 1984, 40, 1767–1769.
(87) Bryce, D. L.; Wasylishen, R. E. Inorg. Chem. 2002, 41, 4131–4138.
2318 Inorganic Chemistry, Vol. 48, No. 5, 2009

Structural Disorder and Aurophilic Bonding
2. NMR Terminology and Theory
on through the bond, is generally less than 20 Hz and thus
too small to affect these spectra.87,98-100 There is also a
We summarize here the interactions that affect the carbon-
13 and nitrogen-15 SSNMR spectra of these aurocyanide
materials, along with the terminology and conventions used
to describe them. Some background theory of the NMR
experiments, as well as the information garnered from each
will also be described. All of the studied samples made use
of 13C,15N dilabeled cyanide ligands, and therefore the
discussion that follows will focus on a heteronuclear pair of
spin-1/2 nuclei. The possibility of coupling interactions with
the relatively small magnetic moment of Au-197 (I ) 3/2,
natural abundance ) 100%) is also discussed.
The observed chemical shift of a nucleus depends on the
orientation of the externally applied magnetic field, B0, with
respect to the molecular features. This orientation dependence
is represented by the second-rank chemical shift tensor,88,89
or CS tensor, generally assumed to be a symmetric ten-
sor.89,90 When represented in its principal axis system, a
symmetric tensor is diagonal and thus has only three
elements. By convention, the values of these three elements
for each CS tensor are labeled in decreasing order as δ11 g
δ22 g δ33.91 The chemical shift of a nucleus is thus fully
described by δ11, δ22, and δ33 in combination with a
specification of how the principal axis system is oriented in
the molecule. For spectra of samples spinning rapidly at the
54.736° magic angle, the anisotropic chemical shift of a
powder is averaged to δiso. If the spinning is at a rate
comparable to the breadth of the magnetic shielding interac-
tion in hertz, the broad static spectrum is replaced by a
combination of spinning sidebands whose intensities encode
the elements of the CS tensor.92-94
through-space interaction, or direct dipolar coupling, de-
101
scribed by the constant RDD
:
µ0p
8π2
-3
RDD
)
γIγS〈r 〉
(1)
IS
(
)
where γI is the strength of the magnetic dipole moment of
nucleus I, 13C or 15N here, and 〈rIS-3〉 is the averaged inverse
cube of the distance between the two nuclei. The magnitude
of this interaction is approximately 2 kHz for the 13C,15N
spin pair in a cyanide ligand and depends strongly on the
internuclear separation, thereby providing a direct measure
of the C-N bond distance. It should be noted that a portion
of the J-coupling tensor has the same geometrical dependence
as the dipolar coupling interaction, and experiments actually
measure a combination of these two effects, termed the
effective dipolar coupling constant, Reff.101,102 As has been
noted previously,87 the difference between Reff and RDD is
expected to be less than 20 Hz for a cyano 13C,15N spin pair.
Since this is less than the experimental error associated with
measuring Reff, we have assumed RDD ) Reff.
A distance measured via the NMR observable Reff has a
slightly different dependence on vibrational averaging than
does an X-ray or neutron diffraction distance measurement,
typically leading to a disagreement of a few percent. This
difference in averaging is known to yield a 1-4% longer
bond distance for C-C distances measured via NMR as
compared to diffraction distance determinations.103-105 In
a study of AgCN, it was argued that a value for the C-N
bond length more amenable to direct comparison with
diffraction data could be produced by shortening the NMR-
measured length by 2.5 ( 1.5%.87 All distance measurements
reported here include this vibrational correction of shortening
the NMR-derived bond lengths and applying the concomitant
increase in error ranges.
Spectra from a stationary sample of a 13C,15N-labeled
cyanide, observing either 13C or 15N, manifest effects from
the anisotropy of both the chemical shift and dipolar coupling
to the neighboring spin-1/2 nucleus. Analysis of such spectra
therefore allows one to obtain both interaction tensors as well
as their relative orientations from the observed line
It is often convenient to describe the CS tensor using an
alternate set of three elements: the isotropic or average shift,
δiso ) (δ11 + δ22 + δ33)/3, the span, Ω ) δ11 - δ33, and the
skew, κ ) 3(δiso - δ22)/Ω.91 For any nucleus at a site with
higher than 2-fold rotational symmetry, two of the principal
components are equal to each other, that is, κ ) (1, and
the CS tensor is axially symmetric.89,95,96 It should be noted
that the procedure of fitting sideband intensities has difficulty
in differentiating axially symmetric CS tensors from those
that deviate slightly from axial symmetry.94,97
Energy levels of a 13C nucleus in a 13C,15N-labeled cyanide
are perturbed by the magnetic field generated by the proximal
15N nucleus (and vice versa). The familiar J coupling, passed
(98) Friesen, K. J.; Wasylishen, R. E. J. Magn. Reson. 1982, 48, 152–
154.
(88) Grant, D. M. Chemical Shift Tensors In Encyclopedia of Nuclear
Magnetic Resonance; Grant, D. M., Harris, R. K., Eds.; John Wiley
and Sons: Chichester, U.K., 1996; Vol. 2, pp 1298.
(89) Anet, F. A. L.; O’Leary, D. J. Concepts Magn. Res. 1991, 3, 193–
214.
(90) Schneider, R. F. J. Chem. Phys. 1968, 48, 4905–4909.
(91) Mason, J. Solid State Nucl. Magn. Reson. 1993, 2, 285–288.
(92) Maricq, M. M.; Waugh, J. S. J. Chem. Phys. 1979, 70, 3300–3316.
(93) Herzfeld, J.; Berger, A. E. J. Chem. Phys. 1980, 73, 6021–6030.
(94) Harris, R. K.; Olivieri, A. C. Spinning Sideband Analysis for Spin-
1/2 Nuclei In Encyclopedia of Nuclear Magnetic Resonance; Grant,
D. M., Harris, R. K., Eds.; John Wiley and Sons: Chichester, 2002;
Vol. 9, pp 141.
(95) Buckingham, A. D.; Malm, S. M. Mol. Phys. 1971, 22, 1127–1130.
(96) Robert, J. B.; Wiesenfeld, L. Phys. Rep. 1982, 86, 363–401.
(97) Clayden, N. J.; Dobson, C. M.; Lian, L.-Y.; Smith, D. J. J. Magn.
Reson. 1986, 69, 476–487.
(99) Eichele, K.; Wasylishen, R. E. Solid State Nucl. Magn. Reson. 1992,
1, 159–163.
(100) Barszczewicz, A.; Helgaker, T.; Jaszun´ski, M.; Jørgensen, P.; Ruud,
K. J. Magn. Reson., Ser. A 1995, 114, 212–218.
(101) Wasylishen, R. E. Dipolar and Indirect Coupling Tensors in Solids
In Encyclopedia of Nuclear Magnetic Resonance; Grant, D. M.,
Harris, R. K., Eds.; John Wiley and Sons: Chichester, U.K., 1996;
Vol. 3, pp 1685.
(102) Wasylishen, R. E. Indirect Nuclear Spin-Spin Coupling Tensors In
Encyclopedia of Nuclear Magnetic Resonance; Grant, D. M., Harris,
R. K., Eds.; John Wiley and Sons: Chichester, U.K., 2002; Vol. 9,
pp 274.
(103) Millar, J. M.; Thayer, A. M.; Zax, D. B.; Pines, A. J. Am. Chem.
Soc. 1986, 108, 5113–5116.
(104) Nakai, T.; Ashida, J.; Terao, T. Mol. Phys. 1989, 67, 839–847.
(105) Ishii, Y.; Terao, T.; Hayashi, S. J. Chem. Phys. 1997, 107, 2760–
2774.
Inorganic Chemistry, Vol. 48, No. 5, 2009 2319

Harris and Wasylishen
first half of the sequence excites a double-quantum coherence
using a long train of composite 180° pulses, which is then
allowed to evolve in the indirect dimension for a time t1,
before being converted back to longitudinal magnetization
by the second block of composite 180° pulses, which is
finally observed using a π/2 read pulse.
In each compound, the carbon, and in AuCN, the nitrogen,
atoms are directly bonded to gold, internuclear coupling with
which can affect the spin-1/2 NMR spectra. Gold-197 is a
quadrupolar nucleus with a small magnetic moment (<1/
50th that of 1H) and a large nuclear quadrupole moment (Q
) 54.7 fm2).118 The 197Au nucleus may perturb the energy
levels of neighboring 13C or 15N nuclei through both
spin-spin coupling pathways described above: J coupling
and direct dipolar coupling. As mentioned above, the direct
dipolar coupling between two spin-1/2 nuclei is averaged to
zero when the sample is spun rapidly at the magic angle.
For a spin-1/2 nucleus coupled to a nucleus with a large
quadrupolar coupling constant, however, the direct dipolar
coupling between the two cannot be averaged to zero by
magic-angle spinning, and a residual dipolar coupling
remains.119 For many systems studied at room temperature,
197Au relaxes so rapidly that no effects on the spin-1/2 NMR
spectra are observed; however, there have been examples in
11
Figure 1. Symbolic display of the SR264 pulse sequence. Pulses in the
11
R264 block are shown using the (Pulse Angle)Phase convention.
shape.106-109 This relative orientation is given here using
the angle ꢀ to define the angle between r(C,N) and δ33, while
R is used to define the angle between δ11 and the projection
of r(C,N) in the δ11,δ22 plane. In many cyanides, 13C and 15
N
CS tensors have δ33 along the CN bond, in which case the
value of ꢀ is zero, and R may be taken as zero as well (when
ꢀ ) 0°, r(C,N) is coincident with δ33 and creates no projection
on the δ11,δ22 plane, making R unnecessary). In spectra of
AX spin systems spun at the magic angle at a rate less than
Ω in frequency units, spinning sidebands are generated,
whose intensities are modulated both by the chemical shift
tensor and the dipolar coupling to the neighboring spin-1/2
nucleus.94 Spectral simulations therefore require inclusion
of both interactions, and we applied the program SPINEVO-
LUTION to determine CS tensor elements.110
which spin-spin coupling to 197Au has been observed in 31
P
NMR spectra.120-126
As an example of the effects expected from residual
dipolar coupling to 197Au, we here estimate the 13C line shape
of K[Au(CN)2] expected from such an interaction. In this
approximately linear molecule, the 197Au EFG tensor is likely
axially symmetric with Vzz along the Au-C bond. An
approximate value of CQ(197Au), 1.27 GHz, has been obtained
from Mo¨ssbauer data for K[Au(CN)2] at 4.2 K.127 However,
because of the low Larmor frequency of 197Au, the simulated
spectra are nearly invariant to the magnitude of CQ as long
as it is greater than ∼300 MHz. With the orientation and
principal components of the EFG tensor so defined, spectral
simulations of the line shape need include only two variables:
Reff and Jiso.119 The published crystal structure allows an
initial estimate of 69 Hz for Reff to be set from the Au-C
bond length. An estimate of the expected Jiso(197Au,13C) can
be determined through comparison to an analogous system,
There is a library of techniques available which make use
of the resolution afforded by MAS but retains information
from homonuclear dipolar couplings.111,112 We applied the
supercycled SR26411 sequence, which has been tailored, using
symmetry principles,113 for the combination of a small Reff
and a large Ω required to recouple intermolecular 13C cyanide
sites.114,115 In the present study, we use this dipolar-
recoupling pulse sequence to investigate proximity, that is,
to determine whether pairs of peaks arise from sites within
the same material. A one-channel, two-dimensional version
of this pulse sequence, see Figure 1, was applied.116,117 The
(106) VanderHart, D. L.; Gutowsky, H. S. J. Chem. Phys. 1968, 49, 261–
271.
(107) Zilm, K. W.; Grant, D. M. J. Am. Chem. Soc. 1981, 103, 2913–
2922.
(108) Power, W. P.; Wasylishen, R. E. Annu. Rep. NMR Spectrosc. 1991,
23, 1–84.
(109) Eichele, K.; Wasylishen, R. E. J. Magn. Reson., Ser. A 1994, 106,
46–56.
(118) Harris, R. K.; Becker, E. D.; Cabral de Menezes, S. M.; Goodfellow,
R.; Granger, P. Pure Appl. Chem. 2001, 73, 1795–1818.
(119) Olivieri, A. Solid State Nucl. Magn. Reson. 1992, 1, 345–353.
(120) Barron, P. F.; Engelhardt, L. M.; Healy, P. C.; Oddy, J.; White, A. H.
Aust. J. Chem. 1987, 40, 1545–1555.
(121) Attar, S.; Bearden, W. H.; Alcock, N. W.; Alyea, E. C.; Nelson,
J. H. Inorg. Chem. 1990, 29, 425–433.
(122) Berners-Price, S. J.; Colquhoun, L. A.; Healy, P. C.; Byriel, K. A.;
Hanna, J. V. J. Chem. Soc., Dalton Trans. 1992, 3357–3363.
(123) Baker, L.-J.; Bowmaker, G. A.; Healy, P. C.; Skelton, B. W.; White,
A. H. J. Chem. Soc., Dalton Trans. 1992, 989–997.
(110) Veshtort, M.; Griffin, R. G. J. Magn. Reson. 2006, 178, 248–282.
(111) Schnell, I. Prog. Nucl. Magn. Reson. Spectrosc. 2004, 45, 145–207.
(112) Griffiths, J. M.; Bennett, A. E.; Griffin, R. G. Homonuclear
Recoupling Schemes in MAS NMR In Encyclopedia of Nuclear
Magnetic Resonance; Grant, D. M., Harris, R. K., Eds.; John Wiley
and Sons: Chichester, U.K., 1996; Vol. 4, pp 2390.
(113) Levitt, M. H. Symmetry-Based Pulse Sequences in Magic-Angle
Spinning Solid-State NMR In Encyclopedia of Nuclear Magnetic
Resonance; Grant, D. M., Harris, R. K., Eds.; John Wiley and Sons:
Chichester, U.K., 2002; Vol. 9, pp 165.
(114) Kristiansen, P. E.; Carravetta, M.; Lai, W. C.; Levitt, M. H. Chem.
Phys. Lett. 2004, 390, 1–7.
(115) Kristiansen, P. E.; Carravetta, M.; van Beek, J. D.; Lai, W. C.; Levitt,
M. H. J. Chem. Phys. 2006, 124, 234510.
(116) Brouwer, D. H.; Kristiansen, P. E.; Fyfe, C. A.; Levitt, M. H. J. Am.
Chem. Soc. 2005, 127, 542–543.
(117) Brouwer, D. H.; Darton, R. J.; Morris, R. E.; Levitt, M. H. J. Am.
Chem. Soc. 2005, 127, 10365–10370.
(124) Angermair, K.; Bowmaker, G. A.; de Silva, E. N.; Healy, P. C.; Jones,
B. E.; Schmidbaur, H. J. Chem. Soc., Dalton Trans. 1996, 3121–
3129.
(125) de Silva, E. N.; Bowmaker, G. A.; Healy, P. C. J. Mol. Struct. 2000,
516, 263–272.
(126) Healy, P. C.; Loughrey, B. T.; Bowmaker, G. A.; Hanna, J. V. Dalton
Trans. 2008, 3723–3728.
(127) Faltens, M. O.; Shirley, D. A. J. Chem. Phys. 1970, 53, 4249–4264.
2320 Inorganic Chemistry, Vol. 48, No. 5, 2009

Structural Disorder and Aurophilic Bonding
usingthereducedcouplingconstant,102,128K(I,S))4π2h-1[J(I,S)/
γIγS]. Analogous molecules where one nucleus is replaced
by a heavier congener are typified by a larger K because of
increased electron density at the nucleus. For example, in
the analogous series of linear ClM(PR3) (where M ) Cu,
Gold monocyanide was prepared via three synthetic routes.
AuCN(A) was synthesized using the method given in U.S Patent
#4,222,996.132 Metallic gold was dissolved in aqua regia, diluted
with distilled water, and treated with aqueous KOH to a pH of
∼12, then reduced with 0.75 equiv of Na2SO3, followed by
treatment with 50% 13C,15N-labeled KCN and acidification with
HCl to a pH of ∼2. It was found that complete removal of the
starting materials required soaking this material in distilled water
for 1 week at room temperature. This wash water was then removed
by filtration, washing with several volumes of water, and finally
with MeOH. A second preparation of AuCN was also performed,
with all details the same as in AuCN(A), except for the use of a
10% excess of the reducing agent, Na2SO3. This method produced
AuCN(A′), a material with a greenish color instead of the reported
yellow of AuCN. A third synthetic method produced yellow-colored
AuCN(B) from 50% 13C,15N-labeled K[Au(CN)2], prepared as
described above, by treatment with HCl at 60-70 °C.130 Charac-
terization of these materials by SSNMR, IR, and X-ray diffraction
is discussed below.
1
Ag, Au), Kiso(M,P) approximately doubles for each step
down the periodic table.123 Also, the value of 1Kiso(M,C) in
AgCN is approximately double that of CuCN. This same
numerical relationship between J couplings in analogous
molecules does not hold across the entire periodic table but
seems to be useful within the group 11 metals. The values
of 1Kiso(Au,C) in [Au(CN)2]- anions are therefore likely close
to double the value of 1Kiso(Ag,C) in · · ·CNAgCNAgCN· · ·,
and a value of Jiso(197Au,13C) = 180 Hz is predicted.
Lineshape simulations, using WSOLIDS,129 show that the
expected spectrum is composed of four peaks separated by
60 to 160 Hz, covering a total range of ∼400 Hz. Even if
the J coupling is 0, Reff alone is enough to cause the four
peaks of the simulated spectrum to cover a range of ∼250
Hz. We therefore expect that residual dipolar coupling to
197Au will have a noticeable effect on the 13C spectra, unless
the gold nucleus is self-decoupled by rapid relaxation.
Powder XRD measurements used an Inel powder diffractometer
with copper KR1 radiation. IR spectra were collected on a Nic-
Plan FTIR Microscope, including a reinvestigation of AgCN
samples discussed in a previous publication.87,156
3.2. SSNMR. Carbon-13 NMR spectra were acquired at 50.3
MHz (4.70 T) using a Chemagnetics CMX Infinity 200, at 75.5
MHz (7.05 T) using a Bruker Avance 300, or at 125.8 MHz (11.75
T) with a Bruker Avance 500. All 13C experiments used powdered
samples packed in 4-mm-outer-diameter rotors, were recorded using
4 µs π/2 pulses, and were referenced to TMS by setting the high-
frequency peak of an external adamantane sample to 38.56 ppm.133
For all samples except [(n-C4H9)4N][Au(CN)2], Bloch-decay experi-
ments were used for MAS spectra, while Hahn-echo experiments
were used for spectra of stationary samples. Experiments used pulse
delays of 25 min for K[Au(CN)2], 30 min for Tl[Au(CN)2], 30 min
(but using a 45° pulse) for K[Au(CN)4], and 20 min for AuCN.
Variable-amplitude134 cross polarization135-137 from the protons
in the tetra-n-butylammonium cation was used to collect 13C NMR
spectra of [(n-C4H9)4N][Au(CN)2], under the Hartmann-Hahn
match condition, and TPPM decoupling138 at a power level of ∼60
kHz was applied during acquisition. Under these conditions, a 10 s
pulse delay was utilized for both the CPMAS and static CP Hahn-
echo experiments.
Two-dimensional 13C homonuclear dipolar-recoupling experi-
ments used the SR26411 pulse sequence.114-117 Phase cycling and
pulse timings were those of the published sequence.117 Pure
absorption spectra were obtained using the States method, and
recoupling-field power levels were experimentally optimized. The
pulse sequence was benchmarked using 60% 1-13C-labeled glycine,
Raylo Chemicals, in which intermolecular distances are short
enough to allow 13C,13C recoupling.115 Setup spectra of 60% 1-13C-
labeled glycine were obtained at a MAS rate of 8.097 kHz, using
two excitation blocks totaling 3.95 ms; DQ recoupling efficiencies
3. Experimental Section
3.1. Sample Preparation. A sample of AuCN used as starting
material was prepared from commercial K[Au(CN)2] through
treatment with HCl according to literature methods.130 The sample
of 13C,15N-labeled K[Au(CN)2] was prepared via dissolution of
AuCN in a hot aqueous solution of 99% 13C,15N-labeled KCN,
followed by recrystallization from hot water.130 The 13C,15N-labeled
[(n-C4H9)4N][Au(CN)2] was prepared by dissolution of AuCN in a
hot aqueous solution of 60% 13C,15N-labeled KCN, followed by
precipitation with an aqueous solution of tetra-n-butylammonium
chloride, and then recrystallization from hot water.75 We found it
necessary to spin the sample at less than ∼6 kHz during MAS
experiments to prevent the sample of [(n-C4H9)4N][Au(CN)2] from
melting. The sample of 13C,15N-labeled Tl[Au(CN)2] was prepared
by dissolution of AuCN in a hot aqueous solution of 60% 13C,15N-
labeled KCN, followed by precipitation with an aqueous solution
of Tl2SO4 and recrystallization from hot water.77 A sample of
K[AuBr2(CN)2] for use as a starting material was prepared by the
oxidation of commercial K[Au(CN)2] with a solution of Br2 in
MeOH followed by recrystallization from hot water.131 This
material was then added to a methanol solution of 99% 13C,15N-
labeled KCN, and the result recrystallized from hot water to yield
K[Au(CN)4], which was then dehydrated under a high vacuum at
room temperature.131 The identities of K[Au(CN)2], Tl[Au(CN)2],
and [(n-C4H9)4N][Au(CN)2] were confirmed by matching the
experimental and theoretical X-ray powder diffraction patterns.75-77
Only the hydrate of K[Au(CN)4] has a known crystal structure,
but this material was found to dehydrate too rapidly for identification
by powder XRD. Therefore, the identity of K[Au(CN)4] was
confirmed by matching the 2189.5 cm-1 CN stretch to the published
2189 cm-1 value,131 and the fact that only a single 13C NMR peak
is observed in the solid state.
(132) Okinaka, Y.; Smith, L. E. U.S. Patent 4,222,996, 1980.
(133) Earl, W. L.; VanderHart, D. L. J. Magn. Reson. 1982, 48, 35–54.
(134) Peersen, O. B.; Wu, X.; Kustanovich, I.; Smith, S. O. J. Magn. Reson.,
Ser. A 1993, 104, 334–339.
(128) Jameson, C. J. Spin-Spin Coupling In Multinuclear NMR; Mason,
J., Ed.; Plenum Press: New York, 1987; pp 89.
(129) Eichele, K.; Wasylishen, R. E. WSOLIDS, version 2.0.18; University
of Alberta: Edmonton, Canada, 2000.
(130) Bradley, R. S.; Munro, D. C.; Spencer, P. N. Trans. Faraday Soc.
1969, 65, 1920–1926.
(131) Jones, L. H.; Smith, J. M. J. Chem. Phys. 1964, 41, 2507–2511.
(135) Hartmann, S. R.; Hahn, E. L. Phys. ReV. 1962, 128, 2042–2053.
(136) Pines, A.; Gibby, M. G.; Waugh, J. S. J. Chem. Phys. 1972, 56,
1776–1777.
(137) Pines, A.; Gibby, M. G.; Waugh, J. S. J. Chem. Phys. 1973, 59,
569–590.
(138) Bennett, A. E.; Rienstra, C. M.; Auger, M.; Lakshmi, K. V.; Griffin,
R. G. J. Chem. Phys. 1995, 103, 6951–6958.
Inorganic Chemistry, Vol. 48, No. 5, 2009 2321

Harris and Wasylishen
Table 1. Local Structural Parameters for Gold(I) Cyanides
salt
Au site C-N bond (Å) Au-C bond (Å) C-Au-C angle (deg.) Au-C-N angle (deg.) Au-Au (Å) Au-Tl (Å)
K[Au(CN)2]a
1.17(20)
1.03(4)
2.12(14)
1.96(4)
180
180
180
180
180
180
180
177.1(2)
173(7.5)
174(3)
180
180
180
180
176.6(4)
175.8(4)
4@3.64
4@8.05
2@3.560
[(n-C4H9)4N][Au(CN)2]b
Tl[Au(CN)2]c
1
2
3
1.148(2)
1.148(2)
1.145(2)
1.146(2)
1.145(2)
1.148(2)
1.970(2)
1.971(2)
1.971(2)
1.970(2)
1.972(2)
1.971(2)
2@3.560
2@3.037
2@3.037
2@3.068
3.463
3.446
a Data from the X-ray diffraction study of Rosenzweig and Cromer.76 b Data from an X-ray diffraction study of Schubert and Range.75 c Neutron diffraction
study by Blom et al.77
similar to reported values were obtained.114,115 Recoupling experi-
ments performed on AuCN used a MAS rate of 9.615 kHz, and
11
three SR264 DQ-excitation blocks totaling 4.992 ms.
Nitrogen-15 NMR spectra were collected at 20.3 MHz (4.70 T)
using a Chemagnetics CMX Infinity 200, at 30.5 MHz (7.05 T)
with a Bruker Avance 300, at 50.7 MHz (11.75 T) using a Bruker
Avance 500, or at 91.2 MHz (21.14 T) using a Bruker Avance II
900. All experiments used 4 mm rotors, and peaks were referenced
by setting the low-frequency peak of an external sample of
ammonium nitrate to 23.8 ppm (with respect to neat liquid ammonia
at 0 ppm).139 The MAS spectra, except those of [(n-C4H9)4N]-
[Au(CN)2], used Bloch-decay experiments. Experiments used a B1
field sufficient to generate a 5 µs π/2 pulse for experiments recorded
at 4.70, 7.05, and 11.75 T, while a 9 µs π/2 pulse strength was
applied for experiments at 21.14 T. Spectra of K[Au(CN)2] used a
30° excitation pulse and 40 min recycle delay. Tl[Au(CN)2] required
a 10 min pulse delay when applying a 30° pulse. K[Au(CN)4]
required a 30 min delay when using a 20° pulse, and AuCN samples
used a 10 min delay and a 20° excitation pulse. Variable-
amplitude134 cross polarization135-137 followed by TPPM decou-
pling138 at a power level of ∼60 kHz during acquisition was used
for the 15N MAS spectra of [(n-C4H9)4N][Au(CN)2], and CP
followed by a 180° refocusing pulse was used for static spectra.
Carbon-13 and nitrogen-15 NMR spectra from stationary samples
were simulated using WSOLIDS,129 while nitrogen-15 spectra from
slow-spinning samples were simulated using SPINEVOLUTION.110
Simulated 13C spectra from MAS samples including residual dipolar
coupling to 197Au were calculated using WSOLIDS.129 We found
the programs SIMPSON140 and SPINEVOLUTION110 invaluable
for theoretically comparing various recoupling conditions used in
Figure 2. Simulated (upper trace) and experimental (lower trace) 13C NMR
spectra of a stationary sample of K[Au(CN)2] at 11.75 T.
The axial symmetry and principal component values of the
CS tensor allow its orientation to be assigned as δ33 along
the C-N bond, and the equivalent δ11,δ22 components
perpendicular to the bond (vide infra). Parameters obtained
from line shape simulations of the 13C NMR spectrum from
K[Au(CN)2] are presented in Table 2. After the vibrational
corrections of Reff(15N,13C) described in section 2 were
applied, a C-N distance of 1.15(4) Å was derived. This bond
length is consistent with the X-ray-diffraction-derived value
of 1.17(20) Å. The NMR-derived bond length is of higher
precision than the X-ray value, whose lower than typical
precision is probably due to the large electron density of the
neighboring gold atom. This example shows how a SSNMR
experiment can supplement an X-ray diffraction study where
it may be difficult to pick out light atoms against a
background of very heavy atoms.
11
the SR264 pulse sequence.
4. Results and Discussion
4.1. 13C NMR of Gold(I) Dicyanides. A sample of
K[Au(CN)2] was prepared and investigated using 13C SS-
NMR. The structure, reported in 1959 by Rosenzweig and
Cromer,76 contains alternating sheets of potassium and gold
atoms with cyanide ligands above and below the gold plane.
The local structure around the single unique CN moiety,
reported in Table 1, has the unique Au atom surrounded by
four Au neighbors at 3.64 Å, close enough for aurophilic
bonding to be present.141,142 Presented in Figure 2 is the 13
C
NMR spectrum of a stationary sample of 50% 13C,15N-labeled
K[Au(CN)2], the shape of which evinces the axial symmetry
of the CS tensor. Also shown in Figure 2 are the specific
effects of Reff(15N,13C), δ11 ) δ22, and δ33 on the spectrum.
The CS tensor for a carbon-13 nucleus in a linear molecule
is defined by two components: δ⊥, measured perpendicular
(141) Ahrland, S.; Dreisch, K.; Nore´n, B.; Oskarsson Å. Mater. Chem. Phys.
1993, 35, 281–289.
(139) Bryce, D. L.; Bernard, G. M.; Gee, M.; Lumsden, M. D.; Eichele,
K.; Wasylishen, R. E. Can. J. Anal. Sci. Spectrosc. 2001, 46, 46–82.
(140) Bak, M.; Rasmussen, J. T.; Nielsen, N. C. J. Magn. Reson. 2000,
147, 296–330.
(142) Pathaneni, S. S.; Desiraju, G. R. J. Chem. Soc., Dalton Trans. 1993,
319–322.
(143) Jameson, A. K.; Jameson, C. J. Chem. Phys. Lett. 1987, 134, 461–
466.
2322 Inorganic Chemistry, Vol. 48, No. 5, 2009

Structural Disorder and Aurophilic Bonding
Table 2. Parameters Derived from Simulations of Carbon-13 SSNMR Spectra
salt
Au site δiso (ppm) LWHHa (Hz) Ω (ppm)
κ
Reff(15N,13C)b (Hz) C-N bondc NMR (Å) C-N bond X-ray (Å)
K[Au(CN)2]
158.7(1)
151.90(5)
167.5(4)
161.3(4)
151.7(2)
167(1)
173(1)
107.8(1)
154.1(5)
85
25
255
315
250
360(2)
341(2)
361(50)
352(50)
338(50)
346
353
337(3)
357(5)
1
1
1
1
1
1850(100)
1800(150)
1.15(4)
1.16(5)
1.17(20)
1.03(4)
[R4N][Au(CN)2]d
Tl[Au(CN)2]
1 or 2
1 or 2
3
Tl[Au(CN)2]
Tl[Au(CN)2]
Pb[Au(CN)2]2(H2O)e
>0.95
>0.95
0.485(10)
1
e
Pb[Au(CN)2]2
K[Au(CN)4]
110
450
1925(200)
1925(100)
1.14(6)
1.14(4)
AuCNf
1.150(28)
a LWHH measured at 7.05 T. b From best-fit simulated spectra. c Corrected for vibrational effects (see text). d R ) n-butyl. e SSNMR study from Katz
et al.38 f AuCN(B) as representative of the main site in all samples, see text.
to the long axis of the molecule, and δ|, measured along it;
furthermore, it is well-known that δ| will be approximately
-90 ppm.144,145 In brief, the reason for this is that σparamagnetic
,
the term responsible for most of the change in shielding
between nuclei in differing environments, is identically zero
along a C∞ axis.146 K[Au(CN)2] has CS tensor components
δ11 ) δ22 ) 279 ppm and δ33 ) -81 ppm, but the orientation
of the PAS cannot be determined from symmetry arguments
because the molecule is not quite linear, see Table 1.
However, the value of δ33 is similar enough to that expected
for δ| in a linear molecule to imply that it is oriented along
the CN bond. The two remaining components, δ11 and δ22,
must therefore be perpendicular to the CN bond.
A relaxation study was also performed on K[Au(CN)2],
at 4.70 T, in order to investigate the origins of the 13C NMR
peak’s line width. The 13C longitudinal relaxation time
constant,T1 =5min,wasmeasuredusinganinversion-recovery
experiment in which 10 data points with delays of 121 µs to
1200 s were measured. The value of the transverse relaxation
time constant, T2 = 4.3 ms, was measured using a
90°-τ-180°-τ-ACQ Hahn-echo pulse sequence, in which
τ was increased from 0.25 to 10.90 ms in seven steps. For
a Lorentzian line shape, the line width at half-height, LWHH,
caused by T2 is given by (π × T2)- 1; therefore, the LWHH
of the potassium salt expected from T2 relaxation alone is ∼
75 Hz, close to the experimentally measured value of 85
Hz. It can therefore be concluded that residual dipolar
coupling with 197Au does not affect the 13C spectrum, and
197Au is probably self-decoupled because of efficient 197Au
quadrupolar relaxation at room temperature.
The structure of [(n-C4H9)4N][Au(CN)2] differs from that
of K[Au(CN)2] in that the [Au(CN)2]- units are separated
by large tetra-n-butylammonium cations,75 with the closest
Au-Au contact well beyond the reach of aurophilic bonding
at 8.05 Å.141,142 Local structural parameters for the single
crystallographically independent site are provided in Table
1. The line width obtained for samples under MAS, 25 Hz,
is much less than the value expected from residual dipolar
coupling with 197Au, see section 2; therefore, the quadrupolar
nucleus is clearly self-decoupled at room temperature.
Carbon-13 NMR parameters derived from theoretical simula-
tion of experimental spectra from the 30% 13C,15N-labeled
Figure 3. Local structure of Tl[Au(CN)2] depicting the intersection of the
two Au chains, where Au atoms 1, 2, and 3 are numbered, and metallophilic
bond lengths are given in Ångstroms. Cyanide ligands are oriented with
carbon bonded to the gold atoms.
material are given in Table 2. These spectral simulations
show an axially symmetric CS tensor and a value of -75
ppm for δ33. Following the reasoning from above, we assign
an analogous CS tensor orientation, that is, δ33 along the
C-N bond. The published value for the C-N bond length,
1.03(4) Å,75 appears anomalously short when compared to
typical values. Deriving a C-N bond directly from Reff yields
1.194(33) Å, and after the vibrational corrections discussed
above are applied, a final bond length of 1.16(5) Å is found.
Whether correcting for vibrational effects or not, the NMR-
derived values are more consistent with a typical C-N bond
length than the 1.03(4) Å X-ray value. The most likely
explanation for the discrepancy is the inherent difficulty of
fitting the small C,N electron densities against the back-
ground of the large Au atom when analyzing the X-ray
experiment. Further discussion of this C-N bond length is
presented below.
There have been several diffraction studies performed on
the complicated Tl[Au(CN)2] structure,77,147,148 in which
there are gold-gold as well as gold-thallium metallophilic
bonds. The unit cell is composed of alternating layers of Tl
atoms and Au(CN)2 groups; a small section of the unit cell
is reproduced in Figure 3. There are three crystallographically
distinct Au atoms, and these are linked together into two
perpendicular chains of Au atoms with bond lengths as noted
in Figure 3 and Table 1. There is an interesting range of Au
(144) Beeler, A. J.; Orendt, A. M.; Grant, D. M.; Cutts, P. W.; Michl, J.;
Zilm, K. W.; Downing, J. W.; Facelli, J. C.; Schindler, M. S.;
Kutzelnigg, W. J. Am. Chem. Soc. 1984, 106, 7672–7676.
(145) Dickson, R. M.; McKinnon, M. S.; Britten, J. F.; Wasylishen, R. E.
Can. J. Chem. 1987, 65, 941–946.
(147) Fischer, P.; Ludi, A.; Patterson, H. H.; Hewat, A. W. Inorg. Chem.
1994, 33, 62–66.
(148) Fischer, P.; Mesot, J.; Lucas, B.; Ludi, A.; Patterson, H.; Hewat, A.
Inorg. Chem. 1997, 36, 2791–2794.
(146) Ramsey, N. F. Phys. ReV. 1950, 78, 699–703.
Inorganic Chemistry, Vol. 48, No. 5, 2009 2323

Harris and Wasylishen
the original neutron diffraction study noted that the observed
Pbcn structure of Tl[Au(CN)2] displays pseudosymmetry:
all positions in the crystal deviate by at most 0.3 Å from a
higher-symmetry Imcb structure.77 In the higher-symmetry
space group, C1 to C4 are all equivalent, and C5 is equivalent
to C6. One point of note is that, within the Pbcn structure,
Au(1) and Au(2) are differentiated by the fact that Au(1) is
involved in an extra aurophilic bond, being at the intersection
point of the two perpendicular Au chains shown in Figure
3. In the theoretical Imcb structure, Au(1) and Au(2) of the
Pbcn structure converge to a single site, but only every
second occurrence of the site equivalent to Au(3) from the
Pbcn structure is occupied. This alternating occupation of
Au(3) splits into two the otherwise equivalent C1 through
C4 atoms in the Imcb structure. The pseudosymmetry
reported by Blom et al.77 demonstrates that the six non-
equivalent carbon sites in the Pbcn structure are geo-
metrically very close to being only three sites. It is also
apparent from this argument that C1 through C4, shown
above to be split into two NMR peaks 6.2 ppm apart, are
different from each other in two ways: small geometrical
differences and different aurophilic bonding of the directly
bonded gold atom.
The pseudosymmetry present in Tl[Au(CN)2] provides a
hint that the presence of two distinct chemical shifts from
carbons C1-C4 is due to the presence of the different
arrangements of aurophilic bonds. However, the experimental
structure contains deviations of up to 0.3 Å from the higher-
symmetry Imcb structure, so it is necessary to explicitly
compare the local structure around each carbon atom. The
geometry of each Au(CN)2 anion in the structure is strikingly
similar: all bond lengths are equal within 0.003 Å, and all
bond angles within 4.2° (see Table 1). Therefore, it seems
unlikely that the local structure of each anion is responsible
for producing the two different chemical shifts observed for
C1-C4. The first coordination sphere around the Au(CN)2
anions containing C1 to C4 includes thallium cations and
neighboring Au(CN)2 anions. It is important to evaluate the
thallium contacts because this heavy atom can have a large
effect on the chemical shift of nearby nuclei, an effect known
as the relativistic heavy-atom light-atom (HALA) ef-
fect.149,150 The HALA effect involves a relativistic induction
of electron spin density on the heavy atom, the outcome
being an altered chemical shift at nearby nuclei. The Au(CN)2
anions containing C1-C4 are coordinated at each end by
two Tl atoms, the average N-Tl bond length being 2.871
Å. While the Tl-N bond length is different for each nitrogen
atom, the maximum deviation from this average is only 0.055
Å. It therefore seems unlikely that the very small changes
in anion geometry and cation coordination are responsible
for splitting the chemical shifts of C1-C4. We can then
ascribe one of the two high-frequency peaks to C1 and C2,
and the other peak to C3 and C4. The 6.2 ppm difference in
Figure 4. Upper spectrum: 13C NMR spectrum of Tl[Au(CN)2] under MAS,
νrot ) 12 kHz, acquired at 11.75 T. Lower spectrum: 13C NMR spectrum
of Tl[Au(CN)2] under MAS, νrot ) 10 kHz, acquired at 7.05 T.
coordination environments displayed in the structure: Au(1)
shares bonds with four gold atoms, Au(2) is bound to two
gold atoms, and Au(3) is coordinated to two gold and two
thallium atoms. The two CN ligands bound to each Au atom
are not related by any symmetry elements, so that there are
a total of six distinct carbon atoms in the structure. In the
following discussion, we label the two carbons bonded to
Au(1) as C1 and C2, those bonded to Au(2) as C3 and C4,
and those bonded to Au(3) as C5 and C6; the nitrogen atoms
will be labeled analogously. The occurrences of Au(1) to
Au(2) and Au(3) in the unit cell are in the ratio 1:1:2, and
there is therefore the same total number of C5 and C6 atoms
as there is for C1 through C4 atoms.
Spectra from 30% 13C,15N-labeled samples of Tl[Au(CN)2]
undergoing MAS are shown in Figure 4, where the 1:1:2
ratio of peak heights is apparent; it is interesting that the six
symmetry-unique carbon atoms in Tl[Au(CN)2] produce only
three 13C peaks in the MAS NMR spectrum at 7.05 T. In
the analogous spectrum at 11.75 T, the two high-frequency
peaks are asymmetric, which demonstrates that these two
peaks arise from at least two carbon sites each. Given that
at least four of the total six carbon sites are associated with
the two high-frequency peaks, and that there must be
twice the intensity produced from C5 and C6 as from C1-C4
sites, the only possible assignment is that the low-frequency
peak is the sum of peaks from carbon atoms C5 and C6. It
follows that the two high-frequency peaks are each composed
of signals from one pair of the C1 through C4 atoms.
The structure of Tl[Au(CN)2] provides valuable insight
into the isotropic chemical shifts observed. The authors of
(149) Nomura, Y.; Takeuchi, Y.; Nakagawa, N. Tetrahedron Lett. 1969,
639–642.
(150) Autschbach, J.; Ziegler, T. Relativistic Computation of NMR
Shieldings and Spin-Spin Coupling Constants In Encyclopedia of
Nuclear Magnetic Resonance; Grant, D. M., Harris, R. K., Eds.; John
Wiley and Sons: Chichester, U.K., 2002; Vol. 9, p 306.
2324 Inorganic Chemistry, Vol. 48, No. 5, 2009

Structural Disorder and Aurophilic Bonding
chemical shifts between the two peaks provides a measure
of the effect of the two 3.04 Å aurophilic bonds formed by
the gold atoms bonded to C1 and C2, but not the gold atoms
bonded to C3 and C4.
making initial estimates of the respective spans for the δiso
) 151.7, 161.3, and 167.5 ppm sites equal to 348, 362, and
371 ppm. A simulated spectrum using these parameters
closely matches the shape of the experimental one but is
slightly too broad. Reduction of each span by 10 ppm
produced a theoretical spectrum matching well with the
experimental one, and the chemical shift parameters are
shown in Table 2. Because the three sites are not resolved,
the entire width of the region containing the high-frequency
discontinuites, ∼50 ppm, is reported as an error range.
However, given the above arguments used to set the initial
estimate of Ω, this large error range can be viewed as rather
generous. A stationary experiment was chosen over a slow
MAS experiment in order to limit the total MAS experi-
mental time used for this thallium salt given its expected
toxicity.
The low-frequency NMR peak of Tl[Au(CN)2] is assigned
to carbons C5 and C6 as a corollary to the above arguments
and is shielded by 15.8 and 9.6 ppm with respect to the two
peaks from carbons C1 through C4. Insight into why the
chemical shift for C5/C6 differs from the two chemical shifts
for C1/C2 and for C3/C4 is again provided by consideration
of the structure. This final Au(CN)2 anion is also coordinated
at the nitrogen atom by two thallium cations: N5 at 2.856
and 3.039 Å and N6 at 2.861 and 3.091 Å. There is also a
third, more distant, N-Tl contact at 3.414 Å for N5 and
3.178 Å for N6. Given that these more distant thallium
contacts are significantly different, but C5 and C6 have the
same chemical shift, these longer Tl-N bonds are probably
not too influential on the shift. The remaining distances are
similar to the range of 2.816-2.904 Å N-Tl bond lengths
of N1 through N4, but the deviations from average are
slightly larger. Comparison of the Au(CN)2 anion containing
Au(3) to the anions containing Au(1) and Au(2) shows that
the anion geometry is nearly identical and the thallium
coordination to the nitrogen atoms is also similar; however,
there is a significant difference in the metallophilic bonding
environment at Au(3), where two Au-Au and two Au-Tl
bonds are present (see Figure 3). This difference in metal-
lophilic bonding is therefore assigned as the cause of the
difference observed for the chemical shifts of carbon atoms
C5/C6 as compared to those for C1 through C4. There are
somewhat larger variations in local structure around C5 and
C6 as compared to C1 through C4, namely, bond angles in
the anion and N-Tl contacts, so assignment of the chemical
shift change to solely alteration of the metallophilic bonding
near C5 and C6 is less certain than for the case of C1 and
C2 versus C3 and C4. However, the difference in metallo-
philic bonding in the anion is likely the dominant effect
causing the different chemical shifts found for C5/C6 versus
those of C1 through C4. In summary, a variation in isotropic
13C chemical shifts of 6.2 ppm from differences in aurophilic
bonding has been observed, while differences in metallophilic
gold-thallium bonding have been found to cause a variation
of 15.8 ppm.
The 13C MAS NMR spectrum of Tl[Au(CN)2] differs from
those of the potassium and tetra-n-butylammonium auro-
cyanides in that a larger line width is observed, see Table 2.
The observed line widths match well with the values
predicted from coupling to the neighboring 197Au nucleus,
vide supra, but this is not the only possible source of line
broadening. Other sources include chemical shift broadening,
spin-spin coupling to 205/203Tl, or rapid relaxation. Compar-
ing the line shapes of the two high-frequency peaks at 4.70
and 11.75 T shows that the two underlying sites are
beginning to separate. In contrast, the line shape of the 151.7
ppm peak is nearly exactly superimposable despite the
magnetic field strength more than doubling. Such a field-
independent line width shows that the cause is not a
distribution of chemical shifts, which would lead to super-
imposable peaks plotted on a parts per million rather than a
frequency scale. We also performed relaxation measurements
on Tl[Au(CN)2] to measure the 13C T1 and T2 relaxation time
constants of the 151.7 ppm sites using the same methods
applied to K[Au(CN)2]; however, the measurements on
Tl[Au(CN)2] were performed at 11.75 T. The T1 was
determined to be ∼17 min, using nine data points spanning
1-1200 s. Hahn-echo experiments yielded T2 as ∼7.3 ms,
where the long T1 limited acquisition to only four data points,
spanning 0.15-10.05 ms (at which point the signal had
decayed to <10% of its original value). In this case, the
contribution from T2 to the LWHH is only ∼45 Hz, whereas
the experimental value of the LWHH is much larger at 250
Hz. Coupling with other nuclei in the sample is the only
remaining explanation of the broad 13C line width in
Tl[Au(CN)2]. However, it is not possible at this time to
differentiate between coupling to the directly bonded 197Au
and 205/203Tl nuclei. This ambiguity arises because, while one
would expect 205/203Tl,13C dipolar interactions to be removed
by MAS, J(205/203Tl, 13C) would not be.
The 13C NMR spectrum of a stationary sample of
Tl[Au(CN)2] was also acquired; unfortunately, multiple sites
and broad line widths make deconvolving the line shape
difficult. An axially symmetric CS tensor is expected from
the investigation of K[Au(CN)2] and [(n-C4H9)4N][Au(CN)2],
and the line shape of this static spectrum is consistent with
that symmetry. Therefore, a value of κ ) 1 was fixed in the
simulations. Spectral overlap prevented determination of the
three Reff values; the simulations instead used the average
value of Reff, 1815 Hz, from K[Au(CN)2] and [(n-
C4H9)4N][Au(CN)2]. With the geometry, Reff, δiso, and κ fixed,
the only remaining parameter is the span, Ω. An initial
estimate of Ω can be obtained from the fact that δ33 is not
expected to change very much from -80 ppm (vide supra).
For an axially symmetric CS tensor, Ω ) 1.5 (δiso - δ33),
4.2. 15N NMR of Gold(I) Dicyanides. The sample of 30%
13C,15N-labeled [(n-C4H9)4N][Au(CN)2] discussed above was
also investigated using 15N SSNMR. A CP Hahn-echo
spectrum from a stationary powder of this sample together
with the best-fit simulation is displayed in Figure 5, while
data determined from spectral simulations are given in Table
3. Just as in the 13C data, an axially symmetric 15N CS tensor
Inorganic Chemistry, Vol. 48, No. 5, 2009 2325

Harris and Wasylishen
overinterpreted in terms of being an absolute determination
of axial symmetry.
Nitrogen-15 NMR spectra of Tl[Au(CN)2] were also
acquired. The isotropic peak of the MAS spectrum is
displayed in Figure 7, where it may be observed that none
of the six unique cyanides are clearly resolved. Unfortunately,
the low receptivity of 15N combined with the broad lines
and site multiplicity prevented determination of the CS
tensors from slow MAS experiments. If the hypothesis that
13C chemical shifts of Tl[Au(CN)2] are insensitive to small
changes in local geometry but dependent mainly on bonding
around the Au atom is correct, the further-displaced 15N
nuclei would be expected to show less sensitivity in their
chemical shifts. Conversely, if it were the thallium contacts
that instead controlled the 13C chemical shifts, the 15N nuclei
would be expected to display a larger spread in chemical
shifts, as they are all closer to the Tl atoms. The above
assignment of metallophilic bonding controlling 13C chemical
shifts is thus borne out in the observation of clearly separated
13C peaks, while nonequivalent 15N sites remain unresolved.
4.3. 13C and 15N NMR of a Gold(III) Tetracyanide. A
sample of 50% 13C,15N-labeled potassium tetracyanoaurate-
(III), K[Au(CN)4], was prepared in order to investigate how
the formal oxidation number of Au affects the cyanide NMR
parameters. A structure for the hydrate has been published,151
but we found the dehydrate more amenable to NMR study,
as the hydrate is not stable to dehydration131 and also yielded
fewer NMR resonances than expected for the low-symmetry
hydrate. Neutron diffraction data for the hydrate evidence
an approximately square-planar [Au(CN)4] anion, no Au-Au
contacts < 4.8 Å, and bond lengths similar to those of the
gold(I) cyanides.151 For an ideal D4h [Au(CN)4] anion, two
of the mirror planes would restrict the principal axis systems
of both 13C and 15N CS tensors: one component would be
along the CN bond, one perpendicular to the plane of the
molecule, and the third perpendicular to the first two
components.
Figure 5. Simulated (upper trace) and experimental (lower trace) 15N cross-
polarized NMR spectra of a stationary sample of [(n-C4H9)4N][Au(CN)2]
at 4.70 T.
is found, so the orientation of δ33 is assigned as coincident
with the C-N bond. The derived value of Reff, 1790(75) Hz,
agrees with the value determined from the 13C spectrum
within experimental uncertainty. Correcting for vibrational
effects leads to a final bond length of 1.17(3) Å. Both NMR
measurements of Reff, by observing 13C or 15N, support the
value of a typical C-N bond length, in opposition to the
X-ray study, where an anomalously short distance was
reported.
The sample of 50% 13C,15N-labeled K[Au(CN)2] was also
studied using 15N SSNMR; however, the long T1 values and
lack of protons for CP enhancement prevented acquisition
of spectra from stationary samples. Values for the CS
parameters were therefore obtained from a slow-spinning
MAS experiment. As mentioned above, the spinning-
sideband intensities depend on the CS tensor span and skew,
as well as the dipolar coupling to the neighboring 13C
nucleus. The spinning sideband intensities were analyzed as
an AX spin pair using the nonlinear least-squares fitting mode
of the SPINEVOLUTION software package.110 The value
of Reff(15N,13C) used in this analysis was fixed to that found
from analysis of the 13C NMR spectrum, and the orientation
of δ33 was not varied but affixed along the C-N bond axis.
Presented in Figure 6 is the experimental spectrum and best-
fit simulation, while the parameters used in the simulation
are given in Table 3. Error values shown in Table 3 are 99%
confidence intervals determined in the fitting procedure.
However, the procedure does not incorporate error values
in the experimental peak heights, so the reported error values
should be considered conservative. Also, for isolated spins,
the difficulty in obtaining accurate values of κ from slow
MAS data in systems where this parameter is near 1 has
been well documented,94,97 and this difficulty probably
persists in the present AX spin system. The fact that the
fitting procedure determines a value of approximately 1 for
κ, and reports a small error range, should therefore not be
Only one 13C isotropic peak is observed in the MAS NMR
spectrum for the unknown dehydrate structure, evidencing
a close similarity to D4h symmetry. We therefore expect the
CS tensor to have an orientation similar to that described
for a D4h tetracyanoaurate anion. When the spectrum of a
stationary sample is observed, Figure 8, it is clear that the
value of δ33 is very similar to that observed for the gold(I)
dicyanides. This similarity provides further evidence that one
CS tensor component is aligned along the C-N bond and
allows assignment of that component as δ33. Therefore, the
value of ꢀ was assigned as 0°, which allowed the best-fit
simulated spectrum displayed in Figure 8 to be generated,
using the parameters reported in Table 2. The remaining task
is to assign which of the principal CS tensor components is
perpendicular to the (approximate) molecular plane. When
δ11 and δ22 are compared, it may be seen that the value of
δ11, 246 ppm, is similar to the ∼270 ppm found for δ11 and
δ22 values in the dicyanoaurate(I) salts, while the value of
δ22, 162 ppm, is quite different. This comparison to the
(151) Bertinotti, C.; Bertinotti, A. Acta Crystallogr., Sect. B 1970, 26, 422–
428.
2326 Inorganic Chemistry, Vol. 48, No. 5, 2009

Structural Disorder and Aurophilic Bonding
Table 3. Parameters Derived from Simulations of Nitrogen-15 SSNMR Spectra
LWHHa
Reff(15N,13C)b
C-N bondc
NMR (Å)
C-N bond
X-ray (Å)
salt
δiso (ppm)
(Hz)
Ω (ppm)
κ
(Hz)
K[Au(CN)2]
266.0(1)
275.33(5)
291.3(5)
284(2)
266(1)
259(1)
25
9
325
325
475(20)
505(5)
0.993(4)
1
[R4N][Au(CN)2]d
Tl[Au(CN)2]
1790(75)
1.17(3)
1.03(4)
Tl[Au(CN)2]
Pb[Au(CN)2]2(H2O)e
Pb[Au(CN)2]2(H2O)e
Pb[Au(CN)2]2
K[Au(CN)4]
480
450
>0.8
>0.7
276(1)
285.31(5)
282.38(5)
224.0(4)
25
25
275
495(25)
495(30)
425(25)
1.000(1)
0.989(1)
1
K[Au(CN)4]
AuCNf
a LWHH measured at 11.75 T. b From best-fit simulated spectra. c Corrected for vibrational effects (see text). d R ) n-butyl. e SSNMR study from Katz
et al.38 f AuCN(B) as representative of the main site in all samples, see text.
Figure 8. Experimental (lower trace) and simulated (upper trace) 13C NMR
spectra of a stationary sample of K[Au(CN)4] at 11.75 T.
Figure 6. Simulated (upper trace) and experimental (lower trace) slow-
spinning 15N NMR spectra of K[Au(CN)2] at 11.75 T (νrot ) 2.5 kHz).
unlikely that δ33 deviates much from the C-N bond
vector.
Nitrogen-15 NMR data for the 50% 13C,15N-labeled sample
of K[Au(CN)4] are summarized in Table 3. The CS tensors
for each of the two sites were determined using the same
method applied for analysis of K[Au(CN)2]: nine peaks from
a slow MAS spectrum obtained at 11.75 T were modeled as
an AX spin system using a least-squares fitting method.110
Again, the value of Reff was fixed to the value determined
from the 13C NMR spectrum, and δ33 was affixed along the
C-N bond vector. There are two isotropic 15N shifts
observed, showing that the dehydrate crystal structure may
deviate at least somewhat from anions of D4h symmetry. It
is perhaps worthwhile pointing out that the presence of two
15N peaks does not prove a break from D4h symmetry, since
there may simply be two molecules in the asymmetric unit
cell of this unknown structure. However, the small 2.9 ppm
difference in isotropic shifts would not require much devia-
tion from D4h symmetry. While the 13C isotropic shift for
this gold(III) cyanide lies some distance from the observed
range of values for gold(I) cyanides, the 15N isotropic shift
of K[Au(CN)4] lies within the range seen for gold(I)
complexes. The carbon atom is directly bound to the gold
atom, and hence its chemical shift is affected greatly by a
change in oxidation state of the metal. The more distant
Figure 7. Experimentally observed 15N NMR spectrum of Tl[Au(CN)2] at
11.75 T under MAS (νrot ) 10 kHz).
gold(I) dicyanides shows that δ22 is the component out of
the molecular plane, as CS tensor components are sensitive
to changes in the planes perpendicular to their direction. The
[Au(CN)4] anion may not be of D4h symmetry, so the
described tensor orientation should be taken as approximate.
However, given the similarity of δ33 ) -88 ppm to the value
of ∼-80 ppm observed for the gold(I) dicyanides, it seems
Inorganic Chemistry, Vol. 48, No. 5, 2009 2327

Harris and Wasylishen
Figure 9. Two unit cells of AuCN; carbon and nitrogen atoms are not
distinguished by the neutron diffraction experiment and are therefore
represented using the same shade. See text for further discussion of the
structure.
nitrogen atom has its chemical shift less affected by the
oxidation state of the metal, less in fact than the differences
in bonding and geometry between dicyanoaurate(I) anions
in different salts.
4.4. 13C NMR of Gold(I) Monocyanide. AuCN is an
interesting material that resisted structural characterization
for a long time due to the unavailability of crystals suitable
for single-crystal diffraction measurements. A combined total
neutron diffraction and X-ray diffraction study of the
powdered material determined the structure but showed
evidence of disorder.82 The structure is depicted in Figure
9, where it can be seen that the infinite · · · CN-Au-CN · · ·
chains are aligned to form sheets of aurophilically bonded
Au atoms (Au-Au distance ) 3.39 Å). Interestingly, one
phase of AuCN monolayers adsorbed on a Au(111) surface
has been shown, through scanning tunneling microscopy and
low-energy electron diffraction, to have the same structure
as a 2D slice of the structure displayed in Figure 9.152,153
An alternative AuCN structure has also been proposed in
the literature and investigated using DFT;154,155 however,
no evidence for this theoretical polymorph was observed in
the powder X-ray diffraction patterns recorded here. While
a structure of AuCN was obtained in the diffraction stud-
ies,78,82 the equivalence of the Au-C and Au-N bond
lengths in AuCN makes it impossible to determine the
orientation of the cyanide ligand in the chains from diffrac-
tion experiments alone. The chain structures of CuCN and
AgCN have been shown via NMR and NQR evidence to
contain disordered CN vectors;86,87 therefore, it is likely that
“head-tail” cyanide disorder also exists in AuCN. Reported
error parameters from the neutron diffraction data fitting
procedure suggest that there may be some slipping of the
chains, i.e., the weak aurophilic bond is broken so that some
neighboring chains do not align at the gold atoms.82 There
are of course other types of disorder, for example, vacancies,
that can occur in such systems, but the two varieties described
above are the most likely.
Figure 10. Upper spectrum: 13C NMR spectrum of AuCN(A) under MAS,
νrot ) 13 kHz, acquired at 11.75 T. Lower spectrum: 13C NMR spectrum
of AuCN(B) under MAS, νrot ) 10 kHz, also acquired at 11.75 T. Peaks
marked with a † in the spectrum of AuCN(B) have been assigned as
impurities (see text).
Table 4. Carbon-13 NMR Parameters of AuCN
AuCN sample
δiso (ppm)
LWHHa (Hz)
area (%)
AuCN(A)
154.1(5)
164.1 (5)
154.6(5)
164.0(5)
136(2)
590
450
545
450
540
650
76
24
84
6.5
6.5
3
AuCN(B)
145(2)
a LWHH measured at 11.75 T.
Figure 10 shows the MAS spectrum of AuCN(A), while
the isotropic shifts, line widths and area are summarized in
Table 4. There are two isotropic shifts visible despite the
fact that the crystal structure contains only one unique site
in the absence of disorder. A 2D homonuclear dipolar-
recoupling experiment, SR26411, was performed in order to
determine if there was an impurity or if there was indeed
two different sites in the material. The ∼5 ms excitation
period used yield spectra sensitive to 13C,13C dipolar
couplings corresponding to C-C distances of ∼5 Å, the
exact distance depending on complicated experimental and
site parameters. This distance is large enough that, even under
the presence of chain slipping and CN head-tail disorder,
11
intersite correlation can be probed. In the SR264 experi-
ment, peaks occur only for dipolar-coupled nuclei and are
observed in the indirect dimension at the sum of the chemical
shifts for the two sites. The format of the experimental
spectrum shown in Figure 11 is such that the self-correlation
peaks expected from the network structure of AuCN occur
on the diagonal and are observed for both peaks. Because
the two sites are at +10 and +20 ppm from the transmitter
in the direct dimension, the two +30 ppm peaks in the DQ
dimension demonstrate intersite connectivity. Experimental
(152) Sawaguchi, T.; Yamada, T.; Okinaka, Y.; Itaya, K. J. Phys. Chem.
1995, 99, 14149–14155.
(153) Yamada, T.; Sekine, R.; Sawaguchi, T. J. Chem. Phys. 2000, 113,
1217–1227.
(154) Hakala, M. O.; Pyykko¨, P. Chem. Commun. 2006, 2890–2892.
(155) Zaleski-Ejgierd, P.; Hakala, M.; Pyykko¨, P. Phys. ReV. B: Condens.
Matter Mater. Phys. 2007, 76, 094104.
(156) The samples used in the 2002 study were still present in our lab,
and a check of their IR spectra revealed a single resonance for each
of the 99% labeled samples (dilabeled 13C, 15N as well as monolabeled
15N and 13C materials).
2328 Inorganic Chemistry, Vol. 48, No. 5, 2009

Structural Disorder and Aurophilic Bonding
Figure 11. Carbon-13 SR26411 homonuclear dipolar-recoupling spectrum
of AuCN(A) acquired under MAS, νrot ) 9.615 kHz, at 11.75 T. The
spectrum resulted from 6 days of experimental time. The corresponding
1D Bloch-decay experiment is placed at the top of the spectrum. In the
presented expansion, self-correlation appears on the diagonal, while intersite
correlations occur as pairs of peaks off the diagonal. Spectral artifacts have
no frequency dependence in t1 and occur at 0 ppm in the indirect dimension.
Figure 12. IR spectra of the CN stretching region, upper spectrum from
AuCN(A) and lower spectrum from AuCN(B)
artifacts have no frequency dependence in the second
dimension and create the small features at 0 ppm. Unfortu-
nately, the similarity of model C-C distances produced by
head-tail disorder or chain slipping precluded us from
attempting to distinguish between them at the resolution
available using this experiment.
Infrared spectroscopy is also useful and serves to confirm
the interpretation of the 13C NMR spectra from this material.
The infrared spectrum of AuCN(A), represented in Figure
12, shows two 12C,14N peaks at 2230.2 and 2167.4 cm-1,
which have a similar intensity ratio to those in the 13C NMR
spectrum. Two peaks from the 13C,15N-labeled cyanides in
the sample are also visible at 2147.8 and 2087.1 cm-1,
positions matching well with those predicted by the simple
harmonic oscillator model. Sensitivity of the IR spectrum
to head-tail disorder versus chain slipping may be found
by investigating the analogous material AgCN. NMR spec-
troscopy has shown that there is head-tail cyanide disorder
in AgCN, while there is only one stretch present in the IR
spectrum.87,156 It seems that vibrational spectroscopy is not
sensitive to head-tail disorder, perhaps because the M-C
versus M-N force constants mimic the indistinguishability
of the bond lengths. In the silver-based material, without
aurophilic bonding, the chains are farther apart than those
in AuCN at 5.9 versus 3.4 Å and are not aligned at the metal
position. Since the chains of AuCN interact more strongly
than those of AgCN, it is quite plausible that IR spectroscopy
is sensitive to chain slipping in AuCN, while no change is
observed in AgCN. Vibrational spectroscopy therefore sup-
ports the 13C NMR argument for the presence of chain
slipping in AuCN.
An alternate synthetic method, vide supra, yielded
AuCN(B), whose 13C NMR spectrum, powder XRD pattern,
and IR spectrum provide further insight into the AuCN
system. Two samples were synthesized using this method,
and the two samples were found to produce nearly exactly
superimposable 13C line shapes at 11.75 T. The 13C NMR
spectrum of AuCN(B) is shown in Figure 10, with the
corresponding line shape information in Table 4. This second
sample has two additional peaks beyond the pair of reso-
nances in common with AuCN(A). When the 13C SSNMR
spectra for the two samples are compared, it may be seen
Two separate 13C isotropic shifts are observed within the
same material; therefore, the chemical shift of the carbon
nucleus must be sensitive to some disorder in the material.
Assignment of the multiplicity as due to head-tail cyanide
disorder or chain slipping can be performed by comparison
to the expected chemical shift effect of each. In a SSNMR
study of CuCN, an upper limit of ∼2 ppm for the difference
in δiso(13C) from a change in neighboring CN orientation was
found; that is, the difference in δiso for 13C-Cu-NC versus
13C-Cu-CN is <2 ppm.86 The analogous effect of cyanide
orientation on the 13C chemical shift in the material AgCN
was placed as approximately 0.8 ppm.87 The carbon-metal
bond distance in AuCN is midway between that of AgCN
and CuCN, so the effect of neighboring CN orientation on
the 13C chemical shift is likely midway between 0.8 and <2
ppm, or at least much less than 10 ppm. Because chain
slipping would be accompanied by the breaking of aurophilic
bonds, the effect of chain slipping on δiso(13C) can be
estimated from the data presented above for Tl[Au(CN)2].
For [Au(CN)2]- moieties with nearly identical geometries,
changes are seen in the 13C isotropic shifts of up to 15.7
ppm because of the difference in the metal identity and length
of metal-metal bonds, while a change of 5.9 ppm was
observed for differences in the number and length of Au-Au
contacts for aurophilic bonding. The observed 10 ppm
difference between the two AuCN(A) sites is consistent with
this range while being noticeably larger than the small range
expected from CN head-tail disorder. This argument does
not rule out the presence of head-tail disorder; it simply
shows that any effect of that type of disorder on δiso(13C) is
likely hidden in the large (5 ppm) line width.
Inorganic Chemistry, Vol. 48, No. 5, 2009 2329

Harris and Wasylishen
that the ratio of the pair of resonances the two samples have
in common differs; there is more of the high-frequency site
in AuCN(A). A sample with more chain slipping would be
expected to yield a broader XRD pattern, whereas a sample
with a differing amount of head-tail disorder would yield
the same XRD peak breadths due to the similar scattering
power of C and N atoms. Line widths in the XRD pattern
of AuCN(B) are narrower than those from AuCN(A),
consistent with an assignment of the 154 ppm site as an
ordered site and the 164 ppm peak as a chain slipped site.
Unfortunately, lower intensity and resolution prevented the
use of an SR26411 experiment to identify the extra two peaks
in the 13C NMR spectrum of AuCN(B). The nature of the
sites producing these two NMR resonances is, however,
hinted at by the XRD pattern of AuCN(B): if the extra peaks
represented new types of disorder superimposed on the
AuCN structure, an XRD pattern broader than that observed
for AuCN(A) would be expected, whereas a narrower pattern
is observed. On the other hand, amorphous impurities could
very easily cause only a broad, diffuse X-ray pattern and
not provide any noticeable change in the overall XRD
pattern. We therefore assign the two low-frequency 13C NMR
peaks of AuCN(B) as amorphous impurities rather than
components in a disordered AuCN structure. The IR spec-
trum of AuCN(B), represented in Figure 12, contains the
same peaks as the spectrum from AuCN(A), but the CN
region is broader. Such an increased breadth of IR resonances
from AuCN(B) is also consistent with the presence of
amorphous impurities in that sample. Identification of the
low-frequency peaks as impurities also seems logical, as it
is unlikely that AuCN(B), which has less chain slipping,
would be the material that displays an increase in the types
of disorder. Investigation of a second sample with a slightly
different composition yielded observations entirely consistent
with the above assertion of measurable chain slipping in
AuCN. Interestingly, the 7 ( 2% fraction of cyanides in
slipped chains observed for AuCN(B) is lower than the 24
( 5% found for AuCN(A).
In an effort to investigate the synthesis used to prepare
AuCN(A), a second material, AuCN(A′), was prepared with
a different amount of reducing agent. Synthesis of AuCN(A)
used less than a stoichiometric amount (75%) of the reducing
agent and was found to yield a sample that contained a very
broad 13C resonance centered at ∼50 ppm to low frequency
of the main peaks. We assign this broad peak as cyanides
bound to Au(III) from the isotropic shift observed for
K[Au(CN)4]. The approximately 30-ppm-broad (at 11.75 T)
peak was found to remain even after the washing process.
Use of a small (10%) excess of the reducing agent produced
AuCN(A′), in which some of the Au(III) ions from solution
were reduced further than to the intended target of Au(I).
Comparison of the XRD pattern to that of a Au(s) powder
showed that there was indeed some metallic gold present.
The presence of Au(s) also affected the color of the material:
the expected lemon yellow color was found to be tinted
green. Clearly, batches of AuCN made via the process in
patent #4,222,996132 without careful control of the amount
of reducing agent will contain either Au(III) or Au(0) as an
impurity. The two samples AuCN(A) and AuCN(A′) were
found to contain the same percentage of cyanides assigned
to slipped chains, 24 ( 5% and 25 ( 5%, respectively.
On the basis of the above analysis of the spectra from
MAS samples, we feel the two isotropic shifts displayed by
AuCN are well understood; however, the broad line shapes
are complicated and require further experiments to fully
understand. The pair of isotropic shifts displayed in the
spectra of AuCN(A) result from chain-aligned and chain-
slipped sites, but the width of these two peaks probably has
a contribution from varying numbers of aurophilic bonding
partners, as well as from head-tail cyanide disorder. The
line width of the 154 ppm peak of AuCN(A), see Table 4,
is approximately 30% smaller in hertz at 7.05 T versus 11.75
T; this smaller line width at lower applied field is consistent
with the presence of a distribution of chemical shifts, which
could result from disorder in the number of aurophilic
bonding partners as well as head-tail cyanide disorder.
Comparing the 154 ppm peaks of AuCN(A) and AuCN(B)
at 11.75 T shows that the two have nearly the same line
width, where the only significant difference is that the peak
from AuCN(B) is slightly narrower at the base. However,
the 154 ppm peak of AuCN(B) does not display the same
narrowing in hertz upon decreasing the applied field. Instead,
this peak is nearly exactly superimposable when plotted on
a frequency scale at 7.05 and 11.75 T, so that it is wider
than the corresponding peak in AuCN(A) at 7.05 T. The
behavior of the 154 ppm peak from AuCN(B) with changing
field points to some field-independent broadening mecha-
nism, see above for further discussion of the possibilities,
but it is unusual that this mechanism seems to be absent, or
at least lessened, in AuCN(A). Perhaps some difference in
T2 relaxation time constants or in residual dipolar coupling
with 197Au causes the discrepancy between the two samples.
However, further experiments would be required to address
this matter properly. One possible method would be to apply
annealing techniques, shown useful in diffraction studies of
CuCN,80 to remove some of the disorder-induced chemical
shift broadening of this peak, or to use high-temperature
experiments to decouple 197Au in order to separate the line
broadening mechanisms.
The 13C NMR spectrum from a stationary sample of
AuCN(B), see Figure 13, has the appearance of that from a
single-site material. Most likely, the powder pattern is
representative of the 85% dominant site, see Table 4, with
the other minor sites contributing to broadening only.
Simulation of this powder pattern therefore provides the CS
tensor and Reff(15N,13C) coupling constant from the main,
aligned-chain, site. If we ignore cyanide head-tail disorder
in neighboring chains, a factor unlikely to affect the CS
tensor, the CN bond lies along a C6 symmetry axis.82
Therefore, the CS tensor is axially symmetric, and the PAS
is oriented with δ33 along the C-N bond.89,95,96 Parameters
from the best-fit simulation are given in Table 2, where it
may be seen that the 1.14(4) Å C-N bond length agrees
well with the neutron-diffraction value of 1.150(28) Å.
4.5. 15N NMR of Gold(I) Monocyanide. Nitrogen-15
NMR of AuCN was also investigated. The MAS spectrum
2330 Inorganic Chemistry, Vol. 48, No. 5, 2009

Structural Disorder and Aurophilic Bonding
number of nearest neighbors. If the argument about increased
sensitivity for 15N over 13C is the true cause of the increased
number of sites, it may be that increased chemical shift
broadening is responsible for the lower relative integrated
intensity of the secondary peaks in the 15N spectrum (see
data for AuCN(A) in Table 4 versus that in Table 5). Also
shown in Figure 14 is the nitrogen-15 spectrum of AuCN(B)
under MAS, an interesting spectrum in that only one peak
is visible despite the multiplicity seen in the carbon-13
spectra. AuCN(B) has less than one-third the amount of
slipped chains as does AuCN(A), which may be the reason
that secondary resonances are only observed in nitrogen-15
spectra of the latter. Particularly given that the 15N NMR
spectra of AuCN(A) display secondary resonances at only
approximately half the integrated intensity seen in the 13C
NMR spectra of this material. We did attempt to observe
secondary peaks in a 15N spectrum of AuCN(B) by increasing
the applied field to 21.14 T (a 91.2 MHz Larmor frequency),
but only a single peak was obtained.
Figure 13. Experimental (lower trace) and simulated (upper trace) 13C NMR
spectra of a stationary sample of AuCN(B) at 4.70 T.
There is some uncertainty in a full assignment of the 15
N
data, but there is little doubt that the main ∼ 224 ppm
resonance is that of the neutron-diffraction structure (i.e.,
aligned gold cyanide chains). This peak is the same in all
the samples, so we used the highest-quality slow MAS
spectrum, that of AuCN(B) at 11.75 T, to determine the 15
N
CS tensor; the results are reported in Table 3. Eight peaks
from the experimental spectrum were used in the least-
squares fit,110 and Reff was set to the value determined from
the 13C NMR spectrum. Due to the C6 symmetry axis at the
nitrogen,82 the model used for analyzing the spectrum had
κ fixed to 1, and δ33 fixed along the C-N bond. The 15N
chemical shift of AuCN lies outside the range found for those
from the other gold(I) and gold(III) cyanides presented above.
As might be expected, δiso(15N) appears more sensitive to
the formation of a new bond directly to nitrogen in this
bridging CN ligand, than to the small geometry changes
between gold(I) dicyanides or even to the change in oxidation
state of the gold atom. Previously reported nitrogen-15 data
from a dicyanoaurate in which the cyanide forms a bridging
bond to lead in Pb(H2O)[Au(CN)2]2 (and its dehydrate),38
see Table 3, exhibits 15N chemical shifts that match better
with the range observed here for nonbridging cyanide groups
than that observed for the bridging cyanide of AuCN. The
reason for the δiso(15N) of Pb[Au(CN)2]2(H2O)x agreeing
better with the nonbridging nitrogen environments may be
that the Pb-N bond is much longer (2.595-2.8 Å) than the
Au-N bond in AuCN (1.9703 Å).
Figure 14. Upper spectrum: 15N NMR spectrum of AuCN(A) under MAS,
νrot ) 13 kHz, acquired at 11.75 T. Lower spectrum: 15N NMR spectrum
of AuCN(B) under MAS, νrot ) 10 kHz, also acquired at 11.75 T.
Table 5. Nitrogen-15 NMR Parameters of AuCN
AuCN sample
δiso (ppm)
LWHHa (Hz)
area (%)
AuCN(A)
224.1(5)
211.5(10)
370(10)
300
300
1500
200
87
6
7
AuCN(B)
a LWHH measured at 11.75 T.
223.40(25)
100
5. Conclusions
of AuCN(A), shown in Figure 14, is nearly ninety percent
one resonance, with a small peak to low frequency and a
broad absorbance to high frequency, see Table 5. Nitrogen-
15 chemical shifts are typically more sensitive to environ-
mental changes than are carbon-13 shifts, which hints at one
possible explanation for the spectrum seen in Figure 14.
While only two distinct values are observed for δiso(13C),
for aligned and shifted chains, the more sensitive 15N nucleus
may allow resolution of more sites, e.g., differences in the
Carbon-13 and nitrogen-15 NMR spectra of several gold(I)
dicyanides and one gold(III) tetracyanide have been acquired
and are useful in providing survey data for the interpretation
of future NMR experiments on cyanoaurate coordination
polymers. Chemical shift tensors for the cyano 13C and 15
N
nuclei are reported, and a range of isotropic shifts for the
various environments are determined. Static 13C, and in some
cases 15N, NMR spectra are used to provide several r(C,N)
distance measurements and allow a value of 1.17(3) Å for
Inorganic Chemistry, Vol. 48, No. 5, 2009 2331

Harris and Wasylishen
the C-N bond in [(n-C4H9)4N][Au(CN)2], more accurate than
the existing X-ray diffraction measurement, to be determined.
The impact of metallophilic bonding on 13C chemical shifts
is isolated, and this bonding is found to cause a range of
approximately 16 ppm in δiso(13C). Also, the influence of
gold’s formal oxidation state on 13C and 15N chemical shifts
are determined.
The disordered material AuCN was investigated by study-
ing several samples using 13C and 15N NMR, powder X-ray
diffraction, and IR spectroscopy. Homonuclear carbon-13
dipolar recoupling experiments are found to be invaluable
in determining whether NMR peaks arise from the same or
differing materials. Furthermore, the 13C isotropic chemical
shift is found to be sensitive to alignment of the one-
dimensional chains with respect to their neighboring chains,
that is, breaking of the aurophilic bonds.
Carbon-13 and nitrogen-15 NMR experiments provide a
clearer picture of the disorder superimposed on the average
structures of the coinage metal monocyanides. These materi-
als are all composed of one-dimensional · · ·NC-M-CN · · ·
chains,78-85 and 65/63Cu NQR studies in combination with
13C and 15N NMR spectroscopy of CuCN and AgCN have
shown that the orientation of the cyanide ligands varies: 30
( 10% of the metal sites were disordered (C-M-C or
N-M-N), while 70 ( 10% were ordered (C-M-N).86,87
Disorder of the cyanide orientations is likely also present in
AuCN but simply not visible given that the expected effect
on the chemical shifts is smaller than the observed line
widths. The major difference between the structures of AuCN
and CuCN/AgCN is that, with the presence of gold, there is
an increased interaction between the metal atoms. Because
only very weak interactions between the chains in CuCN
and AgCN exist, any misalignment of the chains creates no
visible effect on the NMR parameters. It follows logically
that the stronger interchain forces in AuCN allow the
possibility of chain slipping to be investigated. The amount
of this disorder in AuCN is found here to vary with the
synthetic method, the two syntheses applied here producing
samples containing from 7 ( 2% to 25 ( 5% slipped chains.
It is likely that the lowered strength of interchain interactions
in AgCN/CuCN versus AuCN is responsible for an increased
amount of chain slipping in the two lighter systems. This
increased slipping is consistent with the broader XRD
patterns observed for AgCN/CuCN as compared to AuCN.78
The body of research on the group 11 monocyanides all
converges to describe a picture of these materials as infinite
one-dimensional chains, with a significant fraction of the
cyanide ligands disordered and also with a large percentage
of the chains slipped with respect to their neighbors.
Acknowledgment. The authors thank the solid-state NMR
research group at the University of Alberta for many helpful
discussions. Professor David Bryce, University of Ottawa,
is also thanked for his interest in this project, as well as the
acquisition of spectra from initial gold cyanide samples. We
also thank Dr. Victor Terskikh, Canadian National Ultrahigh-
Field NMR Facility for Solids, for acquisition of spectra at
21.14 T. The National Ultrahigh-Field NMR Facility for
Solids (Ottawa, Canada), is a national research facility funded
by the Canada Foundation for Innovation, the Ontario
Innovation Trust, Recherche Québec, the National Research
Council Canada, and Bruker BioSpin and managed by the
University of Ottawa (www.nmr900.ca). The Natural Sci-
ences and Engineering Research Council of Canada (NSERC)
supports operation of this facility through a Major Resources
Support grant. Dr. Darren Brouwer, NRC Steacie Institute
for Molecular Sciences, is also thanked for introducing us
11
to the SR264 experiment, and for helpful discussions.
Funding for this research was provided by the Natural
Sciences and Engineering Research Council of Canada, as
well as the Canada Research Chairs Program (R.E.W. holds
a Research Chair in Physical Chemistry at the University of
Alberta). K.J.H. also thanks the Province of Alberta for
financial support, as well as the University of Alberta for
initial funding of this research.
IC8022198
2332 Inorganic Chemistry, Vol. 48, No. 5, 2009