Photochemical reactivity of two gold(I) dinuclear complexes, cis/trans-(AupNBT)2dppee: Isomerization for the cis-(AupNBT) 2dppee isomer, radical substitution for trans-(AupNBT) 2dppee

Article abstract of DOI:10.1016/j.ica.2012.03.030

Two complexes, cis (1) and trans (2) (AupNBT)₂dppee (pNBT = p-nitrobenzenethiol; dppee = bis(diphenylphosphino)ethylene) have been synthesized and characterized. X-ray diffraction studies reveal that cis-(AupNBT)₂dppee has an intramolecular gold-gold distance of 3.0036(3) ? while the trans-(AupNBT)₂dppee crystal is a dimer with an intermolecular gold-gold distance of 3.1201(4) ?. We have used UV-Vis spectroscopy, ¹H NMR, and ³¹P {¹H} NMR to investigate the photochemical reactivity of both complexes. In a range of organic solvents with λ > 230 nm cis-(AupNBT)₂dppee readily isomerizes to the trans isomer. Under these same conditions, the trans isomer does not isomerize to cis-(AupNBT)₂dppee, but undergoes a photochemical substitution reaction. In chlorinated solvents (CH ₂Cl₂, CH₃Cl) trans-(AupNBT)₂dppee reacts to form trans-(AuCl)₂dppee (3) which crystallizes into an extended chain of molecules linked via intermolecular gold-gold distances of 3.0203(2) ?. TDDFT calculations support the observed experimental results. We propose that at high energies the reaction is initiated by chlorine radicals from direct photolysis of the solvent. At longer wavelengths the excited state of the metal complex is the photoactive species, abstracting a chlorine radical from the solvent.

Full text of DOI:10.1016/j.ica.2012.03.030

Inorganica Chimica Acta 392 (2012) 300–310
Contents lists available at SciVerse ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
Photochemical reactivity of two gold(I) dinuclear complexes,
cis/trans-(AupNBT)2dppee: Isomerization for the cis-(AupNBT)2dppee isomer,
radical substitution for trans-(AupNBT)2dppee
a
b
b,1
Janet B. Foley a, , Angela Herring , Bo Li , Evgeny V. Dikarev
⇑
a Bennington College, Bennington, VT 05201, USA
b University at Albany, SUNY, Albany, NY 12222, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Two complexes, cis (1) and trans (2) (AupNBT)2dppee (pNBT = p-nitrobenzenethiol; dppee = bis(diphenyl-
phosphino)ethylene) have been synthesized and characterized. X-ray diffraction studies reveal that
cis-(AupNBT)2dppee has an intramolecular gold–gold distance of 3.0036(3) Å while the trans-
(AupNBT)2dppee crystal is a dimer with an intermolecular gold–gold distance of 3.1201(4) Å. We have
used UV–Vis spectroscopy, 1H NMR, and 31P {1H} NMR to investigate the photochemical reactivity of both
complexes. In a range of organic solvents with k > 230 nm cis-(AupNBT)2dppee readily isomerizes to the
trans isomer. Under these same conditions, the trans isomer does not isomerize to cis-(AupNBT)2dppee,
but undergoes a photochemical substitution reaction. In chlorinated solvents (CH2Cl2, CH3Cl) trans-
(AupNBT)2dppee reacts to form trans-(AuCl)2dppee (3) which crystallizes into an extended chain of
molecules linked via intermolecular gold–gold distances of 3.0203(2) Å. TDDFT calculations support
the observed experimental results. We propose that at high energies the reaction is initiated by chlorine
radicals from direct photolysis of the solvent. At longer wavelengths the excited state of the metal com-
plex is the photoactive species, abstracting a chlorine radical from the solvent.
Received 24 January 2012
Received in revised form 10 March 2012
Accepted 14 March 2012
Available online 27 March 2012
Keywords:
Gold
X-ray crystal structure
Photochemistry
Gold–gold bonding
Luminescence
Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction
types [8]. Gold–gold interactions, the attraction between gold
atoms in solid state structures or in solution, can provide scaffold-
Gold(I) compounds have been used to treat rheumatoid arthritis
[1] for over 50 years and, more recently, some of the same com-
pounds have shown potential as anticancer drugs [2]. Some studies
propose that these gold complexes inhibit selenoenzymes and that
thiol ligand exchange on the gold plays a crucial role [3]. Some gold
thiol rheumatoid arthritis drugs have shown sensitivity to light [4]
and although these drugs are not widely used now, the nature of
that photochemical instability has implications for other phos-
phine–gold–thiol drugs. The current interest in gold complexes ex-
tends beyond their medicinal applications. The binding of gold to
sulfur is the basis of self assembled monolayers applications [5],
biosensor devices [6], and attachments to gold nanoparticles [7].
Gold(I) complexes have shown promise as catalysts for water sol-
uble addition reactions, rearrangements, and many other reaction
ing for supramolecular constructions [9] as well as affect the lumi-
nescent properties of gold molecules [10]. The ligand replacement
characteristics, the stability of gold–sulfur bonds, and the photo-
physical properties of gold(I) complexes are basic concepts that
directly relate to advancement in these areas of research.
Previous work showed that cis-(AuX)2dppee complexes (X = Cl,
Br, I), that have short intramolecular gold–gold distances, isomerize
to the trans isomer when exposed to light, both UV and k > 320 nm
[11]. Quantum yields, measured at 334 nm for the halides were
found to be
U = 0.204 0.062 (Cl), 0.269 0.092 (Br), 0.363 0.0
55 (I) [12]. Ab initio calculations [13] attribute the isomerization
to a p⁄ excited state accessible to the cis isomer because of the intra-
molecular gold–gold interaction. Because gold–gold bonding is
crucial for the isomerization in the halide series and these weak
interactions are predicted to be enhanced by ‘‘soft’’ ligands [14],
the photochemical properties of a thiol analog of cis/trans-(AuX)2
dppee presented a natural extension of our previous study.
Most of the photochemistry of Au(I) has been redox, resulting in
disproportionation to Au(0) and Au(III). A few extensive reviews
have outlined these photochemical studies [15]. Gold complexes
can also form three coordinate structures by oxidative addition
of halides or alkylhalides to form gold(II) complexes. This
Abbreviations: pNBT, para-nitrobenzenethiol; dppee, bis(diphenylphosphino)
ethylene; p-ClBT, para-chlorobenzenethiol; p-NH2BT, para-aminobenzenethiol;
pTC, para-thiocresol; hfac, 111555hexafluoro-2⁄4-pentanedionato; TG, tetraglyme
2581114-pentaoxapentadecane.
⇑
Corresponding author. Tel.: +1 802 440 4463.
E-mail addresses: jfoley@bennington.edu (J.B. Foley), dikarev@albany.edu (E.V.
Dikarev).
1
Tel.: +1 518 442 4401.
0020-1693/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.ica.2012.03.030

J.B. Foley et al. / Inorganica Chimica Acta 392 (2012) 300–310
301
mechanism is most commonly seen for dinuclear substrates with
2.2. Photolysis reactions
bridging ligands [16], dithiocarbamates [17], and phosphine annu-
lar ions [18]. Radical initiated oxidative addition has also been
reported in both gold complexes and with other metals [19]. Many
metals (Pt [20], Ru [21], Fe [22], Au [15]) have been shown to un-
dergo photochemical reactions in chlorinated solvents. These reac-
tions can be initiated by homolytic cleavage of the solvent, by
excitation of the metal complex, or by CTTS (charge transfer to sol-
vent) transitions. There can also be solvent initiated and metal cen-
tered reactions happening in the same system. The reactivity
depends on the solvent C–Cl bond energy, which decreases along
the series CH2Cl2 > CHCl3 > CCl4.2 An experiment by Hoggard and
co-workers [23] reevaluates some previous examples of iron dithio-
carbamate photochemistry, proposing that at high energies in chlo-
rinated solvents, a solvent-initiated process predominates and at
lower energies a metal-initiated process generates the radicals, both
resulting in the same product. Photosubstitution without change in
metal oxidation state is not common but Hoggard and co-worker
[24] suggest a direct solvent bond homolysis for the photosubstitu-
tion of [Ru(bpy)2(N3)2] to [Ru(bpy)2(Cl)2] with no participation of the
metal complex because the reaction does not happen at longer
wavelengths where the complex absorbs.
Au(I) is a formally closed shell metal ion which eliminates the
option of d–d transitions that might happen at low energies but
ligands, particularly reducing thiols, can push the lowest energy
absorption into the visible, making it likely that the lowest energy
transition is an LMCT (ligand to metal charge transfer) [25]. Relativ-
istic effects [26] cause the 6s and 6p orbitals on the gold to be close in
energy to the formally filled 5d and make metal orbitals potentially
available to the excited state. Many gold(I) complexes luminesce un-
der mild conditions. This property has been the basis of systems that
detect ions in solution [27], the presence of organic solvents [6b],
and many other applications. The possibility of low energy transi-
tions, the electronic effects of gold–gold interactions, and the heavy
atom properties of gold combine to make these complexes ideal sub-
strates for probing the basic chemistry of this unique metal.
Our present study describes (1) the synthesis and characteriza-
tion of the cis and trans-(AupNBT)2dppee complexes, (2) the
photochemical isomerization of cis-(AupNBT)2dppee, (3) the photo-
chemical ligand substitution of trans-(AupNBT)2dppee, and (4) DFT
and TDDFT calculations performed to explain the experimental
results.
All reactions were subjected to a 100 W Oriel photolysis lamp
and monitored by UV–Vis or NMR spectroscopy. Samples were
photolyzed in quartz cuvettes allowing wavelengths of the Hg
Arc lamp greater than 230 nm to illuminate the samples. A
420 nm Oriel cutoff filter: (transmittance 440–2500 nm; 75%
transmittance at 440 nm) was used for the low energy excitations.
Reactions were carried out in air except when nitrogen is specified.
In some cases the photolysis in quartz used a 14 ꢁ 2 mm mask on
the cuvette sample holder to limit photon flux to slow the reaction
enough to observe more detail.
2.3. Synthesis of cis-(AupNBT)2dppee (1)
Cis-(AuCl)2dppee was synthesized by literature methods [28].
The white solid (0.255 g, 0.29 mmol) was degassed and dissolved
in degassed methylene chloride. In a separate flask 2.5 equivalents
(0.1147 g, 0.73 mmol) of p-nitrobenzenthiol were dissolved in eth-
anol and put under a nitrogen atmosphere. Ethanolamine (0.0445
ml, 0.73 mmol) was added to deprotonated the ligand. The solution
turned bright red–orange and was added via cannula to the solu-
tion of the gold complex. After stirring for 1 h in ice, the solvent
was evaporated under vacuum and an orange solid was recovered
The product was recrystallized from methylene chloride/hexanes.
Average yield = 78%. Elemental Anal. Calc. for C38H30Au2N2O4P2S2:
C, 41.53; H, 2.73; N, 2.55. Exp.: C, 41.57; H, 2.68; N, 2.5%. 1H
NMR (CD2Cl2, 300 MHz) d = 7.3–7.65 (m), d = 7.74 (d) {J = 9 Hz}.
31P {1H} NMR d = 19.5 ppm (s); 85% H3PO4 = 0 as reference.
2.4. Synthesis of trans-(AupNBT)2dppee (2)
Trans-(AuCl)2dppee was synthesized according to literature
methods [29]. 0.255 g, (0.29 mmol) of the relatively insoluble
white compound was degassed, put under nitrogen and dissolved
in degassed methylene chloride to form a white slurry. In a
separate flask 2.5 equivalents (0.1147 g, 0.73 mmol) of p-nitro-
benzenthiol were dissolved in ethanol and put under a nitrogen
atmosphere. Ethanolamine (0.0445 ml. 0.73 mmol) was added to
deprotonate the ligand. The solution turned bright red–orange
and was added via cannula to the solution of the gold complex
which turned transparent yellow. After stirring for one hour in
ice, the solvent was evaporated under vacuum and a yellow, air
stable solid was recovered. Product was recrystallized in methy-
lene chloride/hexane to give an average of 67% yield. Elemental
Anal. Calc. C38H30Au2N2O4P2S2: C, 41.53; H, 2.73; N, 2.55. Exp.: C,
41.31; H, 3.09; N, 2.99%. 1H NMR: (CD2Cl2, 300 MHz) d = 7.545–
7.795(m), d = 7.86(d) (J = 9 Hz); 31P {1H} NMR d = 35.6 ppm (s);
85% H3PO4 = 0 as reference.
2. Experimental
2.1. General methods
HAuCl4ꢀ3H2O, p-nitrobenzenethiol and other reagents were
purchased from Alfa Aesar or Aldrich and used as is. Syntheses
were carried out under nitrogen using standard Schlenk line tech-
niques and adapted from previously reported procedures [11]. 1H
NMR and 31P {1H} NMR spectra were recorded on a Varian
200 MHz or a Varian 300 MHz instrument in CDCl3 or CD2Cl2.
Chemical shifts are quoted relative to TMS for the 1H NMR or
85% H3PO4 for 31P {1H} NMR. C, H, N, Cl analyses were conducted
by Desert Analytics, part of Columbia Analytical Services, Tucson,
Arizona. Crystals for X-ray analysis were grown by slow diffusion
of ethyl ether into a methylene chloride solution of the gold com-
plex. Absorption spectra were recorded on an Ocean Optics 2000
spectrometer. Emission and excitation spectra were measured on
a Jobim Hovan fluorimeter in solution (CH2Cl2) or in the solid state
as KBr mixtures at room temperature.
2.5. Characterization of the product, trans-(AuCl)2dppee (3) from
photolysis of trans-(AupNBT)2dppee
Crystals for the X-ray analysis were collected from the slow dif-
fusion recrystallization (methylene chloride/ethyl ether) of the
product, identified as trans-(AuCl)2dppee, of the photolysis in
quartz of trans-(AupNBT)2dppee in methylene chloride. Elemental
Anal. Calc. for C26H22Au2Cl2P2: Calc: C, 36.20; H, 2.56; Cl, 8.25.
Exp.: C, 36.37; H, 2.97; Cl, 8.40%.3 1H NMR: (CD2Cl2, 300 MHz)
d = 7.66–7.53(m), d = 7.114 (t) (J = 20 Hz), 31P {1H} NMR (CD2Cl2):
d = 29.6 ppm(s); 85% H3PO4 = 0 as reference.
2
In addition to the C, H, Cl analysis, there was a small percentage of sulfur (0.11%)
3
which might indicate some residual starting material or disulfide in the product that
C–Cl bond dissociation energy (kJ mꢂ1
) from NIST: CH2Cl2 = 307.8,
was not evident in the 31P 1H NMR or the 1H NMR.
CHCl3 = 299.15, CCl4 = 284.1.

302
J.B. Foley et al. / Inorganica Chimica Acta 392 (2012) 300–310
2.6. Ab initio computations
p-aminobenzenethiol, and p-chlorobenzenethiol, were unsuccess-
ful, although the trans counterparts were made in good yields. The
Ground state electronic structures were calculated using the
density functional (DFT) application in the GAUSSIAN 09 program
[30] on model compounds that substituted hydrogens for the phe-
nyl rings in the phosphine. Singlet and triplet excited states were
calculated using the TDDFT application in the same program. The
B3LYP exchange–correlation function was used with the LANL2DZ
UV–Vis spectrum shows a high energy absorbance with a maximum
at 236 nm (
e
= 4.37 ꢁ 104 Mꢂ1 cmꢂ1) that has been assigned to phe-
nyl transitions in the phosphine based on comparison to similar
compounds [31]. There is also a broad low energy peak with a max-
imum at 400 nm (
e
= 3.25 ꢁ 104 Mꢂ1 cmꢂ1), assigned to the p-nitro-
benzenethiol ligand, based on comparison to spectra of other
molecules in this series [11] as well as other experiments with this
ligand [32]. The extended conjugation of the nitrobenzenethiol li-
gand is likely responsible for this low energy absorbance compared
to other benzene thiol ligands. Ab initio calculations indicate that the
percentage of sulfur in the HOMO of this ligand is greater than for
any other of the common aromatic substituents (48.1%) [33] and
the extreme electron withdrawing nature of the nitro substituent
stabilizes the HOMO of this ligand compared to other benzenethiols.
The stability of the HOMO appears to allow for productive overlap of
the sulfur orbitals with the gold while other ligands (p-ClBT, p-NH2–
BT, pTC) do not.
basis set for gold with an additional f function (af = 0.2) in the va-
lence shell of the gold. A 6-311G(d,p) basis set was used for the non-
metal atoms. Full geometry optimizations were performed for all
molecules in the ground state and single point calculations were
used to examine the excited states. The Polarized Continuum Model
using the integral equation formula (IEF-PCM) was used to simulate
the solvent, methylene chloride, for the trans-(AupNBT)2dppee
calculation because the nature of the reaction suggested more reli-
ance on the nature of the solvent and it gave better agreement with
the experimental results. Solvent parameters were not used in the
cis-(AupNBT)2dppee calculation. No symmetry constraints were im-
posed. GAUSSVIEW 05 was used for visualization.
The primary feature of the crystal structure is the 3.0037(3) Å
gold–gold distance, Fig. 1 implying that there are significant intra-
molecular gold–gold interactions [34]. The linear geometry of the
P–Au–S axis is slightly distorted to 174.10(3)° in order to maximize
the gold–gold attraction. There are no close intermolecular gold–
gold interactions in the unit cell. The ‘‘arms’’ of the molecule are
not parallel, but are offset by 74.4°, minimizing the repulsions that
might occur between the p-nitrobenzenethiol ligands. Table 2 com-
pares specific details with other similar crystal structures.
2.7. X-ray crystallographic procedures
Selected single crystals suitable for X-ray crystallographic analy-
sis were used for structural determination. The X-ray intensity data
were measured at 173(2) K (Bruker KRYOFLEX) on a Bruker SMART
APEX CCD-based X-ray diffractometer system equipped with a Mo-
target X-ray tube (k = 0.71073 Å) operated at 1800 W power. The
crystals were mounted on a goniometer head with silicone grease.
The detector was placed at a distance of 6.14 cm from the crystal.
For each experiment a total of 1850 frames were collected with a
3.2. Photochemistry of cis-(AupNBT)2dppee (1)
When a methylene chloride solution of cis-(AupNBT)2dppee (1)
was exposed to light, in air or under nitrogen, it isomerized to the
trans isomer. The isomerization occurred at wavelengths greater
than 230 nm and when a 420 nm cutoff filter was used. The reaction
was followed by 1H NMR, 31P {1H} NMR, and UV–Visible spectros-
copy. The UV–Visible spectrum in methylene chloride, Fig. 2, shows
the isomerization process as the cis isomer (c = 4.27 ꢁ 10ꢂ5 M), with
scan width of 0.3° in
x and an exposure time of 20 s/frame. The
frames were integrated with the Bruker SAINT Software package
using a narrow-frame integration algorithm to a maximum 2h angle
of 56.54° (0.75 Å resolution). The final cell constants are based upon
the refinement of the XYZ-centroids of several thousand reflections
above 20 r(I). Analysis of the data showed negligible decay during
data collection. Data were corrected for absorption effects using
the empirical method (SADABS). The structures were solved and re-
fined by full-matrix least-squares procedures on |F2| using the Bru-
ker SHELXTL (version 6.12) software package. The coordinates of metal
atoms for the structures were found in direct method E maps. The
remaining atoms were located after an alternative series of least
squares cycles and difference Fourier maps. For alkenes in 1, 2 and
3, hydrogen atoms were located and refined independently while
the remaining hydrogen atoms were included in idealized positions
for structure factor calculations. Anisotropic displacement parame-
ters were assigned to all non-hydrogen atoms, except solvent mol-
ecules in trans-(AuCl)2dppee. Relevant crystallographic data are
summarized in Table 1.
kmax = 400 nm (
e
= 3.25 ꢁ 104 Mꢂ1 cmꢂ1) isomerized to the trans
(kmax = 385 nm;
e
= 3.3 ꢁ 104 Mꢂ1 cmꢂ1
) in a quartz cuvette.
Although the absorptivity of the cis and trans isomers is about the
same, the peak decreases quickly even as the isomerization pro-
ceeds because the trans-(AupNBT)2dppee undergoes a further pho-
toreaction (vide infra). Based on a weighted average calculation, at
40 s there is 28% of the cis-(AupNBT)2dppee remaining while 72%
of the initial 4.27 ꢁ 10ꢂ5 M solution isomerized to the trans isomer.
However, at 40 s, some of the trans compound has reacted further,
evidenced by the decrease in the the absorbance at 385 nm. The
isomerization suggests that the excited state has p⁄ character, sim-
ilar to the other cis complexes in this series, [X = Cl, Br, I] [11]. In the
31P {1H} NMR spectrum the initial peak at 19.5 ppm diminishes as a
broad peak around 35.0 ppm grows in, indicating formation of
trans-(AupNBT)2dppee. As the reaction proceeds several intermedi-
ate peaks are observed at 33, 32, and 31 ppm perhaps representing
intermediates or mixtures generated in the photochemical reaction
of trans-(AupNBT)2dppee. This supports our observation from the
UV–Vis experiment that the trans-(AupNBT)2dppee reacts as soon
as it is produced from the isomerization.
3. Results and discussion
3.1. Synthesis, crystal structure and absorption spectrum of cis-
(AupNBT)2dppee (1)
Synthesis was by chloride substitution of cis-(AuCl)2dppee with
deprotonated p-nitrobenzenethiol in methylene chloride to form
the dinuclear cis-(AupNBT)2dppee, an orange, air stable solid that
luminesced red–orange under UV light and was soluble in most
organic solvents. The product was characterized by elemental anal-
ysis, X-ray crystal analysis, UV–Vis, 1H NMR, and 31P {1H} NMR spec-
troscopy. The synthesis highlights the unique nature of the
p-nitrobenzenethiol because attempts to synthesize similar cis
complexes in our lab using other benzenethiol ligands, p-thiocresol,
3.3. Synthesis, crystal structure and absorbance spectrum of trans-
(AupNBT)2dppee (2)
The trans-(AupNBT)2dppee complex was synthesized by ligand
substitution of the deprotonated p-nitrobenzenethiol ligand with
trans-(AuCl)2dppee in methylene chloride. The product was a yel-
low air-stable solid that luminesced bright orange under UV light.

J.B. Foley et al. / Inorganica Chimica Acta 392 (2012) 300–310
303
Table 1
Crystallographic data and structure refinement parameters for cis-(AupNBT)2dppeeꢀ2CH2Cl2 (1ꢀ2CH2Cl2), trans-(AupNBT)2dppeeꢀCH2Cl2 (2ꢀCH2Cl2), and trans-(AuCl)2dppee (3).
1ꢀ2CH2Cl2
40H34Au2Cl4N2O4P2S2
1268.48
monoclinic
C 2/c
2ꢀCH2Cl2
C39H32Au2Cl2N2O4P2S2
1183.56
3
Formula
C
C26H22Au2Cl2P2
861.21
monoclinic
C 2/c
Formula weight
Crystal system
Space group
triclinic
ꢀ
P1
Unit cell dimensions
a (Å)
b (Å)
c (Å)
18.3627(10)
14.1750(8)
16.5005(9)
90.00
102.5050(10)
90.00
4193.0(4)
4
2.009
12.3682(10)
13.1123(10)
14.6372(12)
66.7240(10)
68.7770(10)
80.1530(10)
2031.6(3)
2
16.5106(6)
15.0625(5)
11.3460(4)
90.00
117.0630(10)
90.00
2512.69(15)
4
2.277
a
(°)
b (°)
c
(°)
V (Å3)
Z
q
l
(g cmꢂ3
(mmꢂ1
)
)
1.935
7.568
7.464
12.017
k (Å)
Mo K
a
(0.71073)
Mo K
a
(0.71073)
MoKa (0.71073)
Transmission factors
T (K)
0.4006–0.5866
173(2)
0.2994–0.7517
173(2)
0.0622–0.4465
173(2)
Data/restraints/parameter
4556/0/251
0.0220, 0.0565
0.0250, 0.0578
1.054
8674/0/455
0.0420, 0.1091
0.0523, 0.1155
1.059
2749/11/178
0.0168, 0.0439
0.0174, 0.0443
1.065
b
R1a, wR2 I > 2
r(I)
All data
Quality-of-fitc
a
b
c
R1 = ||Fo| ꢂ |Fc||/
RFo.
[w(Fo ꢂ Fc2)2]/
R .
[w(Fo2)2]]
2
½
wR2 = [
R
[w(Fo ꢂ Fc2)2]/(Nobs ꢂ Nparams)] , based on all data.
2
½
Quality-of-fit = [
R
Table 2
Selected bond lengths and bond angles for (1)cis and (2)trans and (3)trans-(AuCl)2dppee compared to other gold(I) phosphine complexes.
Complex
Au–Au (Å)
3.0036(3)
P–Au (Å)
2.2717(8)
X–Au (Å)
P–Au–X (°)
Au2–Au1–P1 (°)
Ref.
cis-(AupNBT)2dppee (1)
2.3120(8)
X = S
174.10(3)
82.10(2)
This work
trans-(AupNBT)2dppee (2)
trans-(AuSPh)2dppee
3.1201(4)
intermolecular
2.2665(18)
2.2597(17)
2.308(2)
X = S
171.25(7)
171.41(7)
X = S
98.12(5)
94.92(5)
This work
a
b
c
3.023(2)
intermolecular
3.189(1)
intermolecular
3.05(1)
2.274(9)
2.24(1)
2.242(6)
2.30(1)
X = S
2.314(5)
X = Cl
2.288(7)
X = Cl
2.291(2)
X = Cl
2.5629(9)
X = I
2.5435(8)
X = I
2.2928(7)
X = Cl
174.3(3)
(AuCl)2dppe
dppe = bis(diphenyl)phosphino ethane
cis-(AuCl)2dppee
175.4(2)
175.4(2)
173.5(1)
170.36(6)
174.81(6)
173.92(3)
108.0(2)
2.238(5)
2.235(2)
2.253(2)
2.250(2)
2.2412(7)
d
e
e
trans-(AuCl)2dppee
cis-(AuI)2dppee
3.043(1)
intermolecular
2.9526(5)
94.21(6)
trans-(AuI)2dppee
3.2292(7)
intermolecular
3.0203(2)
101.42(6)
101.823(18)
trans-(AuCl)2dppee (3)
Photoproduct
This work
intermolecular
a
b
c
Onaka et al. [35].
Bates and Waters [45].
Jones [28].
Eggleston et al. [29].
Foley et al. [12].
d
e
The product was verified by elemental analysis, X-ray crystallogra-
phy, 31P {1H} NMR and 1H NMR and UV–Vis spectroscopy. The
main feature of the crystal structure is the dimer unit with a short
between two molecules of trans-(AuCl)2dppee in the crystal. In
the unit cell of trans-(AuI)2dppee [12] there are two dimers, each
with an intermolecular gold–gold distance of 3.2292 Å. A crystal
structure of an analogous thiophenol, trans-(AuSPh)2dppee [35],
shows an infinite chain forming a two-dimensional network held
(3.1201 Å) intermolecular gold–gold distance and some
p–p
stabilization from the interaction of two phenyl rings from
the diphenylphosphine backbone (Fig. 3). There are no other close
gold–gold contacts in the unit cell which includes a methylene
chloride molecule. The p-nitrobenzenethiol ligands are nearly
perpendicular to each other in the dimer. The P(1)–Au(1)–S(2)
and the P(2)–Au(2)–S(1) bond angles (171.27(6) and 171.42(6),
respectively) are distorted from linear, presumably to enhance
the gold–gold interaction. Dimer structures in other trans
complexes with the dppee backbone are not unusual. Eggleston
et al. [29] report an intermolecular gold–gold distance of 3.043 Å
together by intermolecular gold–gold interactions and
p–p
interactions of the SPh ligands. As in our experience, Onaka et al.
[36] could not synthesize the cis-(AuSPh)2dppee by ligand
substitution.
The absorption spectrum is dominated by a broad low energy
peak with
a
maximum wavelength at 385 nm
(
e
= 3.3 ꢁ 104
Mꢂ1 cmꢂ1). As in the cis isomer, the trans-(AupNBT)2dppee has a
high energy absorbance at 236 nm (
common for complexes with phenylphosphine ligands.
e
= 5.1 ꢁ 104 Mꢂ1 cmꢂ1) that is

304
J.B. Foley et al. / Inorganica Chimica Acta 392 (2012) 300–310
When trans-(AupNBT)2dppee was photolyzed in a variety of or-
ganic solvents (ACN, DMF, acetone, THF) a photochemical reaction
occurred evidenced by the decrease in the low energy absorbance
(kmax ꢃ 385 nm) in the UV–Vis spectrum. The peak maximum red
shifted slightly with more polar solvents, a phenomenon also ob-
served by others for gold complexes with the pNBT ligand [32a].
Only when trans-(AupNBT)2dppee was photolyzed in a chlorinated
solvent (CH2Cl2 or CHCl3) were we able to separate and identify
some of the products. We used CH2Cl2 for better solubility of the
complex and easier analysis in the 1H NMR. Based on UV–Vis data,
1H NMR, and 31P {1H} NMR, and X-ray analysis, the products of the
photolysis of trans-(AupNBT)2dppee were identified as trans-
(AuCl)2dppee and various radical thiol products including the
4,40-dinitrophenyldisulfide. Eq. (1) summarizes these results.
h
m
trans-ðAupNBTÞ2dppee
!
trans-ðAuClÞ2dppeeð3Þ
CH2Cl2
þ ðNO2C6H4SÞ2 þ other thiol products and alkylchlorides
ð1Þ
Photolysis of trans-(AupNBT)2dppee was carried out using two
wavelength ranges, (a) in quartz (k > 230 nm), and (b) the
420 nm filter, in order to define conditions necessary for the for-
mation of the trans-(AuCl)2dppee product. Fig. 4 illustrates two
typical experiments: the 385 nm peak of the original trans-
(AupNBT)2dppee complex decreases as a peak at 320 nm grows
in, with an isosbestic point at 335 nm.
Fig. 1. Molecular structure of cis-(AupNBT)2dppee (1). Atoms are represented by
thermal ellipsoids at the 40% probability level. See Table 2 for selected distances (Å)
and angles (°).
3.5. Characterization of the products of the photolysis of trans-
(AupNBT)2dppee
385
400
The product of the photolysis of trans-(AupNBT)2dppee was
collected, recrystallized and separated into a crystal fraction and
a solvent fraction. The 1H NMR of the crystal fraction, (see Supple-
mentary data, Fig. S-1a) identified as trans-(AuCl)2dppee (3),
clearly shows peaks for the dppee phenyl hydrogens and a distinct
triplet ((CD2Cl2, 300 MHz) 7.112 ppm; J = 20.0 Hz) related to the
hydrogens on the ethylene backbone that is typical for these trans
complexes [11]. The 31P{1H}NMR peak is at 29.6 ppm. Both spectra
are identical to an independently synthesized sample of trans-
(AuCl)2dppee and the elemental analysis was consistent with the
formation of trans-(AuCl)2dppee. However, it is interesting to note
that the crystals we obtained appeared as light-yellow rectangles
while the original crystals of trans-(AuCl)2dppee [29] were re-
ported as white needles representing a dimer with intermolecular
gold–gold distances of 3.043 Å. The crystal structure of our trans-
(AuCl)2dppee photoproduct (3) (Fig. 5) sheds some light on this
discrepancy by revealing an infinite chain of trans-(AuCl)2dppee
molecules with gold–gold distances of 3.0203 Å. It is not unusual
for gold complexes to form polymorphs when crystallizing [37]
and the extended chain of gold–gold interactions could result in
a red shift in absorbance causing a slight change in color. In solu-
tion the sample’s 1H NMR spectrum was identical to an authentic
sample of trans-(AuCl)2dppee. The yield of recrystallized trans-
(AuCl)2dppee photoproduct varied from 40% to 60%. There was
some evidence of unidentified insoluble bits in the product. The
crystal structure is similar to that of Eggleston et al. with compar-
ative Au–Cl bond lengths and P–Au–Cl bond angles. There are
slightly different Au–P bond lengths: 2.2413(7) Å for our photo-
product compared to 2.235(2) Å for Eggleston’s crystal structure.
The recrystallized thiyl fraction was examined by 1H NMR but
the interpretation is less definitive than the trans-(AuCl)2dppee
samples (See Supplementary data, Fig. S-1b). There are a series of
doublets, two of which can be identified as the 4,40-dini-
trophenyldisulfide, one at 8.187 ppm (J = 9.0 Hz) and the other at
7.655 ppm (J = 9.0 Hz), identical to an authentic sample of the
disulfide. There are other doublets, one at 8.265 (J = 9.0 Hz),
Wavelength (nm)
Fig. 2. Photolysis of cis-(AupNBT)2dppee (1) in CH2Cl2: t = 0–40 s; k > 230 nm; kmax
cis-(AupNBT)2dppee = 400 nm; kmax trans-(AupNBT)2dppee = 385 nm. Light inten-
sity was reduced by using a mask on the sample holder to slow the reaction and
allow more detail to be observed. Trans-(AupNBT)2dppee (Dashed line) overlay for
comparison only.
3.4. Photochemistry of trans-(AupNBT)2dppee (2)
While investigating the photochemical reaction of cis-
(AupNBT)2dppee, we noted that the very rapid isomerization reac-
tion was followed by a photoreaction of trans-(AupNBT)2dppee.
Other trans complexes in this series, trans-(AuX)2dppee [X = Cl,
Br, I, p-thiocresol, SCN]4 did not exhibit further photochemical reac-
tion in the time frame of the isomerization experiments (several
minutes). However, after prolonged photolysis (18 min in a quartz
cuvette) we did observe evidence (reduced gold) that the trans-
(AuCl)2dppee may have disproportionated.
4
See Ref. [11]; trans-(Auptc)2dppee and trans-(AuSCN)2dppee photolysis are from
unpublished results in our lab.

J.B. Foley et al. / Inorganica Chimica Acta 392 (2012) 300–310
305
Fig. 3. Molecular structure of trans-(AupNBT)2dppee (2). Atoms are represented by thermal ellipsoids at the 40% probability level. See Table 2 for selected distances (Å) and
angles (°).
a 1.6
1.4
1.2
1
b 1.8
1.6
1.4
1.2
1
385
385
320
0.8
0.6
0.4
0.2
0
0.8
0.6
0.4
0.2
0
320
225 275 325 375 425 475 525
225 275 325 375 425 475 525
Wavelength (nm)
Wavelength (nm)
Fig. 4. (a) Photolysis of trans-(AupNBT)2dppee (2) in CH2Cl2: t = 0–20 s; k > 230 nm. Isosbestic point at 335 nm. (b) Photolysis of trans-(AupNBT)2dppee (2) in CH2Cl2: t = 0–
13 min; 420 nm filter. Isosbestic point at 335 nm.
7.600 (J = 9.0 Hz) and 7.496 (J = 9.0 Hz), that correspond to the
hydrogens on the p-nitrobenzenethiol and indicate the formation
of other thiyl compounds. Based on integration, the ratio of forma-
tion of the disulfide to all other products is approximately 1:1.5.
There are also two singlets, one at 4.1 ppm and another at
3.8 ppm, that represent radical recombination of the thiyl and sol-
vent radical fragments. Based on comparison with an authentic
sample, the peak at 3.8 ppm is identified as C2H4Cl2, from the rad-
ical combination of two CH2Cl fragments. The peak at 4.1 ppm
likely represents the hydrogens of the CH2Cl fragment from a
NO2C6H4SCH2Cl radical product formed by combination of the thiyl
radical and radicals from the solvent. The thiyl fraction mixture
was difficult to separate further. Some of these peaks may also rep-
resent breakdown products of the 4,40-dinitrophenyldisulfide as it
undergoes a photochemical reaction of its own.
photolysis before separation containing trans-(AuCl)2dppee and
4,40-dinitrophenyldisulfide. The trans-(AuCl)2dppee (c dash-dot
line) has no absorbance at 320 nm. Trace d (dashed line) illustrates
the spectrum of an authentic sample of 4,40-dinitrophenyldisulfide
for comparison. The isosbestic point implies that one species is
being directly transformed into another with no intermediate.
We suggest that the isosbestic point here indicates the substitution
of a thiyl by a chlorine radical and the formation of the 4,40-dini-
trophenyldisulfide and other thiyl species that absorb at 320 nm.
The trans-(AupNBT)2dppee is a dinuclear molecule and it is unli-
kely that both p-nitrobenzenethiol ligands would be replaced
simultaneously. We considered the possibility of an intermediate
such as trans-(AuCl)(AupNBT)dppee whose UV–Vis spectrum is
similar to the spectrum of trans-(AupNBT)2dppee differing only in
the intensity of the 385 nm peak because of only one p-nitroben-
zene thiol ligand. The existence of transient peaks in the 31P {1H}
NMR during the photolysis reaction make this speculation
reasonable. Further work is being done to identify intermediates
and propose a more complete mechanism for this photochemical
reaction.
Fig. 6 shows individual UV–Vis spectra that are overlaid to
support the assignment of the UV–Vis peaks. The initial spectrum
of the trans-(AupNBT)2dppee (a solid line) shows a maximum
absorbance at 385 nm which decreases as the experiment
progresses. Trace b (dotted line) indicates the product of the

306
J.B. Foley et al. / Inorganica Chimica Acta 392 (2012) 300–310
UV–Visible spectroscopy. When the solvent (CH2Cl2) was photo-
lyzed with UV light for 2 min and the compound added in the dark,
the UV spectrum just showed the peak at 320 nm and no peak at
385 indicating that the reaction went to completion as soon as
the compound was added even though the trans-(AupNBT)2dppee
itself was not photochemically stimulated. This result is consistent
with a mechanism in which the substitution is initiated by a chlo-
rine radical from the solvent that displaces a thiyl radical from the
complex. However, when the solvent was irradiated for 10 s or 30 s
at wavelengths greater than 230 nm, the dark reaction occurred
but did not go to completion, indicating that radical formation
was not a catalytic chain reaction or that the rate of termination
of the radicals was greater than rate of interaction with the sub-
strate (see Supplementary data, Fig. S-2).
Under the same conditions with the 420 nm cutoff filter, there
was no reaction in the dark when trans-(AupNBT)2dppee was
added after photolysis of the solvent for 2 min. When the solution
of methylene chloride and trans-(AupNBT)2dppee was exposed to
the light with the 420 nm cutoff filter, the 385 nm peak diminished
and the 320 nm peak grew in, indicating that the complex was nec-
essary for the reaction to take place. Control experiments (samples
in CH2Cl2 kept in the dark for two days at room temperature) show
that trans-(AupNBT)2dppee does not react with any chlorinated
solvent without photolysis. We concluded from these experiments
that with the 420 nm cutoff filter the excited state of the complex
initiated the reaction but at wavelengths less than 265 nm, the
reaction was solvent initiated. However since the gold complex
also absorbs at 236 nm, both pathways could be available at the
shorter wavelengths.
Fig. 5. Molecular structure of the photoproduct, (trans-(AuCl)2dppee) (3) isolated
from the photolysis of trans-(AupNBT)2dppee (k > 230 nm). See Table 2 for selected
distances (Å) and angles (°).
1.6
1.4
1.2
3.6.2. Photolysis in air and under nitrogen
1
The trans-(AupNBT)2dppee photoreaction as described in Eq. (1)
and monitored by UV–Vis spectroscopy, occurs when the solution
was photolyzed in air or under an inert atmosphere, nitrogen. The
rates of the photochemical reaction in air or under nitrogen are
similar implying that oxygen does not quench the reaction nor
does it accelerate it (see Supplementary data, Fig. S-3). Often oxy-
gen has a quenching effect on metal based photochemical reac-
tions [38] but we do not observe this here.
a.
0.8
d.
b.
0.6
0.4
0.2
c.
3.6.3. Radical trap experiments
0
225
DPPH (a,a-diphenyl-b-picrylhydrazyl) was used to detect if
275
325
375
425
475
525
radicals were formed during the reaction. When a molar excess
of DPPH was added to trans-(AupNBT)2dppee in methylene chlo-
ride and the sample was photolyzed in quartz, there was little
change in the rate of photolysis, 2 ꢁ 10ꢂ6 mLꢂ1 sꢂ1, measured by
the change in absorbance at 385 nm representing the decrease in
concentration of trans-(AupNBT)2dppee (see Supplementary data,
Fig. S-4). If the initiating step in the photolysis is the homolytic
cleavage of the C–Cl bond in the solvent and the solvent is in great
excess compared to the complex, it is not surprising that there is
no measurable change in the rate of the reaction. However there
was a clear decrease in the DPPH absorbance at 550 nm, indicating
that the DPPH was indeed trapping radicals.
When DPPH was added to a sample of trans-(AupNBT)2dppee in
methylene chloride that was photolyzed with the 420 nm cutoff
filter, the concentration of trans-(AupNBT)2dppee was reduced to
4.99e-5M (a 17.5% reduction) after 12 min whereas in the control
sample (without DPPH) the concentration was 7.05e-6M, (a
87.5% reduction) clearly indicating that DPPH was slowing the
photoreaction. The decrease in absorbance at 550 nm, where DPPH
absorbs, was clear as well, indicating that DPPH was reacting with
radicals.
Wavelength (nm)
Fig. 6. Overlay of initial compound and products of photolysis reaction of trans-
(AupNBT)2dppee in CH2Cl2 (a) initial trans-(AupNBT)2dppee (solid line), c = 2.95 e-
5 M (b) final photoproduct, trans-(AuCl)2dppee plus 4,40-dinitrophenyl disulfide
(dotted line), (c) separated trans-(AuCl)2dppee product (dash-dot), and (d) authen-
tic sample of 4,40-dinitrophenyldisulfide (dashed line) for reference.
What is unusual here is that the same products are identified for
the high energy photolysis (k > 230 nm) and the low energy process
(k > 420 nm), despite the fact that methylene chloride does not ab-
sorb at wavelengths greater than 265 nm. We conducted several
experiments to identify the photoactive species and to determine
if radicals were involved as described in the next section.
3.6. Supporting experiments
3.6.1. Solvent experiments
To examine the possibility of chlorine radical formation directly
from the solvent, we photolyzed the methylene chloride (2.5 ml)
solvent first using a quartz cuvette and then with the 420 nm cut-
off filter. The trans-(AupNBT)2dppee (c = 5.4 ꢁ 10ꢂ5 M) was added
in the dark and the progress of the reaction was monitored by
The solvent experiments support our hypothesis that at high
energy chlorine radicals from direct photolysis of the solvent
initiate the reaction. Other researchers have observed this solvent

J.B. Foley et al. / Inorganica Chimica Acta 392 (2012) 300–310
307
initiated photolysis in addition reactions to metals [39]. Examples
also show that metal complexes, excited at wavelengths where sol-
vents do not absorb, can abstract chlorine radicals from the solvent
[40]. Pignolet’s work [41] with photolysis of metal dithiocarba-
mates explains the substitution of Cl for a dithiocarbamate by a
LMCT reduction of the metal and the reduced complex abstracting
a Cl radical from the solvent. It is interesting to note that the
dithiocarbamate disulfide was identified as a minor product of this
photolysis. Van der Graaf [42] discusses the abstraction of a chlo-
rine from the solvent by a rhenium radical formed from excitation
of a bimetallic complex at wavelengths where the solvent does not
absorb. Giuffrida et al. [43] show a solvent initiated process for
Sr(hfac)2TG (hfac = 1,1,1,5,5,5,hexafluoro-2,4-pentanedionato and
TG = tetraglyme = 2,5,8,11,14-pentaoxapentadecane) at high ener-
gies, but a solvent-assisted reaction involving electron transfer
from an excited triplet state of the complex to the solvent at lower
energies: both processes giving the same product, SrCl2, by differ-
ent mechanisms.
3.7. Summary of trans-(AupNBT)2dppee photoprocess
Formation of the trans-(AuCl)2dppee is not that interesting in it-
self, in fact; it is the starting material for all our syntheses and it is
easily prepared by reduction of HAuCl4 and addition of trans-
bis(diphenylphosphino)ethylene. The interesting point is that we
are producing the same product by two different pathways depen-
dent on the wavelength of excitation. A Au–S bond is stronger than
a Au–Cl bond [44], and the p-nitrobenzenthiol is a much ‘‘softer’’
ligand and is expected to have a better orbital overlap with the
gold than the ‘‘harder’’ chloride ligand. Thermodynamically, we
are driving the reaction ‘‘uphill’’ by absorption of energy, even with
visible wavelengths. If we look at a rough bond breaking/bond
making scenario, considering that cleavage of one C–Cl bond in
the solvent will initiate other radicals, we look at putting energy
into break one C–Cl bond (307.8 kJ mꢂ1) plus two Au–S bonds
(836 kJ mꢂ1) and getting energy from formation of two Au–Cl
bonds (686 kJ mꢂ1) and one disulfide bond (230 kJ mꢂ1) giving an
Fig. 7. Frontier orbitals, HOMO(a) and LUMO(b), of the model complex (C14H14Au2-
N2O4P2S2) of cis-(AupNBT)2dppee based on the TDDFT calculation using LANL2DZ
basis set, 6-311(d,p) basis set for non metal atoms, and the B3LYP correlation
function.
overall
D
H of +227.8 kJ mꢂ1 which is compensated even by ener-
gies of k > 420 nm (E = 285 kJ mꢂ1).
3.8. Computational results
gold atoms and the obvious necessity of weakening the gold–gold
attraction for the isomerization.
DFT and TDDFT calculations performed on the model of com-
plexes, cis and trans-(AupNBT)2dppee C14H14Au2N2O4P2S2, in GAUSS-
IAN 09 were used to explain the observed photochemical reactions
and the emission spectra of the complexes. The lowest energy
geometry of the cis isomer model complex used coordinates given
by the crystal structure for cis-(AupNBT)2dppee, including the
3.00 Å gold–gold distance. The HOMO and the HOMO ꢂ 1 of cis-
(AupNBT)2dppee are clearly on the thiol ligand, mostly S pz with
some electron density on the gold atoms (see Fig. 7). The LUMO is
primarily ethylene p⁄ with some bonding and antibonding interac-
tions on the gold atoms. The lowest energy singlet excitation for cis-
(AupNBT)2dppee is at 542.71 nm and is a HOMO to LUMO transition.
There are several low energy triplet states (at 556.56, 540.76 nm)
that are also HOMO (or HOMO ꢂ 1) to LUMO that could initiate
the isomerization. Because of heavy metal spin orbit coupling, it is
likely that a triplet state is involved in the photoprocess, so popula-
tion of higher energy states available with the excitation in quartz
could readily undergo intersystem crossing to a lower energy triplet
with ethylene p⁄ character. Higher excited states, the LUMO + 1 or
LUMO + 2 do not have any ethylene p⁄ character so it is unlikely that
these states would initiate the isomerization. These data suggest
that the lowest energy transition in the cis-(AupNBT)2dppee is a li-
gand to ligand transition (LLCT) affected by a ligand to metal–metal
(LMMCT) transition because of the electron density between the
It is interesting to note that results of a different DFT calculation
on the model cis-(AupNBT)2dppee structure that did not confine
the gold–gold distance to 3.0 Å had a slightly higher energy and
a gold–gold distance of 3.379 Å, close enough for weak gold–gold
interactions. Both calculations indicated a similar orbital composi-
tion but the initial excited states (1–5) of the 3.379 Å structure are
markedly blue shifted relative to the structure with gold–gold dis-
tance of 3.0 Å (see Supplementary data; S-7). Gold–gold interac-
tions are weak, about the same energy as hydrogen bonds, and
although the distance may vary in solution compared to the crys-
tal, the closer gold–gold distance not only facilitates the isomeriza-
tion but appears to allow absorption and emission at lower
energies.
DFT and TDDFT calculations were done on the model complex
of trans-(AupNBT)2dppee using the IEP-PCM for the solvent meth-
ylene chloride to investigate the basis for our experimental results
(see Supplementary data; S-8). The HOMO and the HOMO ꢂ 1 are
close in energy alternating electron density on the two p-nitroben-
zenethiol ligands of the dinuclear complex, primarily on the sulfur
pz but with some minor electron density on the gold atoms (see
Fig. 8). The LUMO and the LUMO + 1 of the complex are pzp⁄ on
the NO2 and around the phenyl ring of the p-nitrobenzenethiol.
This suggests that the lowest energy excited state is an intraligand
transition where the sulfur donates electron density to the NO2 of

308
J.B. Foley et al. / Inorganica Chimica Acta 392 (2012) 300–310
than our experimental value of 3.3 ꢁ 10ꢂ4 M cmꢂ1. This agreement
suggests that the calculation is an adequate model for our system.
3.9. Photophysical properties of the cis and trans-(AupNBT)2dppee
isomers
Although there are similarities between the two isomers (see
Table 3 for specific data), such as absorbance in the visible and
the red shift in the emission going from solution to solid state,
the geometry and the nature of the excited states are clearly differ-
ent. Many of the similar features can be attributed to the p-nitro-
benzenethiol ligand that dominates the HOMO of both of these
complexes. The differences in reactivity are likely a result of the
cis geometry facilitating the gold–gold interaction, making the
LUMO ethylene p⁄. In contrast, in the trans isomer intramolecular
interactions are not possible and the donor/acceptor relationship
of the p-nitrobenzenethiol lowers the energy of the LUMO enough
so the ethylene p⁄ orbital is higher energy. Studies are underway to
evaluate the reactivity of other thiol ligands with this system to
determine if the p-nitrobenzenethiol is unique in its photochemi-
cal reactivity.
Fig. 8. Frontier orbitals, HOMO(a) and LUMO(b), of the model complex (C14H14Au2-
N2O4P2S2) of trans-(AupNBT)2dppee based on the TDDFT calculation using LANL2DZ
basis set, 6-311(d,p) basis set for non metal atoms, and the B3LYP correlation
function.
3.9.1. Emission
When excited at 385 nm, trans-(AupNBT)2dppee emits at
505 nm at room temperature in methylene chloride and at
600 nm in the solid state (with KBr) (see Fig. 9). Based on the
TDDFT calculation, the emission in solution for trans-(AupNBT)2dp-
pee is likely from a triplet state that absorbs at 496.99 nm and is a
LUMO or the LUMO + 1 to the HOMO or HOMO ꢂ 1, all p-nitroben-
zenethiol transitions. Unlike other assignments for gold(I) thiol
complexes where the lowest energy transition is characterized as
an LMCT, this emission is dominated by a LLCT, an assignment that
has been observed before with this ligand [32b].
the same ligand. The three lowest energy triplet transitions in the
trans-(AupNBT)2dppee are all HOMO to LUMO (or H ꢂ 1 to L + 1)
which are also p-nitrobenzenethiol centered transitions. Popula-
tion of the LUMO would take electron density from the sulfur,
weakening the gold–sulfur bond, allowing substitution at the gold,
which is what we see experimentally. The transition with the high-
est oscillator strength (0.8069) is a singlet transition at 390.41 nm
and the excitation with the highest coefficient is HOMO to LUMO
(see Supplementary data; S-8). The reaction happens also at wave-
lengths greater than 420 nm and there are two triplet states at
496.99 and 496.50 nm that are also HOMO to LUMO, and could ini-
tiate the same substitution reaction. As in the cis isomer, spin–orbit
coupling makes occupation of triplet states likely. The calculations
support the assertion that the excited state involves an intraligand
Table 3
Photochemical and photophysical data for cis-(AupNBT)2dppee (1) and trans-
(AupNBT)2dppee (2).
Absorbance kabs(nm)
, M1 cmꢂ1
Emission
charge transfer (LLCT) from the pzp
of the sulfur to the pzp⁄ of the
(e
)
kem
(nm)
NO2. In the excited states, there is no electron density on the gold
until the LUMO + 2 or above, making the possibility of an LMCT less
probable.
1
2
400 nm (32500)
236 nm(43700)
385 nm (32900)
236 nm (51500)
510 nm (in CH2Cl2)
588 nm (solid state in KBr)
505 nm (in CH2Cl2)
Computations were done on the mixed ligand complex, trans-
(AuCl)(AupNBT)dppee in methylene chloride to explore the exis-
tence of this species as a possible intermediate. The first singlet
transition is HOMO to LUMO and both orbitals are very similar to
the frontier orbitals of the trans-(AupNBT)2dppee as expected be-
cause of the dominance of the p-nitrobenzenethiol ligand. There
is a singlet transition at 387.41 nm, with an oscillator strength of
0.4758 that is HOMO to LUMO. Again there are several triplet
states that populate the LUMO at lower energies that would ex-
plain the photolysis with the 420 nm filter.
600 nm(solid state in KBr)
800000
700000
600000
500000
400000
300000
200000
100000
0
3.8.1. Absorbance
The theoretical absorption spectra derived from the TDDFT cal-
culation of both the cis and trans-(AupNBT)2dppee reflect what we
see experimentally (see Supplementary data; S-5). The absorbance
spectrum of cis-(AupNBT)2dppee shows the most favored transi-
tion with an oscillator strength of 0.4260 is
a singlet at
393.70 nm compared to our experimental absorbance maximum
at 400 nm. The theoretical molar absorptivity of 3.0 ꢁ 104 is close
347
397
447
497
547
597
to our experimental value of 3.25 ꢁ 104 M cmꢂ1
. For trans-
Wavelength (nm)
(AupNBT)2dppee the transition with the highest oscillator strength
occurs at 390.41 nm, close to our experimental maximum at
385 nm. The molar absorptivity is slightly higher at 3.6 ꢁ 10ꢂ4
Fig. 9. Emission and excitation spectra of cis-(AupNBT)2dppee (dashed line:
emission at 510 nm), and trans-(AupNBT)2dppee (solid line: emission at 505 nm.).

J.B. Foley et al. / Inorganica Chimica Acta 392 (2012) 300–310
309
As another way of characterizing the excited state, we followed
References
the photolysis reaction by UV–Vis and emission spectroscopy as
the reaction was progressing. When trans-(AupNBT)2dppee was
photolyzed in quartz, the emission intensity of the 505 nm peak
initially increased, reaching a maximum at t = 45 s, then decreased
rapidly to a minimum at t = 60 s while a peak grew in at 437 nm
(see Supplementary data, Fig. S-6). By t = 60 s the 385 peak in the
UV–Vis had decreased to virtually zero and the peak at 320 had
grown in, indicating that the reaction was complete. The initial in-
crease in emission intensity is interesting, suggesting a different
emitting species from the initial trans-(AupNBT)2dppee, perhaps
an intermediate not apparent in the UV–Vis spectrum. This species
then changes quickly, forming the characterized products. This
experiment supports the hypothesis that the state responsible for
the low energy emission is the same state that initiates the photo-
chemical ligand substitution.
[1] C.F. Shaw, Chem. Rev. 99 (1999) 2589.
[2] (a) F. Caruso, C. Pettinari, F. Paduano, R. Villa, F. Marchetti, E. Monti, M. Rossi, J.
Med. Chem. 51 (2008) 1584;
(b) I. Ott, X. Qian, Y. Xu, D.H.W. Vlecken, I.J. Marques, D. Kubutat, J. Will, W.S.
Sheldrick, P. Jesse, A. Prokop, C.P. Bagowski, J. Med. Chem. 52 (2009) 763.
[3] (a) V. Gandin, A.P. Fernandes, M.P. Rigobello, B. Dani, F. Sorrentino, F. Tisato, M.
Björnstedt, A. Bindoli, A. Sturaro, R. Rella, Biochem. Pharmacol. 79 (2010) 90;
(b) K.P. Bhabak, G. Mugesh, Inorg. Chem. 48 (2009) 2449.
[4] (a) D.A. Harvey, W.F. Kean, C.J. Lock, D. Signal, Lancet (1983) 470;
(b) M.C. Grootveld, P.J. Sadler, J. Inorg. Biochem. 19 (1983) 51.
[5] (a) B.J. Henz, T. Hawa, M.R. Zachariah, Langmuir 24 (2008) 773;
(b) W.R. Browne, T. Kudernac, N. Katsonis, J. Areephong, J. Hjelm, B.L. Feringa, J.
Phys. Chem. 112 (2008) 1183.
[6] (a) M.J. Katz, T. Ramnial, H. Yu, D.B. Leznoff, J. Am. Chem. Soc. 130 (2008)
10662;
(b) J. Schneider, Y. Lee, J. Perez, W.W. Brennessel, C. Flaschenriem, R. Eisenberg,
Inorg. Chem. 47 (2008) 957.
[7] (a) C. Chen, Y. Lin, C. Wang, H. Tzeng, C. Wu, Y. Chen, C. Chen, L. Chen, Y. Wu, J.
Am. Chem. Soc. 128 (2006) 3709;
The complex, cis-(AupNBT)2dppee, has a 510 nm emission in
methylene chloride at room temperature when excited at 400 or
420 nm and a 588 nm emission in the solid state. The lowest en-
ergy emission from the TDDFT calculation of cis-(AupNBT)2dppee
with the 3.0 Å gold–gold distance was at 556.56 nm, in between
the solution and the solid state values. The discrepancy between
experimental and theoretical emission could be explained by this
structure being more similar to the solid state crystal with the
gold–gold distance of 3.00 Å rather than a more flexible gold–gold
distance that might exist in solution. The TDDFT results of the
more relaxed cis structure (gold–gold distance = 3.379 Å), has the
lowest triplet emission at 536.11 nm, closer to the experimental
value.
(b) J.D.E.T. Wilton-Ely, Dalton Trans. (2008) 25.
[8] (a) S. Sanz, L.A. Jones, F. Mohr, M. Laguna, Organometallics 26 (2007) 952;
(b) N. Meyer, E. Schuh, F. Mohr, Annu. Rep. Prog. Chem. 106 (2010) 255.
[9] (a) S. Cha, J. Kim, K. Kim, J. Lee, Chem. Mater. 19 (2007) 6297;
(b) W.J. Hunks, M.C. Jennings, R.J. Puddephatt, Inorg. Chem. 41 (2002) 4590;
(c) B. Tzeng, H. Yeh, Y. Wu, J. Kuo, G. Lee, S. Peng, Inorg. Chem. 45 (2006) 591.
[10] (a) V.W-W. Yam, C-L. Chan, C-K. Li, M-C. Wong, Coord. Chem. Rev. 216–217
(2001) 173;
(b) C-M. Che, S-W. Lai, Luminescence and photophysics of gold complexes, in:
F. Mohr (Ed.), Gold Chemistry. Applications and Future Directions in the Life
Sciences, Wiley-VCH, Weinheim, 2009.
[11] J.B. Foley, A.E. Bruce, M.R.M. Bruce, J. Am. Chem. Soc. 117 (1995) 9596.
[12] J.B. Foley, S.E. Gay, M.J. Vela, B.M. Foxman, A.E. Bruce, M.R.M. Bruce, Eur. J.
Inorg. Chem. 31 (2007) 4946.
[13] P. Schwerdtfeger, A.E. Bruce, M.R.M. Bruce, J. Am. Chem. Soc. 120 (1998) 6587.
[14] P. Pyykkö, J. Li, N. Runeberg, Chem. Phys. Lett. 218 (1,2) (1994) 133.
[15] (a) P.E. Hoggard, Coord. Chem. Rev. 159 (1997) 235;
(b) A. Vogler, H. Kunkely, Coord. Chem. Rev. 230 (2002) 243.
[16] N. Meyer, C.W. Lehmann, T.K.-M. Lee, J. Rust, V.W-W. Yam, F. Mohr,
Organometallics 28 (10) (2009) 2931;
4. Conclusion
(b) F. Mohr, S.H. Privér, S.K. Bhargava, M.A. Bennett, Coord. Chem. Rev. 250
(2006) 1851.
[17] (a) R. Usón, A. Laguna, M. Laguna, J. Jiménez, P.G. Jones, J. Chem. Soc., Dalton
Trans. (1991) 1361;
(b) H.H. Murray, J.P. Fackler Jr, Inorg. Chim. Acta 115 (1986) 207;
(c) H. Schmidbauer, P. Jandik, Inorg. Chim. Acta 74 (1983) 97.
[18] (a) R.G. Raptis, H.H. Murray, R.J. Staples, L.C. Porter, J.P. Fackler Jr., Inorg. Chem.
32 (1993) 5576;
We have synthesized and characterized a pair of cis and trans
dinuclear gold isomers and investigated their photochemical and
photophysical properties. The complex, cis-(AupNBT)2dppee, has
an intramolecular gold–gold distance of 3.0036(3) Å and isomeriz-
es to the trans isomer with light of wavelengths greater than
230 nm as well as k > 420 nm. The complex, trans-(AupNBT)2-
dppee, crystallizes as a dimer with an intermolecular gold–gold
distance of 3.1201(4) Å and undergoes a wavelength dependent
photochemical reaction in chlorinated solvents. This reaction
produces the same products, trans-(AuCl)2dppee, 4,40-dini-
trophenyldisulfide, and thiyl radical products, at k > 230 nm and
at k > 420 nm, but through different radical mechanisms. The
crystal of the trans-(AuCl)2dppee photoproduct is an infinite chain
polymer with intermolecular gold–gold distances of 3.0203(2) Å.
DFT and TDDFT calculations support the assignment of a LLCT
(Spz ? ethylene p⁄) modified by a LMMCT (Spz ? Au–Au) transi-
tion for the cis-(AupNBT)2dppee and an LLCT (Spz ? NO2 pzp⁄) for
the trans-(AupNBT)2dppee. The photochemical stability of
gold–thiol bonds has implications for the use of gold–thiol drugs
as well as the many surface applications that rely on gold–thiol
attachments.
J.D. Basil, H.H. Murray, J.P. Fackler Jr, J. Tocher, A.M. Mazany, B. Trzcinska-
Bancroft, H. Knachel, D. Budis, T.J. Delord, D.O. Marler, J. Am. Chem. Soc. 107
(1985) 6908.
[19] (a) R.H. Hill, R.J. Puddephatt, J. Am. Chem. Soc. 107 (1985) 1218;
(b) S. Sathiyabalan, P.E. Hoggard, Inorg. Chem. 34 (1995) 4562;
(c) T.H. Nguyen, P.J. Shannon, P.E. Hoggard, Inorg. Chim. Acta 291 (1999) 136.
[20] (a) E.M. Jaryszak, P.E. Hoggard, Inorg. Chim. Acta 282 (1998) 217;
(b) P.E. Hoggard, A.J. Bridgeman, H. Kunkely, A. Vogler, Inorg. Chim. Acta 357
(2004) 639.
[21] (a) C.C. Tong, M. Winkelman, A. Jain, S.P. Jensen, P.E. Hoggard, Inorg. Chim. Acta
226 (1994) 247;
(b) N.E. Leadbeater, C. Jones, Transition Met. Chem. 25 (2000) 99;
(c) G.L. Miessler, L.H. Pignolet, Inorg. Chem. 18 (1979) 210.
[22] (a) C.R. Bock, M.S. Wrighton, Inorg. Chem. 16 (1977) 1309;
(b) P.E. Hoggard, M. Gruber, A. Vogler, Inorg. Chim. Acta 346 (2003) 137.
[23] J.M. Stegge, S.M. Woessner, P.E. Hoggard, Inorg. Chim. Acta 250 (1996) 385.
[24] K.W. Lee, P.E. Hoggard, Inorg. Chem. 32 (1993) 1877.
[25] (a) R. Narayanaswamy, M.A. Young, E. Parkhurst, M. Ouellette, M.E. Kerr, D.M.
Ho, R.C. Elder, A.E. Bruce, M.R.M. Bruce, Inorg. Chem. 32 (1993) 2202;
(b) W.B. Jones, J. Yuan, R. Narayanaswamy, M.A. Young, R.C. Elder, A.E. Bruce,
M.R.M. Bruce, Inorg. Chem. 34 (1995) 1996;
(c) J.M. Forward, D. Bohmann, J.P. Fackler Jr., R.J. Staples, Inorg. Chem. 34
(1995) 6330.
Acknowledgements
[26] (a) P. Pyykkö, Chem. Rev. 97 (1997) 597;
(b) H. Schmidbauer, Chem. Soc. Rev. (1995) 391;
(c) D.J. Gorin, F.D. Toste, Nature 446 (2007) 395.
[27] C-K. Li, X-X. Lu, K.M-C. Wong, P.G. Chan, N. Zhu, W-W. Yam, Inorg. Chem. 43
(2004) 7421.
[28] P.G. Jones, Acta Cryst. B36 (1980) 2775.
[29] D.S. Eggleston, J.V. McArdle, G.E. Zuber, J. Chem. Soc., Dalton Trans. (1987) 677.
[30] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M.
Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery Jr., J.E. Peralta, F. Ogliaro,
M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J.
JBF thanks Vermont EPSCoR for partial funding of this project.
JBF is grateful to Dr. Alice E. Bruce for helpful discussions. EVD is
grateful to the National Science Foundation (CHE-1152441) for
financial support.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.ica.2012.03.030.

310
J.B. Foley et al. / Inorganica Chimica Acta 392 (2012) 300–310
Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. (c) D.R. Smyth, B.R. Vincent, E.R.T. Tienkink, Cryst. Growth Des. 1 (2) (2001)
Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C.
Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth,
P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, O. Farkas, J.B. Foresman,
J.V. Ortiz, J. Cioslowski, D.J. Fox, GAUSSIAN 09, Revision A.02, Gaussian, Inc.,
Wallingford CT, 2009.
113.
[38] D. Sandrini, M. Maestri, V. Balzani, L. Chassot, A. Von Zelewsky, J. Am. Chem.
Soc. 109 (1987) 7720.
[39] (a) S. Whang, T. Estrada, P.E. Hoggard, Photochem. Photobiol. 79 (2004) 356;
(b) A.L. Le, Patrick E. Hoggard, Photochem. Photobiol. 84 (2008) 86.
[40] (a) R.M. Laine, P.C. Ford, Inorg. Chem. 16 (1977) 388;
(b) Y.A. Mikheev, V.P. Pustoshnyi, D.Y. Toptygin, Russ. Chem. Bull. 25 (1976)
1962;
(c) C. Aprile, M. Boronat, B. Ferrer, A. Corma, H. Garcia, J. Am. Chem. Soc. 128
(2006) 8388.
[41] (a) G.L. Miessler, E. Zoebisch, L.H. Pignolet, Inorg. Chem. 17 (1978) 3636;
(b) G.L. Miessler, G. Stuk, K.W. Given, M.C. Palazzotto, L.H. Pignolet, Inorg.
Chem. 15 (1976) 1982.
[31] C. King, J.-C. Wang, M.N.I. Khan, J.P. Fackler Jr., Inorg. Chem. 28 (1989) 2145.
[32] (a) S.M. Bessey, M. Aghamoosa, G.S.P. Garusinghe, A. Chandrasoma, A.E. Bruce,
M.R.M. Bruce Inorg, Chim. Acta 363 (2010) 279;
(b) C-H. Li, S.F. Kui, I.H.T. Sham, S.S-Y. Chui, C-M. Che, Eur. J. Inorg. Chem.
(2008) 2421.
[33] R.S. Sengar, V.N. Nemykin, P. Basu, New J. Chem. 27 (2003) 1115.
[34] (a) H. Schmidbauer, A. Schier, Chem. Soc. Rev. 37 (2008) 1931;
(b) P. Pyykkö, Chem. Rev. 88 (1988) 563.
[42] T. van der Graaf, A. van Rooy, D.J. Stufkens, A. Oskam, Inorg. Chim. Acta 187
(1991) 133.
[43] S. Guiffrida, L.L. Costanzo, G.G. Condorelli, G. Ventimiglia, I.L. Fragalà, Inorg.
Chim. Acta 358 (2005) 1873.
[35] S. Onaka, Y. Katsukawa, M. Yamashita, Chem. Lett. (1998) 525.
[36] S. Onaka, Y. Katsukawa, M. Shiotsuka, O. Kanegawa, M. Yamashita, Inorg. Chim.
Acta 312 (2001) 100.
[37] (a) E.M. Gussenhoven, J.C. Fettinger, D.M. Pham, M.M. Malwitz, A.L. Balch, J.
Am. Chem. Soc. 127 (2005) 10838;
[44] Bond energies are from Lange’s Handbook of Chemistry, 15th ed., John Dean, p
4.45. Bond energies are for gas state dinuclear molecules and are meant as an
approximation.
(b) R.L. White-Morris, M.M. Olmstead, S. Attar, A.L. Balch, Inorg. Chem. 44
(2005) 5021;
[45] P.A. Bates, J.M. Waters, Inorg. Chim. Acta 98 (1985) 125.