Electron-transfer-coupled ligand dynamics in Cu(I/II)(TTCN)2 complexes in aqueous solution

Article abstract of DOI:10.1021/ja9627757

One-electron oxidation of copper(I) bis(1,4,7-trithiacyclononane), [Cu(I)(TTCN-κ³)(TTCN-κ¹)]⁺, 1, a coordination complex with a tetrahedral CuS₄ core, to [Cu(II)(TTCN-κ³)₂]²⁺, 2, with an octahedral CuS₆ core, has been studied by pulse radiolysis and electrochemistry in aqueous solution at various pH values. In addition to the geometry change about the metal ion in this oxidation, the nonchelating 1,4,7-trithiacyclononane (TTCN) ligand in 1 changes conformation on becoming chelated in 2. However, pulse radiolysis reveals that this process does not occur intramolecularly but affords a bimolecular reaction in which the oxidized copper incorporates an external TTCN. Evidence for this mechanism is drawn from corresponding experiments with a variety of related Cu(I) complexes in which the monodentate TTCN has been replaced by other sulfur-containing ligands and which have been structurally characterized by X-ray crystallography. From all these studies it is concluded that oxidation of 1 and all these other complexes of Cu(I) is accompanied by immediate loss of the monodentate ligand generating [Cu(II)(TTCN-κ³)(H₂O)₃]²⁺, 3. Transient 3 is characterized by an optical absorption with λ(max) = 370 nm and ε ~ 2000 M^-1 cm^-1 which depends on pH because this transient participates in three acid/base equilibria. Deprotonation of the three water ligands associated with Cu(II) results in increasingly blue-shifted absorptions. Undeprotonated transient 3 prevails at pH ≤ 6, and converts directly into the stable Cu(II) complex 2 via reaction with an unoxidized molecule of 1 or free TTCN. The corresponding bimolecular rate constants are 5.2 (± 0.5) x 10⁵ and 8.4 (± 1.0) x 10⁵ M^-1 s^-1, respectively. For the deprotonated forms of 3 this process is increasingly slowed down and at higher pH (≤ 9) the formation of 2 is completely prevented. The formation of transient 3 is also consistent with the pH dependence of the electrochemistry of 1. Under electrochemical conditions the conversion into 2 follows first-order kinetics due to a relatively high TTCN concentration available near the electrode surface after oxidation of 1. All the results require rapid ligand exchange in 1 and a particularly labile monodentate TTCN ligand. This has been corroborated by ¹H NMR spectroscopic studies on 1.

Full text of DOI:10.1021/ja9627757

2134
J. Am. Chem. Soc. 1997, 119, 2134-2145
Electron-Transfer-Coupled Ligand Dynamics in CuI/II(TTCN)2
Complexes in Aqueous Solution
Sanaullah,† Hartmut Hungerbu1hler,‡ Christian Scho1neich,§ Martha Morton,†
David G. Vander Velde,† George S. Wilson,*,† Klaus-Dieter Asmus,*,| and
Richard S. Glass*,
Contribution from the Departments of Chemistry and Pharmaceutical Chemistry, The UniVersity
of Kansas, Lawrence, Kansas 66045, Bereich Physikalische Chemie, Abteilung Strahlenchemie,
Hahn-Meitner Institut Berlin, Glienicker Strasse 100, 14109 Berlin, Germany, Radiation
Laboratory and Department of Chemistry and Biochemistry, UniVersity of Notre Dame,
Notre Dame, Indiana 46556, and the Department of Chemistry, The UniVersity of Arizona,
Tucson, Arizona 85721
ReceiVed August 8, 1996X
Abstract: One-electron oxidation of copper(I) bis(1,4,7-trithiacyclononane), [CuI(TTCN-κ3)(TTCN-κ1)]+, 1, a
coordination complex with a tetrahedral CuS4 core, to [CuII(TTCN-κ3)2]2+, 2, with an octahedral CuS6 core, has
been studied by pulse radiolysis and electrochemistry in aqueous solution at various pH values. In addition to the
geometry change about the metal ion in this oxidation, the nonchelating 1,4,7-trithiacyclononane (TTCN) ligand in
1 changes conformation on becoming chelated in 2. However, pulse radiolysis reveals that this process does not
occur intramolecularly but affords a bimolecular reaction in which the oxidized copper incorporates an external
TTCN. Evidence for this mechanism is drawn from corresponding experiments with a variety of related CuI complexes
in which the monodentate TTCN has been replaced by other sulfur-containing ligands and which have been structurally
characterized by X-ray crystallography. From all these studies it is concluded that oxidation of 1 and all these other
complexes of CuI is accompanied by immediate loss of the monodentate ligand generating [CuII(TTCN-κ3)(H2O)3]2+
,
3. Transient 3 is characterized by an optical absorption with λmax ) 370 nm and ꢀ ∼ 2000 M-1 cm-1 which depends
on pH because this transient participates in three acid/base equilibria. Deprotonation of the three water ligands
associated with Cu(II) results in increasingly blue-shifted absorptions. Undeprotonated transient 3 prevails at pH
e6, and converts directly into the stable CuII complex 2 via reaction with an unoxidized molecule of 1 or free
TTCN. The corresponding bimolecular rate constants are 5.2 ((0.5) × 105 and 8.4 ((1.0) × 105 M-1 s-1, respectively.
For the deprotonated forms of 3 this process is increasingly slowed down and at higher pH (g9) the formation of
2 is completely prevented. The formation of transient 3 is also consistent with the pH dependence of the
electrochemistry of 1. Under electrochemical conditions the conversion into 2 follows first-order kinetics due to a
relatively high TTCN concentration available near the electrode surface after oxidation of 1. All the results require
1
rapid ligand exchange in 1 and a particularly labile monodentate TTCN ligand. This has been corroborated by H
NMR spectroscopic studies on 1.
Introduction
The redox chemistry of [CuI/II(TTCN)2] presents a fascinating
system for detailed studies because the change in oxidation state
of the copper ion is accompanied by substantial geometric
changes in the coordination sphere of the metal ion as well as
one of the ligands. The conformation and coordination versatil-
ity of the 1,4,7-trithiacyclononane (TTCN) ligand shown in
Figure 1 has resulted in a large variety of transition metal
complexes with novel structural, spectroscopic, and redox
properties.6,7 While the TTCN-κ3 ligand in [CuI(TTCN-κ3)-
(TTCN-κ1]+ (1; κn ) n-dentate ligand),8 shown in Figure 2,
adopts the endodentate [333] conformation, found in the solid
state for the free ligand,9 the nonchelating TTCN-κ1 has a unique
conformation different from both the [333] and [12222]10
The redox chemistry of CuI/II is intriguing because there is a
change in coordination geometry associated with the change in
oxidation state: d10 CuI is preferentially tetrahedral and d9 CuII
typically adopts tetragonal or square-planar geometry.1 The role
of ligated protein and copper ion coordination geometries is
important in copper-containing redox proteins.2,3 For example,
the key to the remarkable properties of “blue” copper proteins4,5
may be the distorted geometries of the metal center, involving
(2N, 2S) donor sets, which is intermediate between the preferred
CuI and CuII coordination geometries.5
† Department of Chemistry, The University of Kansas.
‡ Hahn-Meitner Institut.
§ Department of Pharmaceutical Chemistry, The University of Kansas.
| University of Notre Dame.
(6) Cooper, S. R.; Rawle, S. C. Struct. Bonding (Berlin) 1990, 72, 1.
(7) Blake, A. J.; Schro¨der, M. AdV. Inorg. Chem. 1990, 35, 1.
(8) Sanaullah; Kano, K.; Glass, R. S.; Wilson, G. S. J. Am. Chem. Soc.
1993, 115, 592.
(9) Glass, R. S.; Wilson, G. S.; Setzer, W. N. J. Am. Chem. Soc. 1980,
102, 5068.
The University of Arizona.
X Abstract published in AdVance ACS Abstracts, February 15, 1997.
(1) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 5th ed.;
Wiley: New York, 1988; pp 755-775.
(2) Biological & Inorganic Copper Chemistry; Karlin, K. D., Zubieta,
J., Eds.; Adenine Press: Guilderland, NY, 1986; Vols. 1 and 2.
(3) Bioinorganic Chemistry of Copper; Karlin, K. D., Tykla´r, Z., Eds.;
Chapman & Hall: New York, 1993.
(4) Messerschmidt, A. AdV. Inorg. Chem. 1993, 40, 121.
(5) Kaim, W.; Schwederski, B. Bioinorganic Chemistry: Inorganic
Elements in the Chemistry of Life; Wiley: Chichester, 1994; pp 187-214.
(10) The [12222] conformation is adopted by the central, bridging TTCN
I
molecule in [Cu2 (TTCN)3] (Clarkson, J. A.; Yagbasan, R.; Blower, P. J.;
Cooper, S. R. J. Chem. Soc., Chem. Commun. 1989, 1244) and the
monodentate TTCN molecule in [CuI(TTCN)2] (Blake, A. J.; Gould, R.
O.; Greig, J. A.; Holder, A. J.; Hyde, T. I.; Schroeder, M. J. Chem. Soc.,
Chem. Commun. 1989, 876).
S0002-7863(96)02775-8 CCC: $14.00 © 1997 American Chemical Society

CuI/II(TTCN)2 Complexes in Aqueous Solution
J. Am. Chem. Soc., Vol. 119, No. 9, 1997 2135
for the importance of ligand dissociation, and 3 is identified as
the structure of the Cu(II) intermediate formed en route to
octahedral 2. The essential aspects of this newly adduced
mechanism are summarized in Scheme 2, which serves as a
guide through the Results and Discussion section for the reader.
Experimental Section
All the experiments were done at ambient temperature (≈20 °C).
The solutions were prepared in highly pure deionized (18 MΩ) water.
Adjustments of pH were made with HClO4 or NaOH.
Instrumentation and Techniques. All the NMR spectra were
acquired on a Bruker AM-500 or Varian XL-300 NMR spectrometer
and were referenced to the residual proton impurity in the deuterated
solvent. The X-ray crystallographic data were collected according to
previously described procedures.8 Details of the apparatus and
procedure for cyclic voltammetry have been described elsewhere.8 The
voltammograms were recorded with a glassy carbon working electrode
against a Ag/AgCl (saturated NaCl) reference electrode.
Figure 1. Drawings and labels of the ligands.
conformations. In [CuII(TTCN-κ3)2]2+ (2),11,12 shown in Figure
2, both TTCN ligands have the endodentate [333] conformation.
The octahedral11 or Jahn-Teller distorted12 coordination ge-
ometry of the copper ion in 2 becomes tetrahedral in 1.
In an effort to understand how these redox-associated
geometry changes occur, the electrochemistry of this system
has been studied in considerable detail.8 Evidence was obtained
for a chemical step which was fast on the CV time scale
following electron transfer. The simplest interpretation of the
data was an ECEC square mechanism8,13,14 as shown in Scheme
1. In this proposal, oxidation of tetrahedral 1 first gives transient
tetrahedral [CuII(TTCN-κ3)2]+ which rearranges rapidly to
octahedral 2 with an accompanying conformational change of
one of the TTCN molecules. Reduction of octahedral 2, on
the other hand, affords transient octahedral [CuI(TTCN-κ3)-
(TTCN-κ1)]2+ which then quickly yields tetrahedral 1 with an
accompanying conformational change in the TTCN-κ1 ligand.
The principle of microscopic reversibility also requires that
tetrahedral 1 may first undergo a geometry change to octahedral
1, albeit unfavorable, followed by oxidation to octahedral 2 and,
similarly, octahedral 2 may first isomerize to tetrahedral 2
followed by reduction as shown in Scheme 1. Although this
square mechanism accounted for the originally obtained data,
additional studies were warranted to establish the structures
proposed for the transient intermediates: tetrahedral 2 and
octahedral 1. The present investigation was undertaken to
further study and characterize the proposed short-lived inter-
mediate in the oxidation of 1 to 2 by using time-resolved optical
and conductometric measurements with the radiation chemical
technique of pulse radiolysis.15,16 Similar studies of other [CuI-
(TTCN-κ3)(L)]+ complexes, where L is monodentate thiolane
(TL), thiane (TN), or 1,4-dithiane (DTN), 5-7, respectively,
as well as [CuI(18-ane-S6)]+ (8; 18-ane-S6 ) 1,4,7,10,13,16-
hexathiacyclooctadecane) also provided insight into this reaction.
The structures of all of the ligands and their labels are shown
in Figure 1 and the structures of all of the complexes are given
in Figure 2. Additional results on the dependence of the
electrochemical redox process on pH afforded further evidence
on the identity of the transient which correlates with the time-
resolved measurements. These new results and NMR spectro-
scopic studies reveal the oxidation of 1 to 2 to be more complex
than previously proposed. In particular, evidence is presented
In the pulse radiolysis experiments N2O-saturated aqueous solutions
of the analytes were irradiated with high-energy pulses of electrons to
•
generate OH radicals. The energy was administered through either
1-2-µs pulses of 1.55-MeV electrons from a van de Graaf accelerator
or 50-ns pulses of 15-MeV electrons from a linear accelerator (LINAC).
Typically a total energy of 1-3 Gy [1 Gy (Gray) ) 1 J kg-1] was
released into the system during the pulses. This corresponds to a total
•OH radical concentration on the order of 10-6 M. The radical
concentration was generally smaller than the complex or any other
solute concentration. Further details of the pulse radiolysis technique
and dosimetry can be found in the literature.15,16
Reagents. All the chemicals employed were from Aldrich (Mil-
waukee, USA) and were used as obtained. [Cu(TTCN)2](BF4)2,11 [Cu-
(18-ane-S6)]PF6,17 and [Cu(CH3CN)4]PF618 were synthesized according
to the cited literature procedures.
[Cu(TTCN)(TL)]PF6, 5. Exactly the same procedure and quantities
of reactants were used for the synthesis of the thiolane (TL) complex
as those for the synthesis of the TN complex, described below. The
yield was 41%.
Anal. Calcd for C10H20S4CuPF6: C, 25.18; H, 4.23; S, 26.88; P,
6.49; F, 23.90. Found: C, 25.16; H, 4.16; S, 26.95; P, 6.50; F, 23.71.
[Cu(TTCN)(TN)]PF6, 6. [Cu(CH3CN)4]PF6 (410 mg, 1.1 mmol)
was added to a solution of TTCN (200 mg, 1.11 mmol) and of thiane
(TN) (205 mg, 2 mmol) in methanol (40 mL) which was prepurged
with argon. The solution was stirred for about 3 h. The reaction
mixture was concentrated on a rotary evaporator until crystallization
started and then left overnight at -40 °C. The crystalline product was
collected on a sintered glass funnel under argon, washed with diethyl
ether, and desiccated. The yield was 56%.
Anal. Calcd for C11H22S4CuPF6: C, 26.91; H, 4.52; S, 26.12; P,
6.31; F, 23.21. Found: C, 26.81; H, 4.48; S, 26.44; P, 6.52; F, 23.01.
[(TTCN)Cu(DTN)Cu(TTCN)](PF6)2, 7. To a degassed solution
of TTCN (200 mg, 1.11 mmol) and 1,4-dithiane (70 mg, 0.58 mmol)
in methanol (25 mL) was added [Cu(CH3CN)4]PF6 (410 mg, 1.1 mmol)
under stirring. After the solution was stirred for 2 h under Ar,
nitromethane (10 mL) was added to dissolve the precipitate. After
filtration under Ar the filtrate was concentrated on a rotary evaporator
until crystallization started. After overnight refrigeration, the crystalline
product was removed by filtration and washed with diethyl ether. The
filtrate was then layered carefully with diethyl ether and refrigerated
overnight. A second crop of crystals was recovered and processed
under Ar. The overall yield was 42%.
(11) Setzer, W. N.; Ogle, C. A.; Wilson, G. S.; Glass, R. S. Inorg. Chem.
1983, 22, 266.
(12) Glass, R. S.; Steffen, L. K.; Swanson, D. D.; Wilson, G. S.; de
Gelder, R.; de Graaf, R. A. G.; Reedijk, J. Inorg. Chim. Acta 1993, 207,
241.
(13) Bernardo, M. M.; Schroeder, R. R.; Rorabacher, D. B. Inorg. Chem.
1991, 30, 1241.
(14) Bernardo, M. M.; Robandt, P. V.; Schroeder, R. R.; Rorabacher,
D. B. J. Am. Chem. Soc. 1989, 111, 1224.
Anal. Calcd for C16H32S8Cu2P2F12: C, 21.40; H, 3.59; S, 28.56; P,
6.90; F, 25.39. Found: C, 21.32; H, 3.60; S, 28.43; P, 6.83; F, 25.44.
[Cu(TTCN)(TTCNO)]PF6, 9. To a solution of TTCN (1,4,7-
trithiacyclononane, 80 mg, 0.443 mmol) and TTCNO19 (1,4,7-trithia-
cyclononane 1-oxide, 90 mg, 0.458 mmol) in degassed methanol (30
mL, Fisher, HPLC Grade) was added [Cu(CH3CN)4]PF6 (165 mg, 0.443
mmol). The reaction mixture was stirred for 3 h under argon. The
(15) Asmus, K.-D. Int. J. Radiat. Phys. Chem. 1972, 4, 417.
(16) Asmus, K.-D.; Janata, E. In The Study of Fast Processes and
Transient Species by Electron Pulse Radiolysis; Baxendale, J. M., Busi,
F., Eds.; NATO ASI; D. Reidel: Dordrecht, 1982; pp 91-113.
(17) Hartman, J. R.; Cooper, S. R. J. Am. Chem. Soc. 1986, 108, 1202.
(18) Kabus, G. J. Inorg. Synth. 1979, 19, 90.
(19) TTCNO was synthesized following the procedure reported by L.
K. Steffen, Ph.D. Thesis, The University of Arizona, 1990.

2136 J. Am. Chem. Soc., Vol. 119, No. 9, 1997
Sanaullah et al.
Figure 2. Drawings of the structures of the copper complexes and transients.
Table 1. Crystal Dataa for [(TTCN)CuI(DTN)CuI(TTCN)](PF6)2
(7) and [CuI(TTCN)(TTCNO)]PF6 (9)
The structure was solved by direct methods.20 The non-hydrogen
atoms were refined anisotropically. For 9 the structure refinement was
achieved with the bond angle and bond distances restrained to ideal
values because the TTCNO ring is disordered. The structures were
refined by full-matrix least-squares techniques using neutral-atom
scattering factors,21 and anomalous dispersion terms22 were included
in Fc; the values for ∆f ′ and ∆f ′′ were those of Cromer.23 All
calculations were performed using the TEXSAN24 crystallographic
software package of the Molecular Structure Corporation and the
KUDNA crystallographic software package developed at the University
of Kansas.25 The final cycle of full-matrix least-squares refinement
was based on 1993 observed reflections (I > 0.01σ(I)) and 246 variable
parameters for 7 and 1389 observed reflections (I > 0.01σ(I)) and 234
variable parameters for 9.
7
9
empirical formula
fw
CuS4C8H16PF6
448.96
CuS6C12H24OPF6
mw
space group
a, Å
b, Å
c, Å
â, deg
Z
897.92
585.19
C2/c (No. 15)
19.832(6)
13.396(2)
15.815(4)
133.57(1)
8
P21/n (No. 4)
7.939(2)
8.688(3)
15.964(4)
102.94(4)
2
crystal color, shape
R
white, needle
0.041
white, prism
0.078
Rw
GOF
0.044
1.08
0.075
1.12
Complex 7 has a center of symmetry. The full complex can be
obtained by performing the symmetry operation 0.5 - x, 0.5 - y, -z
-
a Standard deviation of the least significant figure is given in
parentheses.
on the coordinates of the half complex. Two half PF6 counterions
are found in the asymmetric unit. A full PF6- ion can be obtained by
applying the 1 - x, y, 1.5 - z symmetry operation. Several short C-H-
- -F contacts are observed which are probably weak hydrogen bonds.
Further structural details are given in the Supporting Information.
Binding Constant Determination by NMR Spectroscopy. Titra-
tions of Cu(I) with TTCN monitored by NMR spectroscopy were carried
out in acetonitrile as well as in CD3CN-D2O mixtures under an Ar
atmosphere. Titration of [Cu(TTCN)2]PF6 with [Cu(CH3CN)4]PF6 in
CD3CN-D2O gave similar results. There were no differences in the
chemical shift data in CD3CN versus CD3CN-D2O. The binding
constant for the first equivalent of TTCN is extremely high and cannot
be determined by NMR spectroscopic methods. The binding constant
for the second equivalent of TTCN was calculated26 using linear and
curve fitting equations for the following equilibrium: [CuI(TTCN)]+
+ TTCN h [CuI(TTCN)2]+. There was no consistent change in
broadening over the titration in the line shape or width.
white powdery product was removed by filtration under argon and
washed sequentially with methanol and diethyl ether and dried under
vacuum (77% yield): 1H NMR (CD3CN, 297 K) δ 3.88 (ddd, 2Ha),
3.36 (ddd, 2Hb), 3.27 (ddd, 2Hc), 2.99 (ddd, 2Hd) [(Jab ) 14.3 Hz), (Jac
) 2.6 Hz), (Jad ) 9.9 Hz), (Jbc ) 7.4 Hz), (Jbd ) 2.8 Hz), (Jcd ) 15.4
Hz)], 2.95 (m, 16 H); 13C NMR (CD3CN, 297 K) δ 53.5, 35.7, 32.8,
26.2.
Anal. Calcd for C12H24S6OCuPF6: C, 24.63; H, 4.13; S, 32.87; P,
5.29; F, 19.48. Found: C, 24.48; H, 4.04; S, 32.99; P, 5.22; F, 19.33.
X-ray Single-Crystal Structure Studies. A single crystal of 7
(having approximate dimensions of 0.20 × 0.20 × 0.50 mm) or 9
(having approximate dimensions of 0.20 × 0.30 × 0.40 mm) was
mounted on a glass fiber and placed in a Rigaku AFC5R diffractometer.
Cu KR radiation (λ ) 1.54178 Å) was used with a graphite mono-
chromator and a 12-kW rotating-anode generator. Cell constants and
orientation matrices were obtained from least-squares refinement using
the setting angles of a number of carefully centered reflections in an
approximate range of 2Θ and are given in Table 1. The monoclinic
space group based on the systematic absences at hkl for h + k * 2n
and at h0l for l * 2n was determined to be C2/c (No. 15) for 7, and
based on the systematic absences at 0k0 for k * 2n to be P21/n (No. 4)
for 9.
Results and Discussion
•OH Radical Induced Oxidation. In order to obtain time-
resolved information about the possible intermediates in the
(20) Beurskens, P. T. Technical Report 1984/1; Crystallography
Laboratory: ToernooieVeld, 6525 Ed; Nijmegen: The Netherlands, 1984.
(21) Cromer, D. T.; Waber, J. T. International Tables for X-ray
Crystallography; The Kynoch Press: Birmingham, England, 1974; Vol.
IV, Table 2.2A.
(22) Ibers, J. A.; Hamilton, W. C. Acta Crystallogr. 1964, 17, 781.
(23) Cromer, D. T. International Tables for Crystallography; The Kynoch
Press: Birmingham, England, 1974; Vol. IV, Table 2.3.1.
(24) TEXSAN-TEXRAY Structure Analysis Package; Molecular Structure
Corp.; Woodlands, TX, 1985.
(25) Takusagawa, F. Crystallographic Computing System: KUDNA;
Department of Chemistry, The University of Kansas, 1984.
(26) Connors, K. A. Binding Constants; John Wiley: New York, 1987;
pp 189-200.
The data were collected at a temperature of -160 ( 1 °C for 7 and
23 ( 1 °C for 9. For 7 ω-scans of several intense reflections, made
prior to collection, had an average width of 0.41° at half height with a
take-off angle of 6.0°. Scans of (1.57 + 0.30 tan Θ)° were made at a
speed of 32.0 deg/min (in ω). Weak reflections (I < 10.0σ(I)) were
rescanned. A total of 2188 reflections were collected for 7 and 1667
reflections for 9.

CuI/II(TTCN)2 Complexes in Aqueous Solution
J. Am. Chem. Soc., Vol. 119, No. 9, 1997 2137
Scheme 1. Originally Proposed Square Mechanism for the
Electron-Transfer Reaction of the [CuII/I(TTCN)2] System
Scheme 2. Reaction Scheme for the Oxidation of 1 to 2
oxidation of 1 to 2 pulse radiolysis experiments were conducted
with dilute aqueous solutions of the copper(I) complex and a
number of related model compounds. The actual oxidation was
•
achieved by reaction of the complexes with OH radicals.
The results obtained upon pulse radiolysis of aqueous, N2O-
saturated, 5.0 × 10-5 and 1.0 × 10-4 M solution of 1 at pH 4.0
are displayed in Figures 3 and 4, respectively. The time-
resolved spectra and individual absorption traces, recorded on
the microsecond to second time scale after a 1-µs pulse, reveal
the ultimate formation of a stable product with an absorption
peaking at 445 nm (Figures 3b and 4d) whose optical charac-
teristics are identical with those of the well-characterized
octahedral complex 2 with both TTCN ligands in the [333]
conformation.
Figure 3. UV/vis spectra of a 5 × 10-5 M, N2O-saturated aqueous
solution of 1 at pH 4 obtained upon pulse irradiation: (a) within 200
µs at steps of 10 µs, and (b) within 100 ms at steps of 5 ms.
The pulse radiolysis experiments further show the formation
of two transients in the investigated time interval. The first is
a shorter-lived intermediate with λmax at 540 nm, 4, which
converts into a longer-lived species with λmax at 370 nm, 3, as
shown in Scheme 2. This is evidenced by an isosbestic point
at 435 nm (Figure 3a) and identical, exponential decay and
formation kinetics (t1/2 ) 13.7 ( 0.7 µs for 10-4 M 1) at the
respective maximum wavelengths (Figures 4a and 4b). The
370-nm transient is considerably longer-lived and appears to
be the direct precursor of the final and stable oxidation product
2, as can be deduced from Figure 3b (isosbestic point at 395
nm) and the time-resolved traces in Figures 4c and 4d. The
associated kinetics are also described by an exponential rate
law with t1/2 ) 7.65 ( 0.30 ms for 10-4 M 1. The spectra in
Figure 3a and the trace displayed in Figure 4b show that a
considerable amount of the 370-nm absorption has already
developed immediately after the pulse. Since the short-lived
540-nm transient 4 does not exhibit any significant absorption
in the 370-nm range (as revealed by other experiments presented
later), it is concluded that the 370-nm transient 3 is formed by
multiple pathways as shown in Scheme 2. One proceeds
through the 540-nm species and, on a much shorter time scale,
an •OH adduct of 1, accounting for about one and two thirds of
the 370-nm yield, respectively.
containing high concentrations (10-3 M) of Me2S, Et2S, or free
TTCN in addition to that of 1 at e10-4 M. In all of these cases
the final product 2 (λmax 445 nm) and its immediate precursor,
the 370-nm transient, are identical to the •OH-induced oxidation.
In each system there is also a precursor of the 370-nm species
which, however, shows solute-specific characteristics. For Me2S
and Et2S these initially formed transients can be identified with
the well-known respective σ2σ*1 three-electron bonded radical
cations (R2S SR2)+.27,28 Under the experimental conditions
they are generated within the duration of the 1-µs pulse via the
well-established overall reaction shown in eq 1 and exhibit
•OH + 2R2S f (R2S SR2)+ + OH-
(1)
absorption maxima at 465 (R ) Me) and 485 nm (R ) Et). No
540-nm absorption is observed in these systems.
In the case of solutions containing added TTCN the primary
transient exhibits the same absorption features as the 540-nm
•
species in the direct OH-induced oxidation of 1 except that
the total absorption yield at this wavelength is about three times
higher. Furthermore, the same absorption with the same high
yield is observed in solutions containing only free TTCN and
no copper complex. This and the analogy to numerous other
examples on the oxidation of polythia compounds29,30 suggest
To deduce the structures of the species giving rise to the two
transient absorptions various pulse radiolysis experiments were
carried out with other oxidizing radicals and with complexes
in which the monodentate (TTCN-κ1) trithia ligand was
substituted by heterocycles containing only one sulfur atom.
Change of Oxidant. In order to change the identity of the
(27) Bonifacˇic´, M.; Mo¨ckel, H.; Bahnemann, D.; Asmus, K.-D. J. Chem.
Soc., Perkin Trans. 2 1975, 675.
(28) Go¨bl, M.; Bonifacˇic´, M.; Asmus, K.-D. J. Am. Chem. Soc. 1984,
106, 5984.
(29) Asmus, K.-D.; Bahnemann, D.; Fischer, Ch.-H.; Veltwisch, D. J.
Am. Chem. Soc. 1979, 101, 5322.
•
oxidizing radicals the OH radical was generated in solutions

2138 J. Am. Chem. Soc., Vol. 119, No. 9, 1997
Sanaullah et al.
Figure 4. Absorption Vs time traces at different wavelengths and times of a pulse-irradiated 10-4 M, N2O-saturated aqueous solution of 1 at pH
4: (a) 540 nm (entire time range covered: 200 µs), (b) 370 nm (200 µs), (c) 370 nm (100 ms), and (d) 445 nm (100 ms).
that the 540-nm absorption in the free TTCN containing
solutions is due to the (TTCN)•+ radical cation 4 formed in the
reaction shown in eq 2. The 540-nm absorption in the copper-
bimolecular process depending on the copper(I) complex
concentration. Moreover, the rate constant of 4.1 ((1.0) × 108
M-1 s-1 is practically the same as that for the process in
solutions with added TTCN which unambiguously involves free
4, suggesting that the oxidizing species is, in fact, the same in
both cases. This suggests that the oxidized TTCN moiety does
not remain attached to the copper ion after oxidation of the
complex. Because if it were still bound to the copper, then, in
view of the ultimate formation of 2, an intramolecular rather
than a bimolecular reaction should have occurred, i.e. [CuI-
•OH + TTCN f (TTCN)•+ + OH-
(2)
(I) complex solution, accordingly, is concluded to result from
the oxidation of the TTCN ligand without any direct oxidation
of the metal ion. Alternatively, the monodentate TTCN complex
may, in fact, dissociate to give [CuI(TTCN-κ3)(H2O)]+, 10, and
TTCN. The free TTCN may then be oxidized to 4.
(TTCN-κ3){(TTCN-κ1)•+}]2+ f [CuII(TTCN-κ3)(TTCN-κ1)]2+
.
Since the longer-lived intermediate 3 absorbing at 370 nm is
formed irrespective of the nature of the oxidant, and in all cases
with oxidants other than OH, exclusively via one secondary
process, its generation is attributed to the following general
reaction shown in eq 3. In our specific cases with the added
The experimental finding, therefore, strongly refutes such a
mechanism and indicates detachment of the (TTCN-κ1) ligand
before the 540-nm intermediate can engage in a secondary
process. Since Cu(II)-S bonds involving thioethers are ex-
ceptionally weak, dissociation of the monodentate TTCN ligand
is reasonable.
•
Ox + (1) or (10) f 3
(3)
Change of Monodentate Ligand. Further insight into the
structure of the transient species is provided by experiments on
the •OH-induced oxidation of copper(I) complexes in which the
monodentate (TTCN-κ1) ligand was replaced by thiolane (TL),
thiane (TN), or 1,4-dithiane (DTN), 5-7, respectively (under
otherwise identical experimental conditions as for the oxidation
of 1). The structures of the original Cu(I) complexes in which
thiolane or thiane replaces the monodentate TTCN ligand are
supported by their elemental analyses. The structure of 7, in
which each of the two sulfur atoms of 1,4-dithiane is mono-
coordinated with each of the two copper ions and thereby
bridges them, is established by an X-ray crystallographic study.
The crystal data for this complex are given in Table 1, a PLUTO
drawing is shown in Figure 5, and complete lists of bond lengths,
bond angles, fractional coordinates, and thermal parameters are
provided in the Supporting Information. Two results in the •OH-
induced oxidations of these complexes are particularly note-
worthy: (i) in no case is the short-lived 540-nm absorption
thioethers and free TTCN “Ox” represents (Me2S SMe2)+,
(Et2S SEt2)+, or 4. The fact that 3 is indeed formed via a
bimolecular process is proved by the pseudo-first-order kinetics
for the decay of the “Ox” and the formation of the 370-nm
absorptions. Absolute rate constants derived from the respective
concentration dependencies are 2.0 ((1.0) × 108 M-1 s-1 for
the reaction of both (R2S SR2)+ and 3.5 ((1.0) × 108 M-1
s-1 for the reaction of 4 (for a detailed kinetic evaluation of
these and other reactions referred to in this paper see ref 31).
The underlying reactions involve metal ion oxidation because
oxidation of the ligand molecules alone does not yield any
comparable absorption in the 370-nm range. Accordingly the
oxidation product 3 is a copper(II) complex.
Most interestingly, even in the solutions containing only 1
(no added TTCN) the 540 nm f 370 nm conversion is a
(30) Asmus, K.-D. Acc. Chem. Res. 1979, 12, 436.

CuI/II(TTCN)2 Complexes in Aqueous Solution
J. Am. Chem. Soc., Vol. 119, No. 9, 1997 2139
Figure 6. Kinetic analysis of (a) the decay of the transient absorbing
at 370 nm and (b) of the formation of 2 at 445 nm in terms of first-
order rate constant kobs Vs [1], both at pH 4.
Figure 7. pH dependence of the transient absorptions recorded in a 5
× 10-5 M, N2O-saturated aqueous solution of 1 200 µs after the pulse:
(a) at 380 nm and (b) at 340 nm.
to react with an unoxidized copper(I) complex, and/or the free
TTCN which is liberated in the possible dissociation equilibrium
Figure 5. PLUTO drawing of 7.
with
it.
observed, and (ii) in all cases, however, 3 is formed and
eventually converted into the octahedral 2. The first finding
corroborates our conclusion that the 540-nm species results from
oxidation of the monodentate or free TTCN ligand. The second
finding demonstrates that the nature of the monodentate ligand
is irrelevant for the optical properties of 3. This could, in
principle, be explained by a complete lack of electronic
contribution from the monodentate ligand to the relevant optical
transition. However, a more likely explanation is that the same
transient 3 is produced in all cases. This can result by
detachment of the monodentate ligand from the central copper
ion in the oxidation process, or prior to it. In the case of
oxidation of 1 either the oxidized (TTCN-κ1) (as suggested based
on the kinetic results above) or the unoxidized (TTCN-κ1) may
detach.
Dependence of the Formation of 2 on the Concentration
of 1. Further support for dissociation of a TTCN ligand can
be derived from the kinetics for the conversion of 3 into the
final, stable complex 2. In the “square mechanism” an assumed
tetrahedral [CuII(TTCN-κ3)(TTCN-κ1]2+ intermediate with both
TTCN rings still attached to Cu(II) would only have to undergo
a conformational change to form the stable product and this
should be a true first-order process. Under the conditions of
the pulse radiolysis experiments, however, the decay of the
absorption at 370 nm and the corresponding formation at 445
nm is only of pseudo first order and linearly depends on the
concentration of the unoxidized copper(I) complex as shown
in the plot of the measured first order rate constant as a function
of the concentration of 1 in Figure 6. At concentrations g10-4
M the bimolecular process is clearly dominating and any true
first-order contribution, potentially associable with the intercept,
becomes of lesser importance. Formation of 2 thus requires 3
The bimolecular rate constant for this process, calculated from
the slope of the straight line in Figure 6, is 5.2 ((0.5) × 105
M-1 s-1. This low value indicates relatively high energetic and/
or entropic demands for the formation of 2. That such a
bimolecular process, nevertheless, prevails over a presumably
much faster intramolecular conformational rearrangement strongly
suggests that the latter is, in fact, not possible simply because
the second TTCN ligand is not present in 3.
Alternatively, the TTCN ligand required for the transforma-
tion of 3 into 2 can be supplied by purposely added free,
uncomplexed TTCN. In this case, the rate constant for the
conversion of 3 into 2 becomes somewhat larger. It is found
to be 8.4 ((0.8) × 105 M-1 s-1
.
pH Dependence. Another relevant observation is that the
optical absorption spectrum of 3 varies with pH. The spectrum
displayed in Figure 3a with λmax 370 nm shows these optical
characteristics only at pH e6, and the maximum is gradually
shifted to lower wavelengths with increasing pH. At pH 10.1,
for example, it peaks at 345 nm. The entire pH dependence,
recorded at 380 and 340 nm, respectively, is plotted in Figure
7. It shows two practically complementary curves with a clear
indication for one breakpoint at pH ≈10, and at least one other,
possibly even two, in the pH 5-7 range. Since it is reasonable
to assign such breakpoints to pKa values, the question is at which
sites the protonations occur. Because the thioether sulfur is not
sufficiently basic, particularly if attached to the positive copper
ion, the sites of protonation must be copper-coordinated OH
moieties.
Such sites can be formed most simply by association of an
•OH radical with the copper(I) and subsequent electron transfer
followed by the introduction of an oxo ligand into the coordina-

2140 J. Am. Chem. Soc., Vol. 119, No. 9, 1997
Sanaullah et al.
Figure 8. Variation of the cyclic voltammograms of 1 with pH (0.1
M NaBF4, glassy carbon electrode Vs Ag/AgCl, saturated NaCl): (O)
pH3; (4) pH 6; (0) pH 7; (s) pH 8.5.
Figure 9. Conductivity Vs time trace at pH 3.85 (a) and pH dependence
of maximum yield of conductivity changes (b), obtained upon pulse
irradiation of a 5 × 10-5 M, N2O-saturated aqueous solution of 1.
tion sphere of Cu(II), as shown in eq 4.
•OH + 1 f [CuI(TTCN-κ3)(TTCN-κ1)]+(•OH) f
[CuII(OH-)(TTCN-κ3)(TTCN-κ1)]+ (4)
dissociation of the monodentate TTCN ligand and acidic water
ligands coordinated to Cu(II). The similar characteristics of
the electrochemical metastable species generated by oxidation
of 1 and the transient precursor of 2 formed by pulse radiolysis
of this complex argue that these are the same species.
However, to account for the apparent further pKa values
additional protonation sites must be introduced. Indeed, dis-
sociation of the monodentate TTCN ligand and coordination
of a total of three OH- groups provides octahedral Cu(II) with
three pKa values expected in the range observed. Furthermore,
this proposal for dissociation of the monodentate TTCN ligand
is consistent with the kinetic results described above. (See the
section on conductivity as well.)
Conductivity Experiments. The suggestion that hydroxyl
ligands are involved in the UV intermediate and possibly even
replace the monodentate (TTCN-κ1) is further supported by time-
resolved conductivity experiments. Figure 9a shows, as an
example, the respective trace recorded from a pulsed, pH 3.85,
N2O-saturated solution of 5 × 10-5 M 1. The negative signal
developing after the pulse is indicative of a conductivity loss.
With time it attains a constant value32 of G×∆Λ ) -900 S
cm2. With increasing copper(I) complex concentration the
conductivity loss, and consequently the yield, further increases
a bit but eventually levels off at G×∆Λ ) -1200 S cm2 at g2
× 10-4 M. To interpret these results it is important to note
that neither the conversion of 4 to 3 nor the conversion of the
latter to the final complex 2 is accompanied by any conductivity
change. Furthermore, as can be seen from Figure 9b, the change
in the loss of conductivity becomes smaller at pH >5, with
practically no change between pH 6 and 9, and becomes
negative again at pH >9 with a leveling-off value of G×∆Λ )
-700 S cm2 at pH g10.5. For the high pH range the overall
negative change is attained by a process which starts from a
small positive value, present immediately after the pulse (G×∆Λ
) 150-300 S cm2), and then parallels the 4 f 3 conversion,
quite in contrast to the conductivity kinetics on the acid side.
Qualitatively, the conductivity data obtained on the acid side
would be compatible with the original suggestion that a
tetrahedral [CuII(TTCN-κ3)(TTCN-κ1)]2+ transient is formed on
oxidation of the Cu(I) complex. In this case, the initial
conductivity changes would be accounted for by the following
hypothetical reactions referring to oxidation of either the copper
or the (TTCN-κ1) ligand in 1 as shown in eqs 5 and 6, or the
Note also that with increasing pH the rate constant for the
formation of 2 decreases. At pH 7, for example, it has dropped
by one order of magnitude to ≈7 × 104 M-1 s-1 as compared
to that of 5.2 × 105 M-1 s-1 at pH e5. Above pH 8, in fact,
less and less of the stable 2 is formed and another process
appears to take over, in which the transient decays independent
of the copper(I) complex concentration, and a different, non-
absorbing product is formed. (The latter process may involve
hydrolysis, but no product assignment is possible at present.)31
The redox chemistry of 1 at a solid electrode also varied
substantially with increasing pH. Figure 8 shows a number of
voltammograms of this complex run under different pH condi-
tions. The shift in the formal potential toward less positive
values with increasing pH with a large break in the potential-
pH curve at pH 6.5 indicates the involvement of an acid-base
reaction in the electron-transfer mechanism of the complex. Also
the incremental suppression and finally the disappearance of
the reverse (cathodic) peak with increasing pH suggest that
successive formation of Cu(II) hydroxo complexes makes their
subsequent displacement by free TTCN increasingly more
difficult. When the cyclic voltammograms of 1 were run in
the presence of free TTCN at pH 3, both the anodic and cathodic
peaks shifted to a less positive potential by about 10 mV with
an apparent increase in the slope of the peaks. Thus the
dependence of the electrochemistry of this complex on both
the pH and the concentration of TTCN is consistent with
(32) G represents yield in terms of the number of the species formed
per 100 eV of absorbed energy. G ) 1.0 corresponds to 0.1036 µM species
produced per joule of absorbed energy. ∆Λ is the change in conductivity
in Siemens (S).
(31) Hungerbu¨hler, H.; Guldi, D.; Asmus, K.-D.; Sanaullah; Scho¨neich,
C.; Wilson, G. S.; Glass, R. S. To be submitted for publication.

CuI/II(TTCN)2 Complexes in Aqueous Solution
J. Am. Chem. Soc., Vol. 119, No. 9, 1997 2141
copper in 10 as shown in eq 7.
complex is regenerated.] As formulated, both overall mecha-
•OH + 1 f OH- + (TTCN-κ1)•+
+
•OH + 1 f [CuII(TTCN-κ3)(TTCN-κ1)]2+ + OH- (5)
free/complexed
[CuI(TTCN-κ3)] (10)
•OH + 1 f [CuI(TTCN-κ3)(TTCN-κ ) ]
1 •+ 2+ + OH- (6)
(7)
(TTCN-κ1)•+free/complexed + 1 + H+ + 2H2O f 3 + 2TTCN
•OH + 10 f [CuII(TTCN-κ3)(H2O)]2+ + OH-
(11)
nisms, i.e., via the •OH adduct or TTCN oxidation, are in accord
with the conductivity data observed in the acidic solution. To
accommodate the data in the neutral and basic pH range
consecutive deprotonation is inferred as formulated in eqs 12-
14.34
In the acid environment all of these reactions would be followed
by immediate neutralization of the hydroxide ion as shown in
eq 8. The overall net result in either case would be the
OH- + Haq+ f H2O
(8)
[CuII(TTCN-κ3)(H2O)3]2+
[CuII(TTCN-κ3)(OH-)(H2O)2]+ + Haq (12)
h
replacement of one “normal” cationic charge (Λ ) 45 S cm2 at
18 °C) at the expense of a highly conducting Haq+ (Λ ) 315 S
cm2 at 18 °C), i.e., ∆Λ ) -270 S cm2. Since a conversion of
[CuI(TTCN-κ3)({TTCN-κ1}•+)]2+ into [CuII(TTCN-κ3)(TTCN-
κ1)]2+ would not lead to any significant conductivity changes,
the combined yield of the two initial reactions would equal the
+
[CuII(TTCN-κ3)(OH-)(H2O)2]+ h
+
[CuII(TTCN-κ3)(OH-)2(H2O)] + Haq (13)
•
yield of OH radicals, namely, G(•OH) ) 6. The actual yield,
[CuII(TTCN-κ3)(OH-)2(H2O)] h
calculated from the conductivity measurements, amounts,
however, only to G ) 4.45 with an error limit of no more than
(0.5. This discrepancy alone would not pose a serious problem,
however, because some of the hydroxyl radicals could abstract
a hydrogen atom from the TTCN moieties, a reaction not
involving changes in conductivity and well known for thioether
oxidations.27 However, the above scheme is incompatible with
the conductivity results in basic solutions where, in contrast to
the experimental result, all the potential OH- releasing reactions
(eqs 5-7) would call for an increase in conductivity because
the hydroxide ion would not be neutralized anymore. In
essence, like all the other pulse radiolysis experiments the
conductivity experiments also require a structure for 3 which
is different from a simple tetrahedral [CuII(TTCN-κ3)(TTCN-
κ1)].
+
[CuII(TTCN-κ3)(OH-)3]- + Haq (14)
From the conductivity point of view the absence or presence
of H2O ligands is, of course, irrelevant and, in principle, a second
TTCN could also still be attached. In order to account for the
observed conductivity signals in basic solutions at higher pH it
is, however, necessary to invoke a total of three OH- ligands.
The proposed structure of 3 with its three acid-base forms is,
therefore, the most reasonable possibility because it is the
simplest one which fits the data.
This conclusion is also consistent with a number of non-pulse-
radiolysis observations described below and in the following
sections. It involves hydrolysis of the monodentate (TTCN-
κ1) in the unoxidized copper(I) complex as shown in eq 15.
Identity of 3, the Precursor of 2. A structure of 3
compatible with all of the experimental data and conclusions is
that shown in Figure 2. Its formation is suggested to arise via
an initial •OH radical addition to the Cu(I) complex in the overall
reaction shown in eq 9. To identify a possible •OH adduct pulse
1 + OH- h [CuI(TTCN-κ3)(OH-)] + TTCN (15)
This is compatible with the instability of 1 in basic solution,
and with titration experiments which indicate a pK ≈ 6.5 ((0.2)
for this equilibrium. Assuming that [CuI(TTCN-κ3)(OH-)] is
the prevalent form at the high pH range of our experiments,
i.e. at pH g10.5, and the reaction of the •OH radical with [CuI-
(TTCN-κ3)(OH-)] and free TTCN (which under these circum-
stances are present at equal concentration) occur with the same
rate constant, the yield of both reaction routes would be equal,
i.e. occur each with G ≈ 3. Quantitatively, this translates into
the following experimental data to be expected at high pH (for
detailed calculations see ref 31): (i) Oxidation of the (TTCN-
κ1) ligand, to yield 4, results in a conductivity increase of G×∆Λ
≈ +700 S cm2 and oxidation of [CuI(TTCN-κ3)(OH-)], to yield
[CuII(TTCN-κ3)(OH-)3]-, leads to a loss of G×∆Λ ≈ -500 S
cm2. Since both processes occur fast the combined signal
observable immediately after the pulse amounts to G×∆Λ ≈
+200 S cm2. (ii) The secondary oxidation of [CuI(TTCN-κ3)-
(OH-)] by 4 involves a loss of G×∆Λ ≈ -1100 S cm2, and
the overall signal should thus attain a final value of G×∆Λ ≈
-900 S cm2 with kinetics entirely controlled by this secondary
process. Considering the various possible uncertainties involved
(yield and specific conductances) this proposed scheme is in
excellent agreement with the experimental results.
+
H
1 + •OH f [CuI(TTCN-κ3)(TTCN-κ1)]+(•OH) 2H O8
2
2+ + TTCN (9)
[CuII(TTCN-κ3)(H2O)3]
3
radiolysis was extended into the nanosecond time scale. Indeed,
a very short-lived absorption with λmax at 260 nm was observed
which decays by a true first-order process (k ) 1.4 × 106 s-1
)
directly into 3 and accounts for the fast-formed portion on the
microsecond time scale experiment in Figure 4b. (As will be
discussed in detail elsewhere31 this absorption is likely to be
•
due to an OH adduct to any one of the copper-bound sulfur
atoms in the TTCN ligand rather than a hydroxyl adduct to the
copper ion.)
The second, slower process of the formation of 3 in Figure
4b is accordingly assignable to an oxidation of the copper(I)
complex by 4 (irrespective of whether this is detached from
the metal or not), followed or accompanied by a fast TTCN T
OH- ligand exchange and coordination of water, as shown in
eqs 10 and 11. [The [CuI(TTCN-κ3)] complex fragment is likely
to coordinate a water molecule as fourth ligand33 unless it
encounters a free TTCN in which case an original copper(I)
(34) Copper complexes are also known to form µ-oxo species under basic
conditions. However, these are omitted from consideration for simplicity
because they would not affect the conclusions drawn from the experiments
described in this paper.
(33) Olmstead, M. M.; Musker, W. K.; Kessler, R. M. Transition Met.
Chem. 1982, 7, 140.

2142 J. Am. Chem. Soc., Vol. 119, No. 9, 1997
Sanaullah et al.
The three equilibria (eqs 12-14) also provide an excellent
basis to explain the observed optical changes. In fact, the blue-
shift with increasing (H2O) f (OH-) ligand conversion probably
makes good sense from the electronic point of view. Finally,
both our optical and conductivity data allow the extraction of
at least two of the equilibrium constants with a reasonable degree
of confidence, namely, pKa ≈ 5.6 ((0.3) for this first and ≈9.7
((0.5) for the third equilibrium. The value for the second is
more difficult to evaluate, but it should be close to that of the
hydrolysis of 1, i.e., ≈7 to account for the overall near zero
conductivity change in the pH e9 range.
Compatibility with the Formation Kinetics of 2. The
equilibria formulated in eqs 12-14 also afford a satisfactory
explanation for the pH dependence of the conversion kinetics
of the copper(II) transient into the final, stable complex 2. Since
the transient contains only one TTCN ligand it needs to acquire
the second one from outside in a bimolecular process. In acid
solutions the source of this ligand is unoxidized 1 and/or
possibly free TTCN. The corresponding reaction with 1, shown
in eq 16, is envisaged to proceed via an initial interaction of
the softer sulfur. In fact, it seems to need the entire entropic
advantage of chelation of a tridentate TTCN ligand to stabilize
the structure of 2. How sensitive this system responds even to
the subtle differences in the complex geometry emerges from
an experiment with [CuI(18-ane-S6)]+, 8. The cyclic 18-ane-
S6 ligand contains six sulfur atoms, each two of them separated
by a -(CH2)2- unit as in TTCN. In the copper(I) complex
four of the sulfurs are coordinated with the copper, satisfying
the tetrahedral preference of Cu(I), while the remaining two
are not. This complex could be an ideal system for a fast and
intramolecular hexacoordination once the copper is oxidized.
•
Indeed, in the OH-induced oxidation there is no indication of
a long-lived intermediate absorbing at 370 nm, and only a fast
(within ≈5 µs) formation of an absorption around 450 nm,
seemingly characteristic for hexasulfur-coordination, is observed.
What is interesting, however, is the limited lifetime of this
product. Even in acid solution it decays with t1/2 ≈ 7 ms. (In
basic solution the decay is further accelerated, e.g., t1/2 ≈ 750
µs at pH 8.3.) This finding suggests that the coordination
geometry of 2 cannot be achieved by 8. The initial fast
formation of a hexasulfur-coordinate intermediate in the oxida-
tion of 8 is expected because the additional two sulfur atoms
ligated after the copper oxidation would be the two originally
non-coordinated sulfurs in 18-ane-S6. However, the geometry
attainable in this complex apparently does not afford the
thermodynamically most favorable structure. It is conceivable,
and in agreement with crystal structure analysis,17 that 2 does
not adopt octahedral CuS6 coordination geometry of highest
symmetry but involves two different sulfur-sulfur distances,
one within and another between the respective [333] units. In
18-ane-S6 ligand all sulfurs are equally spaced by two methylene
groups and in order to attain the apparently most favorable 2
structure the two linking -S-(CH2)2-S- units need to be
constrained, and this would make at least one (or possibly even
more) sulfur(s) susceptible for eventual replacement by oxo
ligands. Similarly in the oxidation of [CuI(15-ane-S5)]+, in
which the metal ion coordinates only four sulfur atoms of the
ligand, coordination of the fifth sulfur atom becomes extremely
rapid once the CuI is oxidized to CuII.35 Nevertheless, this
additional Cu-S bond is suggested35 to be weak.
[CuII(TTCN-κ3)(H2O)3] + 1 f 2 + other products (16)
the copper(II) intermediate with the two free sulfurs in the
monodentate (TTCN-κ1) of the copper(I) complex.8 Associated
with this first step is a displacement of two water ligands. This
is then followed by the complete detachment of the (TTCN-κ1)
ligand from the Cu(I) ion, possible structural change to an all-
endodentate [333] conformation, and ultimate displacement of
the third water ligand in the Cu(II)species. The complexity of
the overall process provides a most reasonable rationale for high
steric demands and/or activation energy, and thus for the
relatively low rate constant. The fact that the reaction of the
copper(II) intermediate with added free TTCN occurs only
marginally faster (8.4 × 105 M-1 s-1 Vs 5.2 × 105 M-1 s-1
)
would indicate that the detachment of the monodentate ligand
and the possibly necessary adoption of the [333] conformation
of the incoming TTCN are energetically of less importance than
the actual ligand exchange process in the copper(II) complex.
This conclusion is also supported by the NMR spectroscopic
data presented below.
Ligand Dynamics in 1. The pulse radiolysis experiments
show that 1 on oxidation undergoes loss of the monodentate
TTCN ligand or that the monodentate TTCN ligand dissociates
and it is 10 which is oxidized.
The proposed identity of the copper(II) transient not only
explains the independence of its properties from the nature of
the monodentate ligand in the unoxidized copper(I) complex,
it equally well explains the much slower conversion of the
transient into the final complex 2 for all [CuI(TTCN-κ3)(L)]
systems with L * TTCN because in these systems the only
TTCN available is a relatively tightly bound (TTCN-κ3). Of
course, when free TTCN is added the rate of reaction becomes
correspondingly faster and attains, as expected, a common value
for all systems.
The suggested structure of the copper(II) transient also
provides a plausible rationale for the pH dependence of the
formation of 2. Thus the fastest reaction occurs at pH e5 when
the incoming TTCN has to replace three water ligands. At
higher pH the rate decreases practically parallel to the successive
conversion of these water ligands into hydroxide ligands which,
because of their negative charge, are more strongly bound to
the positive Cu2+ ion and consequently require higher activation
energy to be replaced. Eventually two, and finally all three,
water ligands will be hydroxides and their replacement by
(TTCN-κ3) becomes, in fact, too slow to compete with other
decay routes of the transient.
NMR spectroscopic studies provide some insight into the
ligand dynamics in 1 that are not redox coupled. The 1H NMR
spectrum of TTCN shows a singlet at δ 3.08 ppm. Thus all of
the hydrogen atoms in this molecule are equivalent. Assuming
that the lowest energy conformer in solution for TTCN is
[333],9,36,37 the equivalency of the hydrogen atoms is accom-
modated by rapid interconversion of the gauche conformers as
suggested by Lockhart and Tomkinson,38 and ring inversion of
the endodentate [333] conformers on the NMR time scale.
Titration of TTCN with Cu(I) in a 1:1 mixture of CD3CN/D2O
monitored by 1H NMR spectroscopy gave the results shown in
Figure 10. As Cu(I) is added the signal broadens and undergoes
an upfield shift. At a 1:1 ratio of Cu(I) to TTCN two broad
peaks are observed and beginning with a 3:2 and continuing
with higher ratios of Cu(I) to TTCN an unchanging AA′BB′
(35) Vande Linde, A. M. Q.; Juntunen, K. L.; Mols, O.; Ksebati, M. B.;
Ochrymowycz, L. A.; Rorabacher, D. B. Inorg. Chem. 1991, 30, 5037.
(36) Setzer, W. N.; Coleman, B. R.; Wilson, G. S.; Glass, R. S.
Tetrahedron 1981, 37, 2743.
(37) Blom, R.; Rankin, D. W. H.; Robertson, H. E.; Schro¨der, M.; Taylor,
A. J. Chem. Soc., Perkin Trans. 2 1991, 773.
The overall low bimolecular rate constant for the incorpora-
tion of the second TTCN ligand also makes sense in view of
Cu2+ being a hard acid, preferring the hard oxo ligands over
(38) Lockhart, J. C.; Tomkinson, N. P. J. Chem. Soc., Perkin Trans. 2
1992, 533.

CuI/II(TTCN)2 Complexes in Aqueous Solution
J. Am. Chem. Soc., Vol. 119, No. 9, 1997 2143
Figure 12. 1H NMR spectra of [Cu(TTCN)2]PF6 in CD3CN at various
temperatures.
Figure 10. Titration of TTCN with Cu(I) in CD3CN/D2O monitored
by H NMR spectroscopy.
1
by variable-temperature NMR spectroscopic studies on [CuI-
(TTCN)2]+ in CD3CN. As shown in Figure 12, as the
temperature is lowered the peak broadens and then separates
into three resonances. Below 225 K the solution freezes but at
this temperature the ratio of the three peaks determined by
integration is 2:1:1. This suggests that the interconversion of
tridentate and monodentate TTCN ligands has been frozen out.
The monodentate TTCN absorbs at δ ∼3.05 with its twelve
magnetically equivalent protons and at δ ∼2.86 and ∼2.68 by
the tridentate TTCN with its marginally resolved axial and
equatorial protons, respectively. The equivalence of the protons
of monodentate TTCN suggests that not only is this ligand
conformationally mobile but that each of the sulfur atoms may
be bonded to the CuI, i.e., there are metallotropic shifts.39,40 Such
fluxional behavior is analogous to that suggested for [Pd-
(TTCN)2] (BF4)2 which shows a singlet in its 1H NMR spectrum
even at 238 K but its solid state structure shows a square plane
of four sulfur donors and two weak apical interactions.41 Thus
metallotropy is suggested in this case to account for the NMR
spectroscopic equivalence of the ligand hydrogen atoms.
To reduce metallotropy in the monodentate TTCN and
interchange of the monodentate and tridentate TTCN ligands,
the [CuI(TTCN-κ3)(TTCNO-κ1)]+ complex 9, in which the
monodentate ring is the mono-S-oxide of TTCN, was prepared
and studied. The desired mixed ligand complex selectively
crystallized from a 1:1:1 molar mixture of TTCN, TTCNO, and
[CuI(CH3CN)4]PF6. Its structure was determined by X-ray
crystallographic methods. The crystal data for this complex
are given in Table 1, a PLUTO drawing of this complex is given
in Figure 13, and complete lists of bond lengths, bond angles,
fractional coordinates, and thermal parameters are provided in
the Supporting Information. The CuS4 coordination geometry
is tetrahedral with three sulfur atoms from one TTCN coordi-
nated in the endodentate [333] conformation and, in the fourth
coordination site, with one of the thioether not sulfoxide-sulfur
atoms in the monodentate TTCNO ring. This complex under-
goes irreversible oxidation at a peak potential more positive
than that for [CuI(TTCN-κ3)(TTCN-κ1)]+ as shown in Figure
14a. It will be noted that the reduction of the Cu(II) complex
resulting from oxidation of 9 occurs at exactly the same potential
Figure 11. 1H NMR spectrum of a CD3CN/D2O solution containing
a 3:2 ratio of Cu(I):TTCN.
pattern is obtained as exemplified by Figure 11. Under these
conditions there is apparently no free TTCN in solution and
the tridentate TTCN ligand gives rise to the observed spectrum.
This suggests that the first TTCN molecule coordinated to Cu-
(I) binds strongly as a tridentate ligand but that the second TTCN
ligand bound to [CuI(TTCN-κ3)] exchanges with unbound
TTCN. In addition the spectrum shown in Figure 11 suggests
that coordination of TTCN to Cu(I) as a tridentate ligand
prevents ring inversion but not interconversion of the two gauche
forms. In addition, under these conditions when the ratio of
Cu(I) to TTCN is 2:3, there is still no free TTCN because
stoichiometry requires that the additional TTCN be bound to
[CuI(TTCN-κ3)]+. This additional TTCN is monodentate owing
to the preferential four-coordination of Cu(I) as shown in the
X-ray crystallographic structure of [CuI(TTCN)2]PF6.8 How-
ever, at room temperature the 1H NMR spectrum shows
equivalent TTCN ligands. Consequently, the tridentate and
monodentate TTCN ligands must interconvert rapidly on the
NMR time scale. More insight into this process was obtained
(39) Abel, E. W.; Beer, P. D.; Moss, I.; Orrell, K. G.; Sik, V.; Bates, P.
A.; Hursthouse, M. B. J. Organomet. Chem. 1988, 341, 559.
(40) Bernardo, M. M.; Heeg, M. J.; Schroder, R. R.; Ochrymowycz, L.
A.; Rorabacher, D. B. Inorg. Chem. 1992, 31, 191.
(41) Blake, A. J.; Holder, A. J.; Hyde, T. I.; Roberts, Y. V.; Lavery, A.
J.; Schro¨der, M. J. Organomet. Chem. 1987, 323, 261.

2144 J. Am. Chem. Soc., Vol. 119, No. 9, 1997
Sanaullah et al.
Figure 13. PLUTO drawing of 9.
Figure 15. 1H NMR spectra of (a) TTCNO and (b) 9 in CD3CN.
Figure 16. Binding constant graph for the binding of monocoordinated
TTCN to 1 according to a procedure given in ref 26.
process in which the electron transfer is preceded by the
substitution of ligated TTCNO by the free TTCN added to the
solution.
As pointed out above, 3 either is formed by oxidation of 1
with loss of the monodentate TTCN ligand or this ligand
dissociates first and 10 is oxidized. The relative importance of
these two pathways depends in large part on the relative
concentrations of 1 and 10. These concentrations depend on
the binding constant for monodentate TTCN to [CuI(TTCN-
κ3)]+. Although this binding constant could not be determined
in water alone, it could be measured in aqueous acetonitrile by
1H NMR spectroscopic monitoring of the titration of [CuI(CH3-
CN)4]PF6 with TTCN or of [CuI(TTCN)2]PF6 with [CuI(CH3-
CN)4]PF6. From the dependence of the chemical shift difference
between bound and unbound TTCN on the concentration of
TTCN shown in Figure 16 a binding constant of 650 L/mol
can be derived26 for the association of monodentate TTCN with
[CuI(TTCN-κ3)]+. This binding constant means that [CuI-
(TTCN-κ3)L]+, L ) H2O or CH3CN, and TTCN are the
preponderant species at an initial concentration of 1 of 5 × 10-5
M as used in the pulse radiolysis experiments. This argues for
the formation of 3 principally by oxidation of 10. However,
this binding constant may be significantly greater in water than
in aqueous acetonitrile resulting in a greater importance of the
formation of 3 by oxidation of 1 followed by loss of the
monodentate TTCN ligand. The fact that attachment of a “soft”
sulfur ligand to the “soft” Cu+ in the unoxidized complex
provides such a highly labile system serves as a convincing
Figure 14. Cyclic voltammograms of various Cu(I) species in pH 3
buffer (0.1 M NaBF4, glassy carbon electrode Vs Ag/AgCl, saturated
NaCl): (a) 9 (4), (b) 9 + TTCN (0), (c) 1 (O).
as that for 1. This suggests that the Cu(II) ligands must be
either the same or similar in both cases. If excess TTCN is
added to a solution of 9, then TTCN is seen to displace some
of the TTCNO ligand, resulting in a split oxidation wave.
The 1H NMR spectrum of 9 as well as that of uncomplexed
TTCNO is shown in Figure 15. Since the spectrum of 9 is the
same as that obtained by superposition of the spectra of
uncomplexed TTCNO and a 1:1 mixture of Cu(I) and TTCN,
it appears that TTCNO does not bind to [CuI(TTCN-κ3)]+ in
CD3CN solution. The TTCNO ligand is apparently replaced
by acetonitrile. Thus TTCNO is a much poorer monodentate
ligand42 than TTCN or acetonitrile toward [CuI(TTCN-κ3)]+.
The poor ligand character of TTCNO compared to TTCN is
also demonstrated by the cyclic voltammetry of [CuI(TTCN-
κ3)(TTCNO-κ1)] in the presence of free TTCN as shown in
Figure 14b. The appearance of an anodic wave corresponding
to that of [CuI(TTCN-κ3)(TTCN-κ1)] clearly suggests a CE
(42) That TTCNO is a poorer monodentate ligand than TTCN with Cu-
(I) is unexpected because of the known importance of ligand π-acidity for
complexes of TTCN with low valent transition metal ions6,43 and TTCNO
should be a better π-acid than TTCN.
(43) Blake, A. J.; Holder, A. J.; Hyde, T. I.; Schro¨der, M. J. Chem. Soc.,
Chem. Commun. 1989, 1433.

CuI/II(TTCN)2 Complexes in Aqueous Solution
J. Am. Chem. Soc., Vol. 119, No. 9, 1997 2145
rationale for an even weaker interaction with the “hard” Cu2+
and monodentate TTCN which is compatible with this case.
Correlation of Electrochemical and Pulse Radiolysis
Studies. As already pointed out, the metastable intermediate
formed on electrochemical oxidation of 1 and the transient
variety of complementary experimental techniques. It also
includes higher pH regions in which the initial copper(I)
complex appears to be partially hydrolyzed. The transient 3
formed on oxidation of 1 and which serves as the immediate
precursor to 2 is [CuII(TTCN-κ3)(H2O)3]2+. Depending on pH,
it undergoes one or more deprotonations. The structure of these
hexacoordinate transients is likely to be octahedral. This
replaces the previous suggestion, which was the simplest
interpretation of the electrochemical data, that this species is
tetrahedral [CuII(TTCN-κ3)(TTCN-κ1)]2+. The key features of
the mechanism for this oxidation proposed in this paper are the
high lability of monodentate TTCN and the pH dependence of
this process. This study shows the value of using complemen-
tary experimental techniques for discerning the chemistry of
complex systems.
•
precursor of 2 obtained by OH-induced oxidation of 1 are
suggested to be the same species, namely 3, based on their
similar properties. However, the conversion of this species into
2 depends on the concentration of 1 in the pulse radiolysis
experiment but not in the electrochemical experiment. This
major discrepancy may be explained on the basis of reaction
and diffusion kinetics and the specific characteristics of the two
techniques.
In pulse radiolysis only a small fraction (typically only a few
percent at most) of the solute molecules are converted in a redox
process and, furthermore, any TTCN released from the oxidized
copper can freely diffuse into the bulk of the solution.
Conversion of the transient 3 to 2 will, therefore, almost
exclusively occur by encounter with an unoxidized Cu+ complex
(or free TTCN). In electrochemistry, it can be anticipated that
all Cu+-complexed molecules formed at the electrode surface
will be oxidized in a fast and efficient electron-transfer process
so that the double layer volume and immediate vicinity of the
electrode become depleted of intact 1 but may still contain a
high concentration of those (TTCN-κ1) ligands which are
detached from the copper in the oxidation process. This
facilitates reattachment of the latter to the intermediate 3 which
is presumed to exist also in electrochemistry. It also explains
its independence from the Cu+ complex concentration. The
comparatively high free TTCN concentration near the electrode
would also allow fast reconstitution of 1 after a possible and
presumably fast reduction of any intermediate 3, providing the
rationale for reversibility irrespective of the actual stoichiometry
of the copper(II) complex.
Acknowledgment. Partial support of this work by NATO
Research Grant No. 910917 is gratefully acknowledged. S.,
R.S.G., and G.S.W. thank the U.S. National Institutes of Health
(Grant HL-15104) for financial support of this work. Part of
this work was supported by the Office of Basic Energy Sciences
of the Department of Energy and, as such, it is contribution
No. NDRL-3919 from the Notre Dame Radiation Laboratory.
We thank Dr. Fusao Takusagawa for his assistance in the X-ray
crystallographic structure determinations.
Supporting Information Available: Table of crystal data,
experimental details, tables of fractional coordinates and thermal
parameters, and complete listings of bond distances, bond angles,
selected torsion angles, ORTEP drawings of the molecules, and
PLUTO stereoviews of the unit cells for 7 and 9 (29 pages).
See any current masthead page for ordering and Internet access
instructions.
Conclusion
In conclusion, a consistent picture has emerged on the
oxidation mechanism of 1 to 2 accommodating the results of a
JA9627757