Facile determination of the spectra of unstable electrode products using simultaneous fiber-optic chronoabsorptometry and chronoamperometry

Article abstract of DOI:10.1021/ic1012654

A widely applicable fiber-optic UV-vis method to determine the spectra of in situ generated redox products and intermediates at or near an electrode surface is described mathematically and implemented experimentally. The quantitative spectral information obtained gives extinction coefficients (absorptivities) as a function of wavelength, requires no arbitrary subtraction of the spectrum of the starting material, and is relatively insensitive to path length and concentration during the spectroelectrochemical measurements. We demonstrate proof-of-concept of this methodology by reproducing the expected spectrum of the ferrocenium ion from electrooxidation of ferrocene in MeCN, and by reproducing the spectrum that reveals π-radical cation formation from the electrooxidation of (T(p-OMe)PP)Co(NO) ((T(p-OMe)PP = 5,10,15,20-tetra(p- methoxyphenyl)porphyrinato dianion). Importantly, we demonstrate its use for the facile detection of unstable redox products not previously detected by current spectroelectrochemical methods. We obtain, for the first time, the experimental UV-vis spectrum of the short-lived fac-[(dppe)Mn(CO)₃Br]⁺ cation, a hitherto uncharacterized intermediate that forms during the archetypal redox-induced fac-to-mer isomerization of (dppe)Mn(CO)₃Br (dppe = diphenylphosphinoethane). Spectral features of the Mn-containing species have been verified by comparison to theoretical spectra calculated by time-dependent density functional theory methods.

Full text of DOI:10.1021/ic1012654

9590 Inorg. Chem. 2010, 49, 9590–9598
DOI: 10.1021/ic1012654
Facile Determination of the Spectra of Unstable Electrode Products Using
Simultaneous Fiber-Optic Chronoabsorptometry and Chronoamperometry
Michael J. Shaw,*,† Daniel L. Cranford,† Kenneth W. Rodgers,† James E. Eilers,† Bradley Noble,‡
Adam J. Warhausen,§ and George B. Richter-Addo§
†Department of Chemistry, Box 1652, Southern Illinois University at Edwardsville, Edwardsville, Illinois 62026,
‡Department of Electrical and Computer Engineering, Box 1801, Southern Illinois University at Edwardsville,
Edwardsville, Illinois 62026, and §Department of Chemistry and Biochemistry, University of Oklahoma,
620 Parrington Oval, Norman, Oklahoma 73019
Received June 25, 2010
A widely applicable fiber-optic UV-vis method to determine the spectra of in situ generated redox products and
intermediates at or near an electrode surface is described mathematically and implemented experimentally. The
quantitative spectral information obtained gives extinction coefficients (absorptivities) as a function of wavelength,
requires no arbitrary subtraction of the spectrum of the starting material, and is relatively insensitive to path length and
concentration during the spectroelectrochemical measurements. We demonstrate proof-of-concept of this methodol-
ogy by reproducing the expected spectrum of the ferrocenium ion from electrooxidation of ferrocene in MeCN, and by
reproducing the spectrum that reveals π-radical cation formation from the electrooxidation of (T(p-OMe)PP)Co(NO)
((T(p-OMe)PP = 5,10,15,20-tetra(p-methoxyphenyl)porphyrinato dianion). Importantly, we demonstrate its use for
the facile detection of unstable redox products not previously detected by current spectroelectrochemical methods. We
obtain, for the first time, the experimental UV-vis spectrum of the short-lived fac-[(dppe)Mn(CO)3Br]þ cation, a
hitherto uncharacterized intermediate that forms during the archetypal redox-induced fac-to-mer isomerization of
(dppe)Mn(CO)3Br (dppe = diphenylphosphinoethane). Spectral features of the Mn-containing species have been
verified by comparison to theoretical spectra calculated by time-dependent density functional theory methods.
Introduction
this methodology. For example, Heath and co-workers have
utilized OTTLE cells to obtain spectral data for a number of
intermediates generated from fullerenes,3 alkynyl-bridged
organometallics,4 alkynyl-bridged porphyrin complexes,5 co-
ordination compounds with M-M multiple bonds,6 and other
systems.7 These examples include high-quality determina-
tions of extinction coefficients from chemical species in
multiple oxidation states at low temperatures.8 For some
systems, detailed information about non-linear third-order
optical properties are available from OTTLE studies.9
Commercial OTTLE cells for UV-vis spectroelectro-
chemistry are now readily available.10 However, some limita-
tions remain. For example, factors such as (i) inaccuracies in
cell path length because of the presence of a metallic electrode
grid and non-reproducible spacer length after multiple cell
The accurate identification of redox products is an im-
portant goal in chemistry, and significant effort has been
placed on the use of spectroelectrochemical methods to aid in
this goal. Successful approaches have been reported for the
generation and identification of short-lived redox products,
and these include methods involving the use of optically
transparent thin layer electrochemical (OTTLE) cells,1 bulk
electrolyses and sampling of aliquots of product solutions,
and chemical generation of redox products using chemical
redox reagents.2 The use of OTTLE cells has revolutioni-
zed the field of spectroscopic identification of electrode
products, and numerous advances have been achieved using
*To whom correspondence should be addressed. E-mail: michsha@
siue.edu.
(5) Arnold, D. P.; Heath, G. A. J. Am. Chem. Soc. 1993, 115, 12197–
12198.
(6) Heath, G. A.; Raptis, R. G. J. Am. Chem. Soc. 1993, 115, 3768–3769.
(7) Baldas, J.; Heath, G. A.; Macgregor, S. A.; Moock, K. H.; Nissen,
S. C.; Raptis, R. G. J. Chem. Soc., Dalton Trans. 1998, 2303–2314.
(8) Webster, R. D.; Heath, G. A.; Bond, A. M. J. Chem. Soc., Dalton
Trans. 2001, 3189–3195.
(9) Morrall, J. P.; Powell, C. E.; Stranger, R.; Cifuentes, M. P.;
Humphrey, M. G.; Heath, G. A. J. Organomet. Chem. 2003, 670, 248–255.
(10) The cell is available from BASi, W. L., IN 47906.
(1) Hawkridge, F. M. In Laboratory Techniques in Electroanalytical
Chemistry; Kissinger, P. T., Heineman, W. R., Eds.; Marcel Dekker, Inc.:
New York, 1996, pp 280-289.
(2) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877–910.
(3) Heath, G. A.; Mcgrady, J. E.; Martin, R. L. J. Chem. Soc., Chem.
Commun. 1992, 1272–1274.
(4) Powell, C. E.; Cifuentes, M. P.; Morrall, J. P.; Stranger, R.;
Humphrey, M. G.; Samoc, M.; Luther-Davies, B.; Heath, G. A. J. Am.
Chem. Soc. 2003, 125, 602–610.
r
pubs.acs.org/IC
Published on Web 09/13/2010
2010 American Chemical Society

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9591
assemblies, (ii) influence of thin-layer resistance on the
accuracy of the applied potential and the resulting need to
change the applied potential slowly (increasing the chance of
decomposition of electrode products), and (iii) other pro-
blems such as non-reproducible results because of the non-
trivial assembly (and leakage) of the cell for the inexperienced
researcher, limit the easy use of these OTTLE cells.
electrolyte NBu4PF6 was obtained from Acros Chemicals,
and was recrystallized from boiling ethanol and dried at 100 ꢀC
before use. Ferrocene was obtained from Acros Chemicals
and was sublimed before use. Mn2(CO)10 was obtained from
Alfa-Aesar and used as received. Literature methods were
used to prepare H2T(p-OMe)PP,18 (T(p-OMe)PP)Co,19 (T(p-
OMe)PP)Co(NO),20 and fac-(dppe)Mn(CO)3Br21 (T(p-OMe)-
PP=5,10,15,20-tetra(p-methoxyphenyl)porphyrinato dianion;
dppe = diphenylphosphinoethane).
A new spectroelectrochemical method that overcomes, in
part, the problems normally associated with OTTLE cells
was reported in 1996.11 Here, the authors made use of a
commercial fiber-optic infrared (IR) reflectance instrument
to monitor the progress of low-temperature bulk electrolysis
experiments, and used this method to generate and charac-
All calculations were performed with the Amsterdam
Density Functional (ADF) programfrom ScientificComput-
ing and Modeling (SCM; Netherlands).22,23 A TZ2P basis of
Slater orbitals was employed for all calculations. The geometry
optimizations used medium size effective core potentials and
the GGA:BLYP XC functional. The spectral calculations
(done at the optimized geometries) employed the SOAP24
model XC potential and used the Davidson Method25,26 to
determine the low lying excited states and oscillator strengths.
Simultaneous Chronoabsorptometry and Chronoamperometry
of Ferrocene. Visible-NIR data were collected under anaerobic
(dinitrogen atmosphere) conditions using an Ocean-Optics USB
2000 detector (360-1000 nm), an Ocean Optics LS-1-tungsten
lamp, and a T-300-dip-probe. The absorbance spectrum of a
5.59 mM solution of ferrocene in 0.10 M NBu4PF6/MeCN was
recorded using the same concentration of a blank support
electrolyte solution as reference. The path length was 1.00 cm.
The fiber-optic spectroelectrochemical cell configuration was
near-identical to the one previously described for mid-IR experi-
ments.12 Specifically, the light from the dip probe was directed
through 2-3 mm of a solution of ferrocene (ca. 3.0 mM) in 0.10 M
NBu4PF6/MeCN onto a polished 3.0 mm BAS Pt electrode. This
configuration allows the beam to reflect off of the electrode back
into the T-300-dip-probe and to the detector. A Pt wire was used as
the auxiliary electrode. For MeCN solutions, a bare Ag wire was
used as a quasi-reference electrode; for other solvents, a Ag wire
electrocoated with AgCl was used as the reference electrode.
The Ocean Optics OOIBASE or Spectrasuite program was set to
record and save spectra automatically upon receiving trigger
signals (external software trigger). A 10-15 ms integration time
was used.
terize reactive redox products in non-aqueous solvents.11
A
modification of this method was reported in which the mirror
in the transmission head was replaced by a Pt disk electrode
that served as a reflecting mirror, allowing for the IR spectral
identification (by difference spectral analysis) of electroche-
mically generatedredox products at theelectrode surfaceon a
typical cyclic voltammetry time scale.12 This allowed for the
collection of spectra at shorter reaction times than were
possible during bulk electrolysis experiments (e.g., collection
of spectra during a single cyclic voltammetry scan at scan rates
up to 0.4 V/s),13 and allowed a simple modification to
collect spectra of reactive species generated at variable
(low) temperature.14,15 In using this modified method, how-
ever, information on the relationship between current passed
and the build-up of electrode products was lost.
We have developed a new method to recover quantitative
spectral information on redox products generated at elec-
trode surfaces. Specifically, we used chronoamperometry and
chronoabsorptometry simultaneouslyto recover quantitative
spectral information on the build-up of redox products at or
near the electrode surface. Using this method, we can obtain
the UV-vis spectra of electrode products without the use of
arbitrary spectral subtraction such as that commonly used
in the manual generation of difference (product-minus-
reactant) spectra. We use mathematics to unambiguously
determine plots of absorptivity (ε, extinction coefficient)
versus wavelength of the electrode product(s). This method
allows us to spectrally characterize electrode products with
half-lives of ∼0.3 s or longer. In this article, we demonstrate
proof-of-concept for this relatively simple new method and
use it to determine the spectral characteristics of a previously
invoked but elusive electrogenerated intermediate along a
complex reaction pathway. We show that this new method is
general and applicable to a wide variety of chemical systems.
(18) Adler, A. D.; Longo, F. R.; Finarelli, J. D.; Goldmacher, J.; Assour,
J.; Korsakoff, L. J. Org. Chem. 1967, 32, 476.
(19) Adler, A. D.; Longo, F. R.; Kampas, F.; Kim, J. J. Inorg. Nucl.
Chem. 1970, 32, 2443–2445.
(20) Richter-Addo, G. B.; Hodge, S. J.; Yi, G.-B.; Khan, M. A.; Ma, T.;
Caemelbecke, E. V.; Guo, N.; Kadish, K. M. Inorg. Chem. 1996, 35, 6530–
6538 [Erratum: 1997, p 2696].
(21) Bond, A. M.; Grabaric, B. S.; Grabaric, Z. Inorg. Chem. 1978, 17,
1013–1018.
(22) te Velde, G.; Bickelhaupt, F. M.; van Gisbergen, S. J. A.; Fonseca
Guerra, C.; Baerends, E. J.; Snijders, J. G.; Zeigler, T. J. Comput. Chem.
2001, 22, 931–967.
ꢀ
(23) Baerends, E. J.; Autschbach, J.; Bashford, D.; Berces, A.; Bickelhaupt,
F. M.; Bo, C.; Boerrigter, P. M.; Cavallo, L.; Chong, D. P.; Deng, L.;
Dickson, R. M.; Ellis, D. E.; van Faassen, M.; Fan, L.; Fischer, T. H.;
Fonseca Guerra, C.; Ghysels, A.; Giammona, A.; van Gisbergen, S. J. A.;
Experimental Section
All solvents were dried under appropriate conditions and
freeze-pump-thawdegassedbeforeuse.16,17 Thesupporting
€
Gotz, A. W.; Groeneveld, J. A.; Gritsenko, O. V.; Gruning, M.; Harris,
F. E.; van den Hoek, P.; Jacob, C. R.; Jacobsen, H.; Jensen, L.; van Kessel,
G.; Kootstra, F.; Krykunov, M. V.; van Lenthe, E.; McCormack, D. A.;
Michalak, A.; Mitoraj, M.; Neugebauer, J.; Nicu, V. P.; Noodleman, L.;
Osinga, V. P.; Patchkovskii, S.; Philipsen, P. H. T.; Post, D.; Pye, C. C.;
Ravenek, W.; Rodriguez, J. I.; Ros, P.; Schipper, P. R. T.; Schreckenbach,
G.; Seth, M.; Snijders, J. G.; Sola, M.; Swart, M.; Swerhone, D.; te Velde,
G.; Vernooijs, P.; Versluis, L.; Visscher, L.; Visser, O.; Wang, F.; Wesolowski,
T. A.; van Wezenbeek, E.; Wiesenekker, G.; Wolff, S. K.; Woo, T. K.;
Yakolev, A. L.; Ziegler, T. http://www.scm.com; Vrijie Universiteit:
Amsterdam, The Netherlands, 2009.
(24) Gritsenko, O. V.; Schipper, P. R. T.; Baerends, E. J. Chem. Phys.
Lett. 1999, 302, 199–207.
(25) van Gisbergen, S. J. A.; Snijders, J. G.; Baerends, E. J. Comput. Phys.
Commun. 1999, 118, 119–138.
(26) Wang, F.; Ziegler, T. Mol. Phys. 2004, 102, 2585–2595.
(11) Shaw, M. J.; Geiger, W. E. Organometallics 1996, 15, 13–15.
(12) Shaw, M. J.; Henson, R. L.; Houk, S. E.; Westhoff, J. W.; Jones,
M. W.; Richter-Addo, G. B. J. Electroanal. Chem. 2002, 534, 47–53.
(13) Shaw, M. J.; Afridi, S. J.; Light, S. L.; Mertz, J. N.; Ripperda, S. E.
Organometallics 2004, 23, 2778–2783.
(14) Carter, S. M.; Lee, J.; Hixson, C. A.; Powell, D. R.; Wheeler, R. A.;
Shaw, M. J.; Richter-Addo, G. B. Dalton Trans. 2006, 1338–1346.
(15) Zahran, Z. N.; Shaw, M. J.; Khan, M. A.; Richter-Addo, G. B. Inorg.
Chem. 2006, 45, 2661–2668.
(16) Shriver, D. F.; Drezdzon, M. A. The Manipulation of Air-Sensitive
Compounds, 2nd ed.; Wiley-Interscience: Toronto, 1986.
(17) Wayda, A. L.; Darensbourg, M. Y. Experimental Organometallic
Chemistry: A Practicum in Synthesis and Characterization; American Chemi-
cal Society: Washington, D.C., 1987.

9592 Inorganic Chemistry, Vol. 49, No. 20, 2010
Shaw et al.
Electrochemical Data Processing and Methodology. For the
redox reaction shown in eq 1, the integrated Cottrell equation
(eq 2)28 describes the
A þ ne- h B
ð1Þ
rffiffiffiffiffi
Dt
Q ¼ 2nFAeCAꢀ
þ Qdl þ nFAeΓA
ð2Þ
π
chronocoulometric charge that flows through an electrode as a
function of time after a potential is applied. This equation results
from the application of the Nernst equation and Fick’s laws of
diffusion to a planar electrode where linear diffusion is the only
means by which redox-active material is transported to the
electrode surface.29 The applied potential has to be such that
virtually all of the material that diffuses to the electrode surface
undergoes electron transfer on contact. In eq 2, the charge (Q) is
expressed as a function of time (t), the number of electrons (n),
Faraday’s constant (F ), the electrode area (Ae), the diffusion
coefficient of the species present in the bulk solution (D), and the
concentration of the redox species in the bulk solution which
undergoes electron transfer (e.g., C*A for species A). Q also
depends on a term due to adsorbed surface species (nFAeΓA,
where ΓA is the fraction of the surface covered) and double layer
charging (Qdl). For species that do not adsorb to the electrode
surface, the values of nFAeΓA and Qdl are very small, and these
processes are often complete within 50 ms of the application of
the pulse.30-32
Figure 1. Schematic representation of the spectroelectrochemical cell
(analyte volume of ∼15 mL). A photo of the cell is supplied in the
Supporting Information, Figure S1.
A Labview 8.0 program was written to control a National
Instruments USB-6251 interface so as to apply a potential to the
working electrode via the external input of an EG&G PAR
263A potentiostat, to record the current and potential analog
output of the potentiostat, and to apply digital trigger signals at
regular known intervals to the USB-2000 detector. Labview
programs developed for these application are posted on the
National Instruments Developer Zone Web site (http://decibel.
ni.com/content/people/roadchem?view=documents; accessed
9/1/10). In these experiments, the data collection rate was typically
set to 500 Hz, the potential was set either 100 or 300 mV positive
of ferrocene’s Eꢀ value (as determined by an initial cyclic
voltammogram), and 20 spectra were recorded at 0.200-0.500 s
intervals. The files and the current versus time data were saved
for subsequent processing.
A second Labview 8.0 program was written which performs
an optional “edge correction” on the current versus time data to
account for radial-edge diffusion at the electrode.27 The current-
time data was corrected, then integrated, and the charges that
corresponded to times when the spectra were recorded were
noted. At every wavelength used to construct the spectrum, a
linear least-squares analysis was performed which yielded a
slope that was then converted to an absorptivity (extinction
coefficient) value. The standard deviation of each absorptivity
at each wavelength was also calculated.
Equation 3 shows the relationship of the absorbance (AbsB(λ,t)
)
of species B generated at an electrode surface with time,
assuming that the values of D for the various species present
are similar.28 This equation is the basis of chronoabsorpto-
metry at wavelengths where the starting material does not
interfere.
rffiffiffiffiffi
Dt
AbsBðλ,tÞ ¼ 2εBðλÞCAꢀ
ð3Þ
π
A similar equation can also be derived for the variation in the
starting material’s absorbance with time. If adsorption and
double layer charging are small, these two equations can be
combined with eq 2 to yield eq 4 (see Supporting Information).
Q
Measurements on (T(p-OMe)PP)Co, (T(p-OMe)PP)Co(NO),
and (dppe)Mn(CO)3Br. Beer’s Law plots of these complexes were
obtained using a 1.0 cm cuvette held in an Ocean-Optics CUV-
ALL-UV 4-way Cuvette Holder, using an Ocean-Optics USB 4000
fiber-optic spectrophotometer (200-850 nm) and a LS-1 tungsten
lamp. Simultaneous chronoabsorptometry/chronoamperometry
measurements were performed as described above, but employing
the spectroelectrochemical cell sketched in Figure 1 and shown in
Supporting Information, Figure S1. This cell was custom-made
by Chemglass (www.chemglass.com), and assigned part numbers
SIUE-0803-041MS (cell body, specify closed bottom) and SIUE-
0803-042MS (cell cap).
Electron Paramagnetic Resonance (EPR) Spectroscopy. A
commercial EPR spectroelectrochemical cell (Wilmad) was used
for in situ EPR spectroscopic measurements. Typically, a 1.0 M
solution of analyte in 0.5 M NBu4PF6 in CH2Cl2 was injected
into the cell in a drybox. The electrodes were positioned, and the
cell sealed and secured in the EPR cavity. A potential sufficient
to cause current to flow through the cell was applied, and an
EPR spectrum was recorded.
ΔAbs
¼
ðεBðλÞ - εAðλÞÞ
ð4Þ
ðλ,tÞ
nFAe
Equation 4 shows how the change in absorbance (ΔAbs(λ,t)) in
a set of difference spectra varies with the charge Q passed
through the electrode. From eq 4, it is evident that a plot of
change in absorbance versus charge at any specific wavelength
for a set of difference spectra should yield a straight line with
slope of (εB(λ) - εA(λ))/nFAe. Thus, if εA(λ), n, and Ae are known,
then the value of εB(λ) can be extracted. This calculation may be
repeated at each wavelength to obtain the complete absorptivity
versus wavelength spectrum of the redox product. This approach
renders quantitative information about electrode products for
reactions which exhibit complete chemical reversibility, or complete
chemical irreversibility.
Absorptivity information is also available for electrode reac-
tions that exhibit partial chemical reversibility. Consideration of
the chronoabsorptometry equations for an EC electrode me-
chanism (eq 5) with a first order or pseudo-first order chemical
(29) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals
and Applications, 2nd ed.; Wiley: New York, 2001.
(27) Heinze, J. J. Electroanal. Chem. 1981, 124, 73–86.
(30) Schwarz, W. M.; Shain, I. J. Phys. Chem.-Us 1965, 69, 30–40.
(31) Christie, J. H.; Osteryoung, R. A.; Anson, F. C. J. Electroanal. Chem.
1967, 13, 236–244.
(28) Kissinger, P. T.; Heineman, W. R. In Laboratory Techniques in
Electroanalytical Chemistry, 2nd ed. (revised and expanded); Kissinger, P. T.,
Heineman, W. R., Eds.; Marcel Dekker, Inc: New York, 1996; pp 57-68.
(32) Anson, F. C.; Osteryoung, R. A. J. Chem. Educ. 1983, 60, 293–296.

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9593
0
step yields eq 6 (derived in Supporting Information). The γ(k ,t)
function (eq 7) describes
eqs 4 and 6) include the following: (i) there is no arbitrary
spectral subtraction of the starting material’s spectral
features to enhance the spectrum of the intermediate,
(ii) the approach is independent of knowledge of the
accurate path length of the electrochemical cell, and
(iii) quantitativeknowledgeof the absorptivityvaluesfacil-
itates assignment of the spectra features for the electrode
species. In previous work by others, an approach using
eq 4 was applied by measuring εB(λ) values one at a time
over a given spectral window;35,36 this approach proved
to be quite tedious, requiring >60 manual trials at each
separate λ. To the best of our knowledge, eq 6 has not
been used previously to yield absorptivity versus wave-
length data from a single experiment. Given the current
availability of inexpensive diode-array-detector-based
fiber-optic spectrometers, fast analog/digital I/O devices,
and fast computers, it is now possible to use techniques
based on these equations to generate the UV-vis spec-
trum of a product generated at an electrode surface
during a chronoabsorptometric experiment at multiple
wavelengths simultaneously.
f
kf
s
f
A þ ne- h B
C
ð5Þ
ð6Þ
0
γ
ΔAbs
1
ðkf ,tÞ
nFAe
ðλÞ
¼
ðεBðλÞ - εCðλÞÞ þ
ðεCðλÞ - εAðλÞÞ
Q
nFAe
n
¥
0
X
0t
ðkf tÞ
γðk ,tÞ ¼ e- k
ð7Þ
f
0
f
ð2n þ 1Þn!
n ¼ 0
0
the variation of species B and C with time. Since γ(k ,t) converges
f
within a few terms (10 iterations are convenient), it can be easily
calculated. Rate constants kf0 can be obtained by voltammetry.
From eq 6, it can be deduced that plots of (ΔAbs(λ))/Q versus
0
γ
(k ,t) should yield linear plots where the slope gives the differ-
f
ence between the absorptivity of the intermediate B and the
product C, and the intercept yields the difference between the
absorptivity of the product C and the starting material A. Thus,
if the starting material’s absorptivity is known, then the values
of εC(λ) and εB(λ) can be obtained from the intercept and slope,
respectively, when both are multiplied by nFAe. Repetition at
each wavelength should then yield complete absorptivity versus
wavelength curves for the intermediate and product. However,
it is important to note that the quality of the final data depends
on the quality of the ελ values known for the starting material
since the method relies on differences in absorptivity. Thus, a
Beer’s Law plot of the starting material should be performed to
ascertain absorptivity values in the concentration range where
data collection is performed, and the spectroelectrochemistry of
compounds with non-linear Beer’s law plots should be ap-
proached with caution.
The electrochemical cell is simple in design and is
sketched in Figure 1 (see also Supporting Information,
Figures S1 and S2). In this configuration, a waveguide
brings light through the analyte solution to and from the
electrode surface, and provision is made for stirring. The
final equation utilized in this proof-of-concept demon-
stration is given by eq 4 (Experimental Section, and see
Supporting Information for derivation).35 Note that this
eq 4 does not include terms for concentration or path
length as variables;
Q
ΔA
¼
ðεBðλÞ - εAðλÞÞ
ð4Þ
ðλ,tÞ
Results and Discussion
nFAe
We demonstrate the use of new spectroelectrochemical
methodology employing simultaneous chronoabsorptome-
try and chronoamperometry for the facile detection of
unstable redox products not previously detected by current
spectroelectrochemical methods. This section is divided into
two parts: (i) development of the method and proof-of-
concept using the well-studied ferrocene/ferrocenium couple
and application to the detection of a porphyrin π-radical
cation, and (ii) application of the method to generate and
characterize, for the first time, an unstable intermediate in a
redox-induced fac to mer reaction pathway for a coordina-
tion compound of manganese.
Proof of Concept. The methodology employed in this
work is based on established electrochemical principles.
The theory33,34 and application35,36 of chronoabsorpto-
metry have been reported. In our continuing efforts to
improve the spectroelectrochemical methodology for the
ready generation and detection of reaction intermediates,
we have employed the fundamental mathematics of
chronoabsorptometry to develop readily accessible in-
strumentation to generate UV-vis absorption spectra of
intermediates not observable by current techniques that
employ OTTLE or other spectroelectrochemical cells.
The advantages of this new methodology (e.g., using
these terms, however, are needed for the determination of
the required Beer’s law plots (see later).
Ferrocene. We chose ferrocene (Cp2Fe; Cp=(η5-C5H5),
cyclopentadienyl) as the prototype to demonstrate this
method as it is known to undergo a well-characterized
one-electron oxidation in various solvents (eq 8),37-39
and is a widely accepted electrochemical redox standard.2
Cp2Fe h ½Cp2Feꢁþ þ e-
ð8Þ
Importantly, the UV-vis spectra of authentic samples
of the neutral ferrocene and its oxidized derivative of eq 8
have been reported previously.40,41
Figure 2 shows the final plots of absorptivity (ε) versus
wavelength for both Cp2Fe and [Cp2Fe]þ in acetonitrile
obtained from this work; the sample was kept under
rigorously anaerobic conditions so as to avoid follow-
up reactions with oxygen. The spectrum for the starting
Cp2Fe compound was obtained from Beer’s law plots,
while that for [Cp2Fe]þ was obtained after a four-second
data collection experiment using our new methodology
(from 20 spectra; each 15 ms integration time at 0.2 s
(37) Tsierkezos, N. G. J. Solution Chem. 2007, 36, 289–302.
(38) Crooks, R. M.; Bard, A. J. J. Electroanal. Chem. 1988, 243, 117–131.
(39) Kadish, K. M.; Ding, J. Q.; Malinski, T. Anal. Chem. 1984, 56, 1741–
1744.
(40) Prins, R. J. Chem. Soc. D: Chem. Commun. 1970, 280b–281.
(41) Scott, D. R.; Becker, R. S. J. Organomet. Chem. 1965, 4, 409–411.
(33) Li, C.-Y.; Wilson, G. S. Anal. Chem. 1973, 45, 2370–2380.
(34) Evans, D. H.; Kelly, M. J. Anal. Chem. 1982, 54, 1727–1729.
(35) Jones, D. H.; Hinman, A. S. Can. J. Chem. 1996, 74, 1403–1408.
(36) Jones, D. H.; Hinman, A. S. Can. J. Chem. 2000, 78, 1318–1324.

9594 Inorganic Chemistry, Vol. 49, No. 20, 2010
Shaw et al.
Figure 2. Plots of absorptivity versus wavelength for (A) ferrocenium
ion (0) generated at an electrode surface over a four-second period from
electrooxidation of 3.4 mM ferrocene in 0.10 M NBu4PF6/MeCN at
298 K, and (B) ferrocene ([).
intervals). Notably, the value of λmax at 619 nm and
associated absorptivity of 450 cm-1 are similar to reported
literature values.40,41 The processes involved in computing
the spectrum for [Cp2Fe]þ are shown schematically in
Figure 3. We used chronoamperometry (Figure 3A,B)
to determine the current response as a function of time,
and the plot obtained was as expected from the Cottrell
equation. Simultaneously, spectra were collected every 0.2
s in the 380-800 nm range along the decay curve for the
data shown in Figure 3B; in general, however, we sampled
every 0.5 s for the Co and Mn systems (see later). Correc-
tion for edge diffusion for small electrodes is necessary to
obtain accurate values of Q by integration of current
data.19 To perform edge correction for chronoamperome-
try, we applied the method of Heinze (Supporting Infor-
mation, Figure S3)27 noting the near-perfect fit of the
correction factor to eq S24 (see Supporting Information).
The edge correction requirement introduces two readily
obtained variables into the experiment, namely, the geo-
metric radius of the electrode (easily obtained) and the
diffusion coefficient of the compound being studied
Figure 3. (A) Potential versus time for a typical chronoamperometry/
chronoabsorptometry experiment. (B) Current response versus time
recorded during the potential step shown for 3.9 mM ferrocene in 0.1 M
NBu4PF6. Times at which spectra were recorded are indicated by ꢂ’s.
(C) Change in absorbance vs charge recorded at 619 nm ([) with least-
squares best fit (solid line). Charge was obtained by integration of
current after applying a correction for radial diffusion as described by
Heinze (See Supporting Information, Figure S3).19
constant over multiple experiments (i.e., if we did not
redetermine the exact value of electrode area), (ii) we did
not fix the concentrations of analytes to be the same over
the trials, and (iii) we did not take special precautions to
keep the path length identical over several runs. We find
that larger deviations occur where the spectrometer has
lower sensitivity (i.e., <400 nm and >900 nm) mostly
because of the emission characteristics of the tungsten
halide source, and that brief stirring (e.g., for 15 s) between
runs is necessary to ensure reproducibility.
(e.g., Dꢀ (ferrocene) = 2.24 ꢂ 10-5 cm2 s-1 37,39
;
can be
extracted from chronoamperometry data collected during
the experiment). From Figure 3B and the edge correction
data, we obtained values of Q by integration (note that I =
Q/t) along the decay trace.
In summary, we have used the ferrocene-ferrocenium
couple to demonstrate a robust calculation of absorptiv-
ity over a range of wavelengths for electrogenerated
products at an electrode surface. The four-second experi-
mental time frame is comparable to many cyclic voltam-
metry time scales, and the spectra obtained are essentially
independent of path length and concentration. Further,
the method is simple, reliable with no ambiguity over the
exact value of applied potential (c.f. the OTTLE cell
limitation), and yields quantitative spectral information
on electrode products.
Cobalt Nitrosyl Porphyrins. The electrochemistry of
cobalt nitrosyl porphyrins has been reported; these com-
pounds undergo multiple electron transfer events.20,42-45
The square-wave voltammogram for the oxidation of
(T(p-OMe)PP)Co(NO) is shown in Figure 4A, and reveals
A resulting plot of change in total absorbance (i.e., for
starting material and product; see eqs S1-S10 in the
Supporting Information) versus corrected Q for a repre-
sentative λ is shown in Figure 3C; similar plots are
determined for each λ sampled (i.e., >2500 separate
wavelength measurements). From eq 4 and Figure 3C,
the slope is given by Δελ/nFAe for a particular λ, where
Δελ is equal to ελ(B) - ελ(A).
Thus, the value of ελ(B) (i.e., absorptivity of product at
a particular wavelength) can be obtained from knowledge
of the slope of the line in Figure 3C (Δελ/nFAe; if the
electrode area (Ae) and number of electrons (n) are
known) and by calculating Δελ(A) of the starting com-
pound using Beer’s law. Importantly, the computer anal-
ysis of the data collected after this 4 s experimental run to
give the spectrum of [Cp2Fe]þ shown in Figure 2 is
obtained in <1 s on a standard modern computer! We
noted previously that the final eq 4 used was independent
of concentration and path length. Thus, there is no need
for the reproducible placement of the electrode(s) and
waveguide with respect to each other. Indeed, we find that
we obtain good reproducibility for the value of ελ(B) (to
within 6%) even if (i) the electrode area is assumed
(42) Kelly, S.; Lanc-on, D.; Kadish, K. M. Inorg. Chem. 1984, 23, 1451–
1458.
(43) Kadish, K. M.; Mu, X. H.; Lin, X. Q. Inorg. Chem. 1988, 27, 1489–
1492.
(44) Kini, A. D.; Washington, J.; Kubiak, C. P.; Morimoto, B. H. Inorg.
Chem. 1996, 35, 6904–6906.
(45) Kadish, K. M.; Ou, Z.; Tan, X.; Boschi, T.; Monti, D.; Fares, V.;
Tagliatesta, P. J. Chem. Soc., Dalton Trans. 1999, 1595–1601.

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9595
Figure 5. Absorptivity vs wavelength for (A) [(T(p-OMe)PP)Co(NO)]þ
(0), generated at an electrode surface over a 10 s period from 0.22 mM
(T(p-OMe)PP)Co(NO) in 0.10 M NBu4PF6/MeCN at 298 K, and (B)
(T(p-OMe)PP)Co(NO) ([).
Figure 4. Square wave voltammograms of (A) 0.22 mM (T(p-OMe)-
PP)Co(NO) and (B) 0.13 mM (T(p-OMe)PP)Co in CH2Cl2/0.1 M
NBu4PF6 at 298 K. Arrows indicate the potentials applied for the
chronoamperometry/chronoabsorptometry method used to obtain
spectroscopic data. Scan parameters: 2 mV pulse width, 50 mV pulse
height, 50 Hz.
π-radical cations.46,47 The related [(OEP•þ)Co(NO)] is
known to display visible bands in the 520-570 nm region.48
redox processes at Eo = 0.46 and 0.84 V versus the
0
ðTðp-OMeÞPPÞCoðNOÞ h ½ðTðp-OMeÞPP•þÞCoðNOÞꢁ þ e-
ð10Þ
In contrast, the absorptivity plot of the non-nitrosyl-
containing (T(p-OMe)PP)Co compound does not show a
new band at ∼650 nm upon oxidation (Supporting In-
formation, Figure S4). The absence of this feature is
consistent with (T(p-OMe)PP)Co undergoing a metal-
based oxidation rather than a porphyrin-based oxidation
under our conditions (eq 11).43
Fc/Fcþ couple consistent with that reported previously
for this compound.20,44 For comparison, the square-wave
voltammogram of the parent (0T(p-OMe)Co compound is
shown in Figure 4B, with Eo values of 0.31, 0.59, and
0.75 V versus the Fc/Fcþ couple.
The ∼150 mV positive shift of the first oxidation of
(T(p-OMe)PP)Co(NO) relative to that of (T(p-OMe)-
PP)Co is not unexpected, and this magnitude of shift has
been observed for other cobalt nitrosyl porphyrins as
well.43,45
We were interested in the product of the first oxidation
of (T(p-OMe)PP)Co(NO) (eq 9; site of oxidation not
indicated) to better define the location of electron
ðTðp-OMeÞPPÞCoII h ½ðTðp-OMeÞPPÞCoIIIꢁþ þ e-
ð11Þ
ðTðp-OMeÞPPÞCoðNOÞ h ½ðTðp-OMeÞPPÞCoðNOÞꢁþ þe-
ð9Þ
Taken together, the results of eqs 10 and 11 are consistent
with similar results published for other cobalt porphyrin
compounds.20,42,43,45 However, the spectral information
obtained in Figure 5A for the [(T(p-OMe)PP•þ)Co(NO)]
π-radical cation from this short time scale absorptivity
method avoids spectral contamination from byproducts
generated at 298 K when longer time scales are utilized.
To demonstrate the effects of longer time scales and the
effects of thin-layer resistance on the integrity of the
initially generated (T(p-OMe)PP•þ)Co(NO) compound,
we used in situ EPR spectroelectrochemical methodology
employing a commercial air-free sample cell. We note that
Kadish and co-workers have reported the in situ EPR spec-
trum of the oxidation product of the related (TPP)Co(NO)
complex in dichloromethane,43 without obvious side-
products. The 293 K EPR spectrum of the oxidation
product of (T(p-OMe)PP)Co(NO), electrogenerated in
an air-free commercial thin-layer EPR cell, is shown in
Figure 6A; the analogous spectrum of electrogenerated
[(T(p-OMe)PP)Co]þ is shown in Figure 6B. As can be
deduced from a comparison of these spectra, the longer
time scales (up to 600 s) typically utilized in the EPR
spectroelectrochemistry experiments can result in the
unwanted loss of the NO fragment from the singly
oxidized derivative of the cobalt nitrosyl porphyrin to
give [(T(p-OMe)PP)CoIII]þ. Such a loss of the NO frag-
ment has been described by Kadish and co-workers for
other (por)Co(NO) compounds.20,42,43,45
removal (e.g., metal- or porphyrin-based oxidation) on
the short time-scales now available to us. Using the
procedures described in the previous section, we obtained
plots of ε versus λ for the product obtained after the first
oxidation (Figure 5). The electrode was set at a potential
just positive of the first oxidation couple for the cobalt
complex in Figure 4A (at þ0.6 V versus the ferrocene-
ferrocenium couple). The diffusion coefficients used to
perform the edge correction for (T(p-OMe)PP)Co(NO)
and (T(p-OMe)PP)Co were found to be 1.15 ꢂ 10-5 cm2
s-1 and 1.33 ꢂ 10-5 cm2 s-1, respectively. The resulting
plot is shown in Figure 5A and is compared with that of
the parent neutral (T(p-OMe)PP)Co(NO).
The neutral (T(p-OMe)PP)Co(NO) displays a Soret
band at 420 nm and a visible band at 541 nm (ε = 1.95 ꢂ
104 M-1 cm-1). The oxidized product shows the presence of
a new absorption band centered at 658 nm (ε = 3.2 ꢂ 104
M-1 cm-1). This new band is attributed to the formation of
the [(T(p-OMe)PP•þ)Co(NO)] π-radical cation species at
the electrode surface (eq 10), as this band is in the ∼600-
700 nm region generally associated with tetraarylporphyrin
(46) Gross, Z.; Barzilay, C. Angew. Chem., Int. Ed. Engl. 1992, 31, 1615–
1617.
(47) Fuhrhop, J.-H. Struct. Bonding (Berlin) 1974, 18, 1–67.
(48) Fujita, E.; Chang, C. K.; Fajer, J. J. Am. Chem. Soc. 1985, 107, 7665–
7669.
In summary, although the first oxidation product of
(T(p-OMe)PP)Co(NO) has been proposed to be the

9596 Inorganic Chemistry, Vol. 49, No. 20, 2010
Shaw et al.
Figure 8. EC mechanism for the oxidation of fac-[(dppe)Mn(CO)3Br]
which results in the formation of the mer-[(dppe)Mn(CO)3Br]þ cation.
Figure 6. X-Band EPR spectra of the oxidized products of (A) (T(p-OMe)-
PP)Co(NO) and (B) (T(p-OMe)PP)Co, generated by electrochemical in
situ oxidation in 0.5 M NBu4PF6/CH2Cl2 at 293 K. The top figure shows
spectral contamination by the species in (B).
Figure 9. (A) Change in absorbance data recorded at 517 nm during the
simultaneous chronoamperometry/chronoabsorptometry experiment
plotted versus corrected charge (Qcorr) as developed in “method 1” (R2 =
0.9739). (B) Change in absorbance divided by Qcorr, plotted against the
Figure 7. Cyclic voltammogram of 1.0 mM fac-[(dppe)Mn(CO)3Br] in
0.1 M NBu4PF6/CH2Cl2 at 0.40 V/s and 298 K. Arrow indicates initial
direction of scan. Note the reversibility of the merþ/0 feature revealed
upon the subsequent scan.
dimensionless parameter γ(k0f,t) as defined in “method 2” (R2
=
0.9229).
π radical cation [(T(p-OMe)PP•þ)Co(NO)] on the basis
of IR spectroscopy,20,44 this work confirms its assignment
by UV-vis spectroscopy. The successful demonstration
of the use of the simultaneous chronoamperometry/
chronoabsorptometry to provide proof-of-concept for
the determination of UV-vis spectra of redox products
at an electrode surface allowed us to use this method to
determine the UV-vis spectrum of an intermediate which
has never before been characterized.
Application for the Detection of Electrode Intermedi-
ates. The compound fac-[(dppe)Mn(CO)3Br] (referred to
herein as fac) contains a Mn-tricarbonyl moiety in a facial
arrangement. Such fac-[(L2)Mn(CO)3Br] (L2 = chelating
diphosphines/bipyridines) are known to undergo fac
to mer isomerization reactions following oxidation.21
These reactions are of renewed interest because structural
changes that follow electron-transfer reactions are of
fundamental importance to applications such as long-
term digital memory storage.
We thus chose the fac-[(dppe)Mn(CO)3Br] compound
as our prototype for the study of redox-induced con-
figurational lability in transition metal complexes. The
cyclic voltammogram of fac-[(dppe)Mn(CO)3Br] is shown
in Figure 7.
The compound undergoes a well-behaved one-electron
oxidation process with scan rate-dependent chemical re-
versibility, consistent with the formation of the oxidation
product facþ. This initially generated species converts to a
daughter product that shows a redox feature at a less
positive potential. These CV features are consistent with
isomerization of facþ to the corresponding merþ as detailed
in Figure 8.21
A value of k of 1.5 s-1 for the facþ f merþ step in
acetonitrile was determined previously by Bond using
chronoamperometry;21 a similar value of 2.0 s-1 for this
process in CH2Cl2/NBu4PF6 was determined by us by
digital simulation (DigiElch) of background-subtracted
square wave voltammetry data.
Importantly, the scan rate-dependent chemical revers-
ibility of the process fac f facþ f merþ implies the
existence of more than two species at the electrode surface
(i.e., a classic EC electrode mechanism). Thus, although
the absorptivity method described in the previous section
can give information on the absorptivity of merþ, it is not
entirely appropriate for this particular sequence of reac-
tions because it does not take into account the time-
dependent absorbance of the additional species on the
electrode surface. The solution to this issue is fortunately
straightforward. The data collection strategy is exactly
the same as that described in the previous (proof-of-
concept) section, but the analysis of the data is performed
according to eq 6 for the EC case (c.f., eq 4). In eq 6, εA, εB,

Article
Inorganic Chemistry, Vol. 49, No. 20, 2010 9597
Figure 11. Geometry-optimized structures of fac and facþ from DFT
calculations. The data in brackets are from the experimental crystal
structure.50
longer time scales (g30 s) for data collection). The EPR
spectrum of the related fac-[(dppm)Mn(CO)3Cl]þ (dppm =
1,1-bis(diphenylphosphino)methane) has been reported
by Bond;49 this cation was generated from nitrosonium
oxidation at low temperature of the neutral precursor in
acetonitrile, under conditions where the sample tempera-
ture was briefly raised to 202 K and refreezing the sample
to 77 K. The optical method described herein is expected
to yield complementary solution-phase information to
such low-temperature frozen glass EPR studies.
Figure 10. Absorptivity versus wavelength for (A) fac-[(dppe)Mn(CO)3Br]
(solid line), (B) fac-[(dppe)Mn(CO)3Br]þ (O) with data smoothed using a
median filter, and (C) mer-[(dppe)Mn(CO)3Br]þ (0). The cationic species
were generated at an electrode surface over a 10 s period from 1.0 mM
fac-[(dppe)Mn(CO)3Br] in 0.10 M NBu4PF6/CH2Cl2 at 298 K. It is
possible that the low energy shoulders in trace (B) may be artifacts of
curve smoothing. See the Supporting Information for the raw spectral
data for (B) prior to smoothing.
and εC represent the absorptivities of fac, facþ, and merþ,
respectively.
The correct calculation of the spectrum of B (λmax
623 nm; ε = 2.9 ꢂ 103 M-1 cm-1) in Figure 10 requires
knowledge of the rate constant k for the facþ to merþ
isomerization process; plots of the effects of different
arbitrary values of k on the shape of this spectrum are
shown in Supporting Information, Figure S5. This data
has been smoothed with a median filter, the effect of
which can be seen by comparing Supporting Information,
Figures S6 and S7. We find from our DFT calculations
(next section) that the λmax and absorptivity values for the
facþ cation are consistent with the smaller gap between
the highest occupied molecular orbital (HOMO) and the
lowest unoccupied molecular orbital (LUMO) expected
relative to either the neutral fac or the cationic merþ
complexes.
Theoretical Calculations. To verify our spectral assign-
ment for the facþ intermediate, we performed density
functional theory (DFT) calculations to obtain optimized
geometries for the four species fac, facþ, merþ, and mer
(see the Supporting Information, Table S1 and Figures
S12-S15). The geometry optimization of fac is in close
agreement with the geometrical data obtained for this
compound from a single-crystal X-ray structural analysis
reported previously (Figure 11).50 The X-ray crystal
structure of merþ has not been reported.
0
γ
ΔAbs
1
ðkf ,tÞ
nFAe
ðλÞ
¼
ðεBðλÞ - εCðλÞÞ þ
ðεCðλÞ - εAðλÞÞ
Q
nFAe
ð6Þ
This equation takes into consideration the absorbance
of an intermediate generated at the electrode surface. A
plot of ΔAbsλ/Q (i.e., the change in total absorbance at a
given wavelength divided by the corrected charge) versus
γ(k, t), where t represents the time at which the spectrum
was taken, yields a straight line with an intercept of
(ε(C)λ - ε(A)λ)/nFAe (Figure 9). From the value of this
intercept, and with ε(A)λ known, we can extract the
absorptivity of species C (merþ) without it being contami-
nated by the presence of other species. The diffusion
coefficient of fac used for the correction of radial diffusion
was determined to be 1.09 ꢂ 10-5 cm2 s-1 by chronoam-
perometry.
A unique advantage of this method is that it allows, for
the first time, the recovery of information about B (facþ)
from the slope of the line in Figure 9 so that we can
unambiguously characterize it spectrally although it may
be present only in very low concentration. Figure 10
shows the plots of ε versus λ for species A (fac), B (facþ),
and C (merþ). The spectra for species C (λmax 540 nm; ε =
2.8 ꢂ 103 M-1 cm-1) match the spectra reported by others
(ε540 = 2.1 ꢂ 103 M-1 cm-1).21 The spectrum for A (λmax
395 nm; ε395 = 6.8 ꢂ 102 M-1 cm-1), however, differs from
that previously reported (λmax 388 nm; ε388 = 1.3 ꢂ 103
M-1 cm-1). However, the latter literature values were
obtained using a 2 mM solution in a 1.0 mm OTTLE cell.41
The ε395 value for fac-[Br(CO)3(dppe)Mn] obtained in this
work results from a Beer’s law plot (R2 = 0.997) of five
different concentrations for solutions prepared in a drybox
in volumetric glassware measured using a calibrated 1.00
cm quartz cuvette.
The computational results of the facþ species were of
particular interest to us. Single-point calculations on the
geometry-optimized structures reveal that the energy of
the facþ species is 151.4 kcal/mol above that of the fac
species, and 4.6 kcal/mol abovethatoftheisomerizedmerþ
species. TheHOMO of the startingfac compound isshown
in Figure 12. It can be seen from this Figure that the
HOMO comprises contributions from a Mn-Br anti-
bonding interaction and a Mn-CO bonding interaction.
Thus, upon oxidation, the Mn-Br bond is expected to
become stronger and the Mn-CO bonds weaker. This is,
in fact, borne out by the bond distance changes obtained
Returning to Figure 10, we note that the UV-vis
spectral characterization of B has, however, never been
reported previously despite numerous studies of the processes
described by Figure 8 using OTTLE spectroelectrochemical
cells (note that the use of OTTLE cells generally requires
(49) Bond, A. M.; Colton, R.; Mccormick, M. J. Inorg. Chem. 1977, 16,
155–159.
(50) Beckett, M. A.; Brassington, D. S.; Coles, S. J.; Gelbrich, T.; Light,
M. E.; Hursthouse, M. B. J. Organomet. Chem. 2003, 688, 174–180.

9598 Inorganic Chemistry, Vol. 49, No. 20, 2010
Shaw et al.
merþ increases the HOMO-LUMO gap. Thus, λmax for
the short-lived facþ intermediate is expected to occur at
longer wavelengths than either the starting material
(Figure 10C) or the final product (Figure 10A).
For the easily observed neutral 18-electron species, the
calculated λmax values are within 9 nm of the experimental
values (Table 1).21,51 The 17-electron merþ cation is also
easily observed by UV-vis methods, and its calculated
λmax values are within 24 nm of experiment. This larger
variation is acceptable because the open-shell configura-
tion of the cation introduces more complexity to the
calculation. The calculations reproduce the single peak
expected for the neutral fac species and the two peaks
observed for both mer species. The excellent agreement
between the spectra and the calculations for the three
previously observed species suggests that the calculated
λmax value for facþ is also valid. The calculated and
experimental values for facþ obtained from our work
agree to within 27 nm, so these results serve to validate the
method used to obtain and smooth the spectrum of the
short-lived cationic facþ complex.
Figure 12. HOMO of fac-(dppe)Mn(CO)3Br.
Table 1. Experimental and Calculated λmax Values for [(dppe)Mn(CO)3Br]0/þ
experimental
max (nm)a
calculated
λ
max (nm)b
species
λ
fac-[(dppe)Mn(CO)3Br]
mer-[(dppe)Mn(CO)3Br]
fac-[(dppe)Mn(CO)3Br]þ
mer-[(dppe)Mn(CO)3Br]þ
395
398
418, 455 (sh)c
650
456 (sh), 564
424, 446d
623
442, 540
Conclusions
We have developed a rapid and efficient approach using
simultaneous chronoamperometry and chronoabsorptome-
try to obtain UV-vis spectral data on intermediates gener-
ated on electrode surfaces under normal cyclic voltammetry
time scales (0.05-1.5 V/s). We have demonstrated its applic-
ability for the generation of hitherto unreported spectra for
the short-lived fac-[(dppe)Mn(CO)3Br]þ intermediate that
forms during the redox-induced facfmer isomerization of
(dppe)Mn(CO)3Br. The instrumentation is relatively inex-
pensive. The absorptivities measured by the simultaneous
chronoabsorptometry/chronoamperometry methodology
are essentially independent of path length and of the con-
centration of the starting material. The method extracts
absorptivity information on the short-lived intermediates
using proven mathematical relationships.
a CH2Cl2/0.1 M NBu4PF6, 298 K. b 50 nm peak widths assigned.
c Estimated from reference 21 (Figure 4 therein). d Calculated with 20 nm
peak width.
from the DFT calculations on the oxidized facþ species as
shown in Figure 11.
We then performed time dependent DFT (TD-DFT)
calculations to obtain the predicted spectra of the fac and
mer isomers of both the neutral and the cationic
[(dppe)Mn(CO)3Br] complexes to obtain an independent
measure of the quality of the fac-[(dppe)Mn(CO)3Br]þ
spectrum determined from the simultaneous chronoam-
perometry/chronoabsorptometry method. The calculated
spectra for these species are displayed in Supporting
Information, Figures S8-S11. Importantly, the calculated
UV-vis spectra λmax values for the fac, merþ, and mer
compounds match those determined experimentally to
within (27 nm.41 The calculated spectrum for facþ also
matches that obtained by application of eq 6 and shown in
Figure 10.
Acknowledgment. We are grateful to the National
Science Foundation for funding for this work (CHE-
0911537; to G.B.R.-A. and M.J.S.). We thank Dr. Jun Yi
for assistance with the figures.
Supporting Information Available: Derivations of equations,
definitions, additional figures (electrochemical cell, absorptivity
plots, calculated geometries and spectra, molecular orbital
pictures). This material is available free of charge via the
Internet at http://pubs.acs.org.
The LUMO of neutral fac-[Br(CO)3(dppe)Mn] complex
was found to be predominantly dz2 in character (Sup-
porting Information, Figure S16). Both the HOMO and
the LUMO are essentially unchanged in facþ. The HOMO
(Supporting Information, Figure S17) and LUMO
(Supporting Information, Figure S18) of merþ are predo-
minantly dxz and dyz respectively, with significant contri-
butions from the p orbitals of Br. Isomerization of facþ to
(51) Pereira, C.; Ferreira, H. G.; Schultz, M. S.; Milanez, J.; Izidoro, M.;
Leme, P. C.; Santos, R. H. A.; Gambardella, M. R. P.; Castellano, E. E.;
Lima-Neto, B. S.; Carlos, R. M. Inorg. Chim. Acta 2005, 358, 3735–3744.