Rapid excited-state structural reorganization captured by pulsed x-rays

Article abstract of DOI:10.1021/ja017214g

Visible light excitation of [Cu¹(dmp)₂](BArF), where dmp is 2,9-dimethyl-1,10-phenanthroline and BArF is tetrakis(3,5-bis(trifluoromethylphenyl))borate, in toluene produces a photoluminescent, metal-to-ligand charge-transfer (MLCT) e

Full text of DOI:10.1021/ja017214g

Published on Web 08/15/2002
Rapid Excited-State Structural Reorganization Captured by
Pulsed X-rays
Lin X. Chen,*,† Guy Jennings,‡ Tao Liu,† David J. Gosztola,† Jan P. Hessler,†
Donald V. Scaltrito,§ and Gerald J. Meyer§
Contribution from the Chemistry DiVision and Materials Science DiVision, Argonne National
Laboratory, Argonne, Illinois 60439, and Department of Chemistry, The Johns Hopkins
UniVersity, Baltimore, Maryland 21210
Received October 2, 2001
Abstract: Visible light excitation of [CuI(dmp)2](BArF), where dmp is 2,9-dimethyl-1,10-phenanthroline and
BArF is tetrakis(3,5-bis(trifluoromethylphenyl))borate, in toluene produces a photoluminescent, metal-to-
ligand charge-transfer (MLCT) excited state with a lifetime of 98 ( 5 ns. Probing this state within 14 ns
after photoexcitation with pulsed X-rays establishes that a CuII center, borne in a CuI geometry, binds an
additional ligand to form a five-coordinate complex with increased bond lengths and a coordination geometry
of distorted trigonal bipyramid. The average Cu-N bond length increases in the excited state by 0.07 Å.
The transiently formed five-coordinate MLCT state is photoluminescent under the condition studied, indicating
that the absorptive and emissive states have distinct geometries. The data represent the first X-ray
characterization of a molecular excited state in fluid solution on a nanosecond time scale.
terization of ground and long-lived transient states,5-10 have
not been successfully applied to excited states. The main
Introduction
Molecular excited states have important applications as
luminescent probes and in solar energy conversion, photography,
sensing, and displays.1 A wide variety of theoretical and
experimental techniques have been employed to provide insights
into the fascinating structures of excited states and the distortions
that accompany excited-state relaxation.2,3 Despite their impor-
tance, however, identification of the absolute molecular structure
of excited states using X-rays remains elusive. X-ray spectro-
scopic techniques (including X-ray absorption fine structure,
XAFS, and X-ray absorption near edge structure, XANES),4
which have been tremendously valuable for structural charac-
obstacle for excited-state structure determination is the lack of
pulsed X-ray sources that provide sufficient photons. This
obstacle has been circumvented with new generation synchrotron
sources that offer X-ray pulses with about 100 ps pulse duration
and up to 104 times more photons than previous sources.11 In
addition, notable advances in the production of ultrashort, intense
X-ray pulses have recently appeared in the literature.12-17 Here
we exploit these advances and report structural studies of
(4) (a) Sayers, D. E.; Stern, E. A.; Lytle, F. W. Phys. ReV. Lett. 1971, 27,
1204. (b) Rehr, J. J.; Mustre de Leon, J.; Zabinsky, S. I.; Alberts, R. C. J.
Am. Chem. Soc. 1991, 113, 5135.
(5) Chance, B.; Fischetti, R.; Powers, L. Biochemistry 1983, 22, 3820.
(6) Mills, D. M.; Lewis, A.; Harootunian, A.; Huang, J.; Smith, B. Science
1984, 223, 811.
(7) Thiel, D. J.; Livins, P.; Stern, E. A.; Lewis, A. Nature 1993, 362, 40.
(8) Chen, L. X.; Wang, Z.; Burdett, J. K.; Montano, P. A.; Norris, J. R. J.
Phys. Chem. 1995, 99, 7958.
(9) Chen, L. X.; Bowman, M. K.; Wang, Z.; Montano, P. A.; Norris, J. R. J.
Phys. Chem. 1994, 98, 9457.
(10) Chen, L. X.; Lee, P. L.; Gosztola, D. J.; Svec, W. A.; Montano, P. A.;
Wasielewski, M. R. J. Phys. Chem. B 1999, 103, 3270.
(11) (a) Shenoy, G. K.; Viccaro, P. J.; Mills, D. M. Characteristics of the 7-GeV
Advanced Photon Source, Argonne National Laboratory. Report ANL-88-
9; Argonne National Laboratory: Argonne, IL, 1988. (b) Knapp, G. S.;
Beno, M. A.; You, H. Annu. ReV. Mater. Sci. 1996, 26, 693.
(12) Schoenlein, R. W.; Chattopadhyay, S.; Chong, H. H. W.; Glover, T. E.;
Heimann, P. A.; Shank, C. V.; Zholents, A. A.; Zolotorev, M. S. Science
2000, 287, 2237.
(13) Rischel, C.; Rousse, A.; Uschmann, I.; Albouy, P. A.; Geindre, J. P.;
Audebert, P.; Gauthier, J. C.; Forster, E.; Martin, J. L.; Antonetti, A. Nature
1997, 390, 490.
(14) Bressler, C.; Chergui, M.; Pattison, P.; Wulff, M.; Filipponi, A.; Abela, R.
Proc. SPIE-Int. Soc. Opt. Eng. 1998, 3451, 108.
* To whom correspondence and requests for materials should be
addressed. E-mail: lchen@anl.gov.
† Chemistry Division, Argonne National Laboratory.
‡ Materials Science Division, Argonne National Laboratory.
§ Department of Chemistry, The Johns Hopkins University.
(1) (a) Biophysics and Physiobiology of Excitable Membranes; Adelman, W.
J., Ed.; Van Nostrand Reinhold Co.: New York, 1971. (b) Time-ResolVed
Fluorescence Spectroscopy in Biochemistry and Biology; Cundall, R. B.,
Dale, R. E., Eds.; Plenum Press: New York, 1980. (c) Guillet, J. Polymer
Photophysics and Photochemistry; Cambridge University Press: Cambridge,
1985. (d) Fluorescent Chemosensors for Ion and Molecular Recognition;
Czarnick, A. W., Ed.; ACS Symposium Series 538; American Chemical
Society: Washington, DC, 1993. (e) Roundhill, D. M. Photochemistry and
Photophysics of Metal Complexes; Plenum Press: New York, 1994. (f)
Inorganic and Organic Electroluminescence; Mauch, R. H., Gumlich, H.-
E., Eds.; Wissenschaft und Technik Verlag: Berlin, 1996.
(2) (a) Calvert, J. G.; Pitts, J. N. Photochemistry; John Wiley & Sons: New
York, 1966. (b) Parker, C. A. Photoluminescence of Solutions; Elsevier
Publishing Co.: New York, 1968. (c) Turro, N. J. Modern Molecular
Photochemistry; Benjamin Cummings Publishing Co.: Menlo Park, CA,
1978. (d) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd
ed.; Kluwer Academic/Plenum Publishers: New York, 1999.
(3) (a) Kasha, M. Discuss. Faraday Soc. 1950, 9, 14. (b) Strickler, S. J.; Berg,
R. A. J. Chem. Phys. 1962, 37, 814. (c) Freed, K. F.; Jortner, J. J. Chem.
Phys. 1970, 52, 6272. (d) Jortner, J.; Mukamel, S. Preparation and Decay
of Molecular Excited States. In The World of Quantum Chemistry; Daudei,
D., Pullman, B., Eds.; D. Reidel Publishing Co.: Dordrecht, Holland, 1974;
pp 145-209.
(15) Perman, B.; Srajer, V.; Ren, Z.; Teng, T.-Y.; Pradervand, C.; Ursby, T.;
Bourgeois, D.; Schotte, F.; Wulff, M.; Kort, R.; Hellingwerf, K.; Moffat,
K. Science 1998, 279, 1946.
(16) Rose-Petruck, C.; Jimenez, R.; Guo, T.; Cavalleri, A.; Siders, C. W.; Raksi,
F.; Squier, J. A.; Walker, B. C.; Wilson, K. R.; Barty, C. P. J. Nature
1999, 398, 310.
(17) Tomov, I. V.; Oulianov, D. A.; Chen, P.; Rentzenpis, P. M. J. Phys. Chem.
B 1999, 103, 7081.
9
10.1021/ja017214g CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 10861-10867
10861

A R T I C L E S
Chen et al.
molecular excited states by nanosecond pump-probe XAFS in
fluid solution.
This important achievement was made possible with the aid
of X-ray pulses from a new generation synchrotron source with
a significantly higher X-ray photon flux than was previously
available.11,18,19 A simplified time sequence of the experiment
is shown in Figure 1. A laser pulse excites the sample, and a
sextuplet X-ray pulse cluster probes the structure of the excited
state at its optimal concentration. The goal of this approach is
to take a “snapshot” of the thermally equilibrated excited-state
structure rather than to follow excited-state relaxation dynamics.
Figure 1. Experimental time sequence used in this work. The laser pump
pulse (black triangle) was overlapped with the first X-ray pulse of a sextuplet
cluster (white triangles). This time sequence allowed a “snapshot” of the
thermally equilibrated excited-state structure to be taken when its population
was optimal.
The metal-to-ligand charge-transfer (MLCT) excited states
of cuprous diimine compounds were chosen in part because of
the compelling evidence described in the literature for novel
structural reorganization following light absorption.20 These
reorganization processes are relevant to “gated” electron transfer
in proteins21 and model systems22 as well as in new classes of
molecular devices23 and solar energy conversion materials.24
Therefore, these studies not only represent a proof-of-concept
example of excited-state characterization but also provide new
insights into the dynamics for structural reorganization important
in biology and chemistry.
subject to a Jahn-Teller distortion.26 The observations of large
“Stokes-like” shifts between the absorption and photolumines-
cence and “exciplex” quenching represent strong evidence that
the excited state adopts a more Cu(II)-like geometry.20,25 The
Jahn-Teller distortion is also clearly manifest in the crystal
structures of Cu(II) bipyridine and phenanthroline compounds
that reveal distorted square pyramidal or trigonal-bipyramidal
geometries with an additional ligand derived from a solvent
molecule or counterion.25,26
Therefore, for this specific study, light absorption by Cu(I)
diimine compounds instantaneously creates a Franck-Condon
state with a Cu(II) excited state in a Cu(I) geometry. Within
our instrument response time, this state rapidly undergoes
vibrational relaxation and Jahn-Teller distortions to yield an
emissive, thermally equilibrated excited state that lives for about
100 ns, τ ) 98 ( 5 ns. The purpose of the study is to capture
the structure of this MLCT state at its optimal concentration.
In particular, this work investigates whether the MLCT transition
is a whole or partial charge transfer from Cu(I) to the ligand,
and how the resulting thermally equilibrated MLCT-state
structure differs from the corresponding Cu(I) and Cu(II)
compounds in the ground state.
The origin of the photodriven structural change can be
understood by considering light absorption by a cuprous diimine
compound, such as CuI(dmp)2+, where dmp is 2,9-dimethyl-
1,10-phenanthroline, eq 1.
CuI(dmp)2+ + hν f CuII(dmp-)(dmp)+*
(1)
The Cu(I) ground state has a d10 electronic configuration with
pseudo-tetrahedral geometry.25 Absorption of a visible photon
promotes an electron from copper to a dmp ligand, formally
generating a MLCT excited state with a Cu(II) center coordi-
nated to one reduced and one neutral dmp ligand. The Cu(II)
center in the excited state has a d9 electron configuration and is
(18) Chen, L. X.; Ja¨ger, W. J. H.; Jennings, G.; Gosztola, D. J.; Munkholm, A.;
Hessler, J. P. Science 2001, 292, 262.
(19) (a) Chen, L. X. J. Elect. Spectrosc., Relat. Phenom. 2001, 119, 161. (b)
Jennings, G.; Ja¨ger, W. J. H.; Chen, L. X. ReV. Sci. Instrum. 2002, 72,
362.
Experimental Section
Synthesis. The [CuI(dmp)2](PF6) and CuI(dmp)2(NO3) were synthe-
sized according to previously published procedures.27-30 The [CuI-
(dmp)2](BArF) was synthesized via the metathesis of [CuI(dmp)2](PF6)
with sodium tetrakis(3,5-bis(trifluoromethylphenyl))borate (Boulder
Scientific) in a 1:1 ratio in toluene. The [CuII(dmp)2(NO3)](NO3) was
synthesized following a previously published protocol.31
(20) (a) Kalyanasundaram, K. Photochemistry of Polypyridine and Porphyrin
Complexes; Academic Press: London, 1992; Chapter 9. (b) Scaltrito, D.
V.; Thompson, D. W.; O’Callaghan, J. A.; Meyer, G. J. Coord. Chem.
ReV. 2000, 208, 243. (c) Armaroli, N. Chem. Soc. ReV. 2001, 30, 113.
(21) (a) Lumry, R.; Eyring, H. J. Phys. Chem. 1954, 58, 110-120. (b) Vallee,
B. L.; Williams, R. J. P. Proc. Natl. Acad. Sci. U.S.A. 1968, 59, 498-505.
(c) Holm, R. H.; Kennepohl, P.; Solomon, E. I. Chem. ReV. 1996, 96, 2239.
(22) (a) Meagher, N. E.; Juntunen, K. L.; Salhi, C. A.; Ochrymowycz, L. A.;
Rorabacher, D. B. J. Am. Chem. Soc. 1992, 114, 10411-10420. (b)
Flanagan, S.; Dong, J.; Haller, K.; Wang, S.; Scheidt, W. R.; Scott, R. A.;
Webb, T. R.; Stanbury, D. M.; Wilson, L. J. J. Am. Chem. Soc. 1997, 119,
8857-8864.
(26) Coordination numbers greater than five are known for nonluminescent
copper bis-phenanthroline compounds with unsubstituted 1,10-phenathroline
ligands. They are not known for luminescent complexes where substituents
in the 2- and 9- positions are required.
(27) (a) Blaskie, M. W.; McMillin, D. R. Inorg. Chem. 1980, 19, 3519. (b)
McMillin, D. R.; Kirchhoff, J. R.; Goodwin, K. V. Coord. Chem. ReV.
1985, 64, 83.
(23) Collin, J. P.; Dietrich-Buchecker, C.; Gavina, P.; Jimenez-Molero, M. C.;
Sauvage, J.-P. Acc. Chem. Res. 2001, 34, 477.
(24) (a) Breddels, P. A.; Blassie, G.; Casdonte, D. J.; McMillin, D. R. Ber.
Bunsen-Ges. Phys. Chem. 1984, 88, 2. (b) Alonso-Vante, N.; Nierengarten,
J. F.; Sauvage, J. P. J. Chem. Soc., Dalton Trans. 1994, 1649. (c) Ruthkosky,
M.; Kelly, C. A.; Castellano, F. N.; Meyer, G. J. J. Am. Chem. Soc. 1997,
119, 12004.
(25) (a) Stephens, F. S. J. Chem. Soc., Dalton Trans. 1972, 1350. (b) Stephens,
F. S.; Tucker, P. A. J. Chem. Soc., Dalton Trans. 1973, 2293. (c) Kaiser,
J.; Brauer, G.; Schroder, F. A.; Taylor, I. F.; Rasmussen, S. E. J. Chem.
Soc., Dalton Trans. 1974, 1490. (d) Anderson, O. P. Inorg. Chem. 1975,
14, 730. (e) Harrison, W. D.; Hathaway, B. J.; Kennedy, D. Acta
Crystallogr., Sect. B 1979, 35, 2301. (f) Harrison, W. D.; Hathaway, B. J.
Acta Crystallogr., Sect. B 1979, 35, 2910. (g) Harrison, W. D.; Kennedy,
D. M.; Power, M.; Sheahan, R.; Hathaway, B. J. J. Chem. Soc., Dalton
Trans. 1981, 1556. (h) Tran, D.; Skelton, B. W.; White, A. H.; Laverman,
L. E.; Ford, P. C. Inorg. Chem. 1998, 37, 2505.
(28) (a) Cunningham, C. T.; Cunningham, K. L. H.; Michalec, J. F.; McMillin,
D. R. Inorg. Chem. 1999, 38, 4388. (b) Cunningham, C. T.; Moore, J. J.;
Cunningham, K. L. H.; Fanwick, P. E.; McMillin, D. R. Inorg. Chem. 2000,
39, 3638.
(29) Miller, M. T.; Ganzel, P. K.; Karpisihin, T. B. J. Am. Chem. Soc. 1999,
121, 4292.
(30) Felder, D.; Nierengarten, J. F.; Barigelletti, F.; Ventura, B.; Armaroli, N.
J. Am. Chem. Soc. 2001, 123, 6291.
(31) Riesgo, E. C.; Hu, Y.-Z.; Bouvier, F.; Thummel, R. P.; Scaltrito, D. V.;
Meyer, G. J. Inorg. Chem. 2001, 40, 3413.
(32) Antonio, M. R.; Soderholm, L.; Song, I. J. Appl. Electrochem. 1997, 27,
784.
(33) (a) Hamalainen, R.; Algren, M.; Turpeinen, U.; Raikas, T. Cryst. Struct.
Commun. 1979, 8, 75. (b) Van Meerssche, M.; Germain, G.; Declercq, J.
P. Cryst. Struct. Commun. 1981, 10, 47.
9
10862 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Rapid Excited-State Structural Reorganization
A R T I C L E S
Electrolysis for in Situ XAFS. Bulk electrolysis of 2 mM
CuI(dmp)2 in 0.1 M Bu4NPF6 acetonitrile was carried out using a
+
microprocessor-controlled potentiostat (Princeton Applied Research,
model 273) and a standard three-electrode arrangement in a previously
described spectroelectrochemical cell designed for XAFS use.32 The
working electrode was a 3 cm2 platinum mesh, and the auxiliary
electrode was a platinum wire located in a portion of the cell separated
by a porous glass frit. The applied potential was measured versus an
aqueous Ag+/AgCl reference electrode. Prior to oxidation, the solution
was purged with acetonitrile-saturated nitrogen gas. The nitrogen gas
continued to bubble through the solution during electrolysis to aid in
stirring the solution. After approximately 20 min at +0.80 V, the
complex was completely oxidized. During the remainder of the
experiments, the solution was blanketed with a stream of nitrogen gas.
Pump-Probe XAFS. The copper K-edge (8.979 keV) XANES and
XAFS measurements were conducted at a wiggler beamline, 11ID-D,
Basic Energy Science Synchrotron Research Center of the Advanced
Photon Source at Argonne National Laboratory. The details of the
experimental setup were described elsewhere.18,19 Briefly, the pump
laser pulses were from the second harmonic output of a Nd-YLF laser
(λ ) 527 nm, 1 kHz repetition rate, 1 mJ/pulse, and 5 ps fwhm), and
the probe X-ray pulses (a sextuplet X-ray cluster of 14.2 ns, 271 kHz
repetition rate) were from the synchrotron running with an asymmetric
timing mode where a sextuplet X-ray pulse cluster was separated in
time from other pulses in the storage ring.18,19 The time delay between
the pump and the probe was adjusted so that the laser pulse overlapped
with the first X-ray pulse in the sextuplet pulse cluster (Figure 1). The
laser pump and X-ray probe beams were intersected at a continuously
flowing sample stream of 2 mM [CuI(dmp)2](BArF) toluene solution.
The thickness of the sample stream was about 0.5 mm. A Ge detector
array was used to detect the X-ray fluorescence of the sample. The
different repetition rates of the laser pump and X-ray probe pulses
required timing the data acquisition to isolate those signals that
correspond to the sample being simultaneously illuminated with a laser
and X-ray pulses.18,19 Therefore, the Ge detector array signals were
split into two equal parts connected separately to two scaler modules.
The first module processed only X-ray signals from the sextuplet X-ray
probe pulse cluster that coincided with the laser pump pulse, and the
second module processed signals from all X-ray pulses. The output
from the former was used to obtain XANES and XAFS spectra of the
laser-illuminated sample, and that of the latter, the spectra of the ground
state. Therefore, the pump-probe cycles were repeated at a laser pulse
repetition rate of 1 kHz until the XAFS spectrum of the laser-excited
sample had a satisfactory signal-to-noise ratio to establish the oxidation
state, the coordination geometry, and the local fine structure around
the metal center. The total data acquisition time was approximately
20 h.
Figure 2. Absorption difference spectra obtained after pulsed 532 nm light
excitation (5-8 mJ/cm2, 8 ns fwhm) of [CuI(dmp)2](BArF) in argon-
saturated toluene at room temperature. The spectra are shown at the
following delay times: 0 (9), 10 (b), 25 (1), 50 ([), and 100 ns (2). The
inset displays the ground-state absorption spectrum of [CuI(dmp)2](BArF)
in toluene at room temperature.
accordance with its crystal structure.33 Similarly, the coordination
number of five (four N atoms and one O atom) and the average nearest-
neighbor distance of 2.09 Å were used for CuII(dmp)2(NO3)2.33 WinXAS
9734 was used for data analysis following standard procedures.35 The
experimentally collected data were fit to the equation
2
ø(k) )
Fi(k)S0 (k)Ni/(kRi2) exp(-2σi2k2) sin[2kRi + φi(k)]
∑
where F(k) is the magnitude of the backscattering, S0 the amplitude
reduction factor, N the coordination number, R the average distance,
σ2 the Debye-Waller factor, and φi the phase shift; the subscript
indicates the ith atom, and k is the electron wavevector.
Photoluminescence. Corrected photoluminescence (PL) spectra were
obtained with a Spex Fluorolog that had been calibrated with a standard
tungsten-halogen lamp using procedures given by the manufacturer.
Quantum yields were measured by the optically dilute technique with
Ru(bpy)32+ as a standard.36 Time-resolved PL measurements were made
as previously described.37
UV-Vis Absorption. Ground-state absorption spectra were acquired
at ambient temperature in air using a Hewlett-Packard 8453 diode array
spectrometer. Nanosecond time-resolved absorption measurements were
made with an apparatus that has been previously described.38
Results
The absorption spectrum of [CuI(dmp)2](BArF) in toluene
displayed characteristic broad MLCT bands in the visible region
centered at 460 nm (Figure 2, inset). Light excitation into these
bands resulted in room-temperature photoluminescence with a
corrected maximum at 710 nm, an emission quantum yield φem
) 1.12 × 10-3, and a lifetime of 98 ns. Assuming an intersystem
crossing yield of unity, a radiative rate constant, kr, of 1.14 ×
104 s-1 and a nonradiative rate constant, knr, of 1.02 × 107 s-1
were calculated. Time-resolved absorption spectroscopy yielded
difference spectra with positive absorption bands centered at
∼350 and ∼580 nm that were characteristics of the reduced
dmp ligand and a bleach of the ground-state absorption at
A nanosecond transient UV-vis absorption apparatus with the same
laser and similar sample conditions was performed prior to X-ray
experiments to ensure that the MLCT excited state was generated with
the efficiency expected. The sample integrity was also monitored
periodically by UV-vis spectroscopy during the pump-probe XAFS
experiment and was changed every 3-4 h to ensure sample integrity.
Data Analysis of XANES and XAFS Spectra. The measured
XANES spectrum of the laser-illuminated sample contains contributions
from the ground and the MLCT excited states. The XANES spectrum
of the MLCT state alone was calculated by subtracting the appropriate
fraction of ground-state spectrum from the measured spectrum. The
fraction of the ground state that remained in the laser-illuminated sample
was determined by the Beer-Lambert law and the measured irradiance,
sample concentration, and extinction coefficient with an assumed
intersystem crossing yield of unity.
(34) Ressler, T. J. Synch. Radiat. 1998, 5, 118.
(35) Stern, E. A.; Sayers, D. E.; Lytle, F. W. Phys. ReV. B 1975, 11, 4836.
(36) Demas, J. N.; Crosby, G. A. J. Phys. Chem. 1971, 75, 991.
(37) Castellano, F. N.; Heimer, T. A.; Thandasetti, M.; Meyer, G. J. Chem.
Mater. 1994, 6, 1041.
33
For XAFS data analysis, CuI(dmp)2(NO3)33 and CuII(dmp)2(NO3)2
solids were used as references for the phase and amplitude of the ground
state and the MLCT state. The coordination number of four and an
average Cu-N distance of 2.07 Å were used for CuI(dmp)2(NO3), in
(38) Kelly, C. A.; Thompson, D. W.; Farzad, F.; Stipkala, J. M.; Meyer, G. J.
Langmuir 1999, 15, 7047.
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10863

A R T I C L E S
Chen et al.
Figure 3. In situ XANES spectra of [CuI(dmp)2](BArF) in acetonitrile at
room temperature during the electrolysis at 0.8 V versus an aqueous Ag+/
AgCl electrode, where CuI was oxidized to CuII. E is the photon energy of
the X-ray, and µ(E) is proportional to the amount of X-ray absorbed by the
sample. The initial spectrum of CuI(dmp)2+, before the electrolysis, is shown
as open squares, and the final spectrum of the Cu(II) product is shown as
open circles. Eight intermediate spectra are also shown. The arrows indicate
the directions of the changes in the spectra during the electrolysis. The
inset shows the pre-edge region of the starting and ending spectra only.
460 nm (Figure 2). Clean isosbestic points and first-order
kinetics were observed with lifetimes that agree, within experi-
mental error, with those measured independently by time-
resolved photoluminescence spectroscopy. Steady-state UV-
vis absorption data collected before and after these transient
studies revealed no experimental evidence for sample decom-
position.
Figure 4. (A) XANES spectra of [CuI(dmp)2](BArF) in toluene with and
without laser illumination. (B) XANES spectrum of the MLCT state
generated by subtraction of the ground-state contributions to the spectrum
of the laser-illuminated sample followed by normalization to 100% MLCT
state (see text). The XANES spectrum of the Cu(II) oxidation product from
Figure 3 is shown for comparison. ø(E) is the interference function, defined
as [µ(E) - µ0(E)]/µ0(E)], where µ(E) and µ0(E) are absorption coefficients
of [CuI(dmp)2](BArF) and isolated copper atoms, respectively.
To identify the oxidation state of copper in the MLCT excited
state, we first measured copper K-edge XANES spectra of the
ground-state CuI(dmp)2+ and the Cu(II) product obtained by in
situ electrolysis (Figure 3). The following spectral differences
were observed between the CuI and CuII species: (1) a transition
edge shift that was 3 eV higher for CuII than for CuI, reflecting
the higher positive charges in the former; (2) a shoulder feature
at 8985 eV representing a 1s-to-4pz transition that was pro-
nounced in CuI and missing from CuII, typically observed for
transformation from four- to five- or six-coordinate geometry,
because the 4pz orbital was localized in the former, consistent
with a localized-to-delocalized transition of the 4pz orbital;39
(3) an intensity of the 8997 eV peak that was higher for CuII
than for CuI, suggesting a higher coordination number in the
former, because of the proportionality of the oscillation ampli-
tude with the coordination number; and (4) a pre-edge feature
at 8979 eV (Figure 4 inset) attributed to a 1s f 3d transition
that was only present for the CuII (d9) compound with one
vacancy in 3d orbitals. These XANES spectral features of CuI-
8979 eV. Because such spectral changes were absent when the
sample was not illuminated or when the detection was not gated,
the spectral changes were due to generation of the MLCT
excited state.
The XANES spectrum for the laser-illuminated sample shown
in Figure 4A (circles) represents an algebraic sum of both the
MLCT excited and ground states in the solution at the time of
the X-ray probe. As mentioned in the Experimental Section,
the fraction of excited-state molecules was calculated on the
basis of the Beer-Lambert law. For the data shown, the laser-
illuminated spectrum was comprised of 80% ground state and
20% MLCT excited state. Therefore, subtraction of this fraction
of the known CuI(dmp)2 spectrum from that of the laser-
+
illuminated sample followed by normalization allowed abstrac-
tion of the MLCT-state spectrum in Figure 4B (circles).
+
(dmp)2 and the Cu(II) oxidation product agree well with
previously reported spectra where the CuII(dmp)2+ complex in
acetonitrile was identified to be pentacoordinated.40
Bond lengths from the copper to the ligands for the MLCT
excited state of CuI(dmp)2 were also quantified. Figure 5
+
XANES spectra of the ground-state and laser-illuminated CuI-
(dmp)2+ solution are shown in Figure 4A. The laser-illuminated
CuI(dmp)2+ solution had a reduced shoulder feature at 8985 eV
compared to the ground state and a weak pre-edge feature at
displayed XAFS and Fourier-transformed XAFS spectra for the
ground-state and laser-illuminated CuI(dmp)2+. The magnitude
of the Cu-N peak in Figure 5B for the laser-illuminated CuI-
(dmp)2+ was higher and shifted to a longer distance compared
to that in the ground state, suggesting an increase in the
coordination number of copper and the lengthening of the
average Cu-N bond. The results of data analysis for the nearest-
neighbor distances from the copper ion are listed in Table 1.
For the ground-state CuI(dmp)2+, an average Cu-N bond
(39) (a) Smith, T. A.; Penner-Hahn, J. E.; Berding, M. A.; Doniach, S.; Hodgson,
K. O. J. Am. Chem. Soc. 1985, 107, 5945. (b) Kau, L.-S.; Spira-Solomon,
D. J.; Penner-Hahn, J. E.; Hodgson, K. O.; Solomon, E. I. J. Am. Chem.
Soc. 1987, 109, 6433.
(40) Elder, R. C.; Lunte, C. E.; Rahman, A. F. M. M.; Kirchhoff, J. R.; Dewald,
H. D.; Heineman, W. R. J. Electroanal. Chem. 1988, 240, 361.
9
10864 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Rapid Excited-State Structural Reorganization
A R T I C L E S
state molecules. Thus, the average Cu-N bond distance in the
MLCT state increased by 0.07 Å from that of the ground state.
Because the accuracy for coordination number determination
from XAFS analysis is normally 10-20%, it is difficult to
establish whether the copper coordination number is four or
five in the MLCT state from this analysis alone. However, the
XANES features of the MLCT state described above and the
increase of the Cu-N peak height in Figure 5B support an
increase in copper coordination number in the MLCT state.
Discussion
The results presented above have clearly demonstrated X-ray
characterization of a molecular excited-state structure under
experimentally useful conditions of fluid solution and room
temperature. This exciting observation appears to be unprec-
edented but may be extended to other inorganic and organic
molecular excited states. Below we describe the details of the
structure determination for the Cu(I) MLCT excited states and
the implications of these results on the molecular reorganization
novel to Cu(II/I) charge transfer.
X-ray Evidence for the MLCT Excited-State Structure
of CuI(dmp)2+. The transition edge in a K-edge XANES
spectrum is related to the energy required for ejecting a 1s
electron to the continuum and is therefore sensitive to the
oxidation state of the X-ray absorbing atom. The transition edge
position generally shifts to higher energy with increased
oxidation state41 and is also sensitive to the chemical nature of
the coordinating ligands. The XANES fine structure corresponds
to transitions from the 1s core orbital to valence orbitals that
participate in chemical bonding, subject to selection rules.
Therefore, XANES spectra are extremely powerful for the
elucidation of the coordination environment of transition metal
compounds.39,40-42
Figure 5. (A) XAFS spectra of the ground-state and laser-illuminated [CuI-
(dmp)2](BArF) in toluene, where the frequency of the oscillation is related
to the bond distance from the copper atom and the amplitude of the
oscillation is related to the coordination number of the copper as well as
the Debye-Waller factor (see equation in the Experimental Section), where
k is the X-ray wave vector and ø(k) is the interference function in k space.
(B) Fourier-transformed XAFS spectra of (A), where each peak represents
one average distance between the copper atom and its neighbors. R is the
distance, in angstroms, from the central copper atom. The peak correspond-
ing to the nearest neighbors is labeled. The shift of the peak and the
amplitude change reflect structural changes. The light-illuminated CuI-
One of the issues we sought to explore was whether light
+
absorption by CuI(dmp)2 resulted in whole or partial charge
+
transfer from Cu(I) to a dmp ligand. The copper K-edge XANES
(dmp)2 solution has a longer average Cu-N distance and a higher
spectra for CuI(dmp)2 and the in situ oxidation product were
+
coordination number than the ground state. The spectra are not corrected
with phase factors, so the peaks appear at distances smaller than the real
distances. The details of the data analysis are described in the Experimental
Section, and the phase-corrected structural parameters can be found in Table
1.
used for comparisons with the XANES spectrum for the MLCT
excited state in Figure 4B. Partial charge transfer was expected
to manifest itself in a transition edge position intermediate
between the CuI and CuII XANES spectra. However, the MLCT
and the CuII had virtually identical edge positions, indicative
of complete Cu(I)-to-dmp charge transfer in the MLCT state.
Additional evidence for complete charge transfer in the MLCT
state comes from the appearance of a pre-edge peak at 8979
eV. This peak is due to the 1s-to-3d transition induced by the
X-ray photon, which is feasible only for Cu(II) (3d9) with one
vacancy in 3d orbitals, but not for Cu(I) (3d10) with no vacancy.
The coordination geometry of copper in the MLCT excited
state can also be revealed from the XANES spectra in Figure
4B. The distinctive shoulder feature at 8985 eV in the ground
state of CuI(dmp)2+ has been attributed to the 1s f 4pz transition
when the 4pz is localized on the copper with a square-planar or
tetrahedral geometry.39 When the fifth and sixth ligands bind
to the metal along the z direction, the 4pz orbital delocalizes
and the peak due to the 1s f 4pz transition is smeared out,
Table 1. Ground- and MLCT-Excited-State Structuresa
coordination
2 d
σ2 (Å )
number
4.0 ( 0.5
with light (fit 1 bond length)b 4.5 ( 0.5
R (Å)c
ground-state CuI(dmp)2
2.06 ( 0.02
0.0009
+ b
2.07 ( 0.02 -0.0008
with light (fit 2 bond lengths)e 4.0 ( 0.5 (80%) 2.06 ( 0.02
0.004
4.0 ( 1.0 (20%) 2.13 ( 0.04 -0.009
a Only the nearest neighboring atomic shell is presented. b Using
CuI(dmp)2NO3 solid as reference with an average Cu-N distance of 2.07
Å and a coordination number of four. This one-distance model resulted in
a poor fit to the experimental data. c Bond distance. d Debye-Waller factors.
e Using CuI(dmp)2NO3 and CuII(dmp)2(NO3)+ (average Cu-N distance of
2.09 Å and a coordination number of five) as references for the first and
the second distances, respectively.
distance of 2.06 Å and a coordination number of four were
obtained. The Cu-N bond distances for the laser-illuminated
sample could not be adequately modeled with a single distance,
so a two-distance fit was carried out, resulting in a much better
fit with two Cu-N bond distances at 2.06 and 2.13 Å with a
relative ratio of 80:20. Obviously, the former was from the
ground-state molecules and the latter from the MLCT excited-
(41) Koningsberg, D. C., Prins, R., Eds. X-ray Absorption: Principles, Ap-
plications, Techniques of EXAFS, SEXAFE and XANES; John Wiley &
Sons: New York, NY; 1988.
(42) Westre, T. E.; Kennepohl, P.; DeWitt, J. G.; Hedman, B.; Hodgson, K. O.;
Solomon, E. I. J. Am. Chem. Soc. 1997, 119, 6297.
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10865

A R T I C L E S
Chen et al.
resulting in a smooth transition edge.39 Such changes in the
transition edge due to the ligation have been well documented
for metalloporphyrins with and without one or two axial
ligands,18,38 as well as other transition metal complexes.39,41,42
Therefore, the rather smooth rising edge in the XANES spectrum
of the MLCT state in Figure 4B is consistent with a penta- or
hexacoordination geometry of the copper center. The steric
hindrance from the methyl groups on the phenanthroline ligands
inhibits octahedral geometries, so a pentacoordinate MLCT
structure is more likely. Since the observed pre-edge feature at
8979 eV due to the quadrupole-allowed 1s f 3d transition43 is
weak yet detectable, the coordination is likely a distorted
trigonal-bipyramidal or a distorted square-pyramidal geometry.
base addition to copper excited states in dichloromethane forms
a five-coordinate excited-state complex, or “exciplex”.46 Copper
diimine excited states, and MLCT excited states in general,47
are known to follow Jortner’s energy gap law, wherein the
nonradiative rate constant increases exponentially with decreas-
ing ground state-excited state energy separation.3c,20b Exciplex
quenching can be understood on this basis: coordination of a
fifth ligand stabilizes the excited state, thereby decreasing the
energy gap and promoting nonradiative decay. With CuI-
(dmp)2+*, the exciplexes that have been identified decrease the
energy gap to the extent that nonradiative decay is the sole
relaxation pathway and no emission has been detected. Copper
diimine compounds with more sterically bulky ligands, on the
other hand, inhibit exciplex formation, and room-temperature
photoluminescence is often observed, even in Lewis basic
solvents.20b,31,48 For these sterically congested compounds, the
absorption spectra are solvent independent, while the photolu-
minescence spectra red shift and the excited state lifetimes
decrease significantly in more coordinating solvents.31,48 These
facts stated in the literature are consistent with emission from
a five-coordinate excited state with weak excited state-solvent
interactions that stabilize the excited state but not to the extent
that nonradiative decay is the sole relaxation pathway. However,
we emphasize that photoluminescence is an indirect method for
determining changes in coordination number.
The results from the XAFS data analysis listed show an
average Cu-N bond expansion by 0.07 Å in the MLCT excited
state. Previous excited-state Raman studies are consistent with
localization of the excited state on one dmp ligand;44 however,
we are unable to resolve discrete Cu-N distances from the
current experimental X-ray data. The observed Cu-N bond
elongation in the MLCT state agrees with recent theoretical
+
predictions that addition of a water ligand to a CuI(dmp)2
excited state should substantially increase the Cu-N bond
distance.45 For comparison, the X-ray crystallographically
determined bond lengths indicate that the average Cu-N bond
2+
length is 0.02 Å longer for CuII(dmp)2NO3 than for CuI-
Our hypothesis is that Cu(dmp)2+* in toluene behaves in a
manner similar to that reported previously for sterically bulky
copper compounds. A weak adduct is formed in the excited
state, presumably with toluene or the BArF anion, that stabilizes
the excited state but not to the degree where radiative decay is
(dmp)2 .
+ 33 Although the XAFS data fitting procedure does not
allow us to unambiguously distinguish the coordination number
of the excited state, the XANES evidence presented above and
the increased amplitude of the Cu-N peak for the laser-
illuminated sample (Figure 5B) strongly support a five-
coordinate CuI(dmp)2+ MLCT excited state, despite the uncer-
tainly on the coordination number solely from the XAFS data
fitting.
noncompetitive, kr ) 1.14 × 104 s-1 and knr ) 1.02 × 107 s-1
.
There is no experimental evidence that precludes the existence
of a related five-coordinate excited state for CuI(dmp)2+* in
dichloromethane, although our attempts to characterize it by
X-rays were frustrated by irreversible photochemistry.49 We
anticipate stronger ion-pairing in toluene than in dichlo-
romethane; however, it is difficult to predict which solvent and/
or counterion would be more nucleophilic. The higher energy
emission maximum and longer lifetime in toluene only suggest
weaker excited-state interactions and/or less reorganization for
Cu(dmp)2+* than in dichloromethane.20c We emphasize that the
evidence for exciplex formation from previous work is compel-
ling but indirect. Exciplexes were just inferred for cases where
strong excited-state adducts completely quenched the excited
state. This underscores the utility of X-ray techniques that
provide a direct method for determination of the excited-state
coordination environment.
Implications for Copper Diimine Excited States. The
coincidence of the CuI(dmp)2+* lifetimes measured by time-
resolved absorption and photoluminescence spectroscopies
indicates that the same thermally equilibrated MLCT excited
state is observed by both techniques and strongly suggests that
this is the same state probed by X-ray pulses. The X-ray data
provide convincing evidence that there are five nearest neighbors
in the copper coordination sphere. In addition, copper in the
MLCT excited state is formally a 17-electron d9 Cu(II) system
prone to the addition of a fifth ligand, and there exists a
preponderance of indirect literature evidence that indicates a
five-coordinate excited state.20 Below we discuss the relevant
literature results and the implications of an emissive five-
coordinate thermally equilibrated excited state.
The conclusion that the excited state is five-coordinate has
important implications in copper excited states (Figure 6).50
A photodriven increase in the copper coordination number
has previously been invoked to explain the quenching of copper
excited states by Lewis bases.46 Extensive studies by McMillin
and co-workers have provided compelling evidence that Lewis
(47) (a) Kober, E. M.; Caspar, J. V.; Lumpkin, R. S.; Meyer, T. J. J. Phys.
Chem. 1983, 87, 952. (b) Meyer, T. J. Acc. Chem. Res. 1989, 22, 163.
(48) (a) Dietrich-Buchecker, C. O.; Marnot, P. A.; Sauvage, J.-P.; Kirchoff, J.
R.; McMillin, D. R. J. Chem. Soc., Chem. Commun. 1983, 513. (b)
Ruthkosky, M.; Castellano, F. N.; Meyer, G. J. Inorg. Chem. 1996, 35,
6406. (c) Miller, M. T.; Gantzel, P. K.; Karpishin, T. B. Inorg. Chem. 1999,
38, 3414. (d) Riesgo, E. C.; Hu, Y.-Z.; Bouvier, F.; Thummel, R. P.;
Scaltrito, D. V.; Meyer, G. J. Inorg. Chem. 2001, 40, 3413.
(43) Shulman, R. G.; Yafet, Y.; Eisenberger, P.; Blumberg, W. E. Proc. Natl.
Acad. Sci. U.S.A. 1976, 73, 1384.
(44) Gordon, K. C.; McGarvey, J. J. Inorg. Chem. 1991, 30, 2986.
(45) Sakaki, S.; Mizutani, H.; Kase, Y. Inorg. Chem. 1992, 31, 4575.
(46) (a) McMillin, D. R.; Kirchoff, J. R.; Goodwin, K. V. Coord. Chem. ReV.
1985, 64, 83. (b) Palmer, C. E. A.; McMillin, D. R.; Kirmaier, C.; Holten,
D. Inorg. Chem. 1987, 26, 3167. (c) Crane, D. R.; DiBenedetto, J.; Palmer,
C. E. A.; McMillin, D. R.; Ford, P. C. Inorg. Chem. 1988, 27, 3698. (d)
Everly, R. M.; McMillin, D. R. Photochem. Photobiol. 1989, 50, 711. (e)
Stacy, E. M.; McMillin, D. R. Inorg. Chem. 1990, 29, 393-396.
(49) Berger, R. M.; McMillin, D. R.; Dallinger, R. F. Inorg. Chem. 1987, 26,
3803.
(50) A reviewer suggested an alternative model wherein an equilibrium exists
between an emissive four-coordinate compound and a nonemissive or near-
IR emissive shorter-lived five-coordinate compound. For this model to be
consistent with our observations, the photoluminesce and absorption data
must report on the four-coordinate excited state and the X-rays probe the
five-coordinate state.
9
10866 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Rapid Excited-State Structural Reorganization
A R T I C L E S
Conclusions
Structural information about molecular excited states by
X-rays has been realized for the first time in solution on a
nanosecond time scale. The experiments were performed in fluid
solution at room temperature and thus under conditions mean-
ingful for many applications of luminescent excited states. Our
data support previous studies that have inferred large structural
changes in copper diimine excited states. They have also
revealed previously unrecognized behavior on the dynamics of
excited-state structural reorganization. With a 14 ns time
resolution, the X-ray probe pulse cluster captured the molecular
structure of a thermally equilibrated MLCT state generated by
laser illumination. Light excitation initiates complete charge
transfer from copper to a dmp ligand and an inner-sphere
reorganization that changed the coordination number from four
to five and the geometry from tetrahedral to a likely distorted
trigonal bipyramidal. An increase in the Cu-ligand bond
distances was also observed. Since the excited state probed by
X-rays is photoluminescent,50 in accordance with the Franck-
Condon principle, the product of radiative recombination must
yield a five-coordinate Cu(I) compound that is not simply a
vibrationally excited ground state. Therefore, light absorption
and emission involve two distinct geometries, and at least four
states are involved in the photophysics. These studies provide
important new details and direct structural information on the
inner-sphere reorganization that is novel to CuII/I electron
transfer. More generally, this study demonstrates that time
domain XANES and XAFS can be used to quantify transient
light-driven charge-transfer excited-state structural reorganiza-
tion on short time scales.
Figure 6. Four-state model consistent with X-ray, UV-vis, and photolu-
minescent studies of [CuI(dmp)2+](BArF) in toluene. Light excitation of
the ground-state CuI(dmp)2+, Td, produces a Franck-Condon state, Td*,
with the same geometry. Within the time resolution of the experiment, this
state relaxes to a distorted trigonal bipyramidal, TBP*, or square-pyramidal,
SP* geometry with addition of an exogenous ligand, X, presumed to be
derived from toluene or the BArF anion. Concerted nonradiative decay,
knr′, to the Td ground state as well as radiative, kr, and nonradiative decay,
knr, yields a five-coordinate TBP or SP CuI geometry.
Radiative decay is a vertical process where the product must
maintain the same nuclear coordinates and geometry as the
excited state. Nonradiative decay, on the other hand, can lead
directly to ground-state products. Therefore, light absorption
and light emission involve different states with unique geom-
etries that are not simply vibrational excited states of each other,
as is often the case. This requires a minimum of four states, as
shown schematically. Furthermore, radiative decay must yield
a five-coordinate Cu(I) compound that subsequently releases
the fifth ligand and distorts to yield the pseudo-tetrahedral
ground state. This model shares many similarities with the
“square schemes” used to rationalize thermal CuII/I electron
transfer in biomimetic model compounds which support the
entactic state hypothesis.21,22
Acknowledgment. This work is supported by the Division
of Chemical Sciences, Office of Basic Energy Sciences, U.S.
Department of Energy, under Contract W-31-109-Eng-38, and
by the donors of the Petroleum Research Fund. The authors
thank Drs. Wighard Ja¨ger, Anneli Munkholm, Mark Beno,
Jennifer Linton, and staff members at BESSRC-CAT, Advanced
Photon Source, for their technical assistance.
JA017214G
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10867