Is our way of thinking about excited states correct? Time-resolved dispersive IR study on p-nitroaniline

Article abstract of DOI:10.1002/chem.201103235

Low-lying excited electronic states of an important class of molecules known as push-pull chromophores are central to understanding their potential nonlinear optical properties. Here we report that a combination of high-sensitivity nanosecond time-resolve

Full text of DOI:10.1002/chem.201103235

FULL PAPER
DOI: 10.1002/chem.201103235
Is Our Way of Thinking about Excited States Correct? Time-Resolved
Dispersive IR Study on p-Nitroaniline
Sudhakar Narra, Shu-Wei Chang, Henryk A. Witek, and Shinsuke Shigeto*[a]
Abstract: Low-lying excited electronic
states of an important class of mole-
cules known as push–pull chromo-
phores are central to understanding
their potential nonlinear optical prop-
erties. Here we report that a combina-
tion of high-sensitivity nanosecond
time-resolved dispersive IR spectrosco-
py and DFT calculations on p-nitroani-
line (PNA), a prototypical push–pull
molecule, reveals that PNA in the
lowest excited triplet state has a partial
quinoid structure. In this structure, the
quinoid configuration is restricted to a
part of the phenyl ring adjacent to the
NO2 group. The partial quinoid struc-
ture of PNA cannot be explained by a
commonly used hybrid of a neutral
form and a zwitterionic charge-transfer
form. Our findings not only cast doubt
on the general applicability of the clas-
sical way of looking at excited states,
based exclusively on characteristic res-
onance structures, but also provide
deeper insights into excited-state struc-
ture of highly polarizable molecular
systems.
Keywords: density functional calcu-
lations · excited states · IR spectros-
copy · push–pull chromophores ·
time-resolved spectroscopy
Introduction
NO2 group (Figure 1). This finding shows that our usual way
of thinking about excited states of push–pull molecules can
be erroneous, and lead to dubious predictions and intuitions.
Valence-bond resonance structures are a simple but intuitive
concept that is commonly used in various branches of
chemistry. They are hypothetical formulae representing elec-
tronic structure of molecules in an approximate fashion; an
actual representation should be thought of as a linear com-
bination of all conceivable resonance structures that can be
written for a given molecule. Organic chemists often use res-
onances structures to provide a qualitative account of chem-
ical structures and reactivities. Resonance structures have
been applied to describe electronic structures of molecules
not only in the ground state,[1–3] but also in electronic excited
states. For example, excited states of an interesting class of
molecules known as push–pull chromophores, in which an
electron-donating group and an electron-withdrawing group
are linked by a conjugated system, have been understood in
terms of a mixture of neutral and charge-transfer (CT) reso-
nance structures.[4] Herein, we use a combination of nano-
second time-resolved dispersive IR spectroscopy and DFT
to demonstrate that the electronic structure of the lowest
excited triplet (T1) state of p-nitroaniline (PNA) in acetoni-
trile (ACN) is accounted for not by a commonly used
hybrid of neutral form 1 and zwitterionic form 2, but by an
unusual structure similar to 3, in which a quinoid structure
is restricted to a part of the phenyl ring in the vicinity of the
Figure 1. Representative resonance structures of PNA and their hybrids:
neutral form 1, zwitterionic form 2, and partial quinoid form 3.
p-Nitroaniline is a prototypical push–pull molecule exhib-
iting a large first hyperpolarizability[5–11] and nonlinear opti-
cal properties with possible applications for electrooptic and
second-harmonic generation materials. As a model system,
PNA has long been studied both experimentally[4,12–22] and
theoretically.[23–27] A number of ultrafast spectroscopic stud-
ies[28–31] have been done to understand the dynamics of PNA
after photoexcitation. The lowest excited singlet (S1) state of
PNA is well-known to have CT character with a large dipole
moment[27] of about 15 D, and the S1 state undergoes very
[a] S. Narra, S.-W. Chang, Prof. Dr. H. A. Witek, Prof. Dr. S. Shigeto
Department of Applied Chemistry and Institute of Molecular Science
National Chiao Tung University
1001 Ta-Hsueh Road, Hsinchu 30010 (Taiwan)
Fax : (+886)3-5723764
E-mail: shigeto@mail.nctu.edu.tw
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201103235.
Chem. Eur. J. 2012, 18, 2543 – 2550
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2543

rapid (<0.3 ps) nonradiative relaxation to the ground (S0)
state by internal conversion (IC) or to the triplet manifold
by intersystem crossing (ISC).[29,31] In a conventional two-
state valence-bond model, the ground state is dominated by
neutral form 1, whereas the S1 state has a considerable con-
tribution of zwitterionic form 2, which explains the large
change in dipole moment from about 6 D in the ground
state[32] to about 15 D in the S1 state.[27]
In light of the fact that the T1 state is a primary channel
for the photophysical and photochemical processes in the
nanosecond to microsecond time region, it is desirable to
obtain detailed molecular information on this state as well.
Schuddeboom et al.[30] used time-resolved microwave con-
ductivity to determine the lifetime and dipole moment of
the T1 state in benzene and dioxane. It was shown that the
dipole moment of T1 PNA is in the range of 9–11 D, de-
pending on the solvent and the ISC quantum yield.[30] Thom-
sen and co-workers[31] measured the transient absorption
spectrum in the 340–960 nm range of PNA excited at
400 nm, which is assigned to vibrationally hot PNA mole-
cules in the ground state and triplet–triplet absorption
bands. A theoretical investigation was also performed re-
cently by Kosenkov and Slipchenko,[25] who predicted that
the lowest bright triplet state (13A1) has a much smaller
dipole moment (8.7 D) than the CT singlet state (12.9 D).
Although it would be reasonable to explain these observa-
tions on the basis of the two-state model by assuming that
the T1 state of PNA has a CT structure[33] analogous to the
S1 state, this analogy needs to be critically tested by more
direct information on the bonding structure of PNA in the
T1 state. Vibrational spectroscopy is undoubtedly a powerful
approach that permits one to access such information. Here
we aim to provide new insights into the electronic and mo-
lecular structure of T1 PNA on the basis of experimental
and theoretical IR spectra, which will call into question the
applicability of the classical resonance-structure picture for
elucidation of the electronic structure of excited states of
molecules.
Figure 2. a) Ground-state IR spectrum and b) time-resolved IR differ-
ence spectra of PNA in CD3CN (0.25 mm) excited at 355 nm with argon
bubbling. Each time-resolved spectrum is offset by 5ꢁ10ꢀ5 for clarity of
display.
ꢀ
C C stretch, respectively. In the time-resolved IR difference
spectra of PNA in CD3CN (Figure 2b), three positive peaks
at 1250, 1480, and 1605 cmꢀ1 and four negative peaks at the
same positions as in Figure 2a are observed. The positive
peaks correspond to the transient absorption of photogener-
ated molecular species, and the negative peaks to depletion
(bleaching) of the ground-state population of PNA. The
transient absorption arises instantaneously and decays away
within a few microseconds, synchronously with recovery of
the ground-state depletion. Little bleaching signal remains
in the 7.0–8.0 ms spectrum. Previous transient absorption
studies[29–31] revealed that relaxation of the S1 state of PNA
proceeds extremely rapidly to S0 by IC or to T1 by ISC. The
time constants for IC and ISC of PNA in water are 250 fs[29]
and 10 ps,[31] respectively. These timescales are much faster
than the time resolution of our apparatus, so the dynamics
involving the S1 state is completely over within one laser
pulse in our experiment. Therefore, we assign the transient
species detected here to PNA in the T1 state, which is gener-
ated from S1 as a result of ISC.
Results and Discussion
Time-resolved IR study: First, we present experimental data
obtained by time-resolved IR spectroscopy (see Experimen-
tal Section for details) of PNA in CD3CN, with a focus on
characterization of the triplet-state IR spectra of PNA and
its isotopomers. Figure 2 displays the time-resolved IR dif-
ference spectra of PNA in argon-saturated CD3CN solution,
together with the ground-state (S0 PNA) spectrum recorded
with the dispersive spectrometer. The ground-state spectrum
of PNA (Figure 2a) shows four principal IR bands peaking
at 1317, 1503, 1600, and 1636 cmꢀ1. The 1317 cmꢀ1 band is
accompanied by a shoulder at 1332 cmꢀ1. The peak positions
are in good agreement with the literature.[17,24] Our DFT cal-
culations (see below) reveal that these four bands are attrib-
uted predominantly to the NO2 symmetric stretch, NO2 anti-
symmetric stretch, and admixtures of NH2 scissoring and the
To corroborate our interpretation, we performed a singu-
lar value decomposition (SVD) analysis of the time-resolved
IR spectra, which shows that the observed dynamics can be
reproduced by considering a single independent component.
The intrinsic spectrum of the component (Figure S1a in the
Supporting Information) is essentially identical to the 0–
0.2 ms spectrum (Figure 2b) and decays with an exponential
time constant of 430 ns (Figure S1b in the Supporting Infor-
mation). This time constant is consistent with the reported
lifetime of T1 PNA (434 ns in ACN[22] and 275 ns in diox-
ane[30]).
2544
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2543 – 2550

Time-Resolved Dispersive IR Study on p-Nitroaniline
FULL PAPER
Further support comes from
oxygen-quenching effects on
the transient absorption of
PNA. As shown in Figure S2a
in the Supporting Information,
the intensity of the transient ab-
sorption bands decreases due to
oxygen bubbling and decays
much faster than with argon
bubbling. Even the strongest
positive band at 1480 cmꢀ1 com-
pletely
disappears
already
200 ns after photoexcitation. To
examine the quenching effects
on the kinetics, we compared
the time profiles of the intensity
at 1250 cmꢀ1 between argon-
and oxygen-bubbling measure-
ments (Figure S2b in the Sup-
porting Information). The expo-
nential time constant is reduced
by 75% with oxygen bubbling.
The accumulated experimental
evidence leads us to the conclu-
sion that the dynamics we have
observed (Figure 2b) is attrib-
uted to formation of the triplet
state of PNA and its subsequent
relaxation back to the ground
state. Note that, even in the ab-
sence of oxygen, T1 PNA seems
to show self-quenching effects
at higher concentrations of
PNA, possibly resulting in the
formation of PNA oligo-
mers.[22,30] We did find that at
1 mm, only about 80% of the
excited PNA molecules eventu-
ally return to the ground state.
In contrast, at 0.25 mm, little
bleaching signal remains after
the decay of T1 PNA completes.
The observed positive peaks
have been identified as origi-
Figure 3. Upper panels: experimental IR spectra of PNA and its isotopomers, PNA-15NH2 and PNA-15NO2. a–
c) Ground-state (S0) spectra recorded with an FTIR spectrometer. d–f) Lowest excited triplet-state (T1) spectra
obtained by compensating the depletion of the ground-state absorption. Lower panels: IR spectra of PNA,
PNA-15NH2, and PNA-15NO2 calculated by using the PNA+2ACN model with (solid line) and without
ꢀ
(dotted line) elongation of the C NH2 bond by 0.012 ꢂ. g–i) The ground state. j–l) The T1 state. See text for
details.
nating from PNA in the T1 state. The true triplet-state spec-
trum of PNA can be determined by adding an appropriately
weighted ground-state spectrum to the difference spectrum
so that the negative peaks would be compensated.[34,35] The
IR spectrum of T1 PNA obtained with such a compensation
procedure is presented in Figure 3. The IR spectra of
PNA-15NH2 and PNA-15NO2 in the T1 state derived in a sim-
ilar manner are also shown in Figure 3, together with their
ground-state FTIR spectra. The vibrational frequencies de-
termined for all three compounds are listed in Table 1. The
experimental isotope-substitution effects are summarized as
follows. The 15N substitution in the NH2 group does not
affect much the recorded ground- and triplet-state IR spec-
tra, except for modified frequency spacing between the dou-
blet components of the 1317 cmꢀ1 band (compare Figure 3a
and b for S0 and Figure 3d and e for T1). In contrast, 15N
substitution in the NO2 group alters the spectral pattern in
the 1420–1520 cmꢀ1 region and the doublet at 1317 and
1332 cmꢀ1 for S0 PNA (Figure 3c), and shifts the 1250 cmꢀ1
band by ꢀ25 cmꢀ1 for T1 PNA (Figure 3 f).
Theoretical IR spectra: We now attempt to provide a theo-
retical interpretation for the experimentally recorded IR
spectrum of T1 PNA. We are aware that modeling IR spec-
tra in solution phase requires in general a time-dependent
investigation of dynamics of an ensemble of PNA/ACN mol-
Chem. Eur. J. 2012, 18, 2543 – 2550
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2545

S. Shigeto et al.
Table 1. Experimental and calculated vibrational frequencies and assignments of major IR bands (1200–1700 cmꢀ1
PNA-15NO2 in the ground state and the lowest excited triplet state.
)
of PNA, PNA-15NH2, and
Experiment[a]
PNA-15NH2
DFT calculation[b]
Assignment[c]
(approximate description)
PNA
PNA-15NO2
PNA
PNA-15NH2
PNA-15NO2
ground (S0) state
1636
1600
1630
1600
1637
1601
1638
1600
1578
1507
1494
1632
1600
1577
1506
1492
1638
1600
1573
1480
1494
ds(NH2 scissoring)
ꢀ
n
G
n
(NH2)
1503
1494
1503
1494
1475
1506
1475
nasACTHUNGERTNNU(G NO2), nACHTUNGERTN(NUNG C C)
ꢀ
ꢀ
ꢀ
dACHTUNGTERNNN(GU C H), nAHCTUNRTGENNG(UN C C), nCAHTNUGTERN(UNNG C NH2), ds(NH2 scissoring)
1332
1317
1331
1314
1321
1300
1319
1296
1313
1285
ns
ns
N
N
R
ACHTUNGTRENNUNG(C-NO2)
1300
N
ACHTUNGTRENNUNG
triplet (T1) state
1605
1480
1606
1476
1605
1478
1626
1453
1366
1320
1309
1296
1257
1232
1624
1451
1366
1319
1305
1293
1256
1232
1625
1452
1366
1320
1306
1279
1255
1207
ds(NH2 scissoring), nACTHNGUTER(NNUG C-C)
ꢀ
ꢀ
ꢀ
d
G
R
ACHTUGNTRENUN(GN C NH2), ds(NH2 scissoring)
ꢀ
n
(C C), 1
U
ꢀ
ꢀ
ꢁ1320
ꢁ1320
ꢁ1300
ns
ns
N
ACHTUGNTRNEU(NGN C H), in-phase combination of nACHTUNTRGEG(NUNN C NH2) and nACHTNGURTEN(NUGN C-NO2)
ꢀ
ꢀ
ꢀ
ACHTUNGTREN(NUGN C H), out-of-phase combination of nACHUTNTRGEG(NNUN C NH2) and nACHTUNGTRENNUNG(C NO2)
ꢀ
1250
1250
1225
nasACHTUNGTREN(GUNN NO2), nACHTUNTRGEG(NNUN C C), 1CATHNUGTREN(NUNG NH2)
ꢀ
[a] Taken from FTIR spectra for S0. [b] B3LYP/PNA+2ACN with slight elongation (0.012 ꢂ) of the C NH2 distance. Scaling factor=0.972. [c] n:
stretching; d: in-plane bending; 1: rocking.
ecules at some finite temperature. An appropriate study
would consist of running a molecular dynamics trajectory
for such an ensemble and obtaining the IR spectrum via the
Fourier transform of the recorded dipole moment autocorre-
lation function. Unfortunately, at the present level of devel-
opment of computational meth-
reproducing the observed ground-state spectra of PNA and
its isotopomers in a satisfactory manner, we proceeded to
interpret the triplet-state IR spectrum of PNA.
The calculated gas-phase IR spectrum of PNA (Fig-
ure 4b) is in apparent contrast with the experimental result
ods, such a study is not practi-
cally feasible, in particular for
molecules in excited states. In-
stead, we chose to employ tra-
ditional, time-independent in-
struments of quantum chemis-
try, namely, DFT with the
B3LYP exchange correlation
functional (see Experimental
Section for details).
Another important issue here
is a careful choice of an appro-
priate molecular model. Usual-
ly, for the interpretation of IR
spectra of molecules in the
liquid phase, the calculations
are performed for a single gas-
phase molecule, under the as-
sumption that there is no pro-
nounced difference between the
spectra of gas-phase and solvat-
ed molecules. Here we can test
this supposition by studying the
Figure 4. Comparison of experimental and calculated DFT/B3LYP IR spectra of PNA in the ground state (left
panel) and in the lowest excited triplet state (right panel). a,g) Experimentally recorded IR spectra of PNA in
CD3CN (0.25 mm). b,h) A gas-phase molecular model. c,i) PCM model. d,j) Explicitly solvated model with two
CD3CN molecules attached to the NH2 group (PNA+2ACN model). e,k) PNA+2ACN model with slight
ground state of PNA in detail,
because accurate experimental
data are readily available for
the ground state. After finding
ꢀ
elongation (0.012 ꢂ) of the C NH2 bond. f,l) Explicitly solvated model with six CD3CN molecules (PNA+
a molecular model capable of 6ACN).
2546
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2543 – 2550

Time-Resolved Dispersive IR Study on p-Nitroaniline
FULL PAPER
(Figure 4a). First, the 1250–1370 cmꢀ1 region displays two,
well-separated strong bands with reversed relative intensi-
ties. In addition, the 1450–1650 cmꢀ1 region is distinct from
the experiment. These findings suggest that solvation of
PNA by CD3CN has substantial effects on the resulting IR
spectra and that one should not use a bare, gas-phase molec-
ular model for interpretation of the triplet-state IR spec-
trum. It is customary to account for the solvent effects by
using some implicitly solvated model, in which the solvent is
represented, for example, as a dielectric continuum with a
given dielectric constant (e=37.5 for CD3CN). The resulting
spectrum obtained with B3LYP combined with the polariza-
ble continuum model (PCM) does not show any considera-
ble improvement in comparison with the gas-phase spec-
trum. The main difference from experiment for the simulat-
ed B3LYP/PCM IR spectrum of PNA (Figure 4c) concerns
three details: 1) the doublet structure in the 1317 cmꢀ1 band
is missing, 2) the band around 1600 cmꢀ1 is split, and 3) the
band at 1636 cmꢀ1 is missing. The overall comparison of the
computed gas-phase and PCM spectra with experiment sug-
gest that none of these methods can be successfully used for
interpretation of the recorded triplet-state IR spectrum of
PNA.
can have substantial influence on the simulated IR spec-
tra.[28] It turned out that this effect is negligible,[4,16] but in
the course of this investigation, we have found that some
part of the spectrum changes quite dramatically with varia-
ꢀ
tion of the C NH2 distance. Slight elongation (0.012 ꢂ) of
ꢀ
the C NH2 bond from its equilibrium length (1.352 ꢂ) pro-
duces a hump on the single band at 1300 cmꢀ1, and gives
perfect agreement between the simulated spectra and the
experimental results in the whole spectral window studied
(Figures 3g and 4e). The energy of 0.07 kcalmolꢀ1 associat-
ed with such elongation is completely negligible on the mo-
lecular scale.
The origin of the doublet feature at around 1320 cmꢀ1 has
been a subject of debate.[13–15,17,19,21] Two possible mecha-
nisms for this doublet have been suggested: one is a Fermi
resonance[13,17] and the other is the existence of two distinct
solvated forms of PNA.[14,21] However, according to our cal-
culations, the higher- and lower-wavenumber components of
the doublet are attributed to the NO2 symmetric stretch cou-
ꢀ
pled with out-of-phase and in-phase combinations of the C
ꢀ
NH2 and C NO2 stretches, respectively (see Table 1). We
have found that their relative IR intensities are profoundly
affected by the geometry of PNA as well as by explicit sol-
vation.
Fortunately, the problems with the IR spectra obtained
with the gas-phase and PCM approaches can be almost com-
pletely resolved if one employs an explicitly solvated model
of PNA in which two CD3CN molecules are hydrogen-
bonded to the NH2 group (hereafter referred to as the
PNA+2ACN model; see Figure S3a in the Supporting In-
formation for its geometry). The resulting spectrum (Fig-
ure 4d) shows very good agreement with the experimental
data, except for the missing hump at 1332 cmꢀ1. We consid-
ered a multitude of various explicitly solvated models con-
taining one to six CD3CN molecules attached either to the
NH2 group or the NO2 group, or aligned along the ring (see
Figure 4 f for the calculated IR spectrum and Figure S3b in
the Supporting Information for the geometry of six solvent
molecules, i.e., PNA+6ACN). All of the resulting spectra
are quite similar in terms of a strong single band in the
1300–1400 cmꢀ1 region and reasonable agreement with ex-
periment in the 1600–1700 cmꢀ1 region. The main difference
lies in the region around 1500 cmꢀ1, for which the presence
of four or more solvent molecules results in two split bands.
This is an interesting finding, possibly suggesting that the
first solvation shell of PNA consists of maximally three sol-
vent molecules. One may argue that PNA is capable of
forming only two hydrogen bonds with the solvent mole-
cules through the NH2 group. However, we have found that
B3LYP predicts binding energies of CD3CN to the NH2 and
NO2 groups of similar order of magnitude: 5.5 and 3.5 kcal
molꢀ1, respectively.
One may argue that the bond elongation we have devised
is not justified. However, we would like to present a number
of justifications for such an action. First, we note that the
ꢀ
computed C NH2 bond length in the free and solvated PNA
models varies between 1.343 and 1.375 ꢂ depending on the
number and orientation of the explicitly attached solvent
molecules. The elongation we engineered to reproduce the
experimental spectrum yields a bond length of 1.364 ꢂ,
which falls in the above-mentioned interval. Note that the
engineered elongation is comparable in magnitude to a typi-
cal error (approximately 0.01 ꢂ) expected from DFT calcu-
lations of equilibrium bond lengths. The elongation may
give rise to a certain intensity redistribution in the simulated
IR spectrum, but it does not influence the positions of the
vibrational bands. If we reoptimize the molecular geometry
ꢀ
while keeping the new C NH2 bond length frozen and recal-
culate the spectrum, the largest frequency change is only
15 cmꢀ1. One may inquire whether the engineered elonga-
tion has some physical interpretation. This question is readi-
ly answered by taking into account the contrast between the
intrinsic static nature of the model used in our quantum
chemical calculations and the inherently dynamic behavior
of a real vibrating molecule. Quantum chemical methods
usually give us information about the (static) equilibrium
bond lengths re, but it would be more desirable to consider
the distances averaged over the ground vibrational wave
functions r0, which provide more physically meaningful in-
formation about the time-averaged interatomic distances in
a molecule. For a purely harmonic vibration, these two
The optimized structures of PNA with explicitly attached
solvent molecules display small variations in equilibrium ge-
ometry, which concern mainly the out-of-plane angles of the
ꢀ
quantities would be equal. However, the C NH2 bond
ꢀ
ꢀ
NH2 and NO2 groups and the C NH2 and C NO2 distances.
We have anticipated that the dependence of the out-of-
plane, low-frequency motions of the NH2 and NO2 groups
stretch, like nearly any other stretching vibration, has an an-
harmonic vibrational potential for which r0 is larger than re.
ꢀ
Additional support for the C NH2 bond elongation comes
Chem. Eur. J. 2012, 18, 2543 – 2550
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2547

S. Shigeto et al.
from the simulated IR spectra of PNA-15NH2 and
PNA-15NO2. The PNA+2ACN model (dotted line, lower
left panel of Figure 3) correctly predicts the isotope-substitu-
tion effects in the NO2 group but fails to predict the distinct
doublet structure of the 1317 cmꢀ1 band. In contrast, the
metric stretch of order 1.5 is expected to be about
ꢀ1
ꢀ
1550 cm , whereas the single N O bond stretch of oximes
usually appears around 940 cmꢀ1. In our calculations, the
frequency associated with the NO2 antisymmetric stretch is
reduced from 1507 cmꢀ1 in the S0 state to 1232 cmꢀ1 in the
T1 state. The typical length of the NPO bond is approxi-
mately 1.2 ꢂ, whereas that of a single bond is 1.4 ꢂ. As
ꢀ
PNA+2ACN model with slightly elongated C NH2 bond
(solid line, lower left panel of Figure 3) correctly predicts
the isotope substitution effects in both NH2 and NO2
groups.
ꢀ
shown in Figure 5, our simulations reveal N O bond lengths
We have demonstrated above that the PNA+2ACN
ꢀ
model with slightly elongated C NH2 bond is capable of re-
producing all the accumulated spectral evidence for the
ground state of PNA. We are therefore ready to interpret
the IR spectrum of the T1 state of PNA shown in Figure 3d.
The experiment detects four principal bands in the triplet-
state spectrum at 1250, 1320, 1480, and 1605 cmꢀ1. The engi-
neered PNA+2ACN model (Figure 3j) closely resembles
this experimental pattern by reproducing these bands at
1232, 1309, 1453, and 1626 cmꢀ1, respectively. The assign-
ments of these bands are as follows. The band at 1232 cmꢀ1
corresponds predominantly to the NO2 antisymmetric
stretch. The intense composite band at 1309 cmꢀ1 has a com-
plicated vibrational composition, namely, a mixture of in-
ꢀ
ꢀ
ꢀ
ꢀ
plane C H bending, C NH2 and C NO2 stretches, C C
stretch, and NO2 symmetric stretch. The band at 1453 cmꢀ1
ꢀ
ꢀ
ꢀ
is a mixture of in-plane C H bending and C NH2 and C
NO2 stretches. The band at 1626 cmꢀ1 originates from a mix-
ꢀ
ture of the NH2 scissoring motion and C C stretch. An anal-
ogous band appears in the IR spectrum of the ground state
at 1636 cmꢀ1, although with much lower intensity. As can be
seen from Figure 3d–f and j–l, the PNA+2ACN model with
Figure 5. Bond lengths and angles in the S0 state (upper value) and T1
state (lower value in green) derived from the DFT calculations with the
PNA+2ACN model.
ꢀ
slightly elongated C NH2 bond again does an excellent job
of reproducing the experimental triplet-state IR spectra of
all three isotopomers. The frequencies and assignments of
T1 PNA vibrations derived from our calculations are sum-
marized in Table 1. The shift of the IR bands on going from
S0 to T1 revealed by our calculations can be easily seen in
Figure S4 in the Supporting Information.
of 1.230 ꢂ in the S0 state and 1.271 ꢂ in the T1 state. Both
observations support the local character of the T1 S0 exci-
tation. The local excitation mechanism described here is fur-
!
ꢀ
ther confirmed by the shortening of the C NO2 bond upon
excitation (1.448!1.385 ꢂ), which is associated with the
contribution of the single p* electron localized on the nitro-
gen atom to the aromatic ring. Such a contribution leads to
a partial quinoid structure of PNA in the T1 state. Note that
the picture obtained here is quite different from that given
by conventional hybrids of resonance structures such as 1
and 2 in organic chemistry.[3,4] The quinoid structure is ac-
tually restricted only to the part of the ring in the vicinity of
Photoexcitation of PNA to the S1 state has been interpret-
ed in terms of CT from the electron-donating NH2 group to
the electron-withdrawing NO2 group (zwitterionic form
2).[4,27–31] It would be natural to expect that the T1 state
would be of similar character as well. However, our inter-
pretation of the IR spectra of S0 and T1 PNA does not fully
confirm this hypothesis. No significant change in vibrational
frequencies associated with the NH2 group (only ꢀ80 cmꢀ1)
with a simultaneous large frequency shift in the NO2 anti-
symmetric stretch (ꢀ270 cmꢀ1) suggests that the excitation is
rather local and occurs only at the NO2 moiety. It is relative-
ly easy to design a local molecular orbital picture of the
ꢀ
the NO2 moiety, because the C NH2 bond length is practi-
cally identical between the two electronic states (1.352 vs.
1.353 ꢂ, Figure 5). If we are compelled to translate this par-
tial quinoid structure into the language of resonance struc-
tures, we may find that 3 would be the closest with formal
!
T1 S0 excitation. In the ground state, the four p-electrons
charges of +1= at the ortho positions to the NO2 group. It is
2
of the NO2 moiety display the p2p*2p*0 configuration, which
is changed to p2p*1p*1 upon excitation. Such a change leads
also consistent with the experimental evidence[30] that the T1
state shows a less pronounced CT character compared to
the S1 state.
ꢀ
to a reduction of the N O bond order from approximately
1.5 in the ground state to 1.25 in the T1 state. This change
can be observed in a twofold way: vibrational frequencies
and bond lengths. The typical frequency of the NO2 antisym-
2548
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2543 – 2550

Time-Resolved Dispersive IR Study on p-Nitroaniline
FULL PAPER
sample. The excitation wavelength lies very close to the absorption maxi-
mum (364 nm) of PNA in CD3CN, which corresponds to the S1 S0 tran-
Conclusion
!
sition. The changes in absorption spectrum upon photoexcitation were
probed with mid-IR light derived from a ceramic IR emitter. The probe
light transmitted through the 100 mm-thick sample was introduced to a
dispersive spectrometer and detected with a photovoltaic HgCdTe detec-
tor that was coupled to a preamplifier. The AC-coupled output of the de-
tector was further amplified and finally processed with a high-speed digi-
tizer mounted on a computer, yielding a time stream of IR difference
spectra in the fingerprint region (1200–1700 cmꢀ1 in the present case). A
spectral resolution of 8 cmꢀ1 was used. The time resolution of the appara-
tus was approximately 80 ns, which was determined by the temporal re-
sponse of the HgCdTe detector to a 7 ns pulse. The concentration of
PNA in CD3CN was kept as low as 0.25 mm, so that depletion of the
ground-state absorption almost fully recovers at longer delay times and
no long-lived photoproducts of PNA were detected to a measurable
extent. To avoid possible sample degradation and heat accumulation due
to laser irradiation, the sample solution was continuously circulated by a
gear pump through a laboratory-built flow cell consisting of two calcium
fluoride windows. In a normal run, the sample solution in the reservoir
was continuously bubbled with argon gas, whereas in the oxygen-quench-
ing experiment, it was saturated with oxygen gas. The ground-state IR
spectra of PNA and its isotopomers were also recorded on a JASCO FT/
IR-6100 spectrometer with a spectral resolution of 2 cmꢀ1. All measure-
ments were performed at room temperature.
The experimental and theoretical evidence presented here
strongly suggests that the T1 state of PNA has a partial qui-
noid structure, which is in apparent contrast with the con-
ventional two-state models describing excited states of
push–pull compounds. Such a partial quinoid structure may
also be found for other push–pull compounds in excited
states. p-Dimethylaminobenzonitrile (DMABN) is one of
the most extensively studied compounds of this kind. A
number of time-resolved IR[36–39] and Raman[40,41] studies
have been performed to elucidate the CT S1 and T1 states of
DMABN, and provide conclusive evidence for a long-stand-
ing controversy[42] as to whether the S1 state has a twisted in-
tramolecular CT (TICT) or planar CT structure. With re-
spect to the T1 state, Hashimoto and Hamaguchi[36] suggest-
ed the presence of the TICT T1 state. In contrast, Ma
et al.[40] reported that T1 DMABN has a planar or nearly
planar structure with high negative charge localized on the
ꢂ
C N group and conjugation between the ring and the N-
A
(CH3)2 group. Our results demonstrate that a combination
Computational details: DFT calculations were performed with the
B3LYP functional[47,48] and the cc-pVTZ basis set[49] as implemented in
Gaussian 09.[50] Geometry optimization with tight convergence criteria
was followed by frequency and IR intensity calculations in a double har-
monic approximation, in which both positions and intensities of the
bands are determined by retaining only the lowest nonvanishing terms in
the Taylor expansions of the potential energy and dipole moment. The
lack of anharmonicity in the calculated B3LYP spectra was partially com-
pensated by performing a uniform scaling of the computed frequencies
by a factor of 0.972. The polarizable continuum model (PCM) was tested
to see the implicit solvation effects on the solute. Calculated IR spectra
were plotted using a Gaussian spectral envelope with a half-width at
half-maximum of 8 cmꢀ1. In the engineered spectra, we slightly elongated
of time-resolved IR spectroscopy and the DFT/B3LYP
method with the engineered explicit solvation model can be
helpful to resolve the long-standing issue about the structure
of DMABN in the excited state.
Experimental Section
Materials: All chemicals were purchased from Sigma-Aldrich unless oth-
erwise indicated. p-Nitroaniline (>99%) was purified by recrystallization
from benzene (Alfa Aesar) before use. Two 15N-substituted isotopomers
of PNA were synthesized according to published methods. p-
(15N)Nitroaniline (PNA-15NO2) was synthesized by nitration[43] of acetani-
lide (97%) with K15NO3 (98 atom% 15N) and subsequent hydrolysis. The
ꢀ
the C NH2 distance and performed partial geometry optimization (with
ꢀ
the C NH2 distance frozen) followed by frequency determination for
such a nonequilibrium structure. All vibrational modes are nonimaginary
and, differ only slightly (by up to 15 cmꢀ1) from the original spectrum.
product
was
recrystallized
from
ethanol.
p-NitroACTHNGUTERNNUG
(15N)aniline
(PNA-15NH2) was obtained by nitrating (15N)acetanilide (98 atom% 15N)
with reagent-grade KNO3 (>99%) followed by hydrolysis. The product
was separated by column chromatography on silica gel, first with n-
hexane and then with n-hexane/ethyl acetate (80/20 v/v) as eluents. The
synthesized isotopomers were characterized by 1H NMR, UV/Vis, and
FTIR spectroscopy and mass spectrometry. For UV/Vis and FTIR meas-
urements, [D3]acetonitrile (99.8 atom% D, Cambridge Isotope Laborato-
ries) was used as solvent. The UV/Vis spectra of PNA-15NH2 and
PNA-15NO2 evidenced no noticeable difference from that of normal
PNA.
Acknowledgements
This work was supported by the “Aiming for the Top University” plan of
the Ministry of Education, Taiwan, and National Chiao Tung University.
We also acknowledge financial support from the National Science Coun-
cil of Taiwan (Grants NSC98-2113M-009-011-MY2 to S.S. and NSC99-
2113M-009-022-MY3 to H.A.W.).
1H NMR (300 MHz, [D6]DMSO): d=7.94–7.91 (m, 2H), 6.7 (s, 2H),
6.59–6.56 ppm (m, 2H); MS (EI): m/z (%): 139 (100) [M+]; UV/Vis
([D3]acetonitrile): lmax (e)=364 nm (17830 molꢀ1 dm3 cmꢀ1).
[1] D. J. Craik, G. C. Levy, R. T. C. Brownlee, J. Org. Chem. 1983, 48,
1601.
Nanosecond time-resolved dispersive IR spectroscopy: The apparatus for
nanosecond time-resolved dispersive IR spectroscopy has been described
in detail elsewhere.[34,44] This apparatus utilizes a dispersive monochroma-
tor in combination with an alternating current (AC)-coupled detection
scheme.[35,45] Because the grating of the spectrometer must be scanned to
record a spectrum, the dispersive method usually requires much longer
[2] R. R. Fraser, J. Am. Chem. Soc. 1982, 104, 6475.
[3] H. C. Hiberty, G. Ohanessian, J. Am. Chem. Soc. 1984, 106, 6963.
[4] A. M. Moran, A. M. Kelley, J. Chem. Phys. 2001, 115, 912.
[5] S. P. Karna, P. N. Prasad, M. Dupuis, J. Chem. Phys. 1991, 94, 1171.
[6] C. Castiglioni, M. Del Zoppo, G. Zerbi, Phys. Rev. B 1996, 53,
13319.
[7] C. Dehu, F. Meyers, E. Hendrickx, K. Clays, A. Persoons, S. R.
Marder, J. L. Brꢃdas, J. Am. Chem. Soc. 1995, 117, 10127.
[8] J. N. Woodford, M. A. Pauley, C. H. Wang, J. Phys. Chem. A 1997,
101, 1989.
acquisition time than the FTIR method. However, the dispersive method
[44]
has higher sensitivity of up to DAꢁ1ꢁ10ꢀ6
,
which is hardly accessible
with the FTIR method.[46] Furthermore, the dispersive method is less sus-
ceptible to the instability of the measured transient event in comparison
with the FTIR method.[46]
The third harmonic of a Q-switched Nd:YAG laser (355 nm; 7 ns dura-
tion; 56 mJ energy; 500 Hz repetition rate) was used to photoexcite the
[9] F. L. Huyskens, P. L. Huyskens, A. P. Persoons, J. Chem. Phys. 1998,
108, 8161.
Chem. Eur. J. 2012, 18, 2543 – 2550
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2549

S. Shigeto et al.
[10] P.-E. Larsson, L. M. Kristensen, K. V. Mikkelsen, Int. J. Quantum
Chem. 1999, 75, 449.
[34] S. Yabumoto, S. Sato, H. Hamaguchi, Chem. Phys. Lett. 2005, 416,
100.
[35] T. Yuzawa, C. Kato, M. W. George, H. Hamaguchi, Appl. Spectrosc.
1994, 48, 684.
[36] M. Hashimoto, H. Hamaguchi, J. Phys. Chem. 1995, 99, 7875.
[37] C. Chudoba, A. Kummrow, J. Dreyer, J. Stenger, E. T. J. Nibbering,
T. Elsaesser, K. A. Zachariasse, Chem. Phys. Lett. 1999, 309, 357.
[38] H. Okamoto, J. Phys. Chem. A 2000, 104, 4182.
[11] F. Sim, S. Chin, M. Dupuis, J. E. Rice, J. Phys. Chem. 1993, 97, 1158.
[12] T. P. Carsey, G. L. Findley, S. P. McGlynn, J. Am. Chem. Soc. 1979,
101, 4502.
[13] J. Dreyer, V. Kozich, W. Werncke, J. Chem. Phys. 2007, 127, 234505.
[14] K. Mohanalingam, D. Yokoyama, C. Kato, H. Hamaguchi, Bull.
Chem. Soc. Jpn. 1999, 72, 389.
[39] H. Okamoto, H. Inishi, Y. Nakamura, S. Kohtani, R. Nakagaki, J.
Phys. Chem. A 2001, 105, 4182.
[40] C. Ma, W. M. Kwok, P. Matousek, A. W. Parker, D. Phillips, W. T.
Toner, M. Towrie, J. Phys. Chem. A 2001, 105, 4648.
[15] T. Fujisawa, M. Terazima, Y. Kimura, J. Phys. Chem. A 2008, 112,
5515.
[16] T. Gunaratne, J. R. Challa, M. C. Simpson, ChemPhysChem 2005, 6,
1157.
[41] W. M. Kwok, C. Ma, P. Matousek, A. W. Parker, D. Phillips, W. T.
Toner, M. Towrie, S. Umapathy, J. Phys. Chem. A 2001, 105, 984.
[42] Z. R. Grabowski, K. Rotkiewicz, W. Rettig, Chem. Rev. 2003, 103,
3899.
[43] Z. Cao, K. Armstrong, M. Shaw, E. Petry, N. Harris, Synthesis 1998,
1724.
[44] S. Yabumoto, S. Shigeto, Y.-P. Lee, H. Hamaguchi, Angew. Chem.
2010, 122, 9387; Angew. Chem. Int. Ed. 2010, 49, 9201.
[45] K. Iwata, H. Hamaguchi, Appl. Spectrosc. 1994, 48, 1453.
[46] P. Chen, R. A. Palmer, Appl. Spectrosc. 1997, 51, 580.
[47] A. D. Becke, J. Chem. Phys. 1993, 98, 5648.
[17] V. Kozich, W. Werncke, J. Dreyer, K.-W. Brzezinka, M. Rini, A.
Kummrow, T. Elsaesser, J. Chem. Phys. 2002, 117, 719.
[18] V. Kozich, W. Werncke, A. I. Vodchits, J. Dreyer, J. Chem. Phys.
2003, 118, 1808.
[19] K. Kumar, P. R. Carey, J. Chem. Phys. 1975, 63, 3697.
[20] E. D. Schmid, M. Moschallski, W. L. Peticolas, J. Phys. Chem. 1986,
90, 2340.
[21] S. Shigeto, H. Hiramatsu, H. Hamaguchi, J. Phys. Chem. A 2006,
110, 3738.
[22] H. Ma, S. Yao, J. Zhang, C. Pu, S. Zhao, M. Wang, J. Xiong, J. Pho-
tochem. Photobiol. A 2009, 202, 67.
[48] C. Lee, W. Yang, R. G. Parr, J. Phys. Rev. B 1988, 37, 785.
[49] T. H. Dunning Jr., J. Chem. Phys. 1989, 90, 1007.
[50] Gaussian 09 (Revision A.02), M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V.
Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X.
Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Son-
nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J.
Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A.
Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, ꢄ.
Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, and D. J. Fox,
Gaussian, Inc., Wallingford, CT, 2009.
[23] V. Chis¸, M. M. Venter, N. Leopold, O. Cozar, Vib. Spectrosc. 2008,
48, 210.
[24] E. Kavitha, N. Sundaraganesan, S. Sebastian, Indian J. Pure Appl.
Phys. 2010, 48, 20.
[25] D. Kosenkov, L. Slipchenko, J. Phys. Chem. A 2011, 115, 392.
[26] G. Scalmani, M. J. Frisch, B. Mennucci, J. Tomasi, R. Cammi, V.
Barone, J. Chem. Phys. 2006, 124, 094107.
[27] H. K. Sinha, K. Yates, Can. J. Chem. 1991, 69, 550.
[28] V. M. Farztdinov, R. Schanz, S. A. Kovalenko, N. P. Ernsting, J. Phys.
Chem. A 2000, 104, 11486.
[29] S. A. Kovalenko, R. Schanz, V. M. Farztdinov, H. Hennig, N. P.
Ernsting, Chem. Phys. Lett. 2000, 323, 312.
[30] W. Schuddeboom, J. M. Warman, H. A. M. Biemans, E. W. Meijer, J.
Phys. Chem. 1996, 100, 12369.
[31] C. L. Thomsen, J. Thøgersen, S. R. Keiding, J. Phys. Chem. A 1998,
102, 1062.
[32] A. L. McClellan, Tables of Experimental Dipole Moments, W. H.
Freeman and Company, San Francisco, 1963.
Received: October 13, 2011
[33] J. Wolleben, A. C. Testa, J. Phys. Chem. 1977, 81, 429.
Published online: January 25, 2012
2550
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2543 – 2550