Jacek Dobkowski

Time-resolved spectroscopy:

the key for the understanding

of the dynamics of the microworld

Institute of Physical Chemistry

Polish Academy of Sciences

Warsaw 2007

ISBN 978-83-920719-5-2

Przygotowanie do druku z materiałów camera ready

Sekcja Wydawnicza IChF PAN

1

Contents

1. Introduction 2

2. Apparatus 7

A. Time-resolved spectroscopy of porphycene and its derivatives

in low temperature gas matrices 20

3.1 Time-resolved spectroscopy of porphycene embedded in low

temperature matrices 21

3.2 Time-resolved spectroscopy of the 2,7,12,17-tetra-tert-butyl-3,6,13,16-

dibenzo[cde,mno]porphycene (TTPC) 29

B. Intramolecular and solvent relaxation in the case of molecules

revealing charge–transfer excited states. 36

4. An experimental test of C-N bond twisting in the TICT state –

the case of dialkylamino derivatives of m-cyanopyridine 37

5. Intramolecular charge-transfer properties of molecules with small

donor groups: the case of the carbonyl derivatives of

N,N-dimethylaniline and N,N-diethylaniline. 54

6. Intramolecular charge-transfer properties of molecules with a large donor

group.

The case of: 4’-(1-pyrenyl)benzonitrile (Py-BN) and

4’-(1-pyrenyl)acetophenone (Py-AC) 73

7. Intramolecular charge-transfer properties of a molecule with a large

donor group: The case of 4-Acetyl-4’-N,N-dimethylaminobiphenyl

(ADAB) 93

8. Final remarks. 132

9. Annex 140

10. References 143

11. Experimental 148

12. Glossary of abbreviations 150

13 Acknowledgements 151

2

1 Introduction

For photochemical and photophysical investigations (apart from the importance of such

classification for the study) two observables play the leading role: spectral distribution of the

radiation and its temporal evolution. Until the middle of the last century scientists involved in

the research on the field of the interaction of light and matter concentrated their attention on

the registration of a stationary absorption and emission spectra (spectral distribution). The

application of the pulse technique for the time-resolved experiments was performed for the

first time by George Porter and Ronald Norrish [Porter, 1950]. They used two flash lamps, the

first one for the excitation of the sample, the second for the analysis of the transient products.

The temporal resolution of this apparatus was in the millisecond range of time, but it was

enough for the registration of the absorption spectra of the free radicals.

During next five decades the progress in the field of construction of the time-resolved

spectrometers was incredible, however the idea of the registration of the transient absorption

(TA) proposed by G. Porter and R. Norrish did not change up to-day. It means that for

recording of TA spectrum two pulses should be available: the first one for the excitation

(pump) and the second, delayed with respect to the first one, for the monitoring (probe). G.

Porter and R. Norrish were awarded a Nobel prize. Parallel, to the progress in the registration

of the transient absorption the time-resolved spectroscopy in emission was developed.

Two experimental techniques play the leading role in the time–resolved emissive

spectroscopy: single photon counting and the up-conversion method. Both techniques enable

us to record the decay curves of the fluorescence, however its spectral distribution (TRF –

time-resolved fluorescence) is monitored applying the complex analysis of the fluorescence

decays recorded for a number of the observation wavelengths. The direct registration of the

TRF spectra can be performed using optical Kerr shutter. The time-resolved techniques were

described in details in the chapter 2.

The turning point in the development of the time–resolved spectroscopy was the

construction of the first laser [Maiman, 1960]. It was the beginning of the “race against time”:

a decade of seventies - flourishing of the nanosecond techniques (10-9

s), what enabled the

monitoring of the temporal evolution of the population of the triplet states and observation of

the transient products, whose generation was controlled by diffusion.

a decade of eighties – was dominated by the progress in picosecond spectroscopy (10-12

s).

The initiation of investigations of solvation dynamics, monitoring of electron and proton

transfer in excited states, registration of the short living transients.

a decade of nineties – rapid development and spreading of femtosecond techniques (10-15

s).

Investigations of the photosynthesis and vibronic relaxation, registration of the fast

components of solvation dynamics. A. Zewail as a pioneer on the field of the femtochemistry

was awarded Nobel prize in 1999.

These two Nobel prizes show that the development of time-resolved spectroscopy is

crucial for understanding of the dynamical aspects of the processes occurring in microword.

On the other hand it points out the limits of time-resolution within the visible region. The

absolute limit of the pulse duration is set by the uncertainty principle, which holds that the

product of uncertainty in photon energy and pulse duration, must be above a minimum value,

3

E t ħ/2. For pulses spanning the visible spectrum, the limit on duration is a few

femtosecond.

In 2001 Ferenc Krausz reached another crucial step by producing single attosecond

(10-18

s) x-ray pulses of estimated duration of about 650 as [Uiberacker et al., 2007]. Krausz

has used the attosecond pulses to observe rearrangement of electrons in krypton atoms after

removal of an inner-shell electron. In attosecond time domain the absorption can not be

treated as an immediate event.

The progress in the time-resolved spectroscopy was achived not only by the

development of the laser techniques , but also due to the evolution of the detection

technology. The detector systems evolved from the photographic plate, photomultiplier, linear

photodiode array, up to the cooled CCD matrix equipped with the intensifier.

The apparatus constructed in our laboratory

Laser as the source of the pulses determines mainly the temporal resolution of the

picosecond spectrometer. In the spectrometer of the first generation, the Nd:YAG laser was

applied (pulse width 30 ps), in the spectrometer of second generation the Nd:glass laser was

used (pulse width 1.5 ps). Both lasers were delivered by the Lithuanian companies

Eksma/Ekspla and Light Conversion respectively. The optical components such as the delay

line, polychromators were designed by dr Jan Jasny and produced in the workshop of our

Institute. The electronic devices and computer software were done by eng. Andrzej

Ardasiewicz. The spectrometers were described in details in the chapter 2.

The subject of my investigations

Excited state charge separation

The excited state photoinduced charge-transfer phenomena occurring in molecules with

donor (D) and acceptor (A) subunits linked by a single bond are the interesting subject for the

time-resolved study. It was well established that relatively simple DA-molecules having as the

donor unit –N(R)2 groups linked in the para position with the described below classes of

acceptors:

1. {Ar-X} , where Ar- aromatic ring, X- chromophore, (for example CN group)

2. azaaromatic ring (for example pyridine)

undergo excited state relaxation to the TICT state (Twisted Intramolecular Charge Transfer ),

figure 9.5. There were published hundreds of papers supporting or rejecting the TICT model,

however it is not my task to discuss in this moment the arguments for and against TICT

hypothesis, directing the readers to the review article of Z.R. Grabowski, K. Rotkiewicz, and

W. Rettig [Grabowski et al., 2003].

Two aspects of the TICT state can be investigated by the time-resolved spectroscopy:

excited state charge separation,

kinetic relations between primary and secondary TRF and transient absorption bands.

The TICT state hypothesis postulates the full electron transfer between the donor and

acceptor subunits twisted to the perpendicular conformation around the central bond. The

change of geometry in the excited state was confirmed by the model compounds with the

N(R)2 group fixed or pretwisted with respect to the aromatic ring. The new method of the

monitoring of the twist of the donor group is the optically induced NMR time resolved

experiment (see chapter 4). In this method the polarization of the NMR signals is directly

4

connected with the transient non-Boltzmann population of the ground state conformers

induced by the excited state reaction. The obtained results fully support TICT hypothesis.

(see chapter 4 and 5).

Molecules having the extended electron systems of the donor group relax in the

excited state not in accordance with the reaction scheme: locally excited state of planar

geometry TICT state, for example pyrene derivatives: 4’-(1-pyrenyl)benzonitrile (Py-BN)

and 4’-(1-pyrenyl)acetophenone (Py-AC) or 4-acetyl-4’-N,N-dimethylaminobiphenyl.

(ADAB). These molecules are pretwisted in the ground state and after excitation the twist

angle can increase or decrease. The experimental results obtained for Py-BN, Py-AC and

ADAB favour the hypothesis of the excited state flattening of the molecular skeleton, what

can be well explained in the terms of the SIF (solvent induced flattening) model, see chapter 6

and 7.

Irrespectively of the model (TICT or SIF) the increase or decrease of the twist angle is

increasing the excited state dipole moment which generates the solvent cage reorganization.

This conclusion rises the question of the separation of the environmental and intramolecular

relaxation. Unfortunately, this idea was not fully realized.

The relaxation times ( s) are generally evaluated from room temperature solvent

response functions and frequently are shorter than the temporal resolution of my equipment (6

ps). Low temperature solvent relaxation times were reported in literature occasionally, for

example for butyronitrile. Using as the probe cumarine 153 (C153), an effort was done to

measure the s values for diethylether and MTHF as a function of temperature. Unfortunately

the low temperature TRF spectra of C153 exhibited not only time dependent shift generated

by the solvent cage reorganization, but also temporal evolution of the shape of the spectrum.

Indeed the experiment performed in the supersonic jet indicates the presence of two

conformers of C153 which differ significantly in the Franck-Condon activity of one low-

frequency vibrational mode [Pryor et al., 1999]. That is the reason why the low temperature

values of s evaluated from c(t) functions for C153 in diethylether and MTHF were not used

for the interpretation of the experimental results obtained for the molecules being the subject

of the study. This project will be continued for C153 as well as for the DCS probe (4-

dimethyl-4’-cyanostilbene).

Proper evaluation of the time dependent Stokes shift of the primary (Fb) and secondary

(TICT) fluorescence in the case of carbonyl derivatives on N,N-dimethyl and diethylaniline

failed, (Fig 1.5). The maximum of the Fb fluorescence is located near the strong excitation

line, what extorts applying the cut-off filters. Consequently the shape of the primary band was

significantly disfigured and the time dependent shift of its maximum was not properly

assigned (Fig. 7.5). From the other hand, the TRF band of the TICT emission was broad and

due to its long lifetime, very weak, what makes impossible the precise determination of its

maximum (Fig. 7.5).

2) Excited state proton transfer

Excited state proton transfer can be inter- or intramolecular, frequently the domination

of the one of these mechanisms depends on the solvent.

The 2,9-(di-2’-pyridyl)-4,7-di(t-butyl)carbazole [Wiosna-Sałyga et al., 2006] and

pyridyloindole [Dobkowski et al., 1998] were investigated with application of time-resolved

technique.

An interesting example of the intramolecular proton transfer is the 2,5-bis-(6-methyl-

2-benzoxazolyl)phenol (BMP) in which the excited state reaction is promoted by some

5

vibrations [Luzina et al., 2007]. Low temperature TA and TRF spectra as well as the kinetics

of the primary species (decay) and the secondary one (rise) in excited state were recorded.

The scheme of the degradation of the excitation energy was evaluated. These results are the

part of the PhD thesis of Elena Luzina.

The results presented shortly above, however fitting well the profile of my habilitation thesis,

were not included, because in my opinion this data can not be treated as a closed

homogeneous project.

3) Excited state relaxation processes in the case of porphycene and its derivatives

embedded in low temperature rare gas matrices.

Application of rare gas matrices (RGM) for isolation and spectral characterization of

chemical individua has been initiated by Dewar at the end of 19th

century. It took further few

decades until the cryogenic methods were sufficiently improved and until the true value of

RGM for spectroscopic work was recognized [Whittle et al., 1954], [Andrews and Moskovits,

1989]. The RGM are characterized by chemical inertness and spectral transparency from far

IR to vacuum UV. That is the reason why matrix isolation technique is commonly used for

generation, trapping and stationary spectral studies of unstable compounds. For time-resolved

spectroscopy this experimental method has same additional advantage - relaxation processes

occurring after excitation should be slowed down. It should be pointed that accurate

registration of the transient absorption spectra of the molecules incorporated in the RGM, due

to destruction of the matrix by excitation beam and strong scatter radiation, is a difficult and

time consuming task.

Porphycene (PC), due to its large energy gap between the Soret and Q bands (about

11000 cm-1

) was a good candidate for the investigation of the intramolecular relaxation

processes. Additionally, it was well known that for porphycenes and porphyrins at room

temperature the relaxation between excited states occurs in subpicosecond time domain. For

example in the case of the Zn-tetraphenylporphyrin the energy degradation between the Soret

and Q bands took place within a time of 60-90 fs [Gurzadyan et al., 1998].

Monitoring of the stimulated fluorescence originating from the lowest singlet state of PC

incorporated into the RGM allowed me to determine the vibronic relaxation time of about 100

ps, (see chapter 3).

Prospective projects

On the basis of the literature data I may risk the opinion that the picosecond transient

absorption spectroscopy of the individuals incorporated in RGM’s is the speciality of our

laboratory. This statement is due to the fact that for the realization of such experiment two

methods should be well mastered: matrix isolated technique and time-resolved spectroscopy,

what is not a widespread case. Up to-day all experiments were performed at temperatures

above the temperature of the liquid helium. The experiments at temperatures lower than 4.2 K

can help to understand the mechanism of energy degradation and energy transfer from

incorporated molecule to the matrix in the case of porphyrin, porphycene and their

derivatives, and probably opens new fronts of investigations.

The fast progress in the technology of the infrared sensitive CCD detectors stimulates the

progress of the IR time-resolved spectroscopy. The results in nano [Hashimoto and

Hamaguchi, 1995] and picosecond time domain [Kwok et al.,2000], [Kwok et al., 2003],

[Okamoto et al., 2000] obtained for the p-dimethylaminobenzonitrile (DMABN) show

temporal evolution of the vibronic spectra recorded for S1 state. These experimental results

6

supported by the quantum chemical calculations can accurately determine the excited state

relaxation path. In my opinion the progress in the field of time-resolved spectroscopy in

infrared domain will be fast and it is reasonable to develop this experimental technique.

Other project, not connected directly with the picosecond time-resolved spectroscopy, is

the UV induced time-resolved NMR experiment (chapter 2 and 4). This technique permits

monitoring of the non-Boltzmann population of the ground state individuals generated by the

excited state reaction. This type of the experiments can be useful not only in the case of DA

molecules, but also in the case of other excited state reactions, for example proton transfer

processes.

7

2. Apparatus

A. Picosecond absorption spectrometers

Selection of the method

To record the transient absorption, bleaching and stimulated emission the “pump-

probe” method invented by Porter and Norrish [Porter, 1950] more than 50 years ago

is commonly used until to-day, starting from milliseconds to femtoseconds [Zewail,

2000], for review see [Naskręcki, 2000].

Own construction

The first generation Transient Absorption Spectrometer (TAS1) was based on the

Nd:YAG laser (EKSPLA/EKSMA, Lithuania) emitting pulses of duration 25 ps

( =1064 nm, repetition 3Hz); as the detectors photodiode linear sensors were used

(Fig. 1.2) [Dobkowski et al., 1995]. The second generation system (TAS2) was

equipped with the Nd:glass laser (Light Conversion Ltd, Lithuania), emitting pulses

of 1.3 ps duration ( =1055 nm, repetition 33 Hz). As the detectors, CCD matrices

were used.

TAS1

The computer controlled picosecond spectrometer (Fig.2.1) is based on the

passively mode-locked Nd:YAG laser as the light source. The negative feedback

control system stabilizes the pulse train. A single pulse ( 25 ps fwhm, =1064 nm)

was selected by an opto-electronic pulse selector, amplified and split into two beams

(pulse energy about 30 mJ each). The first beam is directed into the KDP crystals to

generate higher harmonic frequencies: second, 532 nm (15 mJ), third, 355 nm (2 mJ)

and fourth, 266 nm (7 mJ). In most experiments the third harmonic was used as the

excitation beam. The system was supplemented with an optical parametric oscillator

(PG411 VIR, EKSPLA, Lithuania), pumped by the 4th

harmonic of the Nd:YAG

laser, which enables tuning of the excitation between 300-500 nm and 570-2300 nm.

The second beam of the fundamental frequency, optically delayed with respect to the

excitation, was focused inside the cell with D2O to generate the pulses of picosecond

continuum (400-800 nm), used next as a probing pulse. As any large variation of the

pulse intensity (1064 nm) can affect the spectral distribution of the continuum

[Rullière, 1983], the intensity of the latter is controlled by a photodiode (PD) at

fixed the 600 nm .

The spectral distribution was monitored by splitting the continuum in two,

probe and reference beam, each focused onto the slit of the corresponding

polychromator, P(1) or P(2), respectively. Each Jasny model polychromator contains

a concave holographic grating (JOBIN-YVON) with a flat focusing field where the

photodiode array (Reticon RL 1024S) was fixed. Spectral resolution is about 0.4 nm/

diode. The analog signals from the diode arrays were digitized and transmitted to the

buffer memory, and next to the PC. The probe pulses were delayed in the specially

designed variable optical delay line with the roof mirrors. The length of the optical

8

path in the loop (maximum delay 3.4 ns, single step t=0.1 ps) was controlled by the

computer. Both, probe and excitation beam, cross nearly collinearly the sample (Fig.

1.2). Finally, the probe beam is focused onto the slit of the probe polychromator P(1),

spectrally resolved and detected by the photodiode array, as in the case of the

reference channel.

During the experiments with multiple excitation laser pulses, there is a danger

of accumulation of any stable photoproduct in the optical path in the sample, specially

in viscous and rigid media. Therefore, a mechanical device was constructed which,

before every new excitation pulse, moves the sample in the plane perpendicular to the

optical path in such way that the focus of the laser beam draws a Lissajoux curve, and

each excitation pulse irradiates a new intact part of the sample volume [Jasny et al.,

1994]. To record spectra in liquid phase at low temperature the cryostats was installed

in the mechanical holder described above, accessible temperature range was 293K-

110K. To perform experiments with compounds incorporated in low temperature inert

gas matrices Displex 202 or Displex DE 202s cryostats were used. These cryostats

can stabilize temperature within the range <40K-5K>.

OPG

computer

Pd

P(1)

P(2)

Sh

F

F

Sample

Nd:YAG

4H

DL

D O2

HG

10

54

nm

Pum

p

P P

EEE

Probe

F

Fig. 1.2 Optical scheme of the picosecond spectrometer for transient absorption

measurements (TAS1). F- optical edge filters, P-quartz plates, Pd-photodiode, E-

electric connections, P(1), P(2) – polychromators with linear photodiode array

detectors, Sh-shutter, DL-delay line, HG-harmonic generator, OPG-optical parametric

generator.

9

Measurements procedure

The sequence of the tasks is fully controlled by computer. First, the intensity (I) of the

50 pulses of the continuum are measured at 600 nm by the photodiode (Fig. 1.2, Pd),

and their mean (IM) and standard deviation ( ) are statistically evaluated by the

computer program. Only the laser shots resulting in continuum intensities within the

range IM or alternatively I IM- are accepted.

1. START,

2. Calculations of the parameters: IM and ,

3. Input of the user-defined parameters,

4. Delay line setup,

5. Calibration procedure,

6. Reset of the photodiode arrays,

7. Laser trigger,

8. Laser Pulse selection,

9. Array data reading,

10. Numerical procedure and

graphic presentation of the spectrum,

11. Data recording.

The transient absorption spectrum at a given delay time td is calculated from the probe

and reference signals (Sp and SR respectively), with (S) and without (S0) excitation.

The transient absorbance is:

A( ,td) = log { SR( ,td) [SP( ,td)]-1

K( )},

Where the calibration curve

K( )=S0

P( )/S0

R( )

The calibration was carried out for each series of experiments. To improve the S/N

ratio, 25-100 spectra are accumulated. The temporal resolution of the system was 30

ps.

TAS2

The optical scheme of the TAS2 was similar to that presented in the figure 2.1 (pump

and probe technique). The differences between TAS1 and TAS2 are pointed below.

A) A negatively feedback controlled active-passive mode-locked Nd:glass laser

generates pulses of 1-1.3 ps. The pulses, then expanded up to 300 ps in a

regenerative amplifier, are amplified to 7 mJ . The amplified pulses are

recompressed to 1-1.3 ps in two-pass pulse compressor. The output pulse energy

exceeds 4 mJ at 1055 nm with 33 Hz repetition The laser is equipped with a

harmonic generator and an optical parametric generator “Topas”. “Topas” is

pumped by the 2nd

harmonic (527 nm) emitting radiation within the tuning range

640 nm – 2600 nm (Emax.~1mJ), with an optional second harmonic generator 320

nm – 640 nm (Emax.~0.1mJ) .

B) Detectors – One-stage TE-cooled, back-thinned CCD, 1024px/128px, (S7031

series) product of Hamamatsu.

C) Measurements procedure: in the case of TAS1 both laser and detection system

were fully controlled by computer. In the case of TAS2 there is “partnership”

relation between the laser and the computer. The laser generates pulses with

repetition 33 Hz producing also temporally related pre-triggering electronic

signals. The first task of the computer program is to recognize the pre-triggering

signal, to establish the time zero as the time, when the next optical pulse is

10

generated by the laser, and to perform all others procedures as in the case of

TAS1. Because of high energy stability of the emitting pulses (<3% standard

deviation), many experiments have been performed using only one CCD detector.

The temporal resolution of the spectrometer was established recording the

transient spectra for molecules that show bleaching and transient absorption as the

primary effect, for example anthracene, pyrene, perylene, copper

octaethylporphyrin (CuOEP). Figure 2.2 presents a typical procedure: construction

of the kinetic curve of the rise and decay of the transient absorption, calculation of

the differential curve, Gaussian curve fit.

Table 1.2

Result of the Gaussian fit: FWHM, Max (10 steps=1 ps).

Compound integration limits [nm] [ps] Max [step]

Pyrene < 413 – 455 > 2 0.2 2495

Anthracene < 497 – 643 > 2 0.4 2524

CuOEP < 550 – 570 > 1.7 0.2 2528

Perylene < 638 – 740 > 2.2 0.1 2542

The effect of the light velocity dispersion (LVD) within the spectral window 400–

760 nm was estimated be equal to 5-6 ps.

Fig. 2.2 Perylene / cyclohexane, T=293K: A, normalized kinetics of the transient

absorption, integration limits <638 nm - 740 nm>. D, normalized first derivative of

A. Curve - normalized Gaussian fit, the maximum of the Gaussian curve corresponds

to 0 ps.

To obtain the kinetic curves of the emission rise and decay, the transient bands

recorded for various delay times were integrated within the selected spectral range.

The time “0 ps” corresponds to the maximum of the excitation pulse, see Fig. 2.2.

-15 -10 -5 0 5 10 15

0.0

0.5

1.0

A

Gauss

D

Perylene/cyclohexane

A.U

.

ps

11

B. Time-resolved spectrofluorimeter (TRSF)

Selection of the method

For excited state reaction, monitoring the of the spectral evolution and associated

dynamics of the emission decay is important. Thus the emission lifetime and its

spectral changes should be measured. To execute these tasks various methods were

proposed and successfully applied:

a) A single-channel method based on frequency-domain analysis of emission excited

by modulated sources. For this method the resolution is limited by electronics to

20-30 ps.

b) Streak camera is a useful tool to use for emission spectra and their kinetic

measurements. Time resolution reaches single picoseconds. The main limitation

is the dynamic range of single-shot measurements.

c) Single photon counting, characteristic by its great dynamics, is a widely used

technique, especially when a high-repetition laser is available. The time resolution

is limited by the laser pulse duration and a response function of the electronic

devices. The response function can be measured directly using a scattering

sample. Applying the deconvolution procedure it is possible to reach time

resolution about 10 ps. Single photon technique is excellent for recording the

decay curves for selected obs. To use this technique for recording the of time

resolved fluorescence (TRF) is possible but not easy. Typically, TRF spectra are

generated from the set of decays taken about at 10 nm intervals spanning the

fluorescence spectrum. To deconvolve the instrument response from decay data

each decay was fit to a sum of exponents. The purpose of these fits is simply to

represent the decay curves and no physical meaning is ascribed to the derived

exponential parameters. The TRF spectrum at a given time t, S( ,t) is obtained

from the fitted decay series by relative normalization of different wavelengths

using the steady-state fluorescence spectrum [Maroncelli and Fleming, 1987].

d) Optical Kerr shutter – the cuvete with Kerr medium (for example CS2) is inserted

between two crossed polarizers light cannot be transmitted through such

arrangement of the optical elements. But when light, after passing through the

first polarizer, is focused into the Kerr medium together with the variable delayed

gate optical pulse, it can cross the second polarizer, - the shutter is open [Duguay,

1969]. The spectrofluorimeter equipped with the optical Kerr shutter working in

the picosecond regime was first constructed by C.Rullière and his students

[Gilabert, 1987] and next improved [Janusauskas, 1994], Janusauskas, 1997].

Presently this technique is used in the femtosecond time scale [Takeda, 2000],

[Kanematsu, 2000], [Nakamura and Kanematsu, 2003], [Arzhantsev and

Maroncelli, 2005].

e) Up-conversion method. The up-conversion is an optical sampling technique which

is based on the nonlinear properties of crystals. The laser beam is divided into two

parts. The first part excites the sample. The fluorescence is collected and focussed

onto a nonlinear crystal. The second part of the laser beam passes through an

optical delay line and is also focused on the nonlinear crystal. When the laser

pulse and emission are present simultaneously in the nonlinear crystal, due to the

frequency mixing, the up-converted signal is generated. Up-converted frequency

is determined by crystal orientation with respect to the incoming beams and by

optical frequencies of these beams. Delaying the second part of the laser beam

12

with respect to the first one (excitation) causes that fluorescence is sampled at

different delay times. The up-converted signal is than dispersed by a

polychromator and detected by a photomultiplier linked to a photon-counting

system. The time resolution of the up-conversion technique is basically limited by

the temporal width of the laser pulse; however, the group velocity dispersion

induced by the nonlinear crystal and by various optical elements may alter the

time resolution. The up-conversion method is predominantly used in femtosecond

scale of time [Kahlow et al., 1988], [Horng et al., 1995,], [Rosenthal et al., 1994].

f) Recently a new method of recording the of time-resolved fluorescence was

reported, based on noncollinear parametric amplification in beta barium borate

crystal [Fita et al., 2005]. This method provides time resolution of the order of

100 fs.

The picosecond spectrofuorimeter equipped with optical Kerr shutter was

constructed because it works well in picosecond time scale and, additionally, the

detection system for low intensity signals monitoring was accessible. The other

advantage of this method is a possibility of recording, after excitation, the whole

fluorescence spectrum (time-resolved fluorescence, TRF). Thus the spectral evolution

of the fluorescence can be easily observed as a function of time. An inconvenience of

this method is that the dynamics of the decay curves is much smaller than in the case

of single photon counting.

Own construction

The first generation system (TRSF1) was based on the Nd:YAG laser

(EKSPLA/EKSMA, Lithuania) emitting pulses of duration 25 ps ( =1064 nm,

repetition 3Hz) [Dobkowski et al., 2004]; the second generation system (TAS2) was

equipped with the Nd:glass laser (Light Conversion Ltd, Lithuania), emitting pulses

of duration 1.3 ps ( =1055 nm, repetition 33 Hz). In both cases CCD matrix ,

((Princenton Instruments Inc.) was used as the detector.

TRSF1

The spectrofluorimeter is based on a passively mode-locked Nd:YAG laser

system (EKSPLA, Lithuania) as the light source (Fig. 3.2).

The negative feedback control system stabilizes the mean pulse energy. A single pulse

( t=25 ps fwhm, =1064 nm, repetition rate 3 Hz) is selected by an opto-electronic

pulse selector, amplified and split in two channels (pulse energy about 20 mJ each).

The first beam is directed into the KDP crystals to generate the third harmonic

( =355nm) used as the excitation beam. The second beam is passed through the

optical Kerr shutter, (Fig. 3.2) consisting of a cell with CS2 or chlorobenzene placed

between two crossed Glan polarizers. Fluorescence emitted by the sample is collected

and transferred by two spherical off axis parabolic mirrors (Janos Technology Inc.).

The laser pulse (1064 nm) opens the Kerr shutter, so that the fluorescence can

be transmitted by the shutter only during the time window of 30 ps. The opening pulse

is delayed with respect to the excitation pulse by an optical delay line (maximum

delay 3000 ps, 0.1 ps/step). Next, fluorescence is transmitted by a quartz fiber to the

detection system. The detection system consists of the: polychromator (Acton

Research Corporation ), an image intensifier, and a N2 cooled CDD detector, products

of the Princeton Instruments Inc.(Fig. 3.2).

13

The signal from the negative feedback unit of the Nd:YAG laser is used for

intensifier triggering. The spectra were accumulated 60 times, and the background

(spectrum recorded when no gating pulse opens the shutter) has been substracted. The

linear correlation between the energy of the opening pulse and the intensity of the

time resolved fluorescence was checked using rhodamine 6G in methanol (see Fig.

4.2).

3H

IntensifierCCD

PC

Trigger

F L M

C

P

PM 1

P

PM 2

Sample

Fig. 3.2 Optical scheme of the TRSF1. 3H,-third harmonic generator,

F-filter,L-lens, M-mirror, PM- off axis parabolic mirror, P- Glan polarizers,

C- cell with CS2 or chlorobenzene; dotted lines-electric connections.

Fig. 4.2 The TRF intensity of rhodamine 6G in MeOH recorded as a function of

the relative energy of opening pulse in the case of CS2 and chlorobenzene as the Kerr

medium. Lines - the result of the fitting applying the linear function (for

0 20 40 60 80 100

0

100000

200000

300000

400000

500000

C S2

F l u

o r

e s

c e

n c

e

I

n t

e n

s i t

y

E / Emax

[%]

0 20 40 60 80 100

0

100000

200000

300000

400000

500000

C h l o r o b e n z e n e

F l u

o r

e s

c e

n c

e

I n

t e

n s

i t

y

E / Emax

[%]

14

chlorobenzene two initial points were neglected). Typical limits of E/Emax was 60% -

80% (Emax=20 mJ)

Temporal resolution ( t) of the spectrofluorimeter was measured by delaying

the opening pulse (1064 nm) with respect to the second harmonic (CS2) or to the third

harmonic (chlorobenzene), see Fig. 2.5. In the case of CS2 WFHM was equal to 30 ps,

in the case of chlorobenzene t =60 ps. This significant difference of the t value is

connected with the optical response of the Kerr medium. For chlorobenzene

relaxation time(s) are not available, for benzene the “slow” response due to the

reorientational dynamics was reported [Horng, 1995] . The chlorobenzene was used

as the Kerr medium because the absorption edge of CS2 starts at 400 nm and

consequently the fluorescence with wavelengths shorter than 400 nm can not be

recorded.

Fig. 5.2 Relative intensity of the second harmonic (532 nm) and the third harmonic

(355 nm) recorded as a function of the delay time of the opening pulse

(1064nm) for CS2 and chlorobenzene used as Kerr media respectivly.

Experimental points were fitted to a Gauss function.

TRF spectra were not corrected for the apparatus spectral response.

To obtain the kinetic curves of the emission rise and decay, the time resolved

emission bands recorded for various delay times were integrated within the selected

spectral range. Time “0 ps” corresponded to the maximum of the excitation pulse, see

Fig. 5.2.

TRSF2

The”philosophy” of the optical scheme of TRSF2 was similar to the TRSF1

(Fig. 3.2), however some modifications were done: the Nd: YAG laser was replaced

by the Nd:glass laser. In the case of TRSF2 the excitation was collinear with

emission. As the Kerr media CS2 and C2Cl4 were used. The temporal resolution of the

-150 -100 -50 0 50 100 150

0.0

0.5

1.0

CS2

chlorobenzene

A.U

.

ps

15

spectrofluorimeter (C2Cl4) was determined by delaying the opening pulse with respect

to the second and third harmonics, see figure 6.2. The temporal resolution in the case

of CS2 was found be equal to 4.7 ps (527nm), 5.5 ps (351 nm). The light velocity

dispersion (LVD) being negligible for TRSF1 for TRFS2 plays an important role (see

Fig. 6.2)..

The TRF spectra of pyrene in cyclohexane recorded just after excitation (2050

step) ad 21 ps latter (2260 step) are shown in Fig. 6.2. Spectral evolution of this band

is due to the DLV occured mainly in CS2 cell and in the first Glan polarizer. To obtain

the kinetic curves of the emission rise, the time resolved emission bands recorded for

numbers of delay times were integrated within the selected spectral range. The shift of

the intensity rise depends on spectral range (Fig. 6.2). Within the spectral region 425-

510 nm the LVD effect was estimated be equal to 3.3 ps.

Fig. 6.2 Left, TRF spectra of pyrene in cyclohexane recorded for delay times 2050

step and 2260 step (10 steps=1 ps). Right, the kinetic of the rise of first and second

fluorescence band.

To establish the relation between the energy of the opening pulse and TRF response

and between the excitation pulse and TRF intensity, coumarine 153 in butyronitrile

was used as the standard.

Fig. 7.2 The TRF intensity of coumarine 153 in butyronitrile recorded as a function

of the opening pulse energy in the case of CS2 and C2Cl4 as the Kerr medium.

Lines, the linear fits. Eexc =0.30 mJ at 355 nm.

0 1 2 3 4 5

0

200000

400000

600000

800000

1000000

1200000

1400000

CS2

<490nm

- 5

30nm

>

mJ1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

0

20000

40000

60000

80000

100000

120000

140000

C2Cl

4

<490nm

- 5

30nm

>

mJ

1900 2000 2100 2200 2300 2400

0

50000

100000

150000

200000

250000

<455nm-485nm>

<425nm-455nm>

< I

nte

nsity >

step

350 400 450 500 550 600

0

5000

10000

R.S.

2260 [step]

2050 [step]

Flu

o. In

tesity

nm

16

Fig. 8.2 The TRF intensity of coumarine 153 in butyronitrile recorded as a function

of the excitation pulse in the case of CS2 and C2Cl4 as the Kerr medium. Lines – the

result of the fitting of the linear function. Energy of the opening pulse 4.7 mJ.

TRF spectra recorded under the same conditions for coumarine 153 (Fig. 8.2)

when CS2 was used as the Kerr medium were ten times more intense than those

recorded in the case of C2Cl4. The efficiency of the Kerr shutter is described by the

formula: Eg = sin2 ( LIg/ ) were - is the nonlinear refractive index of the medium,

L - thickness of the Kerr medium, Ig , – intensity and wavelength of the opening

pulse. In our case the efficiency of the optical Kerr shutter depends on the index of

the medium only.

TRF spectra were not corrected for the apparatus spectral response (ASR). It

should be stressed, however that ASR functions were monitored using the calibrated

lamp and were approximately constant within the spectral ranges selected for the

experiments.

Data fitting procedure (absorption and emission)

The decay times were evaluated applying the deconvolution procedure with

nonequal time intervals (“Origin” or “Igor Pro” programs). Using the techniques

presented above the kinetic curves were constructed with relatively small number of

the points, so that most important was to analyze the residuals. A typical result is

presented below.

0.0 0.1 0.2 0.3

0

1000

2000

3000

4000

C2Cl

4

Flu

o. In

tesity

mJ

0.0 0.1 0.2 0.3

0

10000

20000

30000

CS2

Flu

o. In

tensity

mJ

17

Fig. 9.2 The kinetics of the TA of 4-Acetyl-4’-N,N-dimethylaminobiphenyl in EtOH

at room temperature. Circles, experimental points; top - residuals

Bottom - curve, biexponential fit: pulse profile-gaussian, pulse width-2.5 ps, pulse

maximum-159.4 0.4 ps, 1=21.5 0.4 ps, 2=52 5 ps, amplitudes: A1=41 9,

A2=-37 9.

C. UV induced time-resolved NMR

Fluorescence lifetimes of many organic molecules are usually single

nanoseconds. The acquisition time of NMR spectrum corresponds to the miliseconds

scale of the time. This is the reason why, until today nobody was able to record NMR

spectrum of the excited molecule. Using NMR technique it is possible, however, to

recognize the excited state reaction path by monitoring the temporal non equilibrium

population of the ground state conformers generated by excited state relaxation

process - , 1990].

-10

-8

-6

-4

-2

0

2

O D

500450400350300250200150Delay Time, ps

-2

-1

0

1

ADAB/EtOH, 28.11.03, 570-650 nm

2 exp fitting:

Trsp=2.5 ps (fixed)

Zero=159.4+-0.4 ps

Tau1=21.5+-2.7 ps

Tau2=52+-5 ps

A1=41+-9

A2=-37+-9

adb6k7

fit_adb6k7

Res_adb6k7

18

Apparatus

Fig. 10.2 Time resolved UV-NMR apparatus. L-lens, F-quartz lightguide, Sp-sample,

P-pulse sequence generator.

NMR spectrometer - Varian 500MHz was chosen. The LPx100 excimer laser

product of Lambda Physics was selected as the source of pulses. The excitation

wavelength was 308 nm, the energy per pulse 90 mJ-60 mJ (laser output). Because of

strong magnetic field around the NMR spectrometer laser was located at the distance

about 8 m. Due to the same reason all mechanicals details of the optical part of the

apparatus was constructed from nonmagnetic materials. The care was done to

minimalize the length of the lightguide (in quartz fiber for =308 nm the loss of the

pulse intensity are significant).

Rare gas matrix isolation apparatus

This technique was described in details earlier [Kołos, 2003]

Laser

NMRSpectrometer

Sp

L

P

F

19

Single photon counting apparatus

Fluorescence decays in nanosecond time domain were recorded with a time resolved

single-photon-counting apparatus, Edinburgh FS 900 CDT. Temporal resolution of

this apparatus was 0.2 ns.

Steady–state apparatus

Absorption spectra were measured with a Shimadzu UV3100 spectrometer. Emission

spectra were recorded using an Edinburgh FS 900 CDT or Jasny spectrofuorimeter

[Jasny, 1978].

The presented material was partially published:

1) J.Dobkowski, Z.R.Grabowski, J.Jasny, Z.Zieliński, Acta Phys.Pol., 88(1995)455,

(construction of TAS1).

2) J.Dobkowski, V.A.Galievsky, J.Jasny, I.V.Sazanovich, Polish J.Chem.,

78(2004)961. (construction of TRSF1)

20

A Time-resolved spectroscopy of

porphycene and its derivatives in low

temperature gas matrices.

Transfer of energy from the excited molecule into the matrix can be

slowed down in comparison with the liquid phase [Bondybey et al.,

1983], [Rentzepis et al., 1984], Huppert et al., 1985]. These studies

focused on excess energies below 2000 cm-1

. The relaxation behaviour

for larger excess energies is less known. That is the reason why

porphycene (PC) which reveals a large energy separation between S1,2

and S3,4 singlets ( 11000 cm-1

) was selected as a good candidate for

the time–resolved experiments.

The aim of this work was:

recording (what has been the experimental challenge) of the

transient absorption (TA) and time–resolved fluorescence (TRF)

spectra of porphycene (PC) and its derivatives in low temperature

gas matrices

elucidation of the mechanism of the energy degradation.

21

3.1 Time-resolved spectroscopy of porphycene embedded in

low temperature matrices

Porphycene (PC)

Stationary spectra Stationary absorption and emission spectra of porphycene in terahydrofuran at 293 K and in

nitrogen matrix at 15 K are shown in Fig.1.3. Matrix isolation leads to the appearance of rich

vibronic structure in the region traditionally named Q bands (S1 and S2 transitions) and in

fluorescence.

Fig. 1.3 Stationary absorption and fluorescence of PC in nitrogen matrix at 15 K (top) and

tetrahydrofuran solution at 293K (buttom).

Detailed analysis of the spectra in various gas matrices has led to the assignment of many

vibrational frequencies in S0, S1 and S2 states, moreover exact origins of the S0 S1 and

S0 S2 have been determined [Starukhin et al., 1998]. In argon and nitrogen matrices , the S1-

S2 separation amounts 880 cm-1

and 882 cm-1

, respectively. Two high laying electronic states

(S3 and S4) are named Soret band. The energy separation between Q and Soret is about

300 400 500 600 700

0

1

300 400 500 600 700

0

1

Em

issio

n in

ten

sity (

arb

. u

nits)

Ab

so

rba

nce

nm

S1

S2

S3

S4

N

N

H

N

N

H

22

11000 cm-1

. That is the reason why porphycene is an attractive candidate for study of the

relaxation between excited states.

Time resolved experiments Transient absorption (TA)

The time resolved spectra of PC incorporated in low temperature gas matrices

recorded after excitation into the Soret band (355 nm) consists mainly of the absorption

bleaching. A weak transient absorption band within the spectral region 400 nm-500 nm is also

observed . Both these signals appear instantaneously after excitation. Additionally, negative

bands appear at longer wavelength than the (0-0) band of the S0 S1 transition. The intensity

and shape of these bands change with the time. The transient spectra of PC in Ar and methane

matrices at 30 K recorded as a function of the delay time are shown in the figure 2.3. Identical

temporal evolution of the transient spectra was observed in N2 and Kr low temperature

matrices, what indicate that such time-dependent spectral transformation is typical for PC in

these conditions of environment, and temperature.

Fig. 2.3 Transient spectra of PC in Ar (A) and methane (B) matrices at 27 K recorded as a

function of delay time, resolution (Re) 3 ps. The time dependent evolution of the bands

located to the red of the (0-0) line of S0 S1 transition are indicated by arrows. The spectra

have been shifted vertically upwards with respect to the spectrum recorded 6 ps after

excitation.

350 400 450 500 550 600 650 700 750 800

276 ps

76 ps

36 ps

6 ps

nm

A0.2

0

O.D

.

350 400 450 500 550 600 650 700 750 800

nm

6 ps

36 ps

76 ps

276 ps

0.2

O.D

.

B

0

23

On the contrary, if porphycene is excited into S1 or S2 states the bands at the red side of (0-0)

transition are observed simultanously with the bleaching (Fig. 3.3).

Fig. 3.3 Comparison of the transient spectra obtained for excitation into different electronic

states (a) exc=591 nm –Q band, (b) exc=355 nm – Soret band. Delay times from the moment

of excitation are indicated for each curve. Scattering of the pumping beam is marked by

“exc.”, Re=30 ps.

The signal that is delayed for higher energy excitation but appears simultaneously with

bleaching for excitation into Q bands corresponds to stimulated S1 S0 fluorescence,

generated by the probing pulse. This can be clearly seen by the inspection of figure 4 where

the stationary fluorescence is compared with the difference between the transient spectra

recorded at 260 ps and at 80 ps delay times, using the excitation onto the Soret bands.

Such subtraction leads to practically complete canceling of signals due to ground state

bleaching and S1 Sn transient absorption, because the difference in delay times of the two

transient signals is much shorter than the S1 lifetime ( 15 ns). This procedure also reveals a

strong 0-0 fluorescence line which otherwise have been difficult to detect, since it coincides

with the 0-0 band of S0 S1 absorption, and thus overlaps with the strongest bleaching feature

(Fig. 4.3).

Closer examination of the time evolution of the stimulated fluorescence monitored for

exc=355 nm reveals not only changes in intensity, but also in spectral shape (Fig. 2.3). At

shorter delay times, up to about 70 ps (Re=3 ps), and about 100 ps (Re=30 ps) this emission is

broad end weak, and may be observed as a structureless shoulder on the low energy side of 0-

0 band of the S0 S1 transition. At longer delay times, the two most prominent vibronic bands

of the fluorescence become visible during the time window available (up to 3 ns).

560 600 640

-1.5

-1.0

-0.5

0.0

(b)

*

* * 260 ps

140 ps

120 ps

10 ps

OD

OD

nm

nm560 600 640

-1.0

-0.5

0.0

(a)exc.

*

* *80 ps

30 ps

10 ps

24

Fig. 4.3 Porphycene in N2 matrix: A- the result of subtracting the transient spectrum

recorded at 80 ps delay from the signal recorded at 260 ps delay (Re=30 ps); B- high-energy

portion of stationary fluorescence.

Time resolved fluorescence (TRF)

TRF spectra of PC in solid argon and nitrogen were recorded as a function of the delay

time from the moment of excitation into the Soret band (355nm).The two different matrices

reveal practically the same behaviour. Just after excitation (20 ps) a broad, weak, unstructured

band is observed, showing time dependent blue shift. At about 70-90 ps the vibronic structure

of the fluorescence becomes visible. The spectral evolution is completed for delay times

longer than 140 ps, when the spectrum becomes indistinguishable from the steady state

fluorescence (Fig.5.3).

The experimental kinetic curve of the integrating intensity of the whole TRF spectrum

of PC in nitrogen could be fitted using a two-exponential model (Fig. 6.3). The fit yields a rise

time of 74 ps and a decay of 9200 ps. The latter value may contain a large error, because time

window accessible in picosecond spectrofluorimeter is 3 ns. Fluorescence lifetimes obtained

at low temperatures by single photon counting in various media are about 12-14 ns. The value

of the rise time should be also be treated with caution, because the sample is not perfectly

stable under prolonged irradiation , required to collect many points at various delays.

600 620 640

Em

issio

n (

a.u

.)O

D

nm

nm

B

600 620 640

-0.4

0.0

*

**

A

25

Fig. 5.3 TRF spectra of PC excited into the Soret band (355 nm) in argon (top) and nitrogen

(bottom) matrices at 15 K. Delay times next to the corresponding spectra, Re=30 ps,

exc=355 nm.

On contrary, if PC is excited into the S1 or S2 states, the TRF spectra exhibit the relaxed

structure immediately after excitation (Fig. 7.3). No temporal evolution of the spectra could

be detected. Because the temporal resolution of the apparatus is 30 ps, it means that the

relaxed state is achieved in less than 30 ps after excitation.

600 610 620 630 640 650 660 670

20 ps 40 ps50 ps

70 ps

90 ps

140 ps

Flu

ore

sce

nce

In

ten

sity

(arb

.un

its)

nm

Ar

600 610 620 630 640 650 660 670

140 ps

90 ps

70 ps

50 ps

40 ps20 ps

Flu

ore

sce

nce

In

ten

sity

(arb

.un

its)

nm

N2

26

Fig. 6.3 The kinetic curve of the emission intensity in N2 matrix obtained after integrating

the whole spectrum at various delays, experimental points (squares); biexponential fit (solid

curve) R

=74 ps, =9200ps, for comparison monoexponential fit (dotted curve) =8900 ps.

Fig. 7.3 TRF spectra of PC in solid nitrogen at 15 K excited into the Q band (593 nm),

Re=30 ps.

The low temperature matrix data are compared with the corresponding room

temperature measurements in solution. No significant spectral evolution could be detected in

the TRF spectra recorded for PC in tetrahydrofuran, for excitation into the Soret region (Fig

(8.3). The initial intensity rise is due to the excitation profile. The spectra recorded just after

excitation show a small red shift of the maximum recorded about 1 ns after excitation.. Closer

examination of the shape of the TRF band recorded for short delay times reveals a shoulder

located at blue side of the 0-0 transition. This shoulder, of which the contribution is largest at

0 500 1000 1500 2000

0

1000

2000

3000

4000

Int.

<6

15

nm

- 6

56

nm

>

ps

N2

590 600 610 620 630 640 650 660

990 ps

90 ps

60 ps

40 psFlu

ore

scence inte

nsity (

arb

. units)

nm

27

the short delay times , can be tentatively assigned to the fluorescence originating from the S2

state

Fig. 8.3 Room-temperature TRF spectra of PC in tetrahydrofuran exc=355 nm, Re=30 ps.

Delay time assigned with respect to the maximum of the excitation pulse.

In order to interpret the results, the excited state relaxation path can be divided into

three stages: (a) internal conversion (IC) from S3/S4 into S1/S2, (b) intramolecular vibration

relaxation (IVR), (c) solute-solvent energy transfer; [Dobkowski et al., 1999], [Dobkowski et

al., 2004].

For PC in raer gas matrices the bottleneck for reaching the lowest excited singlet state

is not IC conversion. If the IC process were slow, one would expect to observe fluorescence

in the Soret region. Assuming lifetime for the S3 is 100 ps and the same radiative constant as

for lowest singlet state the value of 0.004 for the expected of S3 S0 fluorescence quantum

yield is predicted. However, big experimental effort was done, the fluorescence originated

from the Soret region was not detected. The lack of the S3 S0 fluorescence indicates that

about a 100 ps delay necessary to evolve S1 S0 emission into a structured spectrum is due to

the vibrational relaxation or/and slow energy transfer to matrix.

The TRF studies have also been carried out for for 2,7,12,17-tetra-t-butylporphycene

(PCTT), (Fig. 9,10.3). This molecule was selected for two reasons:

four t-butyl groups introduce low frequency modes, not present in PC,

geometry as well as the spectral properties of PCTT are similar to PC, so that the possible

differences due to factors other than the number of low frequency modes should not be

significant.

450 500 550 600 650 700 750 800

0

5000

10000

15000

20000

25000920 ps 20 ps

10 ps

-20 ps

nm

TR

F

Inte

nsity

28

Fig. 9.3 TRF spectra of PCTT in argon matrix at 15 K, exc=355 nm, Re=30 ps.

Fig. 10.3 The initial part of the kinetic curve of the emission intensity of PC in Ar at 15 K

obtained after integrating the whole spectrum at various delays, squares – experimental

points, A-biexponential fit, R

=26 3 ps, =2500 ps, B-monoexponential fit, =2600 ps,

exc=355 nm, Re=30 ps.

The time evolution of the TRF spectra of PCTT is similar to that observed for PC (Fig.

9.3). A broad , structureless band recorded just after excitation shifts to the blue; this is

accompanied by the appearance of vibronic structure. The rise time of the emission detected

for PCTT is significantly shorter than in the case of PC, (Fig. 10.3). This indicates that the

low frequency modes of C(CH3)3 groups accelerate the process of the temporal evolution of

the emission.

600 620 640 660

0

5000

N

N

H

N

N

H

140 ps

90 ps70 ps40 ps20 ps

Flu

ore

sce

nce

In

ten

sity

(A.U

.)

nm

R

R

R

R

R=C(CH3)3

-50 0 50 100 150 200

0

5000

10000

15000

20000

N

N

H

N

N

H

TR

F in

ten

sity

<6

10

nm

-64

0n

m>

ps

R =26 ps (3ps)

A

B

RR

RR

R=C(CH3)3

29

3.2 Time resolved spectroscopy of the 2,7,12,17-tetra-tert-butyl-3,6,13,16- dibenzo[cde,mno]porphycene

(TTPC)

TTPC

Fig. 9.3 Stationary absorption spectra of PC and TTPC in hexane at 293 K: (a) and (c) and

in nitrogen matrix at 15 K: (b) and (d).

The absorption spectra of PC and TTPC are shown in the figure 9.3 . The absorption

pattern of TTPC is different from that of PC, with larger number of electronic states observed

in low energy region in the case of TTPC [Waluk et al., 1991], [Starukhin et al., 1998]. Tree

low-lying band system located at 910 15nm (11000 cm-1), 761 3 nm (13100 cm-1

) and

587 2 nm (17000 cm-1

) in the NIR/visible region were assigned [Dobkowski et al., 2005].

N

N

H

N

N

H

10 15 20 25 30

Hexane, 293K

(d)

(c)S

3S

2

S1

Ab

so

rba

nce

(a

.u.)

Ab

so

rba

nce

(a

.u.)

N2, 15 K

TTPC

15 20 25 30

exc.= 591 nm

exc.=355 nm

N2, 15K

Hexane, 293K

(b)

(a)

x103cm

-1

x103cm

-1

S4

S3

S2

S1

PC

exc.=355nm

30

Fig. 10.3 TTPC in Ar matrix at 27 K: A- low energy part of the stationary absorption

spectrum ; B-transient absorption and bleaching observed 13 ps after excitation at 352 nm.

Horizontal bars indicate the integration intervals used to extract the kinetic parameters. An

artifact due to the second-order of the pumping beam is marked by cross, Re=3 ps.

Fig. 11.3 Kinetic curves of transient bands of TTPC in MTHF at 294 K (left) and at

123 K (right). Circles–experimental results, solid curves-monoexponential fit, Re.=3 ps.

Values in brackets indicate error of the calculated decay times.

400 500 600 700 800

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

B

AO

.D.

nm

760 nm

352 nm

-20 0 20 40 60 80-2

-1

0

1

2

3

4

5

Inte

gra

ted

in

ten

sity

[A

.U.]

<4

25

nm

- 5

00

nm

>

=13 ps (2ps)

ps

T=123 K

-20 0 20 40 60 80

-2

-1

0

<5

60

nm

- 6

00

nm

>

ps

Inte

ga

rte

d in

ten

sity

[A.U

.]

=10 ps (2ps)

-20 0 20 40 60 80

-2

-1

0

<7

30

nm

- 7

60

nm

>

ps

=12 ps (2ps)

Inte

gra

ted

in

ten

sity

[A

.U.]

-20 0 20 40 60 80-2

-1

0

1

2

3

4

= 9 ps (2 ps)

ps

<4

25

nm

- 5

00

nm

>

Inte

gra

ted

in

ten

sity [A

.U.] T=294 K

-20 0 20 40 60 80

-2.0

-1.5

-1.0

-0.5

0.0

0.5

= 11 ps (2 ps)

<5

60

nm

- 6

00

nm

>

psInte

gra

ted

inte

nsi

ty

[A.U

.]

-20 0 20 40 60 80

-2.0

-1.5

-1.0

-0.5

0.0

= 11 ps (2 ps)

<7

30

nm

- 7

60

nm

>

ps

Inte

gra

ted

in

ten

sity [A

.U.]

31

Fig. 12.3 The kinetic curves of transient band for TTPC in Ar matrix at 27 K. exc.=355 nm.

Re.=3 ps . Circles-experimental results, solid curves-monoexponential fits, values in brackets

indicate the errors of the calculated decay times.

Table 1.3

Decay times obtained from transient absorption and bleaching of TTPC for different

environment, temperatures and excitation wavelengths.

Environment T [K] [ps]a

<425-500>b

<560-600>b

<730-760>b

Excitation: 352 nm

MTHF 293 9 10 11

MTHF 123 13 10 12

Ar matrix 27 18 16 20

Excitation: 760 nm

Ar matrix 27 17 16 19

Ar matrix 5.4 18 16 19

a – Estimated error: 2-3 ps.

b – Integration range [nm] of the transient bands, see also Fig. 11.3.

The transient spectra of TTPC recorded in different environments and temperatures

show a spectroscopic pattern almost identical to that observed in Ar matrix (Fig. 10.3). These

spectra consist of the transient absorption band with maximum about 440 nm and bleaching

bands with the maxima at 536, 580, 690,and 755 nm. The bleaching reproduces well, as a

mirror reflection, absorption spectrum recorded at the same conditions (Fig. 10.3). No

-20 -10 0 10 20 30 40 50

0

5

10

15

ps

= 18 ps (2 ps)

<4

25

nm

- 5

00

nm

>

Inte

gra

ted

in

ten

sity [A

.U.]

-20 -10 0 10 20 30 40 50 60

-3

-2

-1

0

ps

= 20 ps (2 ps)

<7

30

nm

- 7

60

nm

>

Ine

gra

ted

in

ten

sity

[A.U

.]-20 -10 0 10 20 30 40 50 60

-3

-2

-1

0

1

ps

= 16 ps (3 ps)

<5

60

nm

- 6

00

nm

>

Inte

gra

ted

in

ten

sity

[A

.U.]

32

difference was found between transient spectra recorded for 352 nm and 760 nm excitation.

The shape of the transient spectra does not depend on temperature.

The kinetic curves of the rise and decay of the transient absorption and bleaching are

shown in Figs 11.3 and 12.3. The calculated decay times are collected in Table 1.3. The decay

time of the transient absorption is equal, within the experimental accuracy, to the decay time

of the bleaching, independent of the excitation energy and varies with temperature and nature

of the medium within the factor 2. The independence of decay parameters of excitation

indicates that the internal conversion from higher electronic states and the vibronical

relaxation into S1 are faster than temporal resolution of the equipment , 3 ps.

There is no long-lived component observed in decays, indicating that the intersystem

crossing does not play any significant role on the relaxation path.

Discussion The presented results show that PC and PCTT, incorporated in low temperature

matrices, relax after excitation in different ways. Fluorescence decay times of PC lie in the

range of 10-15 ns, whereas in the case of PCTT no emission is observed, and decay time of

the S1 population evaluated from the TA and bleaching decay is about 18 ps (Tab. 1.3). The

S0-S1 energy gaps for PC and TTPC are 15900 cm-1

and 11000 cm-1

, respectively. This last

value is very similar to the S2-S3 separation in PC. However the S1 S0 relaxation time in

argon in the case of TTPC is not longer than 20 ps and the observed relaxation for S3/4 S1 in

PC takes about 100 ps (Fig.13.3). Somewhat similar results have been obtained for C60

isolated in rare gas matrices for which dissipation about 1.2 eV in less than 20 ps was

reported. This rapid relaxation brings the molecule into S3 state lying only about 100 cm-1

above the lowest excited singlet. The conversion from S3 to S1,2 was found to be slow, about

170 ps in neon and 70 ps in argon [Sassara et al., 1999], [Chergui, 2000], [Stepanov et al.,

2002].

For azulene excited in low temperature gas matrices into S2 state with excess energies

of 2000-5000 cm-1

the zero vibrational level is populated in about 20 ps [Hopkins and

Rentzepis, 1985]. This was interpreted assuming that the initially excited high-lying

vibrational states relax rapidly ( 5 ps) , populating one or more low-lying vibrational states.

These may act as bottlenecks to the further decay.

The energy transfer from the excited molecule to the matrix (bulk) can be treated as a

sequence of consecutive steps:

1 - internal conversion (IC),

2 - intramolecular vibrational relaxation (IVR),

3 - energy flow from the vibrational levels of the molecule into the neighboring shell of

matrix atoms,

4 - transfer of energy into the bulk of the matrix .

The results obtained for TTPC indicate that large amount of energy (about 2 eV) can

be dissipated in solid argon in less than a few ps. Therefore, it seems that for PC in gas

matrices steps (1) and (4) are faster than (2) and (3), and the difference in the relaxation

mechanisms between PC and TTPC in low temperature matrices must involve the transfer of

energy to the neighboring Ar or nitrogen atoms. Apparently, the intermode couplings are

much stronger in TTPC than PC, which is not surprising given the nonplanarity and lower

33

symmetry in the former. The intramolecular energy distribution into the low energy modes

should be fast in the case of TTPC, and the energy transfer from vibrational levels to phonons

can be effective. On the other hand, no evidence for large geometry distortion of PC in the

excited state is found, so that low energy modes may not be efficiently populated and ,

consequently, energy transfer is slow. This explanation is supported by three time faster rise

of the TRF in the case of PCTT, possessing four t-butyl groups which introduce low

frequency modes not present in PC. In the case of PC excitation into S2, which is located

about 900 cm-1

above S1, structured fluorescence is observed just after excitation what

indicates that vibrational relaxation is rapid. Thus, a”bottleneck” lies above S2.

Fig. 13.3 Scheme of the excited state energies and relaxation time scales in argon matrix in

the case of (A)-PC and (B)-TTPC.

The consideration presented above focused on the direct energy transfer in accordance

with the sequence of the events: Intramolecular vibration of PC skeleton phonons , this

mechanism is supported by the analysis of the oscillations of PC indicating that there are low

energy (58, 68, 83 cm-1

) out o plane vibrations [Waluk and Gil, 2007], corresponding well to

the phonon bands of Ar matrix (20-70 cm-1

), [Jodl, 1989]. The librations of PC molecule were

neglected.

The intramolecular vibrational energy dissipation in the case of free-base porphin

(H2P) incorporated in low temperature n-alkane hosts was monitored by the hole-burning

experiments [Dicker et al., 1981], [Dicker and Völker, 1982]. The experimental results were

interpreted in the term of “exchange model” [Harris 1977], [Shelby et al., 1979]. The

schematic diagram of the transitions involved in the “exchange model” is shown in Fig 14.3.

In accordance with this model intramolecular energy is transferred to the host via localized

phonon modes. The modes are attributed to librational motions of H2P molecule in the host

lattice. The energy of these librations is 5-40 cm-1

depending on the tightness of the fit of the

H2P in the host [Dicker, 1982]. Assuming a similar energy of the librations in the case of PC

and having in mind that in argon matrix accessible phonons are in the range of 20-70 cm-1

[Jodl, 1989] the alternative energy degradation scheme can be proposed: Intramolecular

0

5

10

15

20

25

En

erg

y x 1

03 [cm

-1]

ICS

3/4

S2

S0

100 p

s

IC

10

00

0 c

m-1

15 n

s

16

00

0 c

m-1

A

S1 9

00 c

m-1

0

5

10

15

20

25

<3 p

s

S4

S3

IC

IC

ICS2

S1

S0

10

00

0 c

m-1

40

00

cm

-1

20

00

cm

-1

11

00

0 c

m-1

18 p

s

B

34

vibration of PC skeleton Librations (localized phonons) phonons of the matrix (points 2

and 3).

Fig. 14.3 Schematic diagram of the transitions involved in the “exchange model”. The

energies of the low frequency modes are different in the ground and electronically excited

state, =E’-E.

One can argue that the great difference in the relaxation mechanism is only due to lack

of the appropriate low-frequency modes, because time dependent evolution of the TRF

spectra of PC is observed (Fig. 5.3). The evidence about significant transformation of the

excited state geometry of PC was not reported, but experiments performed in a supersonic jet

indicate that the barrier for proton transfer is much higher in S1 than in S0 lifetime [Sepioł et

al., 1998]. The fact that the barrier is larger in S1 may be caused by overall expansion of the

molecule after excitation. Recent fluorescence experiments performed on PC in polymer

films have shown that below 203 K the excited state proton transfer is slower than the

fluorescence lifetime. Consequently double minimum potential typical for PC can be

simplified (Fig. 15.3). To explain the temporal evolution of the TRF spectra the geometry of

the PC must undergo some transformation in the excited state.

Summary: The PC molecule incorporated in rear gas matrix and excited with above 1eV

excess of vibrational energy needs about 100 ps to relax reaching equilibrated S1. This

unexpected slow energy degradation can be explained by (i) lack of the appropriate low

frequency modes or alternatively (ii) by non efficient interaction between intramolecular

vibration and librations of PC.

S1,0>

S0,0>

S0,1>

S1,1>

E'

E

En

erg

y

E00

E00+

35

Fig. 15.3 Schematic cross section along intramolecular and matrix cage coordinate for PC

incorporated in low temperature gas matrix, 2 1 indicate the sequence of the events. The

proton transfer is slower than the fluorescence lifetime.

The presented material was partially published :

1) J.Dobkowski, V.Galievsky, A.Starukhin, J.Waluk, Chem. Phys. Letters, 318(2000)79.

2) J.Dobkowski, V.Galievsky, M.Gil, J.Waluk, Chem. Phys. Letters, 394(2004)410.

3) J.Dobkowski, Y.Lobko, S.Gawinkowski, J.Waluk, Chem. Phys. Letters,

416(2005)128.

0

5

10

15

En

erg

y

[A.U

.]

S1

S0

Geometry transformation, matrix relaxation

36

B. Intramolecular and solvent relaxation in the

case of molecules revealing charge-transfer

excited states.

The aim of this work is the recognition of the excited state relaxation path in

the case of molecules possessing donor and acceptor groups linked by single

bond and characterized by compact (dialkylamino) or extended (pyrene,

N,N-dimethylaniline) -electron systems of the donor unit; as the acceptor

acetophenone, benzaldehyde or m-cyanopyridine were selected. The

sequence of the events: charge separation, excited state structural

transformation and solvent cage reorganization will be monitored by time-

resolved spectroscopy in absorption and in emission and discussed in details.

37

4. An experimental test of C-N bond twisting in the

TICT state: the case of dialkylamino derivatives

of m-cyanopyridine

In this chapter are presented the results obtained for three electron donor-acceptor

molecules, consisting of a m-cyanopyridine electron acceptor and one of three slightly

different N,N-dialkylamino donor groups: N-methyl-N-isopropyl (PAC), N,N-diethyl

(PEC), and N,N-dimethyl (PC). For comparison, the structure of N,N-

dimethylaminobenzonitrile (DMABN) is also shown.

A. Steady-state spectroscopy

Room temperature experiments

Fig. 1 shows the room temperature absorption and emission spectra of PAC,

PEC, PC in nonpolar as well as in polar aprotic and protic solvents: n-hexane,

acetonitrile, and methanol, respectively. Upon passing to a polar environment, the

three compounds reveal qualitatively the same behavior. The absorption band

corresponding to the lowest transition is only slightly affected, whereas the next,

stronger transition exhibits a larger red shift. These two lowest excited singlet states

may be assigned as “Lb” and “La”-like.

The change in emission is much more dramatic: Dual fluorescence is observed

in polar solvents, manifested by the appearance of an “anomalous”, strongly red-

shifted Fa band, similar to, but considerably weaker than that observed in DMABN

[Zachariasse et al.,1997]. The Fa/Fb intensity ratio varies significantly for the three

compounds, and is the largest for PAC and smallest for PC. Actually, in the

acetonitrile solution of the latter, the low-energy band is barely visible. However, it

becomes clearly observed in alcohols, in which the Fa intensity is strongly enhanced

relative to Fb in all three compounds.

N

N

NC

N

N

NC

N

N

NC

N

PAC

i-pr Me Et Et Me Me

PEC PC

CN

MeMe

DMABN

38

Fig. 1.4 Room temperature absorption and fluorescence spectra of PAC,

PEC and PC in hexane, acetonitrile and methanol.

20000 30000 40000-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

PAC

ACN

MeOH

Hex.

Fb

Fa

Hex.

ACN

MeOHA.U

.

cm-1

20000 30000 40000-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Fb

Fa

PEC

ACN

MeOH

Hex.

Hex.

ACN

MeOH

A.U

.

cm-1

20000 30000 40000-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Fa

Fb

PCACN

MeOH

Hex.

MeOH

ACN Hex.

A.U

.

cm-1

39

By analogy with DMABN, we assign the Fa emission to a charge-transfer

(CT) state, whereas the “normal” Fb fluorescence is attributed to a locally excited

(LE) state. Figure 4.2 shows the shifts of the Fb fluorescence of PAC in different

solvents, plotted against the solvent polarity function:

12

1

2

1

12

1),(

2

2

n

nnF

(1)

From the solvatochromic data (Figure 2.4 and Table 1.4), a value of the excited state

dipole moment of the emitting LE state, e, can be calculated [Liptay,1962],

Liptay,1974] using the formula:

.),()(2~

3max constnFhcao

gee

(2)

Fig. 2.4 Plot of the solvatochromic shift for the Fb maximum of PAC;

see Table 1 for the numbering of solvents.

Assuming that g, the ground state dipole moment of PAC is similar to that of

DMABN (6.6 D) [Schuddeboom et al.,1992] and g is parallel to e,, a value of 10.8

D for e was obtained when using an Onsager radius a0 of 4.31 Å (estimated for N,N-

diisopropylbenzonitrile), and a value of 9.9 D for a0 = 3.87 Å (estimated for

DMABN). These results are very similar to the value of ca. 10 D reported for

DMABA. [Schuddeboom et al.,1992], A weak intensity and broad, structureless

character of the anomalous emission Fa preclude a quantitative determination of the

dipole moment of its emitting state. However, comparing the shifts of Fa and Fb in

0.1 0.2 0.3 0.4

26000

26500

27000

27500

28000

8

6

7

54

3

2

1

Fb m

axim

a [cm

-1]

F(n)

40

polar solvents (Table 1), it may be concluded that the Fa emitting state has a

considerably larger dipole moment.

Table 1.4

Spectral position of fluorescence maxima (Fa,Fb ) and total fluorescence quantum

yields () of alkyl derivatives of 5-cyanopyridine in various solvents at room

temperature. Polarity index given in parentheses. The error of is about 10%.

PAC PEC PC

Solvent Fbmax

[cm-1

] Famax

[cm-1

]

1. n-Hexane (1.88) 28000 0.22 0.20 0.19

2. Diethylether (4.34) 27250 0.21 0.22 0.19

3. Ethyl acetate (6.02) 26650 0.15 0.16 0.17

4. THF (7.58) 26700 0.17 0.20 0.20

5. Butyronitrile (20.31) 26500 0.088 0.11 0.16

6. Acetonitrile (37.50) 26300 19600 0.049 0.055 0.12

7. DMF (37.00) 26300 19300 0.037 0.044 0.08

8. Methanol (32.60) 26100 19000 0.008 0.009 0.017

The quantum yields of both emissions in various solvents are given in Tab.1.

A steady drop in the emission yield with increasing solvent polarity is observed,

mostly in the case of PAC. This is due to the increasing population of the CT state vs

the LE state and reflects the forbidden character of the fluorescence from the former.

Low temperature experiment

The absorption spectra of PAC, PEC and PC in methanol did not depend on

the temperature. For example see the case of PAC (Fig. 3 .4).Upon cooling , the TICT

emission becomes weaker, while the intensity of Fb fluorescence increases, for

comparison see PAC in MeOH (Fig. 1.4 and Fig. 3A.4).

41

Fig. 3.4 Absorpton spectra of PAC in MeOH recorded as a function of temperature,

4H indicate the spectral position of the 4th

harmonic generated by the Nd:glass laser

(see next chapter). A, fluorescence of PAC in MeOH at 183 K.

B, fluorescence of PAC in ether at 293K (solid line) and 183K (points).

Figure 4.4 presents the variation of Fa and Fb with temperature for three alkyl

derivatives of m-cyanopyridine in ether and methanol. The emission spectra in ether

consist of only primary band (Fb), so in this solvent the excited state reaction does not

take place. It has been shown that the analysis of the shape of the plots of ln{(T)}

vs. 1/T is useful for the determination of the kinetic regime of the photoreaction

[Grabowski et al.,1979], [Grabowski et al.,1978]. For instance, a minimum in the

ln(Fb) vs. 1/T plot separates the “high-temperature range”, where the LETICT

equilibrium is established, from the “low-energy region”, where the excited state

charge separation becomes irreversible. The character of the curves presented in

Figure 4.4, in particular a monotonic increase of ln(Fb) with decreasing temperature,

suggests in the case of PAC and PEC in MeOH irreversible kinetics for the

LETICT process below 293K. The calculated energy barrier for LETICT

reaction in MeOH is equal to: 720 cm-1

(PAC), 940 cm-1

(PEC). In the case of PC in

MeOH the minimum is observed at about 253K, what can suggest that at room

temperature excited state reaction is reversible. In alcohol solutions, the Fa/Fb ratio

becomes much larger than in aprotic solvents of similar polarity (Figure 1.4). At the

same time, the total emission quantum yield strongly decreases (Table 1.4). The

relative drop in the Fb intensity is much larger than the Fa increase. This shows that

specific interactions with the solvent, such as hydrogen bonding between the alcohol

and pyridine nitrogen, can lead to two effects: (i) an opening of an efficient

radiationless channel of the LE state deactivation; (ii) enhancement of the population

the secondary state. Similar phenomena have been previously observed for

aminopyridines and aminopyrimidines [Herbich et al., 1993] [Herbich et al., 1992],

[Herbich and Waluk, 1994].

25000 30000 35000 40000

0.2

0 .4

0 .6

0 .8

1 .0

4H

203 K

293 K

243 KPAC

OD

cm-1

15000 20000 25000 300000.0

0.2

0.4

0.6

0.8

1.0

A

A.U

.

cm-1

20000 25000 30000

0.0

0.5

1.0

B

A.U

.

cm-1

42

3,0 3,5 4,0 4,5 5,0 5,5 6,0

-0,2

0,0

0,2

0,4

0,6

0,8

Total

TotalPC

ln {(T

) /

(293

K)}

1/T 10-3 [K

-1]

3,0 3,5 4,0 4,5 5,0 5,5 6,0-0,5

0,0

0,5

1,0

1,5

2,0

2,5

Total

Total

PEC

Fa

Fb

ln {(T

) /

(29

3K)}

1/T 10-3 [K

-1]

3,0 3,5 4,0 4,5 5,0 5,5 6,0

-0,5

0,0

0,5

1,0

1,5

2,0

Total

Total

Fa

FbPAC

ln {(T

) /

(293

K)}

1/T 10-3 [K

-1]

Fig. 4.4 Temperature dependence of the fluorescence yield of PAC, PEC and PC

recorded in ether (o) – total fluorescence and MeOH () – total fluorescence,

and (▲)-Fb , (▼)-Fa.

43

Fig. 5.4 Fluorescence and phosphorescence spectra of PAC in diethyl ether at 77K :

a, total emission; b, phosphorescence obtained after elimination of fluorescence by a

phosphoroscope.

In low temperature glasses, the fluorescence of PAC, PEC and PC is

accompanied by phosphorescence (shown in Figure 5.4 for PAC). Interestingly, the

phosphorescence occurs at shorter wavelengths than the anomalous fluorescence Fa.

B. Time-resolved experiments Pure solvents

To excite the alkyl derivatives of cyanopyridine the 4th

harmonic should be

generated (see figure 3.4). The 4th

harmonic (264 nm) pulse can produce “solvated

electrons” even in pure solvents. This effect has been discovered in water three

decades ago and described in many articles [Assel et al., 1999], [Assel et al., 2000],

[Son et al.,2001], [Sobolewski and Domcke, 2002]. This experiment was repeated.

The energy of 0.5 mJ / pulse was enough to observe intense spectra typical for

solvated electrons in water (Fig. 6.4).

Fig. 6.4 Transient spectra recorded in H2O at 293K as a function of delay time,

exc.=264 nm, E=0.5 mJ.

15000 20000 25000 30000

0.0

0.2

0.4

0.6

0.8

1.0

b

aA

.U.

cm-1

400 500 600 700 800

0.00

0.05

0.10

0.15

A

4 ps

3 ps

O.D

.

nm400 500 600 700 800

0.00

0.05

0.10

0.15

0.20

0.25

B

2850 ps

7 ps

O.D

.

nm

44

That was the reason why hexane, acetonitrile and methanol used next for

preparation of the solutions were carefully examined. In pure hexane and methanol,

just after excitation, the strong band located at 400 nm appears. During next few

picoseconds this band broadens and shifts to low energy. For delay time longer than

20 ps this band is so broad that, in accessible spectral window, was detected as the

significant rise of the zero line. This effect does not disappear up to 3 ns. The origin

of this band and its temporal evolution can be understood, per analogy with water, in

terms of solvated electron. The strong dependence of the intensity of this band on the

excitation energy indicated that it is a two photon process. In pure acetonitrile the

effects described above are not observed.

Fig. 7.4 Transient spectra recorded in n-hexane as a function of the delay time,

T=293K, exc.=264 nm, E=0.5 mJ.

Fig, 8.4 Transient spectra recorded in methanol as a function of the delay time,

T=293K, exc.=264 nm, E=0.5 mJ

400 500 600 700 800

0.0

6 ps

5 ps

3 ps

2 ps

A.U

.

nm

400 500 600 700 800

0.0

5 ps

3 ps

1 ps

A.U

.

nm

45

To record the transient spectra of alkyl derivatives of m-cyanopyridine without the

effects presented above, the energy of the excitation pulse was lowered below

0.15 mJ.

Solutions

Transient absorption spectra of PAC, PEC and PC after excitation at 264 nm were

recorded, as a function of delay time, in hexane, acetonitrile and methanol at room

temperature.

i. Hexane: for three compounds being the subject of the study the broad

unstructured band is observed at about 600 nm, no significant temporal

evolution of this band was observed (Fig.9.4.),

ii. Acetonitrile: for PAC and PEC after excitation the absorption band with the

maximum at about 600 nm is observed, for longer delay time weak band

located at 400nm–450nm is rising; in the case of PC this band is not observed

(Fig. 10.4),

iii. Methanol: for PAC and PEC the temporal evolution of the transient spectra is

observed. The band with maximum at 600 nm decreases whereas the band

located at about 420 nm increases as a function of the delay time. In the case

of PC the band located in the blue region is extremely weak, the decay of the

band located at 600 nm is much slower than in the case PAC and PEC.

Fig. 9.4 Transient spectra of PAC, PEC and PC in n-hexane at 293K, exc.-264nm,

400 500 600 700 800

0.00

0.05

0.10

PAC

t = 17 ps

O.D

.

nm

400 500 600 700 800

0.00

0.05

0.10

PAC

t = 117 ps

O.D

.

nm

400 500 600 700 800-0.05

0.00

0.05

t =17 ps

PEC

OD

nm

400 500 600 700 800

0.00

0.05

t = 117 ps

PEC

O.D

.

nm

400 500 600 700 800

0.00

0.05

0.10

t = 17 ps

PC

O.D

.

nm400 500 600 700 800

0.00

0.05

t = 117 ps

PC

O.D

.

nm

46

Fig. 10.4 Transient spectra of PAC, PEC and PC in acetonitrile at 293K,

exc.-264nm.

In the case of excited states characterized by full electron transfer and weak

interaction between donor and acceptor groups the transient absorption bands can be

approximated by the spectra of corresponding radical ion pair D+

and A- [Okada et al,

1987], [Rullière et al, 1987]. For PAC, PEC and PC the acceptor is 5-cyanopyridine.

Its radical anion shows maxima at 306 nm and 402 nm [Shida, 1988]. The donor

group should possess electronic structure similar to triethylamine. Its radical cation

exhibits broad unstructured spectrum within the spectral region 400 nm–650 nm.

[Shida, 1988]. The transient absorption band with maximum at 410 nm corresponds

well to the band of radical anion of 3-cyanopyridine ( see figure 12.4).

400 500 600 700 800

0.00

0.05

t = 10 ps

PACO

.D.

nm

400 500 600 700 800

0.00

0.05

t = 90 ps

PAC

O.D

.

nm

400 500 600 700 800

-0.02

0.00

0.02

0.04

0.06

t = 17 ps

PEC

O.D

.

nm

400 500 600 700 800

-0.02

0.00

0.02

0.04

t = 87 ps

PEC

O.D

.

nm

400 500 600 700 800

0.00

0.05

0.10

t = 17 ps

PC

O.D

.

nm400 500 600 700 800

0.00

0.05

t = 117 ps

PC

O.D

.

nm

47

Fig. 11.4 Transient spectra of PAC, PEC and PC in methanol at 293K, exc.-264 nm.

Fig. 12.4 Transient absorption of PAC in MeOH and low energy absorption band of

3-cyanopyridine radical anion [Shida, 1988]. Squares indicate absorption of

triethylamine cation normalized to the maximum of absorption band of 3-

cyanopyridine anion [Shida, 1988].

400 500 600 700 800

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

PAC

t = 7 ps

O.D

.

nm

400 500 600 700 800

-0.01

0.00

0.01

0.02

0.03

0.04 PAC

t = 27 ps

O.D

.

nm

400 500 600 700 800

0.00

0.02

0.04

t = 67 ps

PAC

O.D

.

nm

400 500 600 700 800

0.00

0.05

0.10

t = 7 ps

PEC

O.D

.

nm

400 500 600 700 8000.00

0.05

0.10

t = 27 ps

PEC

O.D

.

nm

400 500 600 700 800

0.00

0.05

0.10

t = 67 ps

PEC

O.D.

nm

400 500 600 700 800

0.00

0.05

0.10

PC

t = 7 ps

O.D

.

nm

400 500 600 700 800

0.00

0.02

0.04

0.06

0.08

t = 27 ps

PC

O.D

.

nm400 500 600 700 800

0.00

0.02

0.04

0.06

t = 67 ps

PC

O.D

.

nm

400 500 600 700 8000,0

0,5

1,0

3-cyanopyridine(-)

t = 67 ps

A.U

.

nm

48

Kinetics

The kinetics of the rise and decay of the transient absorption bands of PAC,

PEC and PC in methanol at 294 K are shown in the figure 13.4-15.4 respectively. In

the case of PAC and PEC the decay of the band located at about 600 nm is related

with the rise of the shortwavelength band (410 nm). The slightly faster rise of this

band than decay of the longwavelength can be explained by the transient effects

observed in methanol.

Fig. 13.4 The kinetics of the TA bands of PAC in MeOH, experimental points and fit.

Value in brackets – error: (see also Fig. 14 and 15.4).

Fig. 14.4 The kinetics of the TA bands of PEC in MeOH, experimental points and fit.

Fig. 15.4 The kinetics of the TA band of PC in MeOH, experimental points and fit.

-20 0 20 40 60 80 100-5

0

5

10

15

20

= 18 ps (2ps)

= 1000 ps

ps

<500nm

- 6

40nm

>

-20 0 20 40 60 80 100-1

0

1

2

= 15 ps (3ps)

= 1000ps

<40

0nm

- 4

20nm

>ps

-20 0 20 40 60 80 100-10

0

10

20

30

40

50

= 21 ps (3 ps)

= 1000 ps

<497nm

- 6

43nm

>

ps

-20 0 20 40 60 80 100-2

-1

0

1

2

3

4

5

= 14 ps (3ps)

= 1000 ps

<400nm

- 4

20nm

>

ps

-20 0 20 40 60 80 100

0

10

20

30

= 310 ps (30ps)

<5

40

nm

- 6

00

nm

>

ps

49

The fluorescence decay curves of the alkyl derivatives of m-cyanopyridine were

recorded in n-hexane only . The decay times for PAC, PEC and PC were 2.1 ±.1 ns,

2.0±.1 ns and 2.1 ±.1 ns, respectively.

C. Optically induced NMR time-resolved experiment

The NMR spectroscopy is commonly used to recognize the ground state

geometry of the molecules. It is the first case when this relatively slow technique was

used to establish the excited state reaction coordinate. The typical frequency of the

radiation which is used in NMR spectrometers is in the range of hundreds of MHz; in

the time scale it corresponds to single nanoseconds interval. The excited state

reactions, focusing my attention, are much faster: tens of picoseconds to hundreds of

femtoseconds. To recognize the excited state reaction path the transient perturbation

of the ground state conformers population was monitored. Theoretical arguments,

indicating this method as an useful tool for an excited state reaction investigations,

have been suggested by J.Michl [Michl, Bonačić-Koutecký, 1990].

Description of the experiment

The DMABN molecule possess the C2v symmetry, so the 180o twisting of the

–N(CH3)2 group generates the conformer energetically not distinguishable from its

precursor . Replacement of the –N(CH3)2 by a non symmetric N(CH3 , i-pr) group

and one ortho carbon by nitrogen removes the energy degeneration of these two

conformers denoted in the figure 16.4 as b and b’. For PAC molecule, only in the

case when the excited state reaction is governed by internal rotation of the alkyl group

around the N-C single bond, the negative and positive polarization of the 1H signals of

the N(CH3 , i-pr ) group in the light-dark difference NMR spectrum can be observed.

This effect is strictly connected with transient non-Boltzmann population of the

ground state conformers (Fig. 16.4).

Fig. 16.4 Schematic So and S1 energies of PAC as a function of the C-N twisting

coordinate Θ, E=150 cm-1

.

Ek

4

k3

50% 50%

k2

k1

Fb'

Fa

Fb

S1

S0

0

50

If the rotation of the donor group plays a leading role, the longwavelength

emission (Fa) and negative/positive NMR signals should be observed at the same

conditions (solvent, temperature).

Since PAC can exist in two conformational planar forms, two different Fb

emissions are possible, but only one Fa. No spectral changes of Fb fluorescence was

observed varying the excitation energy within the first two absorption bands. This

implies that the energy difference (E) between b and b’ planar conformers of PAC

in S0 is smaller than 200 cm-1

.

Fig. 17.4 1H NMR spectrum of PAC at 198K: spectrum in MeOH (A), Light-dark

difference spectra in MeOH (B) and in THF (C). The crossed–out region is blocked

by the solvent.

In MeOH at 198K where PAC emits double fluorescence (Fig. 18.4) the NMR

difference spectrum shows photodepletion of b and photoenrichment of b’

conformer. The effect disappeared 1 sec after excitation what indicates that relaxation

time between the ground state b and b’ conformers is on the scale of hundreds of

miliseconds. In pure THF at 198K Fa is unobservably weak and NMR effect is absent

(Fig. 17.14,C ). A symmetrically N-substituted molecule PEC in MeOH at 198K

gives no NMR difference signal.

Fig. 18.4 Normalized fluorescence spectrum of PAC in MeOH and THF at 198K.

3.5 3.0 2.5 2.0 1.5 1.0 0.5

x

C

B

A

(ppm)

15 20 25 30

0.0

0.5

1.0

Fa

Fb+F

b'

THF

MeOH

A.U

.

103 cm

-1

51

C. Conclusions

Although TICT is the most commonly accepted structure for the anomalous

emissive state, there was no direct evidence for it. Alternative assignments attribute Fa

to an exciplex [Visser and Varma, 1980] or more often, to a structural distortion other

than twisting [Zachariasse et al., 1993], [Van der Haar et al., 1995], [Schuddeboom et

al., 1992] but any evidence for them was also only indirect. The NMR results

presented above indicate directly that syn-anti isomerization around C-N bond is

responsible for TICT state generation.

The photophysics of all three studied compounds, can be described by the same

TICT model. The quantitative differences in the emission behaviour can be attributed

to energetic, kinetic, and structural factors. For instance, a much lower TICT intensity

observed in PC relative to PAC and PEC can be rationalized by assuming a larger

ground-state pretwisting of the bulkier dialkylamino groups with respect to the plane

of the acceptor in the latter two molecules. This pretwisting should, in turn, be larger

in PAC than in PEC, in line with the observed Fa/Fb ratio. The correlation between

the energy of the CT state and E1/2ox

(D), the oxidation potential of the donor in the

series PAC, PEC and PC, cannot be made in a quantitative fashion, since the Fa

maxima are located practically at the same energies, 19600200 cm-1

for PAC and

PEC in acetonitrile, and 19000200 cm-1

for PAC-PEC in methanol. Actually, this

finding is in accord with an earlier observation [Herbich et al., 1993] that the

oxidation potentials for –NMe2 and –NEt2 donors are very similar

The enhancement of the TICT formation by a specific interaction with an alcohol

solvent remains an intriguing phenomenon. At least two mechanisms can be

proposed. First is the increase of the electron affinity of the acceptor as a result of the

hydrogen bonding to an alcohol molecule at the pyridine nitrogen. Similar effect has

also been observed upon complexation of 4-(dialkylamino)pyrimidines with

Zn2+

[Herbich et al.,1993]. The other possibility is the twisting of the dialkylamino

group resulting from steric interaction with a nearby hydrogen-bonded alcohol

molecule. However it should also be noted that the effect has been observed for 4-

aminopyridines [Herbich and Waluk, 1994] for which the steric factor should be

absent.

In hexane three derivatives of 3-cyanopyridine show transient band with the

maximum at about 600 nm. No spectral evolution of this band depending with the

delay time is observed (Fig. 4.9); for PEC its decay time is 1600 ps 300ps. The

quantum yields, lifetimes, radiative (kr) and radiationless (krn) rate constants obtained

for three derivatives of m-cyanopyridine in hexane are collected in Table 2.4.

In acetonitrile this band is also present, however for longer delay times for PAC

and PEC the rise on new band with the maximum at about 410 nm is observed (Fig.

10.4); for PC the rise of the secondary band is negligible. The precursor (600 nm)–

successor (410 nm) relation between these two bands is easily visible in methanol

(Fig. 11.4), The kinetics of the rise and decay of the transient absorption bands clearly

demonstrate it (Figs. 13-15.4).

52

Table 2.4

Fluorescence quantum yields - , lifetimes - (ns), kr = / (1/109 s

-1) and

knr= -1

- kr (1/109 s

-1) measured for PAC, PEC, and PC in hexane at 293K.

Molecule kr knr

PAC 0.22 2.1 0.1 0.38

PEC 0.20 2 0.1 0.4

PC 0.19 2.1 0.09 0.39

The error of and is about 10%.

Fig. 19.4 Scheme of energy levels for PAC in methanol (cm-1

). Energies of the

excited states were obtained in following way: S2-from absorption maximum, S1-from

absorption / fluorescence intersection, S’1 the energy corresponds to the emission

maximum enlarged by E0. , E0=3500 cm-1

[Wójcik, 2001], E1=720 cm-1

(see text),

separation between energy levels of b and b’ conformers equal to 150 cm-1

; kr (as in

hexane, Tab. 2.4). The k value for LETICT reaction was calculated from the decay

time of the TA band (see Fig. 13.4).

The similarity of the short-wavelength transient band with the absorption band of

radical anion of the acceptor moiety of the PAC and PEC in MeOH (Fig. 12.4) gives

proof of the full separation of the charges in the secondary singlet state. Such effect,

was observed also in the case of carbonyl derivatives on N,N-dimethylaniline

53

[Rullière et al.,1987]. The intensity of transient band located at 410 nm is well

correlated with the intensity of the TICT fluorescence (ee Figs.1.4 and 9-11.4). The

kinetics relation between two transient bands observed in the accessible spectral

window indicates that the band located at 600 nm can be ascribed to the S1 → Sn

transition of the planar conformer. The reaction constant of the excited state process is

5•1010

s-1

(MeOH, T=293K).

The presented material was partly published:

1) J.Dobkowski, J.Wójcik, W Koźmiński, R.Kołos, J.Waluk, J.Michl,

J.Am.Chem.Soc. , 124(2002)2406. (UV induced time-resolved NMR)

2) J.Dobkowski, J.Michl, J.Waluk, Phys. Chem. Chem. Phys. 5(2003)1027.

(steady-state spectroscopy, partly).

54

5 Intramolecular charge-transfer properties of

molecules with small donor groups: The case of the carbonyl derivatives of N,N-dimethyl and

N,N-diethylaniline

The results obtained for N,N-dimethylamino and N,N-diethyaminolbenzaldehyde

(DMABA, DEABA), N,N-dimethylamino and N,N-diethylaminoacetophenone (DMAA,

DEAA) and three model compounds: 5-N,N-dimethylaminoindanone (DMI) ,5-formyl-1-

methylindoline (FMI) and 3-methyl-N,N-dimethylaminobenzaldehyde (mDMABA) are

presented in this chapter.

A Steady-state spectroscopy

In polar aprotic solvents at room temperaturethe alkylamino derivatives of benzaldehyde and

acetophenone (BA-derivatives) emit the weak fluorescence with quantum yields of 10-3

-10-2

(Table 1.5). All the molecules, except FMI, show fluorescence spectra consisting of two

bands (Fig. 1.5). The indoline derivative (FMI), with the immobilized amino group emits

single fluorescence band analogous to the high-energy “b” bands of BA-derivatives (Fig. 1.5).

Contrary to the liquid solutions in low temperature rigid media the compounds show only

single high energy “b” band of the fluorescence (“b”-band) as well as the phosphorescence, as

shown for example in Fig. 2.5.

N

O H

MeMe

DMABA

N

O H

DEABA

EtEt

N

O Me

MeMe

DMAA

N

O Me

DEAA

EtEt

N

O

MeMe

DMI

N

O

Me

FMI

H

N

O H

MeMe

Me

mDMABA

55

Fig 1.5 Room temperature fluorescence spectra of the BA-derivatives in acetonitrile.

15000 20000 25000 30000 35000 40000

0.0

0.2

0.4

0.6

0.8

1.0A

.U.

cm-1

DMABA

Fa

TICT

Fb

15000 20000 25000 30000 35000 40000

0.0

0.2

0.4

0.6

0.8

1.0

A.U

.

cm-1

DMI

15000 20000 25000 30000 35000 40000

0.0

0.2

0.4

0.6

0.8

1.0

A.U

.

cm-1

FMI

15000 20000 25000 30000 35000 40000

0.0

0.2

0.4

0.6

0.8

1.0

A.U.

cm-1

mDMABA

56

Table 1.5

Absorption and fluorescence maxima [cm-1

], total fluorescence quantum yields ( TOT) the

ratio of the quantum yields of the short and long-wave fluorescence ( a/ b) of the

BA-derivatives in acetonitrile at 293 K.

Abs. max. Fb max. Fa max. aTOT a/ b

DMABA 29800 26100 17000 0.006 2.5

DEABA 29400 26000 17000 0.006 3.2

DMAA 30800 26500 17900 0.006 2.6

DEAA 30000 26400 18000 0.013 4.9

DMI 30400 26500 18500 0.018 4.4

FMI 29200 24800 0.13 -

mDMABA 30600 17000 0.005 16.8

a-error 15%

Low temperature luminescences of BA-derivatives consist of phosphorescence with the

maximum about 21000-23000 cm-1

and fluorescence (Fb). The relative intensity of the

phosphorescence and fluorescence bands depends on the solvent polarity (Fig 2.5).

Fig. 2.5 Luminescence spectra of BA-derivatives recorded at 77 K in MTHF (left) and in

ACN (right); phosphorescence (P), fluorescence (F).

Time-resolved Spectra Transient absorption (TA)

Room temperature transient absorption spectra of BA-derivatives recorded twenty

nanosecond after the excitation are shown on the figure 3.5. One band with the maximum

about 450 nm is observed. Its decay time is about 100-200 ns.

The TA spectra recorded in the picosecond domain are presented in the figures 4-6.5

20000 25000

0.0

0.5

1.0

A.U

.

cm-1

DMABA

DMAAFMI

P F

20000 250000.0

0.5

1.0

A.U

.

cm-1DMABA

DMAA

P F

57

Fig. 3.5 Room temperature TA spectra of BA-derivatives in acetonitrile recorded 20 ns after

excitation.

Fig. 4.5 Room temperature TA spectra of DMABA recorded in diethylether, THF

(tetrahydrofuran) and ACN (acetonitrile) as a function of the delay time.

400 500 600 700 8000.0

0.2

0.4

OD

nm

12 ps

2800 ps

Ether

400 500 600 700 800

0.00

0.05

0.10

0.15

OD

nm

5 ps

3000 ps

THF

400 500 600 700 800

0.00

0.05

0.10

OD

nm

ACN

7 ps

2500 ps

400 450 500

0.0

0.5

1.0

A.U

.

nm

DMABA

DMAA

FMI

58

Fig. 5.5 Room temperature TA spectra of FMI recorded in diethylether, butyronitrile (BuCN)

and acetonitrile (ACN) as a function of the delay time.

400 500 600 700 800

0.0

0.1

0.2

OD

nm

15 ps

2200 ps

ether447 nm

400 500 600 700 800

-0.05

0.00

0.05

0.10

OD

nm

50 ps

900 ps BuCN

400 500 600 700 800

-0.2

-0.1

0.0

0.1

OD

nm

ACN

448 nm

10 ps

2500 ps

59

The TA spectra recorded for mDMABA, the model compound with the pretwisted N(CH3)2

group are presented in the figure 5.6.

Fig. 6.5 Room temperature TA spectra of mDMABA recorded in tetrahydrofuran (THF) and

acetonitrile (ACN) recorded for two delay times.

The temporal evolution of the room temperature TA spectra of BA-derivatives

exhibits similar features:

In weakly polar solvent (ether) the TA band with the maximum at about 450 nm is

observed. This band does not exhibit any temporal evolution.

Except FMI, which has a rigidly planar amino group and emits only Fb fluorescence,

for other molecules in polar solvents the band with the maximum 480-510 nm is

observed. This band shows temporal evolution, for the delay times 2 20 ns its

maximum is located at about 450 nm.

The rise times of the TA bands of DMABA and FMI are collected in the table 2.5.

400 500 600 700 800

0.00

0.05

0.10

0.15

OD

nm

12 ps

2800 ps

THF

400 500 600 700 800

0.00

0.02

0.04

OD

nm

9 ps

2600 ps

ACN

60

Table 2.5

The rise times ( R

/ps) of the TA bands evaluated from the kinetic curves obtained for

DMABA and FMI in weakly polar and polar solvents, T=294 K, in brackets the static

permittivity values .

Solvent DMABA FMI

R

Short-wave band

<440 nm-480 nm>

R

Long-wave band

<490 nm-550 nm.>

R

<440 nm-480 nm>

Diethylether (4.1) 8 1

THF (7.6) 7 2 13 4

BuCN (20.3) 6 1.5 4 1 -

ACN (36.5) 4 2 3 1 a

a – decay time of the stimulated emission is 270 ps 25 ps

Amazingly, the rise times of TA bands of BA-derivatives at 203 K are almost identical to

those evaluated at room temperatures (Tab.3.5).

Table 3.5

The rise times (R

/ps) of the TA bands evaluated from the kinetic curves for BA-

derivatives in butyronitrile at 203K.

R Short-wave band

<440 nm-480 nm>

R Long-wave band

<490 nm-550 nm>

DMABA 5 1 4 1

DEABA 7 1.5 7 1

DEAA 5 1 4 1

Time-resolved fluorescence (TRF)

Fig. 7.5 Room temperature TRF spectra of DMI in ACN recorded 3 and 7 ps after

excitation. Dashed line – stationary spectrum recorded for non-perpendicular arrangement

of the polarizers in the Optical Kerr Shutter.

300 400 500 600

0

200

400

600

800

TR

F Inte

nsi

ty

nm

3 ps

7 ps

Stac.

exc.

DMI

Fb

Fa

61

Fig. 8.5 Room temperature normalized kinetic curves of DMI in ACN: circles-excitation

pulse, squares – short-wave fluorescence at integration limits <365 nm-390 nm>,

gaussian fit of the excitation pulse, solid curve. A-kinetic curve of the long-wave

fluorescence, integration limits <500 nm-550 nm>.

Stationary spectrum of DMI recorded for non perpendicular arrangement of the

polarizers in the optical Kerr shutter shows significant contribution of the long-wave

fluorescence (Fa), see Fig 7.5. Contrary to that, TRF spectra exhibit small contribution of

the Fa to the total TRF emission. This experimental finding indicates that the lifetime of

the long-wave fluorescence is significantly longer than the decay time of the short-wave

emission (Fig.8.5). Except FMI, the decay of the short-wave fluorescence for

BA-derivatives at 294 K is faster than the temporal resolution of the apparatus, ( about 6.5

ps), consequently the rise of the long-wave fluorescence can not be properly monitored

(Fig. 8.5A). Its decay curve, due to the small intensity of TRF-Fa band was not recorded.

The decay times of Fb obtained from low temperature experiments are collected in the

Table 4.5

Table 4.5

The decay times of Fb fluorescence [ps] evaluated from the low temperature kinetic curves

obtained for DMABA and DMI in butyronitrile.

Temperature [K] DMABA DMI

203 6 1.5 11 2

173 23 3

163 33 4

180 200 220 240 260

0.0

0.2

0.4

0.6

0.8

1.0

180 200 220 240 2600.0

0.5

1.0

A.U

.

ps

A

A.U

.

ps

62

Discussion

Having in mind the experimental data and the conclusions presented in the chapter 4 the

results obtained for BA-derivatives will be interpreted in the terms of the Twisted

Intramolecular Charge Transfer (TICT) model, see figure 9.5.

Fig. 9.5 TICT state model.

The model of the TICT state postulates the full electron transfer between the donor (D) and

acceptor (A) subunits twisted to the perpendicular conformation around the central bond

[Grabowski et al., 1979].

To analyze the excited state reaction :

locally excited state (LE) of planar geometry TICT state

two assumptions of the TICT model should be discussed:

full electron transfer,

excited state geometry transformation.

Full electron transfer For the perpendicular arrangement of the donor and acceptor subunits

the interaction of their systems is weak, consequently the TICT state can be approximated

by the radical ion pair, D+

+A- [Grabowski and Dobkowski, 1983]. The transient absorption

spectra recorded in polar solvents should correspond to the absorption of the radical ion pair

(somewhat perturbed by their mutual interaction) [Rullière et al. 1987]. Indeed, the spectra

recorded just after excitation for all BA-derivatives in polar solvent (except FMI ) show bands

which are well reproduced by the spectrum of the benzaldehyde or acetophenone radical

anion. For example see figure 10.5. The absorption band of the N(CH3)3 cation radical is

located in UV region [Shida, 1988] .

0

100

200

En

erg

y [

A.U

.]

Fb F

a

D A

0o

90o

TICT

A-

D+

63

Fig. 10.5 Room temperature TA spectra of DMABA and mDMABA in DMF recorded for

the delay times 40 ps (DMABA) and 220 ps (mDMABA), dotted line: benzaldehyde radical

anion in MTHF [Shida, 1988].

Excited state geometry : The comparison of the stationary fluorescence of the model

compounds (DMI, FMI, mDMABA) with the fluorescence spectra of the BA-derivatives

having no restriction for the rotation of the alkyl and carbonyl group gives a proof that in the

case of BA-derivatives the TICT state is generated by the rotation of the donor (alkyl) group

(see Fig. 1.5).

Solvent assisted excited state relaxation

Excited state relaxation processes in the case of BA-derivatives depend on the solvent

polarity. In nonpolar and weakly polar solvents TA band with the maximum at about 450 nm

is observed and due to its long lifetime is assigned as T1 Tn. transition (Fig. 3-5. 5). In polar

solvents the TA spectra of BA-derivatives consisted of two bands: the short ( max =450 nm)

and the long-wave one ( max =500 nm) these bands are assigned as the T1 Tn and

(S1 Sn)TICT

, respectively. The rise time of T1 Tn absorption for BA-derivatives in nonpolar

solvents at 294 K is about 7-13 ps (Tab.2.5), in polar solvents the rise time of two bands is

about 3-6 ps. Amazingly, the rise time of the TA band in polar solvent at 203 K is almost

identical to that reported at room temperature, 4-7 ps (Tab.3.5). The dependence of the TA

spectrum on solvent polarity is directly connected with the sequence of two low laying singlet

states 1

* and 1n * (Fig. 11.5).

400 500 600 700 800

0.0

0.5

1.0

A.U

.

nm

mDMABA

DMABA

benzaldehyde(-)

64

Fig. 11.5 The scheme of the energy levels for BA-derivatives : left – the case of nonpolar

enviroment, right – the case of polar enviroment.

Since the Fa fluorescence is observed in polar liquids, the reorganization of the solvent

cage should play the leading role in the case of S1 TICT reaction. The comparison of the

low temperature decay times of Fb emission with the relaxation times of the solvent is

presented in the Table 5.5.

Table 5.5

Low temperature decay-time of Fb fluorescence of DMABA and DMI in butyronitrile

(BuCN), compared with the relaxation time of BuCN obtained using ultrafast optical Kerr

effect spectroscopy and with the spectral response function of the probe spectroscopy.

BuCN / Kerra

BuCN / probe DCSb

T[K] DMABA DMI 1s

[ps] 2s

[ps]

< >c

[ps]

A1,A2 1s

[ps] 2s

[ps]

< >

[ps]

A1,A2

203 6 1.5 11 2 7.2 27.1 23.3 .19,.81 9 41 12.8 .88,.12

173 23 3 21 95 83.2 .16,.84 20 138 60.1 .66,.34

163 33 4 30d

145d

126.6 .16,.84 26 237 142 .45,.55

a -Data obtained from Optical Kerr measurements [Zhu et al., 2005].

b -Data obtained from Spectral Solvation Response Function of 4-Dimethylamino-

4’cyanostilbene (DCS) [Druzhinin et al., 2006].

c: < >= A1 1+A2 2

d – extrapolation of the data [Zhu et al., 2005].

The decay time of the Fb fluorescence can be satisfactorily correlated with the

relaxation time 1s of BuCN determined by Kerr and probe spectroscopy. The contribution of

the fast component into the total solvent relaxation is 19-16% (Kerr) and 88-45%(probe), see

Anex. The depopulation decay time of the S1 state of BA-derivatives in polar solvents is

governed by ISC and TICT state formations rate constants (Fig 11.5). These two processes

depend on the solvent cage reorganization.

S0

T1

TICT

n

*(La)

S1

(Lb)

Ph

S0

S1 TICT

Fa

Fb

T1

Ph

n *

*(La)

(Lb)

65

Temporal evolution of the triplet state population

The lowest triplet state of BA-derivatives is of *(La) type [Dobkowski et al., 1982] . The

sequence of the energy levels of the two lowest singlet states of BA-derivatives in polar

environment is shown in figure 11 .5 (right): the 1

* is slightly lower than 1n * state. The

vibronic interaction (Hvib) can admix higher lying singlet to the lowest one

S1=a 1n *+

1* where: a = <

1n * Hvib

1*> ( E)

-1 and, E is the energy gap between

the 1n * and

1* state.

The intersystem crossing rate constant (kISC) of the S1 T1 transition is kISC~ MISC2 where:

MISC=<S1 HSO T1>( EST)-1

, HSO – spin-orbit operator; EST - singlet –triplet energy gap.

Since the element: <1

* HSO3

*> is smaller than <1n * HSO

3*> [McGlynn et al.

1972] [Lower and El-Sayed, 1966], the kISC a2{<

1n * HSO

3*> ( EST)

-1}

2=a

2M’.

For BA-derivatives the value of the <1n * HSO

3*> as well as the EST value

(about 7000-8000 cm-1

) does not depend significantly on the solvent polarity and solvent

cage reorganization.

Consequently kISC a2M’ = {<

1n * Hvib

1*> ( E)

-1}

2 M’.

The energy of the 1n

* state, due to its small dipole moment, does not undergo

significant solvent stabilization. Contrary to that, more polar 1

* state is energetically

stabilized by a polar environment, and consequently E value (typically 400-1500 cm-1

)

strongly depends on the solvent polarity and solvent cage reorganization. Because

<1n * Hvib

1*> does not depend on the solvent polarity, the kISC ( E)

-2 M, where

M=<1n * Hvib

1*>

2 M’.

Generally the stabilization of an excited state induced by solvent cage reorganization

is described by a c(t) function [Jarzęba and Barbara, 1990]. In many cases experimentally

evaluated c(t) function can be well approximated by mono- or multi-exponential function.

Assuming the simplest case, it means mono-exponential character of the c(t)=exp(-t/ s), where

s –solvent relaxation time:

kISC(t)~{ E0+ E1[1-exp(-t/ s)]}-2

E1=E (t=0)-E (t= ). The comparison of the kISC and c(t) is shown in the figure 12.5

66

Fig. 12.5 The comparison of the c(t) function with predicted kISC(t) function. E0 energy gap

between 1n * and

1* states just after excitation, Es solvent stabilization of the

1* state.

Temporal evolution of the TICT state population: The dipole moment of the TICT state is

higher than the dipole moment of the primary excited S1 state. This is the reason that TICT

state is relatively more stabilized due to the solvent cage reorganization than the S1 state. The

barrier for the reaction ( E) should be time-dependent (for the visualization see Fig. 13.5).

Consequently the kTICT for the TICT state formation should exhibit inverted, in comparison

with the kISC, relation, it means kTICT(t1)<kTICT(t2) when t1<t2 (Fig. 13.5).

The kTICT(t)~exp(- E)=exp-{(a2- Es(t))/2a}

2;

Es(t)=E1exp(t/ s) has the same meaning as in the case of kISC.

Fig. 13.5 The visualization of the time dependence of the barrier for S1 TICT reaction on

the stabilization of the TICT state due to the solvent reorganization. E(t)={(a2- Es(t))/2a}

2;

t1,t2 –time (t1<t2). A- the temporal evolution of the kTICT calculated for a=1, E1=1, s=2.

The intramolecular nature of the barrier, whose time-dependent evolution was

discussed above, is directly connected with the inversion of 1La and

1Lb states on the

-2 0 2 4

-2

0

2

4

En

erg

y [

A.U

.]

Reaction coordinate [A.U.]

a

Es

E

TICT

t1

t2

S1

0 2 4 6

0.8

1.0

k T

ICT

[A.U

.]

Time [A.U.]

A

0 2 4 6

0.0

0.5

1.0

n *

*t=0

t>>0

E0

Es

K I

S C

[A

.U.]

Time [A.U.]

kISC

={ E0+ E

s[1-exp(-t/

s)]}

-2

exp(-t / s)

67

relaxation path. In the case of BA-derivatives, in polar solvents the La state is the lowest one

and consequently the reaction should be barrierless. This conclusion is not in conformity with

the experiments. Two, kinetically coupled, fluorescence bands observed in polar solvents

indicate that the barrier exists on the relaxation path. The solvent effects studied in the case of

DMABA, applying the solvaton model associated with the AM1 procedure indicate that

environment can generate the barrier along coordinate [Gorse and Pasquer, 1995]. In

accordance with their results the energy (Ea) and dipole moment ( a) of the La state (in gas

phase) increase monotonously with angle increases. It is instructive to discuss the solvent

induced modification of the intramolecular barrierless relaxation path in the case of BA-

derivatives.

Let us assume for:

Ea( ) linear correlation limited by Ea(0 )=30000 cm-1

and Ea(90 )=34000 cm-1

a( ) linear or square correlation limited, in accordance with Gorse and Pasquer, by

a(0)=9.5 D, and a(90 )=16.5 D.

The stabilization of the La state can be simply estimated applying Mataga formula:

Esolv=-(eq

)2

f /(ao)3 (1.5)

where: eq

- dipole moment of the equilibrated excited state,

f =( -1)/(2 +1), -dielectric constant, a0-Onsager radius [Mataga, 1975].

Fig. 14.5 The energy of the La state in the case of DMABA: 1- in gas phase, the linear

correlation was assumed. The calculated energy of La in acetonitrile: 2- linear correlation, and

3-square correlation for ( ) was applied.

This simple calculation indicates that when the excited state energy (in gas phase )

increases with the angle increase, the solute-solvent interaction can generate the barrier on

the relaxation path. The position ( max) and height (H) of this barrier depends on the

intramolecular parameters and solvent polarity, as well as, due to the solvent cage

reorganization, on the time.

In the case of BA-derivatives, due to fast intersystem crossing, the TICT state is

populated from not totally equilibrated S1 state. The proper analysis of the time dependent

shift of the maximum of the time-resolved TICT fluorescence is impossible because of its

significant width and small intensity.

0 20 40 60 80 100

24000

26000

28000

30000

32000

34000

Energ

y [c

m-1]

Esolv.

o

1

3

2

68

The TICT state precursor

The energy of the maxima of the 1

* (La) and 1

* (Lb) absorption bands and the

values of the maxima of the Fa and Fb fluorescence in the case of compounds which emit the

dual fluorescence, are collected in the Table 6.5. Below are shown the formulae of the

molecules not presented before.

Table 6.5

The vavenumbers of the maxima of the Lamax

and Lbmax

bands (absorption) and the maxima of

the Fbmax

and Famax

fluorescence (cm-1

) in acetonitrile at 294 K for selected molecules which

can relax to the TICT state.

Compound La Lb Fb Fa Ref. E(Lb)<E(La)

4DMAP 40350 34500 27200 - Herbich et al., 1989

4DEAP 39950 34500 27400 21500 Herbich et al.,1989

DMAP 38100 35500-36500 29000 21900 Szydłowska et al.,2003

PC 35400 32300-31200 26400 - Dobkowski et al., 2002

PEC 35100 32300-31200 26600 19500 Dobkowski et al., 2002

DMABN 34200 31500-32500a

27700 20300 Galievsky et al., 2005

DEABN 33640 27300 20600 Galievsky et al., 2005

E(Lb)>E(La)

5DMAPd 329500 27500 19850 Svarcov, 2006

PA 31600 26400 17700

DMAA 30800 32000 26500 17900

DMI 30400 26500 18500

DEAA 30000 26400 18000

DMABA 29800 32500 26100 17000

DEABA 29400 26000 17000

a-[Herbich et al., 1989]

N

O

MeMe

5DMAPd

O

N

O Me

MeMe

PA

N

N

DMABN

MeMe

CN

N

DEABN

EtEt

C

N

N

DMAP

MeMe

N

N

DMAP

MeMe

N

4

N

N

MeMe

N

PC

C

N

N

PEC

N

C

N

EtEt

N

DEAP

EtEt

N

4

N

69

For para-substituted molecules of the type: D-Ar-X (D donor, Ar-aromatic ring, X-

chromophore) or D-P (P-pyrimidine or pyridine) having the groups N(CH3)2 or N(C2H5)2 as

donors the phenomenon of the dual fluorescence is observed independently of the sequence of

the two lowest * singlet states (Tab.6.5). The short-wave fluorescence can be emitted from

the 1Lb or

1La state. In some cases the vibronic interaction can mix effectively these two states

and clear classification of the origin of the short-wave emission can be difficult. Having in

mind the presented above conclusions about the origin of the short-wave emission it is worth

to point out that the long-wave emission can not be explained by:

solvent-induced inversion of the 1Lb and

1La states [Lippert et al., 1962],

solvent-induced pseudo Jahn-Teller (PJT) mechanism [Zachariasse et al., 1993].

Using the language of the TICT model the long-wave emission occurs from the low

laying state generated by the rotation of the N(CH3)2 or N(C2H5)2 groups around the C-N

bond. The plot of the maxima of the long-wave fluorescence vs maximum of the La band

(absorption) clearly shows the relation between these values (Fig. 15.5). To sum up: the

planar 1La state is the precursor of the TICT state.

Fig.15. 5 The maxima of long-wave fluorescence plotted as a function of the maxima of the 1La bands for the data collected in the table 6.5. Methyl derivatives-squares, Ethyl

derivatives-circles.

In the case of carbonyl derivatives of N,N-dimethylaniline the 1La state is located

below the 1Lb state and the reaction; S1 TICT should be barrierless. This simple model is

disturbed by the presence of the n * state which, in polar solvents, is located few hundreds

cm-1

above the 1La. state. The vibronic coupling (solvent induced pseudo Jahn-Teller

mechanism) between n * and 1La states can significantly modify the energy hypersurface of

the 1La state creating simultaneously the barrier on the S1 TICT reaction path.

That is instructive to compare the rise times of the secondary emission or secondary TA in the

case of the selected molecules emitting double fluorescence, but characterized by different

sequence of the 1La and

1Lb states of the planar conformer (Tab. 7.5). The value of the rise

time is crucially dependent on the sequence of the 1La and

1Lb. Thus, there exist the

30000 35000 40000

17000

18000

19000

20000

21000

22000

FT

ICT m

axim

um

[c

m-1]

DMABN

DMAP

Maximum of 1L

a band (abs.) [cm

-1]

PA

DMAA

DMI

DMABA

5DMAPd

PA

DEAA

PEC

DEABN

DEABA

4DEAP

70

intramolecular barrier, in the case of DMAP and PEC whereas for DMABA the excited state

reaction should be barierrless (or, alternatively, a solvent induced time dependent barrier

exists on the relaxation path.

Table 7.5

The rise times of the secondary emission/absorption (R

) at 293 K for selected compounds (in

acetonitrile) characterized by different sequence of the two lowest 1

* state

Esec R

[ps] Ref.

4-(Dimethylamino)pyridine (DMAP) E(La)>E(Lb) 25 Szydłowska, et al. 2004

2-(N,N-diethylamino)-5-cyanopiridine (PEC) E(La)>E(Lb) 15a

N,N-dimethylaminobenzonitrile (DMABN) E(La) E(Lb) 4 Druzhinin et al., 2005

N,N-dimethylaminobenzaldehyde (DMABA) E(La)<E(Lb) 3

a- methanol

The strong acceptor limit

For the para-substituted molecules of the type D-Ar-X, where {Ar-X} unit is treated as

the acceptor group, the linear correlation of the TICT fluorescence maxima with the

difference of the oxidation potential of the donor and reduction potential of the acceptor group

was found [Grabowski and Dobkowski, 1983].

The energy of the TICT fluorescence my be estimated as: hc =Eox

(D)-Ered

(A)+Ecoul-

Edest, where Eox

(D), Ered

(A) are the polarographic half-wave potentials of one electron

oxidation of the donor (D) and reduction of the acceptor (A), Ecoul is the coulombic energy of

bringing the opposite charges to the fixed distance. The emission from the highly polar TICT

state leads to the Franck-Condon (FC) nonpolar ground state of the twisted conformer . This

FC state has the higher energy than the solvent equilibrated ground state, by the

destabilization energy (Edest).

For different para-substituted dimethylanilines Eox

(D), Ecoul and Edest are nearly

constant, therefore the observed TICT fluorescence maxima were correlated with the

reduction potentials of the acceptors only [Dobkowski et al., 1989]. This simple correlation

allows us to evaluate the energy of the fluorescence maximum emitted from the TICT state

created by internal rotation of the N(CH3)2 in the case of p-nitro-N,N-dimethylaniline

(NDMA), the estimated value is about 13000 cm-1

(Fig. 16.5).

The room temperature fluorescence of NDMA in polar solvents is undetectable in the

range 13000-25000 cm-1

. The maximum of the first band of the radical anion of nitrobenzene

is located at 456 nm [Shida, 1988]. However, in spite of a great experimental effort, the

transient absorption band of NDMA corresponding to the radical anion of the nitrobenzene

was not observed.

71

Fig. 16.5 Energy of the TICT fluorescence maximum for the series of compounds (Tab. 8.5)

in acetonitrile solution vs the polarographic half-wave potentials of one-electron reduction of

the acceptors (A) in acetonitrile or dimethylformamide solutions. Donor groups: N,N-

dimethylamino (circles), N,N-diethylamino (squares). Solid line – linear fit, correlation

coefficient 0.91, dashed line the extrapolation of the linear fit.

Table 8.5

Polarographic half-wave potentials ( Ered

) of one-electron reduction of the acceptors (A) in

volts vs SCE, in acetonitrile or dimethylformamide, and energies of the Fa fluorescence

(hc /eV) in acetonitrile.

No Acceptor -Ered

(A) References hc References

1 Pyridine 2.66 Meites and Zuman, 1974 2.74 Szydłowska et al.,

2003

2 Pyrimidine 2.34 Meites and Zuman, 1974 2.60a

Herbich et al., 1989

3 Benzonitrile 2.35 Beens and Weller, 1968 2.54 Lippert et al., 1962

4 Acetophenone 1.87 Meites and Zuman, 1974 2.24

5 Benzaldehyde 1.93 Meites and Zuman, 1974 2.13

6 1-indanone 2.01 Fournier et al., 1983 2.31

7 Benzoic acid 2.17,

2.24

Meites and Zuman, 1974 2.58 Cowley and Healy,

1977

8 Methyl

benzoate

2.29,

2.32

Meites and Zuman, 1974 2.58 Cowley and Healy,

1977

9 nitrobenzene 1.01,

1.13

Meites and Zuman, 1974

a) 4-dimethylamino-pyrimidine (4DMAP) emits TICT fluorescence only in polar protic

solvents. For the ortho-methylated derivative of DMAP (5-Me-DMAP) the observed

maximum of the TICT fluorescence in ethanol is at the same spectral position as for

DMAP, that is the reason why hc value measured for 5-Me-DMAP in acetonitrile

was used.

1.0 1.5 2.0 2.5 3.01.0

1.5

2.0

2.5

3.0

O ON

N

O O Me

O HO

O

O H

M eO

N

N

C

N

ET

ICT [e

V]

V/SCE

12

3

87

64

5

Ered

(A)

9

72

The results of the quantum chemical calculations [Dobkowski et al., 1989] indicate that in the

case of p-nitro-N,N-dimethylaniline the lack of the TICT fluorescence can be explained in the

following way:

The TICT state generated by the rotation of the N(CH3)2 group is energetically higher by

about 4000 cm-1

than the 1La state of the planar molecule,

The calculated dipole moment of the planar 1La state ( =21.4 D) is higher than the dipole

moment of the TICT state ( =20.1 D).

The comparison of the dipole moments and energies of the TICT and the planar 1La state

allow to exclude a possibility of the solvent stabilization of the twisted form with respect to

the planar one.

Conclusion

Carbonyl derivatives of N,N-dimethyl(ethyl)-aniline after excitation to the planar S1

state can relax, depending on the solvent polarity, to the triplet only (non polar solvents) or

simultaneously to the triplet and to the TICT state (polar solvents). The probabilities of the

transitions: S1 T1 and S1 TICT, in polar solvents, depend on time.

For the molecules of the type D-A, where A={Aromatic ring-chromophore} and D=

N(CH3)2 or N(C2 H5)2, the excited state reaction S1 TICT does not occur for the “very strong

acceptors” e.g. nitrobenzene. However, I am not able to show for which value of the reduction

potential of the acceptor the excited state reaction changes the character from the exo to endo-

energetic.

73

6. Intramolecular charge – transfer properties of

molecules with a large donor group. The case of:

4’- (1-pyrenyl)benzonitrile (Py-BN) and

4’- (1-pyrenyl)acetophenone (Py-AC)

Py-BN Py-AC

Py-BN

Experimental results

Steady state experiments

Room temperature absorption and fluorescence spectra of Py-BN were

recorded in a number of protic and aprotic solvents characterized by different

polarities. The shape and the spectral position of the bands in the absorption spectrum

are practically independent of the solvent, while fluorescence spectra are solvent-

dependent (Fig.1.6, Tab. 1.6). The increase the polarity of the solvent shifts the

fluorescence of the maximum to the red.

Fig.1.6 Room temperature absorption and fluorescence spectra of Py-BN in

n-hexane (solid curve) and acetonitrile (dotted curve).

C

N

CO

Me

15000 20000 25000 30000 35000

0.0

0.5

1.0

A.U

.

cm-1

ACN

n-hexane

74

The plot of the solvatochromic shifts of the fluorescence maxima versus the

solvent polarity function F( ,n) is presented in Figure 2.6. The value of the dipole

moment e of PyBN the emitting singlet state can be determined from the relation

[Liptay, 1962], [Liptay, 1974]:

2 e ( e - g )

max = max (0) - _________________

F( , n) (1.6)

hc(a0)3

where:

- 1 1 n2 - 1

F( ,n) = - (2.6)

+ 2 2 n2 + 1

This function is usually used for a system that cannot be directly reached by

absorption from the ground state. The value of the ground state dipole moment g was

assumed to be 3.5 D as for benzonitrile [Khalil et al. 1976].

An Onsager cavity of radius a0=5.2 Å was evaluated from the molecular geometry

( = 70 ). The value of the dipole moment in the excited state obtained from the

solvatochromic shift is 12.5 D. Alternatively, assuming g=0 as for 4’-(pyrene-1)-

N,N- dimethylaniline [Herbich and Kapturkiewicz, 1991], [Herbich and

Kapturkiewicz, 1993] results in e equal to 10.6 D.

Table 1.6

Maxima of the first absorption band ( max,a/cm-1

) and fluorescence band ( max/cm-1

),

Stokes shift ( s/cm-1

), fluorescence quantum yields ( F), lifetimes ( /ns),

radiative kf= F/ (109s

-1), and radiationless knr=

-1-kf (10

9s

-1) rate constants,

transitions moments (in emission) M={3hkf/644n

3 ( max)

3}

1/2 (D)

measured for Py-

BN in a number of solvents.

Solv. max,a max s a)

F b)

kf knr M

HEX 29150 25700

3650 0.33 5.8 0.06 0.12 2.0

EE 29150 24250 4900 0.63 5.1 0.12 0.7 3.4

EA 29100 23850 5250 0.62 3.3 0.19 0.12 4.0

THF 29100 23900 5200 0.63 3.0 0.21 0.12 4.2

VCN 29100 23500 5600 0.58 2.8 0.21 0.15 4.4

BCN 29100 23550 5550 0.67

ACN 29100 23200 5900 0.64 2.8 0.23 0.13 4.9

EtOH 29100 23200 5900 0.63 2.7 0.23 0.14 4.9

MeOH 29100 23200 5900 0.65 2.8 0.23 0.13 4.9

a,b – The error being about 10%

Solvents:

n-hexane (Hex), diethylether (EE), ethyl acetate (EA), tetrahydrofuran (THF),

valeronitrile (VCN), butyronitrile (BuCN), acetonitrile (ACN), ethanol (EtOH),

methanol (MeOH).

75

Fig.2.6 Plot of the solvatochromic shift of the fluorescence maxima of Py-BN.

The excited state dipole moment value is 12.5 – 10.6 D.

The room temperature magnetic circular dichroism (MCD) spectrum of Py-

BN was recorded in n-hexane (Fig. 3). The MCD spectrum recorded in acetonitrile

was almost identical to that presented below.

Fig. 3.6 Top, room temperature absorption; bottom, MCD spectrum of Py-BN

in n-hexane.

Temperature dependence of absorption and fluorescence

To check the temperature dependence of relaxation processes, the fluorescence

and absorption spectra of Py-BN at low temperature were recorded. In a nonpolar

solvent, as well as in a polar protic solvent (EtOH) a red shift of the absorption at low

temperature was observed. (Fig. 4.6), reflecting the changes in the population of the

ground state vibrational levels.

0.1 0.2 0.3 0.423000

23500

24000

24500

25000

25500

MeOHACN

EtOH

BuCN

THF

EA

EE

HEXm

ax [

cm

-1]

F( ,n)

20000 30000 40000 50000

-8

-6

-4

-2

0

2

4

6

8

0

1

2

3

4

5

[

cm-1

76

Fig. 4.6 Absorption spectra of Py-BN in ethanol recorded at 293 K (1), 223 K (2),

158 K (3).

Fig. 5.6 Normalized fluorescence spectra of Py-BN in ethanol recorded at 293 K (A),

158 K (B), 77 K (C).

No change in the spectral position of the fluorescence in liquid and rigid

nonpolar solvent (2-methyl–butane) was observed. In a polar solvent (EtOH) the

fluorescence spectra of Py-BN changed upon cooling, as shown in figure 5.6. Starting

from room temperature down to 168 K the fluorescence quantum yield ( F) value

was almost constant (5% increase), simultaneously the 800 cm-1

blue shift of the

fluorescence band was observed.

20000 30000 40000 500000.0

0.2

0.4

0.6

0.8

1.0

cm-1

3

2

1Ab

so

rba

nce

.

77

Time-resolved absorption experiments

The fluorescence decay curves of the Py-BN were recorded in number of

solvents; the evaluated decay times are presented in Table 1.6. It should be stressed

that the decay time measured in hexane is much longer than that measured in polar

solvents.

Fig. 6.6 Room temperature transient absorption spectra of Py-BN recorded in:

cyclohexane (1), diethylether (2), ethyl acetate (3), acetonitrile (4) at delay

time of 110 ps, exc=355 nm, temporal resolution 30 ps.

Transient absorption spectra of Py-BN were recorded in several solvents (Fig.

6.6).The increase of the solvent polarity leads to significant changes in the position of

the S1 Sn absorption maximum and in the band intensity. In cyclohexane, a broad,

weak band with the maximum at 610 nm was observed, contrary to the situation in

acetonitrile, where an intense band with the maximum at 564 nm was recorded. These

findings suggest that the nature of S1 Sn transition changes with solvent polarity.

Calculations

Molecular mechanics calculations performed for PyBN indicate that the most

stable ground state conformation is obtained when benzonitrile subgroup is twisted

with respect to the pyrene moiety by =69 . The comparison of the observed and

calculated absorption spectra of Py-BN ( using =69 for the ground state geometry )

is given in Figure 7.6.

500 600 700

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

O.D

.

nm

1

2

34

78

Fig.7.6 Room temperature absorption spectrum of Py-BN in n-hexane and the

calculated transitions for the twist angle =69o. The height of the bars is

proportional to the oscillator strength.

To investigate the nature of the electronic states of Py-BN upon twisting of the

benzonitrile group, INDO/S calculations have been performed for different values of

the dihedral angle . The gas phase energies calculated as a function of angle for

So, S1(Lb), S2(La) and S5(TICT) states are shown in figure 8.6.

From the analysis of the results of the INDO/S calculations two points are important:

a) The computed energy of the * (La) state decreases upon flattening of Py-BN,

simultaneously the value of the dipole moment increases from 6.7 D ( = 69 ) to

8.5 D ( = 40 ) (Fig. 8.6).

b) the TICT state ( =26 D) was assigned as the 5th

state lying 9600 cm-1

higher than

the lowest excited state of Py-BN ( =69 ). This state can be strongly stabilized

by polar solvents.

As discussed above, for the ground state geometry ( = 69 ) the MCD spectra

clearly demonstrate (Fig. 3.6) that in polar and nonpolar solvents the lowest excited

singlet state is of * Lb type, similarly as for the 1-methylpyrene [Waluk and Michl,

1981].

After excitation, two relaxation paths are possible, described by two different

final geometries corresponding either to flattening ( =40 ) or twisting of the

benzonitrile with respect to the pyrene moiety to the perpendicular conformation

( =90 ). The kf value calculated from the quantum yields and lifetimes and transition

moment in emission (Table 1) depends on the solvent. In nonpolar solvent kf is about

4-5 times smaller than in polar solvents. The transition moment in n-hexane is two

times smaller than in acetonitrile. These results suggest that in nonpolar solvent the Lb

contribution to the S1 state is more significant than in polar solvents. Thus, we can

conclude that in nonpolar solvents fluorescence is emitted from the S1 state of which

the geometry corresponds to the geometry of the S0 state. The kf value calculated for

Py-BN in polar solvents is about 2.3x108

s-1

, very close to the value calculated from

integration of the first absorption band ( 3x108 s

-1). The dipole moment of the

emitting singlet state derived from the solvatochromic shift of the fluorescence

30000 40000 500000.0

0.2

0.4

0.6

0.8

1.0

Arb

itra

ry u

nits

cm-1

79

maxima is about 10.6 –12.5 D. This experimental result is to be compared with the

calculated value of the dipole moment of La state ( = 40 ) equal to 8.5D.

Fig. 8.6 Scheme of the energy levels for Py-BN calculated as a function of twist

angle . The first number given next to the curve is the calculated dipole

moment in (D), while the oscillator strengths are given in parentheses. The

energies of the S3 and S4 states are shown only for =69o.

Fig. 9.6 Absorption spectra of benzonitrile anion radical (1, dashed line) and pyrene

cation radical (2, solid line) [Shida,1988]. (3) Room temperature TA spectrum

of Py-BN in EtOH, recorded for delay time 100 ps.

0 20 40 60 80 100

0

5000

25000

30000

35000

40000

S0

5.7(.006)5.7(.008)

S1 (L

b)

5.7(.009)

S2 (L

a)8.5(.91)

6.7(.76)6(.7)

S5 (TICT)

13.3(.24) 19.4(.11)

26(.0004)

cm-1

400 500 600 700

-0.5

0.0

0.5

1.0

A.U

.

nm

1

2

3

405 nm449 nm

80

The dependence of the transient absorption on the solvent polarity (Fig. 6.6) also

indicates a different nature of the lowest emitting singlet state of Py-BN in nonpolar

and polar solvents. The absence of the prominent transient absorption bands around

400 nm and 450 nm, which are characteristic of the benzonitrile anion and pyrene

cation radicals, respectively (Fig. 9.6), excludes the involvement of a state with full

charge separation, such as a TICT state. The high-lying TICT state, because of its

large dipole moment can be effectively stabilized by polar solvents. The question is if

in the case of PyBN the TICT state can be really stabilized in polar solvents as the

lowest singlet state. A negative answer is provided by the following arguments:

1) The TICT So transition is expected to be strongly forbidden [Grabowski et

al., 1979]. For instance, the kf value of TICT state for a close analogue of

DMABN is, in polar solvents,within the limits of (0.4-1.5 x 107 s

-1)

[Grabowski et al. 1978] , [Kirkor-Kamińska et al., 1978], whereas in the case

of Py-BN in polar solvents the value of the radiative rate constant is about one

order of magnitude higher (Tab. 1.6).

2) The measured kf values are more consistent with the solvent stabilization of 1La state and with the resulting Lb and La inversion in polar solvents with a

simultaneous flattening ( =40o) of Py-BN (Fig. 8.6). Because the La state has

a much larger kf value than the Lb state, the stabilization of the La state with

respect to the Lb state should generate an increase of the kf value. Inspection of

Table 1 shows that this indeed occurs. Another possibility, stabilization of the

highly polar TICT as the lowest singlet state, can be excluded. For such a case,

the TICT state should be lower than the Lb state of pyrene, so that the

inversion of two states having a small radiative transition probability should

take place; consequently, the measured kf value for Py-BN in solvents of

different polarities should be small.

3) The transient absorption spectra of Py-BN in polar solvents cannot be

reproduced by the sum of the spectra of the benzonitrile radical anion and

pyrene radical cation (Fig. 9.6). The corresponding radical bands in the spectra

have been clearly identified for TICT states of carbonyl [Rullière et al., 1987]

and cyano [Okada et al., 1999] derivatives of N,N-dimethylaniline.

81

Py-AC

Experimental results

Steady state experiments

Room temperature absorption and fluorescence spectra of Py-AC were

recorded in a number of protic and aprotic solvents characterized by different polarity

indices. The shape and the spectral position of the first band in the absorption

spectrum (the extinction coefficient at the maximum is 29000 cm-1

/M) are practically

independent of the solvent (Fig.10.6).

Contrary to the absorption, the fluorescence spectra are strongly solvent

dependent ( Fig. 10.6 and Table 1). The increase of the polarity of the solvent shifts

the wavelength of the emission maximum markedly to the red. The comparison of this

effect for protic and aprotic solvents shows that the largest shift is observed in the

case of protic solvents (alcohols). Since the polarities ( ) of butyronitrile and

n-propanol are almost identical, whereas the shifts are very different, this indicates

that a specific interaction between Py-AC and alcohols plays an important role in the

excited state behaviour. This can be attributed to the hydrogen-bond formation

between the alcohol and the carbonyl group.

Table 2.6

The fluorescence band maxima ( max/cm-1

) , Stokes shifts ( /cm-1

) , fluorescence

quantum yields ( F), lifetimes ( /ns) and kf= F/ (109 /s), kn=

--1-kf (10

9/s

-1) and

transition moments in emission M={3hkf/644n

3 3}

1/2 measured for Py-AC in number

of solvents characterized by the polarity index .

Solvent max Fa

a kf kn M

hexane 1 1.89 24400 4900 0.011 0.078b

0.14 12.6 3.6

dibutylether 2 3.08 24050 5250 0.036

diisopropylether 3 3.98 23850 5450 0.04 0.22 0.19 4.3 4.2

diethylether 4 4.08 23700 5600 0.098 0.4 0.25 2.3 4.9

ethylacetate 5 6.11 23150 6150 0.21 0.76 0.28 1.04 5.2

1-chlorobutane 6 7.07 23200 6100 0.17 0.64 0.26 1.3 4.9

tetrahydrofurane 7 7.58 22900 6400 0.25 0.92 0.27 0.8 5.1

butyronitrile 8 20.3 21900 7400 0.46 1.82 0.25 0.29 5.4

acetonitrile 9 36.5 21000 8300 0.58 2.58 0.23 0.16 5.7

Alcohols

n-octanol O 10.3 20250 8950 0.48 3.6c

0.13 0.15

n-pentanol Pe 13.9 19600 9600 0.46 3.45c

0.13 0.16

n-butanol B 17.8 19350 9850 0.41 3.25 0.13 0.18

n-propanol P 20.1 19150 10050 0.35 2.62c

0.13 0.25

ethanol E 24.3 18650 10550 0.13 1.28c

0.1 0.68

methanol M 32.6 17900 11300 0.02 1.26c

0.02 0.77

a-The error being about 10%.

b-the value obtained from the decay curve of the transient absorption band.

c-the decay curve was recorded for the obs equal to max , the decay time was

evaluated from the main component.

82

Fig. 10.6 Normalized room- and low-temperature spectra of Py-AC in :

A-2,2-dimethylbutane:n-pentane (8:3 v/v), B – butyronitrile,

C – n-propanol. Solid line: absorption and fluorescence (295 K), dotted line:

fluorescence and fluorescence excitation (113 K).

15000 20000 25000 30000

0.0

0.5

1.0A

.U.

cm-1

A

27000

15000 20000 25000 30000

0.0

0.5

1.0

A.U

.

cm-1

B

15000 20000 25000 30000

0.0

0.5

1.0

A.U

.

cm-1

C

83

Fig. 11.6 Plot of the solvatochromic shift of the fluorescence maxima of Py-AC in

aprotic solvents at room temperature. Solvents numbered as in Tab.2.6.

The plot of the solvatochromic shift, starting from hexane (1) to acetonitrile

(9), can not be satisfactory approximated by a linear function (Fig.11.6). The values

of the excited state dipole moment evaluated from the linear fit of the experimental

points (1-4) and (4-9) are 9.5 D and 18 D, respectively. The value of ground state

dipole moment was assumed to be equal to 3.5 D; which was the result obtained from

molecular mechanics calculations. An Onsager cavity of radius a0=5.2 Å. was

evaluated from the molecular geometry for =60 .

The fluorescence quantum yields and decay times of Py-AC in a series of

aprotic solvents increase as a function of the solvent polarity (Tab. 2.6). The F/

ratio, which can be treated as the radiative rate constant (provided there is no

reversible excited state reaction) decreases for nonpolar solvents. The meaning of this

ratio for alcohol solutions is less clear, because non-exponential decays of the

fluorescence are observed . The kf values given in the Table 2.6 for alcohols were

calculated using the value of the dominant component of the decay. For this reason

the M values for Py-AC were not calculated. The maximum of the low temperature

absorption of Py-AC in polar protic and aprotic solvent is shifted to the red . The

fluorescence of Py-AC in rigid butyronitrile or in n-propanol is observed in the same

spectral region as the fluorescence emitted in nonpolar solvents (24000cm-1

), whereas

at room temperature the maxima were at 21900 cm-1

and 19150 cm-1

, respectively

(Fig. 10.6).

0.1 0.2 0.3 0.4

21000

22000

23000

24000

Flu

ore

sce

nce

ma

xim

a [

cm

-1]

F( ,n)

1 2

34

567

8

9

84

Time-resolved experiments

Transient absorption spectra of Py-AC in n-hexane recorded as a function of

the delay time are shown in the Fig. 12.6. The decay of the longwave band generates

simultaneous rise of the shortwave band . The kinetic curves indicate directly a

precursor-successor relationship.

Fig. 12.6 Top - TA spectra of Py-AC in n-hexane at 294 K recorded as a function of

delay time. Bottom – the normalized kinetics of the decay and rise of the long-

and short-wavelength bands respectively. Evaluated decay and rise times are

78 5 ps and 86 7 ps, respectively.

TRF spectra of Py-AC in MTHF recorded as a function of the delay time at

294 K and 173 K are compared with the TA spectra of Py-AC in THF at 173 K (see

figure 13.6). No temporal evolution of the TRF spectra was observed at 294 K,

whereas at 173 K the time dependent red shift of the TRF band is easily observed.

Transient absorption spectra consist of the stimulated fluorescence and absorption

located at about 450 nm and 600 nm, respectively. The stimulated fluorescence

500 600 700 800

0.0

A.U

.

nm

10 ps130 ps

2900 ps

0 1000 2000 3000

0.0

0.5

1.0

A.U

.

ps

<450nm-500nm>

<620nm-670nm>

85

exhibits a small time-dependent red shift contrary to the maximum of the transient

absorption with a small time dependent blue shift (Fig. 13.6 C).

Fig. 13.6 TRF spectra of Py-AC in MTHF recorded as a function of delay time at:

294 K (A) and at 173 K (B). TA spectra of Py-AC in THF at 173 K (C).

The comparison of the spectral position of the stationary fluorescence of Py-AC

recorded in protic and aprotic solvents of the similar polarity shows that the largest

300 400 500 600

0

100

200

300

TR

F In

t.

nm

6 ps190 ps

1200 psexc.

A

300 400 500 600

0

100

200

300

TR

F Int.

nm

450 nm

1200 ps

76 ps

exc.

2 ps

437 nm

B

400 500 600 700 800

-0.1

0.0

0.1

0.2

0.3

O.D

.

nm

70 ps

10 ps

1100 ps

620 nm605 nm

453 nm

440 nm

C

86

shift is observed in the case of alcohols (Tab. 2.6). The dielectric indices of

butyronitrile and n-propanol are almost identical, whereas the spectral position of the

stationary fluorescence maxima are different (Fig. 10.6). Room temperature TRF

spectra of Py-AC recorded in butyronitrile show insignificant temporal evolution,

contrary to those recorded in n-propanol which depend strongly on the delay time

(Fig.14.6).

Fig. 14.6 Room temperature TRF spectra of Py-AC recorded as a function of the

delay time in: A – butyronitrile, B – n-propanol..

To analyze the time dependent shift of the TRF maxima the empirical solvation

function c(t) was used [Barbara and Jarzęba, 1990]. The best fit of the experimental

data was obtained for s=46 5 ps (see Fig. 15.6).

Fig. 15.6 Squares - c(t) plot determined for Py-AC in n-propanol at 294 K.

Full line shows the best fit, the evaluated relaxation time being 46 5 ps.

400 500 600 700

0

100

200

300

TR

F Int.

nm

4 ps

50 ps

145 ps

1270 ps

B

300 400 500 600

0

100

200

TR

F In

t.

nm

4 ps

2850 ps

A

0 200 400 600 800 1000 1200

0.0

0.5

1.0

c(t)

flu

ore

scence

max.

ps

s= 46 ps

87

Calculations

Molecular mechanics calculations performed for Py-AC indicate that the most

stable ground state conformation is obtained when acetophenone subgroup is twisted

with respect to the pyrene moiety by =60 . From X-ray experiments the value of this

angle for Py-AC was established to be 62 [Lipkowski and Tabaszewska, 1994]. The

calculations show that COCH3 group is twisted by 29 with respect to the phenyl ring.

The experimentally obtained value is equal to 9 [Lipkowski and Tabaszewska,

1994]. To investigate the nature of the electronic states of Py-AC upon twisting of the

acetophenone group , INDO/S calculations have been performed for different values

of the dihedral angle (Fig.17.6).Comparison of the observed and calculated

absorption spectra of Py-AC (using =60 and =29 for the ground state geometry)

is shown in figure 16.6.

Fig. 16.6 Normalized (maximum of the first band) room temperature absorption

spectra of PY-AC in n-hexane and the calculated transitions for a twist angle

=60 ( =30 ). The height of the bars is proportional to the oscillator strength.

Transitions with f<0.007 are not marked. T – TICT state precursor.

For Py-AC the calculated S1 transition corresponds to an n * excitation (Figs

16,17.6) with electron redistribution mainly localized within the carbonyl group. The

calculated energy of this n * transition is probably much too low. For N,N-

dimethylamino-acetophenone the difference between the experimental and theoretical

values is as high as 3000 cm-1

[Dobkowski et al., 1989]. This reflects the well-known

fact that the INDO/S algorithm yields the values that are too low for the energies of

n * transitions.[Dewey et al. 1982], [Vasak et al.1982].

20000 25000 30000 35000 40000 450000

1

2

A.U

.

cm-1

La

Lb

n *

T

88

Two * transitions were predicted to lie within the spectral range of the first

absorption band of Py-AC (Fig. 16.6). These transitions correspond to the La and Lb

states of pyrene.

Fig.17.6 Scheme of the energy levels of Py-AC calculated by INDO/S as a function

of the twist angle ( =30 ), (see formulae for definition of the angles).The

first number given next to the curve is the calculated dipole moment (in

Debyes) while oscillator strengths are given in parentheses.

It is well established that in pyrene the Lb transition is strongly forbidden

[Thulstrup et al., 1970]. For Py-AC the radiative rate constant (kf) measured in polar

aprotic solvents is about 108 s

-1 (Table 2.6), what indicates that it is an allowed

transition. Thus, we conclude that in polar and weakly polar solvents the lowest

excited state of PY-AC cannot be of Lb character. Strictly speaking, for 50 the

distinction between the La and Lb states is difficult or even impossible, since there is

an interaction between the two states that depends on the value of the angle . Py-AC

can relax after excitation from the initial angular distribution, which may lead to a

decrease or increase of the twist angle. The calculations suggest that a flattening of

the molecule is energetically preferred, in particular in the La state. For a 90 twist,

the latter is described as a pure La state of pyrene with a large oscillator strength and a

small dipole moment. The transition that can be assigned as TICT state lies much

higher in the gas phase. It corresponds to the 6th

excited state with a very small

oscillator strength and a large dipole moment (Fig. 17.6).

0 20 40 60 80 100

0

5000

25000

30000

35000

cm-1

o

cm-1

n ** (Lb)

* (La)

4.5(.009)

5.6 (.8)

16.7(.2)28(.001)

TICT

o

89

Discussion

The possible conformers of Py-AC can be described by two angles: , the

dihedral angle between the acetophenone and pyrene moieties and , the twist angle of

COCH3 group (scheme 1.6). No model compounds, which have been so useful for the

verification of the excited state geometry in the case of cyano [Rotkiewicz et al..,

1976] and carbonyl [Grabowski et al., 1984].or its pyrimidine analogs [Herbich et al.,

1989] are available.

From the analysis of the results of INDO/S calculations two points seem important.

(I) The computed energy of the * (La type) states decreases upon flattening

of the Py-AC skeleton. Simultaneously, the value of the dipole moment

increases from 5.6 D (60 ) to 7 D (30 ) (Fig. 17.6). The minimum energy

of the *(La) is located at =30 - 40 ; whereas the minimum energy of

the ground state is at 60 .

(II) The TICT state ( = 28 D ) was assigned to the 6-th excited state lying

9000 cm-1

higher than the lowest excited * state of pyrene, but it can be

strongly stabilized by solvent.

Let us try to confirm or to negate the assumption that polar aprotic solvents can

stabilize the high laying TICT state of Py-AC as the lowest singlet state. The

experimental kf value calculated from the quantum yields and lifetimes (Table 2.6) is

about 2.5 x 108 s

-1, very close to the value calculated from the integration of the

absorption band (4 x 108 s

-1). Starting from diethylether the radiative rate constant is

practically solvent-independent (Tab.2.6). Its temperature variation is also weak: for

Py-AC in butyronitrile at 183 K kf=(4 1) x 108 s

-1. Transient absorption spectra can

not be reproduced by the absorption spectra of the corresponding radical pair :

pyrene(+) and acetophenone(-), (Fig.18.6). These findings provide arguments against

a switching in polar solvents from the emissive state * (La type) to the TICT state.

Fig. 18.6 Room temperature TA spectra of Py-AC in butyronitrile recorded 10 ps and

1100 ps after excitation. Py(+) – Pyrene radical cation, AC(-) – Acetophenone

radical anion [Shida, 1988}.

400 500 600 700 800-0.2

0.0

0.2

0.4

OD

nm

10 ps

1100 ps

AC(-)Py(+)

A.U

.

90

Fig. 19.6 The energy of the *(La) state of Py-AC calculated by INDO/S as a

function of angle and its estimated solvent stabilization. Squares - energy

levels of the La state, the first number is the calculated dipole moment (in

Debyes) while oscillator strengths are given in parentheses. Solid lines –

solvent stabilization (acetonitrile) for a range of possible Onsager radii

(a0=5.2-6.1 Å) [Mataga, 1975].

The figure 19.6 illustrates the fundamental property of the * (La) state of Py-AC:

decrease of the angle increases the excited state dipole moment. The reorganization

of the solvent cage induces the shift of the energy minimum, in other words the

flattening of the molecule skeleton.(SIF - solvent induces flattening). There are also

other consequence of the SIF model. The c(t) function can not be properly calculated

because the time-dependent shift of the TRF spectra is generated simultaneously by

excited state geometry transformation and solvent cavity reorganization (Fig.20.6).

Fig. 20.6 Left – excited state stabilization due to the solvent relaxation only.

Right- excited state stabilization due to the solvent relaxation and the

geometry transformation. Moment of the excitation is assigned by time 0; the

relation 0 1 2 3 indicates the sequence of the events on the time scale.

0 20 40 60 80 100

28500

29000

29500

30000

30500

7.2 (1)

6.6(.93)

4.9(.76)

cm-1

a0=5.2 A

o

6.1 Ao

0

10

20

30

So

lve

nt

rela

xa

tio

n

Intramolecular coordinate [A.U.]

Energ

y [

A.U

.]

S0

S1

0 1

2

3

o

-200

0

200

En

erg

y [

A.U

.]

Intramolecular coordinate [A.U.]

Solv

ent

rela

xation

Geom

etr

y t

ransfo

rmation

Edest

S0

S1

01

2

o

91

The plot of the solvatochromic shift of the fluorescence maxima can not be

satisfactorily approximated by a linear function (Fig. 11.6). The values of the dipole

moment evaluated from the linear fitting for the experimental points from the interval

0.09 F( ,n) 0.23 and 0.23 F( ,n) 0.4 are equal to 9.5 D and 18 D, respectively

(Fig 11.6). Contrary to that, for Py-CN the plot of the solvatochromic shift is well

approximated by linear function (Fig. 2.6). The sequences of the two low laying *

states (La and Lb) for both molecules are almost identical. The maximum of the first

absorption band in n-hexane is 29100 cm-1

(Py-CN) and 29000 cm-1

(Py-AC). The Lb

state is located at 26500 cm-1

for Py-CN (Fig. 3.6) and 27000 cm-1

Py-AC (Fig. 10.6).

However in the case of the Py-AC the n *, state located mainly on the acetophenone

moiety, contributes to the first absorption band. The energy of this state for N,N-

dimethylaminoacetophenone was established to be equal to 28500cm-1

[Dobkowski et

al, 1982]. The interaction between * and n * states generates anomalies of the

solvatochromic plot of Py-AC. Indeed in n-hexane kf value is about two times smaller

than in polar solvents; contrary to that kn is more than one order bigger. The transition

moment in n-hexane is 3.5, whereas in polar solvents it reaches the value 5.2-5.7 D

(Tab.2.6).The role of n * state on the relaxation path can be monitored by transient

absorption experiment. Room temperature TA spectra recorded in n-hexane just after

excitation show the band with the maximum about 650 nm; for longer delay times the

rise of the band located about 490 nm is observed (Fig. 12.6). The decay time

(78 5 ps) of the longwavelength band is equal to the rise time (86 7 ps) of the

shortwavelength one. The decay time of this band is significantly longer than 3ns. It is

reasonable to assign the primary band as the S1 Sn transition and secondary band as

the T1 Tn transition.

It should be stressed that in accordance with the SIF model the value of the dipole

moment of the excited state evaluated from the solvatochromic plot can be

overestimated due to the ground state destabilization Edest (Fig. 20.6, left).

Py-AC in protic solvents

For Py-AC in alcohols, a more complex emission behavior is observed than in aprotic

solvents (Fig. 14.6).

The comparison of the fluorescence of Py-AC observed in pairs of protic and

aprotic solvents with similar polarity indices (Tab. 2.6), e.g. acetonitrile vs methanol

(I) or butyronitrile vs. propanol (II), indicates that in protic solvents the Stokes shift is

evidently larger: for I the difference is 3100 cm-1

while for II it is 2750 cm-1

. Contrary

to that, in the case of Py-CN the Stokes shift in polar protic and aprotic solvents is

almost identical (Tab. 1.6). This fact indicates that great Stokes shift observed for

Py-AC in protic solvents is generated by hydrogen bonds created between carbonyl

group and alcohol molecule. The stationary absorption of Py-AC recorded in protic

and aprotic solvents is almost identical (Fig.10.6) which suggests that in the ground

state specific interaction between Py-AC and alcohol molecule is negligible. The

relaxation times evaluated from the c(t) functions for Py-AC in n-propanol and

butanol are collected in Tab 3.6 and compared with the relaxation times obtained for

Coumarine 153 used as a probe [ Horng et al., 1995].

92

Table 3.6

Relaxation times s, ns [ps] calculated from c(t) plots for Py-AC and

Coumarin 153 [ Horng et al., 1995].

Py-AC Coumarin 153 Solvent s 1s (A1) 2s (A2) 3s (A3) 4s (A4)

n-propanol 46 5 0.03 (.086) 0.34 (.167) 6.57 (.231) 47.8 (.515)

n-butanol 138 10a 0.243 (.159) 5.03 (.107) 42.6(.393) 133 (.341)

a – [Dobkowski et al., 2002)

The experimental data indicate that strong hydrogen-bonding interaction is

generated in the excited state. The temporal evolution of the solvent cage in the case

of alcohols is described by four relaxation times (Tab. 3.6). Two short relaxation

times describe the inertial and diffusive components of the monomer response. Two

longer relaxation times describe reorganization of aggregates. The relaxation times

calculated from the experimentally obtained c(t) function for Py-AC correspond well

with s4 (Tab.3.6) what indicates that after excitation the generation of the excited

states Py-AC hydrogen bonded complexes is connected with a reorganization of the

alcohol aggregates.

Conclusions

INDO/S calculations performed for Py-BN and Py-AC indicate that a higly

polar TICT state can be generated by rotation of the benzonitrile / acetophenone

moiety. However, it is calculated as a high-lying state, so that its stabilization below

S1 does not occur in polar solvents. The energy of the 1La state of Py-BN and Py-AC

slightly decreases upon flattening of the molecule, and its dipole moment increases.

The experimental results also favour the hypothesis of a solvent induced flattening

(SIF) of Py-BN and Py-AC after excitation and exclude sufficient stabilization of the

TICT state to become the lowest emitting singlet state. For Py-AC creation of the

hydrogen bonded complexes connected with a reorganization of the alcohol

aggregates is inferred.

The material was partially published:

1) J.Dobkowski, J.Waluk, W.Yang, C.Rullière, W.Rettig, New J. Chem.,

21(1997)429.

2) J.Dobkowski, W.Rettig and J.Waluk, Phys. Chem. Chem. Phys., 4(2002)4334.

93

7 Intramolecular charge-transfer properties of a molecule

with a large donor group:

The case of 4-Acetyl-4’-N,N-dimethylaminobiphenyl

(ADAB) The results obtained for ADAB, its model compound I with aromatic rings pretwisted in the

ground state ( 4-Acetyl-4’-p-N,N-dimethylamino-2,2’-dimethyl-biphenyl), and the cyano

analog of ADAB II, (4-N,N-dimethylamino-4’-cyanobiphenyl) are presented in this chapter. .

Scheme 1.7

A. Steady state spectroscopy

Room temperature absorption and fluorescence

Room temperature absorption and fluorescence spectra of 4-Acetyl-4’-N,N-

dimethylaminobiphenyl (ADAB) were recorded in a number of protic and aprotic solvents

characterized by different polarity indices .The shape and the spectral position of the first

absorption band is only weakly dependent of the solvent polarity (see Fig.1.7). Contrary to the

absorption, the fluorescence spectra are strongly solvent dependent (Fig. 1.7, Tab. 1.7).

Increasing the polarity of the solvent shifts the wavelength of the emission maximum

significantly to the red. The comparison of the effect for protic and aprotic solvents shows

that the largest shift is observed in the case of protic solvents .The polarities ( ) of

butyronitrile (20.3) and n-propanol (20.1) are similar, whereas the spectral positions of the

fluorescence spectra recorded in these solvents differ significantly (Fig. 2.7), which indicates

that a specific interaction between ADAB and alcohol plays an important role in excited state

hydrogen-bond formation.

MeMe

MeO

N

MeMe

MeO

N

Me

Me

ADAB I

MeMe

N

II

CN

94

Fig. 1.7 Room temperature absorption and fluorescence spectra of ADAB recorded in

n-hexane (Hex.), tetrahydrofuran (THF) and acetonitrile (ACN).

Fig. 2.7 Absorption and fluorescence spectra of ADAB recorded in butyronitrile (BuCN) and

propanol at 293K and at 77K. Circles - room temperature fluorescence excitation

spectrum, obs.=20000cm-1

. A-Fluorescence spectrum of II recorded in BuCN and

propanol at 293K.

The fluorescence spectra of II (cyano analog of ADAB) recorded in protic and aprotic

solvents are almost identical which indicates directly that red shift of the ADAB fluorescence

observed in alcohols is generated by hydrogen bond formation between carbonyl group and

alcohol (Fig. 2.7).

20000 30000

0.0

0.5

1.0HEX.THFACN

A.U

.

cm-1

HEX.

THF

ACN

15000 20000 25000 30000

0

20

40

60

80

100

Propanol

BuCN

Fluo.

T=293K

Fluo.

T=77K Abs.

T=293K

A.U

.

cm-1

20000 25000

0

50

100

Propanol

BuCN

A.U

.

cm-1

A

95

Table 1/7

The fluorescence band maxima ( max/cm-1

), quantum yields ( F), lifetimes ( /ns), radiative

and nonradiative rates kf = F / , kn=-1

-kf , and transition moments (in emission)

M={3hkf/644n

3( max)

3}

1/2 measured for ADAB in number of solvents characterized by the

polarity index .

F – error 10%

a – TRF technique, monoexponential fit, the error: 15%.

b – Single photon counting, monoexponential fit, obs - maximum of the band, error: 0.1ns.

For kf , kn , and M, the error was estimated as 15%

The plot of the solvatochromic shifts of the fluorescence maxima versus the polarity function

F( , n) is shown in Fig. 3.7. The excited state dipole moment was determined from the

relation [Liptay, 1962], [Liptay, 1974] ( see chapter 4), and is found as 25 3 D.

The values of the ground state dipole moment and the Onsager cavity radius were assumed to

be equal to 5.65D and 6Å, respectively, as for II [Maus et al., 1999].

Fig. 3.7 Left, plot of the solvatochromic shift of the fluorescence maxima of ADAB in

aprotic solvents at room temperature (see Tab. 1.7); e = 25 D 3 D.

Right, plot of thermochromic shift of the fluorescence maxima of ADAB in diethyl

ether, e = 27 D 4 D. Two lowest temperature points were not included into the

fitting procedure. The values of F( ,n) function were taken from [T.Yoshihara et al.

2003].

Solvent max

cm-1

F

ns

kf 10

9 s

-1 kn 10

9 s

-1 M D

1 hexane 1.89 26450 0.003 0.006a

0.5 158 5.7

2 dibutylether 3.08 24350 0.11 0.290a

0.4 3.2 5.6

3 diisopropylether 3.88 23900 0.25 0.7b

0.36 1.1 5.7

4 diethylether 4.34 23000 0.61 1.3b 0.45 0.29 6.8

5 ethylacetate 6.02 21800 0.68

6 terahydrofurane 7.58 21750 0.7 1.8b

0.39 0.17 6.5

7 butyronitrile 20.31 20250 0.64 2.7b

0.24 0.13 6

8 acetonitrile 37.50 19150 0.36 1.9b

0.19 0.34 5.9

0.1 0.2 0.3 0.4

20000

22000

24000

26000

Flu

ore

scence m

axim

a [cm

-1]

F( ,n)

1

23

4

56

7

80.26 0.28 0.30 0.32

21500

22000

22500

23000

23500

Flu

ore

scence M

axi

ma [c

m-1]

F( ,n)

T = 294K

T = 183 K

T = 243 K

96

Temperature dependence of absorption and fluorescence

Absorption spectra of ADAB were recorded as a function of temperature in several

solvents: dibutylether, diethylether, methyltetrahydrofuran (MTHF), and propanol. In all types

of solvents the first absorption band shows at low temperatures a red shift of the maximum,

reflecting changes in the population of the ground state vibrational levels and ground state

geometry changes. The absorption spectra of ADAB recorded in ether and MTHF for several

temperatures are presented in figure 4.7.

Fig. 4.7 The absorption spectra of ADAB in ether and MTHF recorded as a function of

temperature.

Contrary to the absorption, the fluorescence spectra exhibit significant temperature

dependence. In polar solvents, the fluorescence spectra of ADAB change drastically upon

cooling, as shown in figure 2.7. The emission of ADAB in rigid butyronitrile or in n-propanol

is observed in a very similar spectral region about 23000 cm-1

, whereas at room temperature

the maxima are located at 20250 cm-1

and 17800 cm-1

, respectively (Fig. 3.7). The maxima of

the fluorescence of ADAB recorded in liquid solvents at selected monotonically decreasing

temperatures reveal a red shift (Fig. 5.7, 6.7). Near the freezing point a reversed tendency is

observed ( see Fig. 6.7).

Fig. 5.7 ADAB / ether: A, Low temperature emission (exc.=29400 cm-1

); B, dependence

of the fluorescence maxima on temperature. The fluorescence maximum at 77 K is

located at 23500 cm-1

.

20000 30000 400000.0

0.2

0.4

0.6

0.8

1.0

Absorb

ance

cm-1

294K233K

213K

173KEther (4)

28100 cm-1

28900 cm-1

20000 25000 30000 35000 40000

0.0

0.2

0.4

0.6

0.8

1.0

Absorb

ance

cm-1

93K

203K

294K

28800 cm-1

27000 cm-1

MTHF

15000 20000 25000 30000

0

2000000

4000000

6000000

8000000

Flu

ore

scence Inte

nsity

cm-1

294K

233K

183K

163K A

160 180 200 220 240 260 280 300

21600

22000

22400

22800

23200

Flu

ore

sce

nce

m

ax.

[cm

-1]

K

B

97

The e value of 27 4 D was determined from the thermochromic shifts of ADAB

fluorescence in diethylether (Fig. 3.7).

Fig. 6.7 ADAB in MTHF: A - Low temperature emission (exc.=29400 cm-1

),

B – dependence of the fluorescence maxima on temperature.

The results of steady state experiments obtained for ADAB in aprotic solvents indicate that:

radiationless constants depend on solvent polarity (Tab. 1.7) ,

the transition moment (in emission) seems to be constant within the error limits (Tab. 1.7),

absorption spectrum weakly depends on solvent polarity and temperature,

fluorescences reveal strong dependence on polarity and temperature.

To understand the nature of a excited state species the model compound (I) with two aromatic

rings pretwisted in the ground state has been synthesized.

Model compound I

To eliminate or confirm that in the case of ADAB the following excited state relaxation

mechanism occurs:

primary excited state TICT

the model compound with two phenyl rings pretwisted in the ground state was investigated

(see scheme 1.7).

The maximum of the first absorption band of I (32200 cm-1

) is located at considerably

higher energy as compared to ADAB (28300 cm-1

), see figure 7.7. The room temperature

fluorescence of I in protic and nonprotic solvents was undetectable. Low temperature total

luminescence recorded in butyronitrile consists predominantly of phosphorescence (Fig. 8.7).

The low temperature luminescence and phosphorescence excitation spectra are identical,

showing the first band in the same spectral region as the room temperature absorption (Fig.

8.7). It is reasonable to suspect that due to the high energy shift of the * state the n * is the

lowest excited singlet state of I. Indeed, the energy of the 1n * state for acetophenone is

28500 cm-1

[Goodman and Koyanagi , 1972], the maximum of the first absorption band of I is

located at 32200 cm-1

. By analogy with 4-N,N-dimethylamino-2,6-dimethyl-4’-cyanobiphenyl

[Maus et al., 1999], [Maus and Rettig 1996] this band can be assigned to * transition.

The separation between 1n * and

3n * states is small, about 1700 cm

-1 in the case of

benzaldehyde and its derivatives [Hayashi et al., 1974]. The energy gap between 1

* and 3

* states is a few thousands cm-1

. Consequently, for I effective ISC transition between 1n *

15000 20000 25000 30000

0

1000000

2000000

3000000

4000000

5000000

6000000

Flu

ore

sce

nce

In

t.

cm-1

93K

298K

133K

A

100 150 200 250 30020000

20500

21000

21500

22000

22500

23000

23500

Flu

ore

sce

nce

Ma

x.

[c

m-1]

Temperature (K)

B

98

state and 3

* generates high population of the lowest triplet state, what can explain the lack

of fluorescence and the presence of phosphorescence in rigid media.

Fig. 7.7 Room temperature absorption spectra of I in n-hexane and butyronitrile. The arrow

indicates the spectral position of the maximum of the first absorption band of ADAB

in butyronitrile.

Fig. 8.7 I in butyronitrile at 77K: solid line luminescence and luminescence excitation;

dashed line, phosphorescence and phosphorescence excitation; dotted curve, room-

temperature absorption.

Taking into account the experimental data presented above it is reasonable to conclude that

the strong dependence of the Stokes shift on the solvent polarity observed for ADAB can not

be explained by the relaxation to the TICT state.

25000 30000 35000 40000 45000 50000

0.0

0.5

1.0A

.U.

cm-1

Hexane

Butyronitrile

ADAB

20000 25000 30000 35000

0

10

20

30

40

50

A.U

.

cm-1

Ph.Ph. exc.

Lum. Lum. exc.

Abs.

99

B. Time-resolved experiments The results of time-resolved experiments for ADAB in liquid phase are presented starting

from nonpolar solvents (hexane) up to polar ones (acetonitrile).

In n-hexane the time-resolved fluorescence (TRF) spectra are observed within the spectral

region 370-450 nm (Fig. 9.7). There is no observable spectral evolution of this band;

however, it should be pointed that the filters cut seriously the high energy part of this

spectrum.

Fig. 9.7 Top: TRF spectra of ADAB in n-hexane recorded at 294K as a function of the delay

time. Temporal resolution 6.5 ps, excitation 351 nm.

Bottom: points - kinetic curve of the TRF decay, integration limits: <400nm-450nm>;

open circles: residuals; solid curve -monoexponential fit, error in brackets.

Some variation of the residuals from the zero line was observed, but 2-exponential fit was not

performed, because the evaluated decay time was comparable with the temporal resolution of

the apparatus.

300 400 500 600

0

100

200exc.

7 ps

4 ps

2 ps

TR

F I

nt.

nm

A

3000

2500

2000

1500

1000

500

0

A.U.

340320300280

ps

800400

0

6.3 ps (1.6 ps)

100

Transient absorption (TA) spectra recorded for ADAB in n-hexane at room temperature are

presented in figure 10.7

Fig. 10.7 ADAB in n-hexane at 294K A – TA spectra recorded as a function of the

delay time. Horizontal lines indicate integration limits. B – points, the kinetics of the

decay of the blue part of the TA spectra <450nm – 500nm>, decay time=8ps 1ps,

open circles, residuals; C – points, the kinetics of the rise of the red part of the TA

spectra <550nm-650nm>, rise time = 6.3 0.6 ps, open points –residuals. Temporal

resolution 3ps, exc. wavelength=351nm.

400 500 600 700 800

0.00

0.05

0.10

0.15

0.20

2780 ps

78 ps

8 ps

3 ps

OD

nm

A

15

10

5

0

A.U.

30002500200015001000500

ps

-6-4-2024

8 ps (1ps)

Res.

B

40

30

20

10

0

A.U.

30002500200015001000500

ps

420

-2

Res.

6.3 ps (.6 ps) CR

101

In the case of ADAB in-hexane the kinetics of TA consist of a fast decay or a fast rise

depending on the observation and a long decay component. Within the accessible scale of

time this component can be treated as a constant. To find its temporal behaviour the decay

curves of the transient absorption were recorded in the nano to microsecond domain. The

evaluated decay time is equal to 50 5ns showing no dependence on the obs within the long-

wavelength band. The decay / rise times obtained from the emission / absorption time-

resolved experiments for ADAB in non polar solvents are collected in the table 2.7.

Table 2.7

Decay times ( d ) calculated from TRF experiments and decay ( d) / rise times ( r) obtained

from TA experiments for selected obs , T=293 K.

ADAB in TRF

d /ps

< obs>

nm

TA

d /ps

< obs>

nm

TA

r /ps

< obs>

nm

n-hexane 6.3 1.6 380-450 8 1 450-500 6.3 0.6 540-640

cyclohexane 9.4 0.7 390-470 11 1.7 445-485 10.5 2 540-640

dibutylether 295 16a 440-460 280 20

b 450-500

a – for the blue edge of TRF the fast component was detectable; d = 9 ps 1.5 ps.

b – fast component was observed; d = 10 ps 3 ps.

Room temperature TA spectra recorded for ADAB in dibutylether and

diisopropylether show the same features as the spectra recorded in nonpolar solvents. After

excitation stimulated emission and a short-wavelength TA band are observed, for longer delay

times the rise of new band located within the spectral region 550nm – 650 nm appears (Fig.

11.7). Contrary to that, in more polar diethylether the long-wavelength band is not observed.

Fig. 11.7 TA spectra of ADAB recorded in ethers at 294K, resolution 3 ps, exc.=351nm.

400 500 600 700 800-0.2

0.0

0.2

O.D

.

nm

8 ps

78 ps2800 ps

dibutylether

400 500 600 700 800

0.0

0.2 O.D

.

nm

8 ps

78 ps

2800 ps

diisopropylether

400 500 600 700 800

-0.2

0.0

0.2

0.4 A.U

.

nm

10 ps

320 ps

diethylether

102

Fig. 12.7 TRF spectra of ADAB in dibutylether recorded at three temperatures as a function

of delay time, exc.=351nm, resolution 6.5 ps

Room temperature TRF of ADAB in dibutylether recorded as a function of delay time

show maxima at 420 nm; however, it should be stressed that just after excitation the blue edge

of the spectrum shows a small red shift. At low temperature the temporal evolution of TRF

spectra is easily observed (Fig.12.7). The typical kinetic curves of the rise and decay recorded

for spectral intervals corresponding to the blue and red part of the spectra are presented in

figure 13.7. The character of the kinetic curves depends strongly on the observation ( obs).

For the blue part of the TRF spectra the biexponential decay is observed, whereas kinetic

curves obtained by integration of the long-wave part of the TRF exhibit a rise and slow decay.

The calculated decay / rise times are collected in table 3.7. The decay and corresponding rise

time obtained at room temperature are equal within the error limits. At low temperature the

complicated pattern of the decay and rise times obtained for a numbers of obs. is presented

in Table 3.7. The values of the decay times increase in accordance with the increase of the

wavelength of the observation. The same tendency is observed in the case of the rise times

(Tab. 3.7).

300 400 500 600

0

500

1000

TRF

Int

nm

exc. 3 ps

80 ps

300 ps

1000 ps

T=193K

300 400 500 600

0

500

1000

1500

2000

TRF

Int.

nm

3 ps

80 ps

300 ps

exc.

T=203K

300 400 500 600

0

500

1000

1500

TRF

Int.

nm

exc.

3 ps

80 ps

300 ps

T=294K

103

Fig. 13.7 ADAB in dibutylether at 203 K.

Top – points, the kinetics of the decay of the TRF obtained for the integration

limits <390nm-410nm>, full line –biexponential fit 1=26 3 ps, 2=130 30 ps,

circles – residuals.

Bottom – points; the kinetics of the rise and decay of the TRF spectra obtained

for integration limits <500nm – 540nm> , full line - biexponential fit R=109 22 ps, 2 =1760 240 ps, circles – residuals.

To analyze the time dependent shift of the TRF maxima the empirical solvation

function c(t) was used: c(t)={ (t)- ( )}/{ (0)- ( )}; where: (0), (t), ( ) are the

fluorescence maxima immediately after excitation, at some time after excitation, and at the

time sufficiently long to reach the equilibrium, correspondingly [Barbara and Jarzęba, 1990].

The c(t) plot was determined using the fluorescence maximum as well as the center of mass of

the TRF spectra ( Fig. 14.7).

12x10

10

8

6

4

2

0

A.U.

200018001600140012001000800600400

ps

800400

0-400 Res.

26 ps

130 ps

4000

3000

2000

1000

0

A.U.

200018001600140012001000800600400

ps

800400

0-400

Res.

109 ps 1760 psR

104

Table 3.7

The decay / rise times ( / r) evaluated from the decay curves recorded for ADAB in

dibutylether at room and low temperature for selected integration limits. A1/A2 – amplitudes

ratio, a negative value indicates rise.

T = 294 K

Integration limits

[nm] 1

[ps] 2

[ps]

A1/A2

390 - 410 9 2 203 10 0.92

440 - 460 297 20

500 - 540 R10 4 295 20 -0.37

T = 203 K

Integration limits

[nm] 1

[ps] 2

[ps]

A1/A2

390 - 410 26 3 131 30 2.62

410 - 440 140 20 1460 240 1.3

440 - 460 R46 8 1670 100 -0.35

460 - 500 R99 13 2540 230 -0.59

500 - 540 R109 22 1760 240 -0.72

T = 193 K

Integration limits

[nm] 1

[ps] 2

[ps]

A1/A2

390 - 410 37 11 157 54 2.2

500 - 540 R130 38 1920 430 -0.72

R-rise

Fig. 14.7 The c(t) plot determined using the center of mass for ADAB in dibutylether

at 294 K. Solid line-monoexponential fit to the experimental points. The value in

brackets indicates the error of the calculated . Dash and dot curve- the fits obtained

for relaxation times reported by [Ernsting and Kovalenko, 2004]. Zero ps corresponds

to the maximum of the excitation pulse.

0 50 100 150 200 250 300

0

1

c(t)

{c

en

ter

of

ma

ss}

ps

25 ps (4ps)

37.3 ps (0ps)

6.9 ps (0ps)

105

Table 4.7

ADAB in dibutylether, the relaxation times ( s) calculated from the c(t) plots determined

using the TRF maxima .

Temperature / K s / ps

Fluorescence maximum

294 25 4a

203 220 10

193 390 20

a) MC calculated from the c(t) plot determined using center of mass of TRF spectra.

Room temperature solvation dynamics of dibutylether detected by monitoring the time-

dependent Stokes shift of the emission of coumarine 153 has been performed by Ernsting and

Kovalenko.[Ernsting, and Kovalenko, 2004]. Fitting the solvation response function they have

obtained the following relaxation times: 0.4 ps (0.27), 6.9 ps (0.45), 37.3 ps (0.28);

corresponding amplitudes in brackets. For ADAB in dibutylether at 294 K the best fit of the

experimental data was obtained for the relaxation time equal to 25 4 ps (Fig.14.7, Tab.4.7).

The monoexponential simulations performed for two long relaxation times (6.933 and 37.34

ps) reported by Kovalenco and Ernsting are presented in figure 14.7. The similarity of the

fitted curve ( s=25 4 ps) with the curve obtained for relaxation time of 37.3 ps indicate that

the temporal evolution of TRF of ADAB in dibutylether at 294 K is fully or mostly controlled

by the dynamics of the solvent.

The TRF spectra of ADAB recorded in diethylether at 294 K and at 203 K are shown

in figure 15.7. Room temperature TRF spectrum consists of a broad band(s). Just after

excitation the maximum of this band is located at 432 nm. For longer delay times the

maximum shows red shift; for example max=440 nm (1500 ps). The low temperature TRF

spectrum consists of two bands showing significant spectral evolution as a function of delay

time. Three ps after excitation the maximum of the shortwavelength band is located at 445 nm

and the longwave maximum is observed at 498 nm. The spectral position of the short-

wavelength maximum depends strongly on the delay time (463 nm for 300ps), contrary to the

long-wavelength maximum which exhibits insignificant blue shift (about 8 nm). Room

temperature kinetic curves were constructed for a few spectral regions (Tab 5.7).

The kinetic curve obtained for the blue edge <390nm-410nm> of the TRF spectrum

can be properly approximated by a biexponential decay. The decay time obtained for the fast

component is 3 1ps and should be treated cautiously, because the temporal resolution of the

apparatus is about 6 ps. For longer < obs> the monoexponential fit is acceptable (Tab.5.7).

106

Fig. 15.7 TRF spectra of ADAB in diethylether recorded at 294 K and 203 K as a function of

the delay time. A – The kinetic curve of the TRF decay obtained for the

integration limits <390nm – 410nm>

300 400 500 600

0

500

1000

TR

F

In

tensity

nm

1500 ps

300 ps

3 ps 203 Kexc.

300 400 500 600

0

500

1000

0 500 1000 1500

0

4000

<3

90nm

- 4

10n

m>

ps

A

TR

F

In

ten

sity

nm

3 ps

300 ps

1500 ps

exc. 294 K

107

Tab. 5.7

Decay times calculated from TRF experiments for ADAB in diethylether at 293 K

Inegration limits [nm] 1 [ps] 2 [ps]

390 – 410 3 1 640 100

410 - 430 2 1 640 70

430 - 450 720 100

450 - 470 1240 180

470 - 490 1220 70

490 - 510 1260 40

510 - 540 1140 270

Tab. 6.7

Decay times calculated from TRF experiments for ADAB in diethylether at low

temperatures. Decay times longer than 700 ps are assigned as infinity.

203 K 173 K 163 K < obs>

[nm] 1,2

[ps] 1,2,3

[ps] 1,2,3

[ps]

400-410 3 0.5 2 1

5 1

8 1

14 1, inf.

25 2, inf. R5 3, 50 20, inf.

R11 2.5, inf.

R16.5 2.5, inf.

R30 4, inf.

R29 4, inf.

R28 3, inf.

R28.5 4, inf.

R29.5 2, inf.

3 0.5

4 0.5

10 0.5

19 1, inf.

43 3, inf. R6 1, 65 26, inf.

R15.5 2, inf.

R26 3 inf.

R34 2 inf.

R36.5 2 inf.

R36.5 2.5 inf.

R36. 2.5 inf.

R41 2 inf.

410-420 4 0.5

420-430 6.5 0.5

430-440 9.5 1, inf.

440-450 17.5 2.5, inf.

450-460 inf.

460-470 inf.

470-480 R8 1, inf.

480-490 R15 2, inf.

490-500 R17 2, inf.

500-510 R12.5 1.5, inf.

510-520 R16.5 2, inf.

520-550 R15.5 2, inf.

R – rise time

Using four-exponential global fitting procedure the following times were obtained:

T=203 K 0.85 0.88 ps, 16.4 0.7 ps, 409 48 ps, 2.69 0.15 ns.

T=173 K 4.6 0.5 ps, 29.7 1.7 ps, 430 110 ps, 2.6 0.3 ns.

T=163 K 6.7 0.4 ps, 32.4 0.7 ps, 590 140 ps, 3.1 0.3 ns.

The decay times of ADAB emission in diethylether evaluated from the experimental room

temperature kinetic curves recorded for several obs are collected in table 5.7. The kinetic

curves corresponding to the blue side of TRF spectrum exhibit two-component decay

(Fig.15.7 A). The decay of the fast component is about 3 ps (below the resolution of the

108

equipment). The decay time of the main component depends on the observation. To recognize

the kinetics of the decay and rise of the low temperature TRF spectra of ADAB in

diethylether the kinetic curves were plotted for 10 nm intervals starting from 400 nm up to

520 nm and thirty nm interval <520nm-550nm> for the red tail of the spectra (Table 6.7).

For the data collected in Table 6.7 two empirical rules can be formulated:

The blue part of the emission (400 nm – 450 nm) consists of the mono or two-

component decays. The calculated decay times of the fast component increase with

obs .

The red part of the emission exhibits the rise and decay; the rise time depends on the

obs in the same way as the decay time of the fast component.

Although for ADAB in diethylether TRF experiments were performed only at three

temperatures below 273 K (Tab. 6.7), the effort was done to evaluate the barrier of the excited

state process. For the rise times a good linear correlation of the ln r vs. 1/T is observed (see

Fig. 16.7, right). In the case of decay times, due to considerable spread of points, linear

correlation is not satisfactorily fulfilled (Fig. 16.7, left).

Fig. 16.7 ADAB in diethylether: the squares indicate experimental plots of ln (left) and ln r

(right) vs. 1/T,where - decay time, r – rise time. Lines - the linear fits.

The values of the barrier height calculated from the slopes of the linear fits are collected in

Table 7.7. The mean value of the barrier is equal to Eb = 540 40 cm-1

.

Table 7.7

Barrier height (Eb) evaluated from the low temperature TRF experiments for ADAB in

diethylether.

Spectral range [nm] 470-490 480-490 490-500 500-510 510-520 520-550

Eb [cm-1

] 675 489 433 620 452 547

0.0050 0.0055 0.0060

2.0

2.5

3.0

3.5

Ln

( )

1/T [K-1]

<440nm - 450nm>

<420nm - 430nm>

0.0050 0.0055 0.0060

2.0

2.5

3.0

3.5

Ln

(

r )

1/T [K-1]

<470nm - 480nm>

<510nm - 520nm>

109

The temporal transformation of the spectral distribution of the TRF of ADAB in diethylether

was analyzed using the c(t) function.

Fig. 17.7 The c(t) plot (squares and circles) determined using the fluorescence maxima

for ADAB in diethylether at low temperature. Solid curves show the best fits.

Tab.8.7

The relaxation times calculated from c(t) plots determined using the TRF maxima ( s) for

ADAB in diethylether.

Temperature [K] s [ps]

294 <4a

203 9.2 0.4

173 12.8 0.6

163 16.1 0.7

a) the long component was also observed, s=610 50 ps

For ADAB in diethylether at 203 K, 173 K and 163 K the c(t) curves are well

approximated by a monoexponential fit (Fig. 17.7). The values of the decay times decrease

with rising temperature (Tab. 8.7).

0 50 100 150 200 250 300 350-0.2

0.0

0.2

0.4

0.6

0.8

1.0

c(t)

{f

luore

scence

max.

}

ps

s = 9.2 ps (0.4ps)

s = 16.1 ps (0.7ps)

T=203K

T=163K

110

The TRF spectra of ADAB in methyltetrahydrofuran (MTHF) recorded at room and low

temperature are presented in figure 18.7.

Fig. 18.7 TRF spectra of ADAB in MTHF at room and low temperatures were recorded for

selected delay times; exc.=351 nm, resolution 6.5 ps.

Six picoseconds after the excitation ADAB in MTHF at 294 K exhibits a broad TRF

spectrum; it shows some dependence on the delay time (Fig.18.7). At intermediate

temperatures the TRF spectrum is composed of two bands. Just after excitation at 133 K the

maximum of the short-wavelength band is located at about 430 nm. During next fifty ps this

band decays and, simultaneously, its maximum shifts by about 25 nm to the red; 200 ps after

excitation only the long-wave band is observed (Fig 18.7). At 93 K no spectral evolution of

the ADAB emission is observed ( max = 425 nm).

300 400 500 600

0

200

400

600

TRF

Inte

nsity

nm

6 ps

16 ps

200 ps

700 ps

exc.

133K

300 400 500 600

0

200

400

600

800

TRF

Inte

nsity

nm

6 ps

300 ps

1300 ps

exc.

93K

300 400 500 600

0

100

200

300

TRF

Inte

nsity

nm

6 psexc.

1300 ps

T=294K

111

As it was done previously for ethers, the temporal transformation of the spectral distribution

of TRF of ADAB in MTHF was analyzed using the c(t) function constructed for the

fluorescence maxima or the center of mass (Fig. 19.7). The evaluated relaxation times from

low temperature experiments are collected in table 9.7.

Fig. 19.7 The typical experimental c(t) plots (squares) determined from using the

fluorescence maxima and center of mass for ADAB in MTHF at 133 K. Curves, the mono

(dotted) and biexponential (solid) fits obtained for indicated relaxation times (the averaged

values are presented in Tab. 9.7).

Table 9.7

The relaxation times calculated from c(t) plots determined using the TRF maxima ( s) and

TRF center of mass ( MC) for ADAB in MTHF.

Temp. [K] s [ps] MC [ps]

monoexp. MC[ps]

biexp. 203 9 2

173 15 2

153 9 2 26 2 6 2, 48 14

133 29 3 65 6 25 5, 200 70

123 94 13 185 25 62 20, 350 90

113 177 13 284 30

103 875 100

The decay times of ADAB fluorescence in THF calculated from the kinetic curves

obtained for the number of obs are collected in table 10.7. Similarly to the results obtained

for ethers the kinetic curves which correspond to the blue side of the TRF spectrum exhibit,

depending on the temperature, mono or two-component decay. For the red part of the

emission the rise and long decay component was observed. Because the decay times of the

long component were evaluated with poor accuracy, the decay times longer than 1000 ps were

assigned as a infinity. The relations between decay and rise times of the blue and red part of

the spectrum are easily observed.

0 50 100

0.0

0.5

1.0

c(t

) {f

luore

scence m

ax.}

ps

T=133 K

29 ps

0 200 400 600 800 1000

0.0

0.5

1.0

c(t

) {

ce

nte

r o

f m

ass}

ps

MC= 69 ps

MC1= 22 ps,

MC2= 144 ps

T=133 K

112

Table 10.7

Decay / rise times calculated from TRF experiments for ADAB in MTHF at low

temperatures. Decay times longer than 1000 ps are assigned as a infinity.

< obs> [nm] 1, 2 [ps],

173 K 1, 2 [ps],

153 K 1, 2 [ps],

133 K 1, 2 [ps],

113 K 400 - 410 <6 <6 <6 26 7

410 - 420 <6 <6 12 2 17 4, 144 18

420 - 430 <6 6 1 20 2 57 18, 270 90

430 - 440 8 1 12 1 33 2 130 50, 480 140

440 - 450 14 1 24 2 73 2 600 45

450 - 460 30 3 55 4 180 15 R80 24, 900 200

460 - 470 52 7, inf. 106 10 R21 5, 250 30

R110 20, inf.

470 - 480 R6 1, inf.

R18 5, 420 100

R42 7, 830 220

R125 30, inf.

480 - 490 R18 1, inf.

R40 3, inf.

R125 8, inf.

R180 60, inf.

490 - 500 R21 1, inf.

R48 3, inf.

R130 8, inf.

R300 70, inf.

500 - 510 R24 1, inf.

R53 3, inf.

R145 10, inf.

R240 70, inf.

510 - 520 R26 2, inf.

R59 4, inf.

R180 14, inf.

R480 200, inf.

520 - 550 R32 2, inf.

R75 5, inf.

R225 25, inf.

R590 300, inf.

R – rise time

For ADAB in MTHF the experimental points of the ln( r) vs. 1/T obtained for two

< obs> are presented in the Fig. 20.7. The average value of the barrier height calculated from

the slopes of the linear fits for obs:<480nm-490nm>, <500nm-510nm> is equal to

510 20 cm-1

.

Fig. 20.7 ADAB in MTHF: experimental points of ln of the rise times vs. 1/T: circles -

<480nm – 490nm>, squares - <500nm-510nm>, lines –the linear fits.

0.006 0.007 0.008 0.0092

4

6

ln (

r)

1/T [K-1]

113

ADAB in nitriles

The room temperature TRF spectra of ADAB in polar solvents: octanenitrile (OCN),

butyronitrile (BuCN), and acetonitrile (ACN) recorded for selected delay times are presented

in the figures 21.7 and 24.7. Dielectric constant of octanonitrile is 13.9 [Landolt-Bornstein,

1991].

Fig. 21.7 Room temperature TRF spectra of ADAB in A - OCN, B- BuCN,

C – BuCN, TRFblue

=TRF(3 ps)-TRF(10 ps).

The temporal evolution of the room temperature TRF spectra of ADAB recorded in

nitriles depends on their polarity. In octanenitrile, six ps after excitation the shoulder located

at the blue side of the main band is observed (Fig. 21.7A), disappearing during next 10 ps. In

BuCN (Fig. 21.7B) this shoulder is barely perceptible, whereas in ACN it is undetectable

(Fig. 24.7). The spectral position of the main band depends on the solvent polarity: 498 nm

(OCN), 510 nm (BuCN), 528 nm (ACN).

TRF spectra of ADAB recorded in BuCN at 173 K just after excitation exhibit a

shoulder located at the blue side of the main band ( max = 502 nm). The maximum of this

band shows time-dependent red shift (Fig. 22.7). It should be stressed that the long-

wavelength band of the TRF spectra of ADAB recorded in a less polar solvent (MTHF) at 173

K does not undergo any time-dependent spectral shift (see, for comparison, Figs 22.7 and

23.7).

300 400 500 600

0

500

1000

TR

F

Inte

nsity

nm

6 ps

1700 ps

14 ps

exc.

A498 nm

300 400 500 600 700

0

500

1000

TR

F In

ten

sity

nm

6 ps

14 ps

1250 ps

B510 nm

300 400 500 600 700

0

500

1000

TR

F In

ten

sity

nm

C

3 ps 10 ps

TRFblue

114

Fig. 22.7 TRF and TA spectra of ADAB in BuCN recorded as a function of delay time

at 173 K

Fig.23.7 TRF spectra of ADAB in MTHF recorded as a function of delay time at 173 K

300 400 500 600

0

200

400

600

TR

F I

nten

sity

nm

6 ps

30 ps

300 ps

exc.

502 nm514 nm

525 nm

TRF

400 500 600 700 800-1.0

-0.5

0.0

0.5

1.0

OD

nm8 ps

310 ps30 ps

510 nm

532 nm

TA

300 400 500 600

0

100

200

300

400

500

TR

F In

tens

ity

nm

40 ps

400 psexc.

12 ps

6 ps

496 nm

6 ps

ADAB / MTHF

115

TA spectra of ADAB in BuCN recorded as a function of the temperature consist of the

absorption band and stimulated emission located, respectively, within the spectral regions:

400 nm - 470 nm and 470 nm – 650 nm. Additionally, a weak absorption is observed within

the spectral region 650 nm – 750 nm. The time-dependent shift of the maximum of the

stimulated emission is easily observed. The calculated relaxation times correspond well to the

relaxation time evaluated from the c(t) function determined from the TRF maxima, see table

11.7.

Table 11.7

The relaxation times ( s) calculated from the c(t) plots determined using the TRF maxima and

the maxima of the stimulated emission (TA experiments) for ADAB in BuCN at various

temperatures, MC-relaxation time determined from c(t) plot for TRF center of mass.

Temperature [K] TA s [ps] TRF s [ps] TRF MC [ps] 203 26 2 24 3 9 2

173 92 5 83 5 40 3

163 153 10 110 10

153 264 13 184 15

ADAB in ACN

The room temperature TRF and TA spectra recorded for ADAB in ACN are presented

in figure 24.7. No time-dependent evolution of the TRF band is observed.

TA spectra consist of the absorption bands located within the spectral regions

400 nm-470 nm, 650 nm-750 nm and stimulated emission with the maximum at 530 nm. This

last value corresponds well to the maximum of the TRF spectra (528 nm). The spectral

distribution of the TA spectra does not depend on the delay time. The absorption band cannot

be reproduced by the spectra of the acetophenone radical anion and N,N-dimethylaniline

radical cation [Shida, 1988], see figure 24.7.

116

`

Fig. 24.7 ADAB in ACN at 294 K . Top – TRF spectra recorded for selected delay times.

Bottom – TA spectra recorded as a function of delay time and compared with the spectra of

the anion and cation radicals of the acetophenone and N,N-dimethylaniline, respectively.

300 400 500 600 700

0

1000

2000

TR

F

Inte

nsity

nm

6 ps

14 ps

1250 ps

528 nmTRF

400 500 600 700 800

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

OD

nm

N,N-dimethylaniline (+)

Acetophenone (-)

max=475 nm

8 ps

1030 ps

max=485 nm

117

Discussion

The results for the twist angles obtained by molecular mechanics calculations (calc.)

performed for ADAB and crystallographic data (exp) [Lipkowski and Tabaszewska, 1994 ]

are collected in the table below.

Table 12.7

The results of molecular mechanics calculations indicate that the most stable ground

state conformation is the one with the acetophenone subgroup twisted with respect to the

dimethylanilino moiety by =40 . From X-ray experiments the value of this angle was

established to be 33.3 . The calculations show that the N,N-dimethylaniline group is planar,

whereas the -COMe group is twisted by 26 with respect to the phenyl ring. The

experimentally obtained value for both is 8.1 .

For ADAB the quantum chemical calculations were not done, however the

calculations performed for 4-(N,N-dimethylamino)-4’-cyano-biphenyl [Maus and Rettig,

1997] and for 4-N,N-dimethylanilino-pyrimidine [Herbich and Waluk, 1994] indicate that in

the proper analysis of the exited state multidimensional relaxation path the angle plays the

most important role. After excitation the angle can increase or decrease its value. The first

case suggests a relaxation to the TICT when dimethylaniline and acetophenone subgroups are

mutually twisted around of the central bond by 90 . In the second case, after excitation

ADAB reaches a more planar structure than in the ground state.

Model compound I with the aromatic rings mutually pretwisted ( 70 ) does not emit

fluorescence in polar solvents at room temperature. Having in mind theTICT model, it is

reasonable to assume that after excitation the relaxation of I is dominated by radiationless

processes: primary excited state TICT So . In accordance with such hypothesis, primary

fluorescence should be observed in a rigid solvent. It is not the case. The low temperature

luminescence was assigned as the phosphorescence. That is the reason why the excited state

reaction : primary excited singlet state TICT state is not a probable explanation for the

excited state behaviour, only a fast ISC deactivation.

Let us try to analyze the experimental material in terms of the flattening of ADAB

molecule in the excited state. In nonpolar environment the fast decay of ADAB fluorescence

(6 ps) is observed (see Tab. 1.7 and Fig. 9.7). TA spectrum consists of two bands showing the

maxima at 470 nm and 550 nm. (Fig. 10.7A). Just after excitation, only shortwave band is

observed. The temporal evolution of two TA bands (Fig. 10.7 ) indicates that the parent-

daughter relation between the population decay (8 ps) and population rise (6.3 ps) of the

directly excited and the final state occurs. The final state is the long-lived one (50 ns), so it is

exp. calc

33.3 40.0

8.1 26.2

8.1 0.9

MeMe

MeO

N

118

reasonable to assign the long-wavelegth TA band as T1 Tn transition and the short-

wavelength one as the S1 Sn transition. The * singlet in n-hexane is located at 27200 cm-1

(Fig. 1.7). The energy of 1n * state of p-dimethylaminoacetophenone corresponds to 28500

cm-1

[J.Dobkowski at al., 1982]. The separation between 1n * and

3n * states is small, for

benzaldehyde and its derivatives about 2000 cm-1

[Hayashi et al., 1974]. Consequently, for

ADAB in nonpolar solvents the lowest singlet state is * and n * triplet is located about

700 cm-1

below it. Such configuration of the excited states explains the fast ISC process.

The polar solvents stabilize the polar singlet state with respect to the less polar n *

triplet, thus decreasing the efficiency of the ISC transition. For ADAB in solvents

characterized by dielectric index 4.34 (diethylether, MTHF, nitriles) the longwavelength TA

band is not observed (Fig. 11.7), which indicates that the T1 state is not effectively populated

and, consequently, IC is the leading radiationless process.

Room temperature spectral / solvation response functions:

After excitation, the stabilization of the 1

* state of ADAB is controlled by solvation

dynamics. The room temperature relaxation times reflecting time-dependent solvent cage

reorganization occurring around the excited ADAB molecule are collected in the Table 13.7

and compared with the data obtained for Coumarin 153, a commonly used standard for the

solvent relaxation monitoring.

Table 13.7

Room temperature solvation relaxation times s [ps] obtained for ADAB from solvation

response functions and the literature data obtained for Coumarin 153, (Ai –amplitudes).

solvent ADAB Coumarin 153a

s1 s1 (A1) s2 (A2) s3 (A3)

dibutylether 25 4MC 0.4(.27) 6.9 (.45) 37.3 (.28)

diisopropylether 155 22 0.118 (.432) 2.047 (.181) 127.7 (.387)

diethylether 4.7c,d

- - -

MTHF 2c,d

0.147b (.443) 1.406

b (.557)

BuCN 6 2MC 1.5-2.1

e

1.9 (0.29)f

4.5 (0.71)f

ACN 0.068 (.697) 0.600 (.303)

MC - relaxation time determined using the center of mass of TRF spectra.

a - [Ernsting and Kovalenko, 2004].

b - data obtained for tetrahydrofuran.

c - the values obtained from the fitting procedure are below the temporal resolution of the

apparatus.

d – a long component was observed: 610 ps (diethylether), 160 ps (MTHF)

e– [Barbara and Jarzęba., 1990]

f- [Zhu et al., 2005], optical Kerr method.

For ADAB in dibutylether and diisopropylether, the relaxation time evaluated from solvation

response function corresponds qualitatively to the longest solvent relaxation time ( s3) , see

table 13.7. For diethylether the relaxation time with subpicosecond resolution was not

published. In the case of MTHF, and BuCN the relaxation times evaluated from the c(t)

functions were equal or shorter than the temporal resolution of the apparatus. The low-

119

amplitude long component of the spectral response function observed for ADAB in

diethylether and MTHF does not correspond to the solvent relaxation.

The spectral position of the maxima of:

i the first band of the stationary absorption at 294 K

ii the blue part (TRF blue

) of the TRF bands observed just after excitation,

iii TRF spectra recorded 100 ps - 1250 ps after excitation,

iv maxima of the stationary fluorescence recorded in rigid solvents

are collected in the Tab 14.7.

TRFblue

was obtained by subtraction of two TRF spectra recorded just after excitation ( 1)

and for a longer delay time ( 2) :TRFblue

=TRF( 1)-TRF( 2) (see Fig. 21.7).

Table 14.7

Maxima of the absorption and emission bands of ADAB in selected solvents.

abs.-maximum of the first band of the stationary absorption (294 K) ; blue

-maximum of the

TRFblue

band (294 K), see text above , 1,2 the delay times in accordance with: TRFblue

=

TRF( 1)-TRF( 2), . -maximum of the TRF spectra recorded for the delay time when

relaxation was completed (294 K) ; =blue

- ; 77.-maximum of the stationary

fluorescence recorded in a rigid solvent (77 K).

solvent abs

[cm-1

]

blue

,

[cm-1

], (nm) 1 , 2

[ps]

,

[cm-1

], (nm)

[cm-1

] 77

[cm-1

]

dibutylether 28900 24300, (411) 3 , 80 23900, (419) 400

diethylether 28900 23500, (427) 3 , 600 22800, (438) 700 23500

MTHF 28500 22200, (450) 5 , 1300 20600, (485) 1600 23400

OCN 28200 22000, (455) a 20100, (498) 1900

BuCN 28400 21400, (467) 3 , 10 19600, (510) 1800 23400

ACN 28400 - b 18900, (528) -

a – TRFblue

the maximum was directly assigned for d=0 ps (see Fig. 21.7 A).

b – TRFblue

was undetectable.

The spectral position of the maximum of the first absorption band does not depend

significantly on the solvent polarity. Contrary to that, spectral distribution of the TRF

strongly depends on the delay time and solvent polarity. The maximum of the TRFblue

band in

diethylether (Tab. 14.7) corresponds to the spectral position of the stationary fluorescence

recorded in rigid solvent, because in rigid medium:

i. solvent molecules are immobilized and the organization of the cage in the excited

state is roughly similar to that in ground state,

ii. the geometrical transformation of ADAB itself is seriously restricted;

it is reasonable to conclude that TRFblue

in diethylether is emitted from the excited state of the

molecule having similar geometry to that of the ground state and surrounded by non totally

equilibrated solvent cage (see Tab. 14.7).

In rigid MTHF and BuCN the maximum of the stationary fluorescence is located at the same

energy as in diethylether, but the maxima of TRFblue

show a low energy shift. It indicates that

solvent relaxation accompanied by the geometry transformation was significantly faster at

120

room temperature than the radiation transition and, consequently, the high energy part of the

TRFblue

was not detected (Figs.22-24.7.)

Spectral dependence of the fluorescence decay at 294 K:

In dibutylether ( =3.08) the fast transformation of the TRF spectra was observed (Tab. 3.7).

Kinetic curves obtained for the blue edge and the red tails of TRF spectra exhibit (9 2) ps

decay and (10 4) ps rise respectively. (Fig.13.7). These times correspond to the relaxation

time ( s2) of dibutylether (Tab. 13.7).

For ADAB in diethylether the decay of the blue part of TRF is faster than the temporal

resolution of the equipment and probably it was the reason why the rise of the red part of the

emission was not detected. The room temperature kinetic curves of the rise and decay of

ADAB’s emission in MTHF and nitriles are not shown; however, the general pattern

observed in the case of ethers was preserved also in more polar solvents, see Fig 18.7 and Fig.

21.7.

Transient absorption spectra of ADAB in polar solvent

TA band of ADAB in ACN at 294 K cannot be reproduced by the sum of the spectra

of the anion / cation radical pair {Acetophenone (-) / N,N-dimethylaniline(+)}, see figure

24.7. This fact indicates that, after excitation, ADAB molecule does not relax to the TICT

state.

The low temperature temporal evolution of the emission

Rettig and coworkers [Maus et al., 1999], [Maus and Rettig, 1996] suggested that for

II solvent polarity can strongly affect the excited state potential surface. Decrease of

temperature should slow down a solvent equilibration around the excited molecule. Believing

that after excitation the solvent reorganization modifies the excited state potential, the TRF

evolution of ADAB should in some way correspond to the solvent relaxation times.

It was the reason why the low temperature experiments in emission were performed for

ADAB in dibutylether, diethylether, MTHF and BuCN.

ADAB in ethers

For ADAB in dibutylether the TRF spectra recorded at low temperature show the

temporal evolution of their spectral distribution (Fig. 12.7). Comparison of the spectra

recorded for selected delay times at 203 K gives the evidence that emission occurs from

different excited species with maxima located at 420 - 430 nm (depending on the delay time)

and about 500 nm. The solvent relaxation time s evaluated at 203 K (Tab.4.7) is longer than

the decay/rise times of the blue/red parts of the TRF spectra (see Tab.3.7). The decay of the

blue edge of the TRF band is not monoexponential and strongly depends on the integration

limits, the rise times of the red part of the emission also depend on observation and

correspond to the decay times (Tab. 3.7) These experimental facts suggest that at low

temperature reorganization of dibuthylether shell and excited state geometry transformation of

ADAB molecule go within similar time intervals.

For ADAB in diethylether at 203 K two bands are clearly seen (Fig. 15.7). The decay

of the short-wavelength band is accompanied by its simultaneous time-dependent spectral

shift from max=448 nm (3 ps) to max=458 nm (1500 ps). The long-wavelength band (495

nm) does not exhibit any significant time-dependent spectral shift. The small blue

displacement of its maximum observed for a long delay time (1500 ps) can be explained by

increase of the overlap between the two bands. The kinetic interaction between the short and

121

long-wavelength emission is supported by the analysis of the decay curves (Tab. 6.7). The

decay of the blue edge of the emission is not homogeneous (see Tab. 6.7). The decay times

obtained from the short decay components of the kinetic curves increase monotonously within

the spectral region 400-450 nm. The rise times evaluated from the red part of the emission

(450-550 nm) correspond well to the decay times of the fast component of the blue edge (Tab.

6.7). The decay times evaluated from the long components of the kinetic curves decrease for

long-wavelength observations ( obs>500 nm). The values of the decay/rise times depend on

temperature (Tab.6.7); but the general pattern described above is preserved. The analysis of

the wavelength dependence of the decay of ADAB emission was performed for three

temperatures only. This limited experimental material gives a chance to calculate the barrier

of the excited state reaction. Only in the case of the rise times a good linear correlation of the

ln r vs. 1/T was obtained, the energy of the barrier was found to be 540

40 cm-1

(Fig. 16.7,

Tab7.7).

Four-exponent global analysis procedure was performed for the complete set of the

decay curves recorded as a function of obs (for comparison see Tab. 6.7). None of the four

evaluated decay times corresponds to the relaxation time ( s) obtained from the solvent

response function of the short-wavelength band at the same temperature. This fact indicates

that this method of global analysis is a not proper tool for understanding the excited state

relaxation mechanisms in the case of ADAB.

Fig. 25.7 Top – stationary fluorescence of ADAB in MTHF at 294 K and in diethylether at

203 K: Bottom – TRF spectra recorded for ADAB in MTHF at 294 K and in diethylether at

203 K for delay times 6 ps and 32 ps, respectively.

300 350 400 450 500 550 600 650-200

0

200

400

600

800

1000

A.U

.

nm

MTHF, 6 ps

T=294 K

Ether, 32 ps

exc.

T=203 K

456 nm

300 350 400 450 500 550 600 650

0.0

0.2

0.4

0.6

0.8

1.0

A.U

.

nm

Eter, T=203 K

MTHF, T=294 K

457 nm

122

The dielectric constant ( ) of diethylether depends significantly on the temperature:

=7.4 at 203 K, =9.5 at 173 K, =11 at 163 K [Yoshihara et al., 2003]. The room temperature

dielectric constant reported for MTHF is equal to 7 [Furutsuka et al.,.1974]. Consequently the

spectral distribution of the steady state fluorescence of ADAB recorded in diethylether at

203K and in MTHF at room temperature should be almost identical. This conclusion is

supported by experimental results (see Fig. 25.7,top). The relaxation time of MTHF should

be similar to that reported for tetrahydrofuran (0.147 ps, 1.406 ps, Tab.13.7). It means that six

ps after excitation at 294 K the MTHF cage is fully equilibrated. The relaxation time ( s ) of

diethylether at 203 K is unknown, s evaluated from the c(t) function obtained from the time-

dependent shift of the short-wavelength maximum of ADAB is equal to 9.2 ps (Tab. 8.7). The

comparison of the TRF spectra of ADAB in diethylether recorded 3 ps after excitation (Fig.

15.7) and 32 ps after excitation (Fig. 22.7) at 203 K indicates that spectral transformation is

completed during the first 30 ps. These results indicate that excited state relaxation

mechanism crucially depends on the solvent cage reorganization.

ADAB in MTHF

The steady state fluorescence of ADAB in MTHF recorded as a function of the

temperature shows a complex spectral shift. Red shift of the emission is observed within the

temperature range 294 K –133 K; for temperatures lower than 133 K, blue shift is observed

(Fig. 6.7).

TRF spectra recorded for selected delay times at 294 K, 133 K and 93 K are presented

in figure 18.7. Room temperature TRF show insignificant temporal evolution of the spectral

distribution, contrary to that recorded at 133 K. The behavior of the ADAB emission in

MTHF at 133 K shows the same features as discussed above in the case of low temperature

TRF of ADAB in diethylether - it means: the short-wavelength band decays with a

simultaneous red shift of its maximum; the increase of the long-wavelength band occurs

without a significant change of its spectral distribution. TRF spectra recorded in rigid MTHF

do not depend on delay time.

For ADAB in MTHF the decay and spectral shift of the short-wavelength band at

203 K and 173 K was so fast that the determination of the c(t) plot using the fluorescence

maxima was impossible. That is the reason why the complementary c(t) function for the

center of mass was determined. The systematic relation s < MC is observed (Tab. 9.7). In the

case of the spectral evolution of the single band MC should be equal to s.. The time-

dependent shift of MC takes into account simultaneously the temporal evolution of the

spectral distribution of two bands, one of them not showing significant shift as a function of

the delay time. The s describes the temporal evolution only of the short-wavelength band,

which can explain the relation s < MC. To recognize low temperature temporal evolution of

the TRF spectra of ADAB in MTHF, the simulation of two fluorescence bands was

performed. The dipole moments of the primary and secondary form were assumed to be

equal. The shape of the TRF bands was approximated by a gaussian function; sr and re

characterize time-dependent shift of the band maxima and their amplitude variation in

accordance with the formulae presented below:

(t)= 0+ (1-exp(-t/ sr)),

I1(t)=I1exp(-t/ re) for short-wave band and I2(t)=I2(1-exp(-t/ re)) for long wave band,

where:

0 – spectral position of the band at 0 ps

- spectral interval of the shift of the band maxima.

123

In the case when a fluorescence is emitted from the minima of a potential hypersurface

sr characterizes solvent relaxation; however when a fluorescence is emitted on the relaxation

path (slope of the hypersurface) sr characterize solvent relaxation and fast molecular

transformation. It should be stressed that fluorescence which is emitted on the relaxation path

is not a “good measure” of the state population.

solvent reorganization

correlated sequence: excitation new electronic structure

geometry transformation

Typical simulations of the temporal evolution of TRF spectra obtained for the case when

re> sr are presented below.

Fig. 26.7 The simulation of the temporal evolution of the TRF spectra for =20nm,

0=440nm and 500nm; sr and re describe the time-dependent shift of the band

maxima and amplitude variation, respectively (see text). sr is equal to the

experimentally obtained s=29 ps for ADAB in MTHF at 133 K (Tab. 9.7)

The maximum of the long-wavelength band does not exhibit any spectral shift (Fig.

18.7). This experimental fact indicates that solvent relaxation occurs on the relaxation path.

The spectra simulated for sr=29 ps and re=50 ps for delay time 5 ps and 20 ps show the

spectral shift of the short- and long-wavelength bands (Fig. 26.7 A), which indicates that

secondary species is surrounded by not totally equilibrated solvent cage. In the case when

re>> sr, no shift of the long-wavelength band was exhibited. (Fig. 26.7 B). For each

simulated spectrum the center of mass was calculated in the same way as for the

experimentally obtained TRF.

The values of the center of masses evaluated from the experimental TRF spectra

recorded as a function of the delay time and biexponential fits are presented in figure 27.7 and

compared with the c(t) curves obtained from the simulated spectra ( sr=29 ps, and

re=50 ps, 130 ps, 250 ps). The experimental points are distributed between the curves

calculated for re=130 ps and 250 ps. These values correspond well to the main values of the

rise times of the long-wavelength band (Tab. 10.7).

400 500 600

0

20

40

60

80

100

A.U

.

nm

1 ps

5 ps

20 ps

sr=29 ps

re=50 ps

100 ps

500 ps A

400 500 600

0

20

40

60

80

A.U

.

nm

1 ps

20 ps

100 ps

200 ps

500 ps

sr=29 ps

re=250 ps

B

124

Fig. 27.7 Squares - the values of the center of mass evaluated from TRF spectra recorded for

ADAB in MTHF at 133 K as a function of delay time; dotted line.- biexponential fit of the

experimental points ( 1,2=22 7 ps, 145 40 ps). Solid lines – biexponential fit of center of

mass obtained from simulated spectra for sr=29 ps, and =50 ps, 130 ps, and 250 ps.

ADAB in BuCN

The low temperature TRF spectra of ADAB in BuCN consist of the short and long-

wave band. The short-wave band is observed just after excitation as the shoulder on the blue

edge of the long-wave band, which, contrary to the results obtained in MTHF reveal time-

dependent shift (see Figs 22.7 and 23.7). This experimental fact indicates that excited state

geometry transformation is comparable or faster than the solvent cage reorganization;

consequently, the solvent relaxation occurs mainly around the secondary excited state species.

Keeping in mind this idea, the simulation of the temporal transformation of the TRF spectra

was performed for the case when sr> re.(see Fig. 28.7 A). The monoexponential fit of the c(t)

function was determined using a center of mass of the simulated TRF spectra (Fig. 28.7A);

the experimental points and their monoexponential fit are presented in the figure 28.7B.

Fig.28.7. A-simulated TRF spectra, B-squares, the values of the center of mass obtained

experimentally for ADAB in BuCN at 203 K. Solid line-monoexponential fit of the center of

mass obtained for simulated spectra ( MC=12 2ps), dashed line-monoexponetial fit of the

experimental points (9 2 ps), Tab.11.7.

0 200 400 600 800 1000

0.0

0.5

1.0c(

t) {

cente

r of m

ass

}

ps

re=250 ps

re=130 ps

re=50 ps

sr=29 ps

exp. - 1=22 ps,

2=145 ps

400 500 600

0

20

40

60

80

100

A.U

.

nm

1 ps

5 ps

10 ps

20 ps

50 psAsr=29 ps

re=15 ps

0 50 100 150 200 250 300

0.0

0.2

0.4

0.6

0.8

1.0

c(t

)

{ce

nte

r o

f m

ass}

ps

12 ps (sim.)

9 ps (exp.)

B

125

The decay of the c(t)MC

function determined for the simulated two bands spectra

reveal the biexponential character. Due to the scattering of the experimental points often only

monoexponential fits of the experimental c(t)MC

functions were performed (Fig.19.7, Tab. 9.7

and 11.7). That is the reason why MC values obtained from the monoexponential fits of the

simulated c(t)MC

functions are collected in table 15.7.

Table 15.7

The MC values obtained from the monoexponential fits of the c(t)MC

functions determined for

the simulated TRF spectra. The other parameters of the simulated spectra: 0=440 nm and 500

nm for the short and longwavelength band respectively, =20 nm, FWHM=40 nm, sr =29 ps

re [ps] 250 130 50 40 15 5

MC [ps] 100 57 29 25 12 5

The comparison of the results of the simulations with the experimental data indicates that:

A) the simulations performed for sr<< re generate the relation sr< MC. That

is the case of ADAB in MTHF when the following experimental relation

was found s< MC (Tab.9.7, monoexp) ,

B) the decrease of the re with respect to the sr generates the relation sr MC.

That is the case of ADAB in BuCN at low temperatures when s> MC

(Tab.11.7).

The s evaluated from the c(t) function obtained for ADAB in BuCN at low temperature

correlates well with the 2 values obtained from the optical Kerr experiment, which indicates

that the time-dependent shift of the long-wavelength band is controlled by solvent relaxation

(Tab. 16.7).

Surprisingly, the 2 values of BuCN detected using the probe molecule are significantly longer

than corresponding relaxation times obtained from the Kerr experiment.

Table 16.7

The relaxation time ( s) evaluated from the c(t) function determined from the maxima of the

TRF and stimulated (TA) emission for ADAB at low temperatures, compared with the

literature data obtained using the ultrafast optical Kerr effect spectroscopy and the spectral

response function of the probe molecule.

ADAB/BuCN BuCN / Kerra

BuCN / probe DCSb

T[K] s

[ps] 1

[ps] 2

[ps]

< >c

[ps]

A1,A2 1

[ps] 2

[ps]

< >

[ps]

A1,A2

203 26 2TA

,24 3TRF

7.2 27.1 23.3 .19,.81 9 41 12.8 .88,.12

173 92 5TA

,83 5TRF

21 95 83.2 .16,.84 20 138 60.1 .66,.34

163 153 10TRF 30

d 145

d 126.6 16, 84 26 237 142 .45,.55

a- Data derived from Optical Kerr Effect measurements [Zhu et al., 2005]

b- Data obtained from Spectral Solvation Response Function of

4-Dimethylamino-4’cyanostilbene (DCS) [Druzhinin et al., 2006].

c- < >=A1 1 +A2 2

d-monoexponential extrapolation of the data [Zhu et al., 2005].

126

The visualization of the role of the solvent relaxation and excited state geometry

transformation on the relaxation path is shown in the figure 29.7.

Fig. 29.7 Top: the solvent reorganization and excited state geometry transformation relaxation

path. A- the case when solvent reorganization is faster than geometry transformation, B-

inverted case. Bottom: schematic cross section along a reaction coordinate in the potential

hypersurface.1,2,3 indicate time-dependent modification of the energy cross section due to the

solvent cage reorganization. S.R.- solvent reorganization.

The dielectric constant of the BuCN at 213 K ( = 39.0) is similar to that reported for

ACN at room temperature; the comparison of the stationary fluorescence spectra recorded for

ADAB in BuCN and ACN at 213 K and at 298 K, respectively, indicate that their spectral

distributions are almost identical (Fig. 30.7A). Because the TRF spectra of ADAB in BuCN at

213K were not registered, the spectrum recorded at 203 K was the subject of analysis. The

relaxation times of ACN are: 0.068 ps and 0.6 ps (Tab.13.7). The 1,2 relaxation times of

BuCN at 203 K are 7.2 ps, 27.1 ps (Kerr) and 9 ps, 41 ps (DCS probe), see Tab. 16.7.

Consequently the TRF of ADAB recorded just after excitation in ACN at room temperature

and in BuCN at 203 K for delay time 100 ps should be emitted from the relaxed * state.

Indeed, these spectra show a similar spectral distribution (Fig.30.7). These facts indicate that

polarity and time-dependent reorganization of the solvents play a leading role on the excited

state relaxation path.

Solv

ent r

ela

xatio

nA

B1

1

W

Z

Rekaxation path

W Z

Relaxation path

W Z

127

Fig. 30.7 A – the stationary fluorescence of ADAB in ACN and BuCN at 298 K and 213 K.

B - TRF spectra of ADAB: in ACN (Kerr medium: CS2) at 298 K for delay time 6 ps and in

BuCN (Kerr medium: C2Cl4) at 203 K for delay time 100 ps.

The excited state relaxation - the geometry transformation

To understand the temporal evolution of the TRF spectra observed at low temperature

in diethylether, MTHF and BuCN, the relaxation path should be described by applying two

intramolecular coordinates. The first coordinate is the angle which describes the mutual

rotation of the acetophenone and N,N-dimethylaniline subgroups around the central bond. The

second coordinate was selected somewhat arbitrary and will be discussed later.

The schematic cross section along the internal rotation coordinate in the potential

hypersurface of ADAB is shown in figure 31.7. The following feature of the energy cross

section is postulated: the decrease of the angle increases the excited state dipole moment. In

accordance with this postulate the position of the minimum of the energy curve ( 1) depends

on the polarity and relaxation time of a solvent. Only in non polar solvents the 1 0 relation

takes place, while for polar solvents the 0> 1 relation occurs (Fig. 31.7). INDO/S

calculations performed for 4-(1-pyrenyl) acetophenone [Dobkowski et al. 1997] show that for

La state the 20 -30 of the “flattening” generates the changes of the dipole moment from 5.6

D up to 7 D, which supports the model presented above.

Fig. 31.7 The proposed schematic cross section along the internal rotation coordinate (

angle) for ADAB in a nonpolar (A) and a polar (B) solvent.

O

S1

S0

O0O1

Edest

Dipole moment

A

B

15000 20000 25000 30000

0.0

0.5

1.0

A.U

.

cm-1

max=526 nm

ACN, T=298 K

BuCN, T=213 K

A

300 400 500 600 700

0.0

0.5

1.0

A.U

.

nm

6 ps

ACN

BuCN,100 psB

298 K

203 K

128

As it was mentioned earlier, for understanding the low temperature evolution of TRF

spectra of ADAB in diethylether, MTHF and BuCN, the relaxation path should be described

applying two coordinates. In figure 32.7 the cross sections along X coordinate for nonpolar,

and polar solvents are shown. The physical meaning of the X coordinate will be discussed

later. The ground state cross section along the X coordinate exhibits three minima located at

zero and X1. The cross section of the excited state shows also a three minima pattern, but the

energetic relation between them crucially depends on the solvent polarity. This means that

two minima located at X are destabilized with respect to the third one located at 0 by time-

dependent reorganization of the solvent cage.

Fig. 32.7 The proposed cross section along the X coordinate; A – weakly polar, B-polar

solvent. For explanation, see text.

X

0 X1X1-

S 0

S1

A

B

129

Fig. 33.7. Two-dimensional cross section describing the relaxation of ADAB in polar

solvents.

The visualization of the two-dimensional relaxation paths are presented in figure 33.7.

In nonpolar solvents the emission occurs from the excited state of the geometry

approximately similar to the ground state geometry; simultaneously, the fast ISC process

depopulates the singlet state. In weakly polar solvents at low temperatures (diethylether,

MTHF) the relaxation takes place down the ( ,X) slope, starting from 0 ,0 and 0 , X1

ending at 1,0 and 1 , X1 (ether) or 1,0 (MTHF). The fluorescence (short-wavelength

OO

O

X

X

0

S1

S0

1

X1

-X1

130

band) is emitted along the relaxation path. The reorganization of the solvent cage occurs on

the relaxation path generating simultaneously the stabilization of the potential (slope). After

the termination of the excited state relaxation, the fluorescence is emitted from equilibrated

conformers described by 1 ,0 and 1 , X1 coordinate (ether) or one conformer 1 ,0

(MTHF). In BuCN at low temperature the relaxation path is similar as in the case of MTHF.

However, the solvent reorganization is slower than the excited state geometry transformation.

Consequently, the solvent relaxation is continued when the geometry transformation has been

terminated ( 1,0). The barrier for geometry transformation for k , X1 k ,0, transitions

where 0 k 1, is comparable with kT at room temperature or even barrierless

(acetonitrile). It seems that the barrier height is not constant along the relaxation path. For this

reason the barrier reported for ADAB in diethylether and MTHF must be treated as averaged

value of the barrier height.

Few words about the physical meaning of the X coordinate. Because for II in

nonprotic solvents the dependence of the fluorescence on the solvent polarity was also

observed [Maus and Rettig, 1999], it is reasonable to ascribe the X coordinate to the

transformation of the N(CH3)2 group. There are two possibilities: pyramidalization around

the amino N atom or torsion around the C-N bond. For aniline in gas phase it was

experimentally established that in the ground state the NH2 group has pyramidal structure,

whereas in the first excited state this group is coplanar with the aromatic ring [Brand, 1966].

Ground state geometry calculations performed for dimethylaniline more than three decades

ago by Dubroca [Dubroca, 1972], [Dubroca and Lozano, 1974] indicate that in the ground

state the N(CH3)2 group is strongly pyramidalized; however a less stable, planar conformer

can also exist. For the torsion coordinate three-minima ground state potential was not

reported. Consequently X can be treated as the wagging coordinate.

In accordance with the presented model the excited state dipole moment increases on

the relaxation path (Fig. 31.7). Because for ADAB the quantum chemical calculations were

not performed, the increase of the dipole moment on the relaxation path can be estimated per

analogy with 4’-(pyrenyl) acetophenone. The quantum chemical calculations performed for

this molecule indicate that the 20o-30

o flattening of its skeleton increases the dipole moment

of La state about 20%-25% of the initial value [Dobkowski et al., 1997] .Consequently the

increase of the dipole moment and reorganization of the solvent cage can be also the driving

forces of relaxation of ADAB in weakly polar and polar solvents. In weakly polar solvents

(diethylether and MTHF) the solvent relaxation is mainly completed on the relaxation path

and, consequently, the time dependent stabilization of the final form is not observed (Fig. 29.7

A). In polar solvent (BuCN at low temperature) the solvent relaxation occurs only partially on

the relaxation path. That is the reason why the solvent stabilization of the final form is

observed (Fig. 29.7 B).

The intensity of TRF emitted on the relaxation path is a function of the amplitude of

the population package and the velocity of its translocation down the potential hypersurface

slope. Consequently, even when the radiative probability of the S1 S0 transition is constant

the TRF is not a “good measure “ of the population of the state for the time period of the

migration of the population package down the potential slope. The TRF intensity is

proportional to the population of the state only in the case of the emission originating from the

minimum of the potential hypersurface.

For the same reason the time value evaluated from the c(t) plot obtained for the

maxima of the TRF emitted on the relaxation path is a function of the solvent relaxation and

geometry transformation.

The spectral position of the maximum of the fluorescence which is emitted on the

relaxation path depends on the geometry transformation and solvent cage reorganization. For

these reasons the relaxation time of the solvent evaluated from the spectral solvation response

131

function c(t) which was determined from the maxima of the fluorescence emitted on the

relaxation path is not a proper measure of solvent relaxation.

There is also another consequence of the proposed scheme: the spectral position of the

maxima of the stationary fluorescence recorded in polar solvents depends not only on the

solvent stabilization of the S1 state but also on the excited state geometry transformation and

ground state destabilization (Edest.) (see Fig. 25.7). That can be the reason why the values of

the excited state dipole moment calculated from solvato or thermochromic shifts are so great,

25 D and 27 D respectively (Fig. 3.7), whereas the calculated value of the excited state dipole

moment for II is only 15.7 D [Maus and Rettig, 1997].

Conclusions

The experimental results support the hypothesis of flattening of ADAB molecule after

excitation and exclude the relaxation to the TICT state. The energy of the lowest 1

* state

decreases upon flattening of the molecule skeleton and its dipole moment increases. The

solvent relaxation and the increase of the dipole moment play the leading role in the excited

state relaxation mechanism.

132

8. Final remarks

The excited states characterized by significant or full transfer of electron from a donor

to an acceptor group can be categorized applying the minimum/maximum overlap rule

[Grabowski and Dobkowski, 1983]. This rule postulates that the overlap between the donor

and acceptor molecular orbitals involved in the electron transfer should reach maximum or

minimum value. The realization of the minimum overlap postulate can be achieved by:

location of the node of the orbital on one of the atoms creating a single bond between

donor/acceptor group with relatively small geometry transformation [ Zilberg and Haas,

2002],[Cogan et al., 2006].

significant geometry transformation of the molecular skeleton, which leads to the donor and

acceptor orbital decoupling.

The experimental material presented in Section A indicates that for a number of

molecules having dialkylamino groups as the donors the secondary emitting state can be well

classified as the TICT state (minimum overlap of the -electron systems generated by the

twist of the donor and acceptor subunits around the central bond). Contrary to that, in the case

of primary states, the systems of the donor and acceptor groups reach the maximum overlap

arrangement.

For molecules having, in comparison with the dialkylamino unit, more extended

electron systems of the donor group (pyrene, N,N-dimethylaniline), and acetophenone as the

acceptor the flattening of the molecular skeleton in excited state increase the overlap between

the donor and acceptor systems.

Increase or decrease of the overlap between the donor/acceptor orbitals occurring on the

relaxation path increases the value of the dipole moment of the excited state and

simultaneously generates the reorganization of the solvent cage.

Time-resolved and stationary spectroscopy as well as the optically induced time-

resolved NMR experiment are sources of information concerning the geometry transformation

and excited state electron transfer. The solvent relaxation can be monitored applying time-

resolved fluorescence technique.

The experimental data obtained for the molecules being the subject of the study in section A

can be analyzed applying a general scheme of the excited state reaction:

S0 X Y

1) For dialkylamino derivatives of m-cyanopyridine (CP-derivatives) in methanol excited

state reaction is inreversible (Fig. 4.4). UV induced time-resolved NMR spectra (Fig.

17.4) and TA spectra (Figs 11,12.4) give the evidence that that the product of the reaction

is generated by the 90o twist of the dialkylamino group with simultaneous electron transfer

from donor (dialkylamino) to acceptor (cyanopyridine). These experimental results are

unquestionable proofs that Y TICT. The rise time ( R

) of the TA absorption band

ascribed to S1 Sn transition of the twisted conformer is 14 3 ps (PEC in methanol at 294

K, Fig. 14.4). For methanol the spectral-solvation response function c(t) can be well

approximated by a four-exponential fit and, consequently, four relaxation times are

reported: s1-4=0.03 ps (0.1), 0.28 ps (0.34), 3.2 ps (0.3) and 15.3 ps (0.26), corresponding

amplitudes in brackets [Horing et al., 1995]. The relaxation of the solvent cage is faster

133

than the excited state reaction : 1/kxy=R

> s1-3 (only for s4 comparable to R

);

consequently, the secondary state is populated from the precursor surrounded by almost

totally equilibrated local environment. Because the excited state dipole moments of the

primary and secondary form fulfill the relation Y> X the reorganization of the solvent

cage occurs also in the final state (Y) (see Fig. 1B. 8).

Summing up: in the case of PC-derivatives the excited state reaction is controlled by

solvent relaxation.

Fig. 1.8 One coordinate, double minimum excited state potential for a non-reversible excited

state reaction: xy< s (A), xy> s (B). Shadow areas indicate solvent stabilization.

2) For the class of para substituted molecules of the type: D-Ar-X (D-donor, Ar—aromatic

ring, X-chromophore) or D-P (P-pyridine or pyrimidine) having the N(CH3)2 or N(C2H5)2

groups as donors in polar solvents the dual fluorescence is observed independently of the

sequence of the 1Lb and

1La states. As it was established earlier, the

1La state is the precursor

of a TICT state (Fig.14.5)

An important difference between para substituted dialkylamino derivatives of m-

cyano-piridine (CP-derivatives) and dialkylamino derivatives of benzaldehyde and

acetophenone (BA-derivatives) exists in the sequence of the two lowest * states in

absorption. In the case of the planar form of the CP-derivatives 1Lb state is located below

1La.

The twist of the donor group with respect to the acceptor unit induces inversion of these

states, what generates the intramolecular barrier on the potential hypersurface cross section

along coordinate.

In the BA-derivatives the 1Lb state is above the

1La, and additionally, depending on the

solvent polarity, the 1n * state is located above or below

1La. In other words: the polarity of

the environment induces the inversion of the *(1La) and

1n * states (chapter 5). In polar

solvents 1La is the lowest singlet state and, consequently,

1La TICT reaction should be

barrierless. The denotation Y TICT was done on the basis of stationary and time-resolved

spectroscopy of the model compounds (see chapter 5).

The barrierless relaxation scheme, however, is not supported by experiments. Two,

kinetically coupled, fluorescence bands observed for BA-derivatives in polar solvents indicate

that a barrier exists on the relaxation path.

0

En

erg

y

[A.U

.]

FX F

Y

Twist angle

YX

Abs. So

lve

nt

sta

bili

za

tio

n

A0

En

erg

y

[A.U

.]

FX F

Y

Twist angle

YX

Abs.

B

134

1n *

1*(La) interaction

To understand the origin of this barrier two coordinate intramolecular potential should

be analyzed for the assignment of the relaxation path (Fig. 2.8). Let us assume that:

in the gas phase, the energy of 1La( ) decreases monotonously as a function of angle ,

the energy of the nonpolar 1n * state does not depend on and solvent polarity.

The energy gap ( E) between 1n * and the lower lying

1*(La) state is small and depends

on the solvent polarity and on the solvent cage relaxation time. The vibronic ―pseudo-Jahn-

Teller (p-JT)‖ interaction between them can significantly modify the potential hypersurface

along an out-of-plane vibronic coordinate [Hochstrasser and Marzzacco, 1969], [Lim,

1977]. This interaction can induce time-dependent quenching of the fluorescence via fast ISC,

but can not generate the intramolecular barrier along coordinate (Fig. 2.8).

Fig. 2.8 Modification of the potential energy curves resulting from pseudo-Jahn-Teller

interaction. A, – interaction between two nearly degenerate states; B – interaction between

two close –lying excited states. C - two-dimensional cross section, assuming that energy of

the 1La. state decreases monotonously along coordinate.

-2 0 2

0

10

Ene

rgy

[A.U

.]

out of plane coordinate [A.U.]

S0

S1

AS

2

o

-2 0 2

0

10E

nerg

y [A

.U.]

out of plane coordinate [A.U.]

S0

S1

S2

B

o

XY

X

Twist angle

0

Cou

t o

f p

lan

e c

oo

rdin

ate

0 /2

135

In the case when:

the energy of the 1La( ), in the gas phase, increases monotonously as a function of angle

(Fig. 3.8),

the energy of non polar 1n * state does not significantly depend on angle and solvent

polarity,

the P-JT interaction perturbs the energy hypersurface not along ( , =0 ) coordinate, but

along ( , = x), when 0 < x<90 . Because the La( ) state, due to its high dipole moment is

stabilized by solvent, the dependence of x on time ( s) is expected. Solvent-induced barrier

can be generated along the coordinate.

Solvent induced-barrier

The solvent effects studied in the case of DMABA applying the solvaton model

associated with the AM1 procedure indicate that environment can generate the barrier along

coordinate [Gorse and Pasquer, 1995]. On the basis of their results it is instructive to

discuss the role of the solvent on the modification of the relaxation path in the case of the BA-

derivatives (for details see chapter 5, formula 1.5, Fig. 14.5). Let us consider, in the gas phase,

the linear increase and decrease of the energy of La as a function of with a simultaneous

parabolic dependent increase of the dipole moment, limited by: (0o)=9.5 D, (90

o)=16.5 D

(Fig. 3.8). Only in the case when the twist of the donor group increases the energy of the La

state the solute–solvent interaction can generate barrier on the relaxation path. It should be

noticed that calculated energetic stabilization in accordance with Mataga formula corresponds

to the equilibrium condition, which means that solvent cage reorganization was completed.

The time-dependent evolution of the energy of the La( ) induced by solvent cage

reorganization was calculated in accordance with the formulae {E(gas phase)-Esolv}[1-exp(-

t/ s)], where s is the solvent relaxation time (Fig. 4.8)

Fig 3.8 Energy dependence of the La state on for BA-derivatives: A,A’-assumed linear

destabilization and stabilization respectively (gas phase). The calculated energy of La in

acetonitrile:E(gas phase)-Esolv where Esolv=-{eq

( )}2f /(ao)

3 formula 1.5,[Mataga, 1975].

0 20 40 60 80 100

0.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

En

erg

y

[AU

]

o

Esolv

Esolv

Esolv

A'

A

136

Fig. 4.8 The time –dependent evolution of the energy of the La state, s=10 [AU].

In the case of BA-derivatives the barrier for the 1La TICT excited state reaction is generated

by solvent. The parameters of this barrier (height and max) depends on the time ( s)

The depopulation of the 1La state of planar geometry in polar solvents is dominated by

two processes: the fast ISC transition between 1La

3 * states and the

1La TICT reaction.

It was proved that both processes are controlled by solvent relaxation and both reaction rate

constants, kISC and kTICT , depend on time. In accordance with the model of the energy

degradation presented above the correlation between lifetime of the primary fluorescence and

relaxation time of the solvent should be observed. Indeed, the lifetimes of the Fb fluorescence

of DMABA and DMI measured in butyronitrile at low temperatures are equal to the

relaxation time ( s1) of butyronitrile reported in literature (Tab.5.5).

0 20 40 60 80 100

0.80

0.85

0.90

0.95

1.00

1.05

1.10

1.15

En

erg

y [A

U]

o

t=0 [AU]

t=5 [AU]

t=10 [AU]

t>>s

137

3) Contrary to the molecules having dialkylamino groups as the donor units the excited state

relaxation path of the 4’-(1-pyrenyl) benzonitrile or acetophenone (Py-BN, Py-AC) can be

well described applying the solvent-induced flattening (SIF) model (see figure 5.8 and chapter

6).

X and Y are the conformers characterized by 1 and 0 angles fulfilling the relation 0> 1.

The relaxation to the TICT state was eliminated on the basis of:

quantum chemical calculations,

independence of the fluorescence rate constant (kf) on the solvent polarity,

impossibility to reproduce the TA spectra recorded in polar solvents by the sum of the

corresponding anion/cation radical absorption bands.

Fig. 5.8 Solvent-induced flattening (SIF) model, exc –excited state dipole moment.

Moment of the excitation is assigned by 0; the relation 0< 1< 2 indicates the sequence of the

events.

The excited state flattening is associated with simultaneous increase of the dipole

moment and solvent cage reorganization. Consequently, one broad band of the stationary

emission is observed, of which the spectral position depends on the solvent polarity. Due to

the ground state destabilization (Edest) the excited state dipole moment evaluated from the plot

of the solvatochromic shift of the stationary fluorescence can be significantly overestimated.

The empirical solvation function c(t) can not be properly calculated, because time-

dependent shift of the TRF spectra is generated simultaneously by excited state geometry

transformation and solvent cage reorganization.

4) Complicated temporal evolution of the TRF spectra of ADAB in polar solvents at low

temperatures cannot be explained in terms of one-dimension SIF model, chapter 7.

The relaxation to the TICT state was eliminated because:

S1 S0 transition is allowed independently of the solvent polarity,

a model compound with donor/acceptor group pretwisted did not show any fluorescence

that can be classified as primary/secondary (TICT) emission,

Sol

vent

rel

axat

ion

Geo

met

ry tr

ansf

orm

atio

n

Edest

S0

S1

01

2

/200

Abs.

Twist angle

exc

Ene

rgy

[A.U

.]

XY

1

138

the transient absorption spectra can not be reproduced by the sume of cation/anion

absorption bands

For the interpretation of the experimental material two dimensional SIF model was applied

(Fig. 6.8). For detailed discussion see chapter 7. The coordinate can be assigned as the out

of plane vibration of the dialkylamino group (wagging). The correlation between and

coordinates results from the postulated model, it means that the flattening of the biphenyl

rings induce the flattening of the dialkylamino group.

Fig. 6.8 Proposed energy hypersurface cross section along: { , 0}-A, { , 1}-B,

{ =0, }-C; 1,2,3 indicate the modification of the energy hypersurface cross section induced

by the increase of the solvent polarity.

Excited state geometry transformation associated with an increase of the excited state

dipole moment induces solvent cage reorganization. In weakly polar solvents (diethylether,

MTHF) at room and low temperatures solvent relaxation occurs mainly during migration of

the population packet along , valley and, consequently, the time-dependent stabilization of

the final form is not observed (Fig. 23.7). In polar solvents (butyronitrile at low temperatures)

solvent relaxation is not completed on the relaxation path and, consequently, the time

dependent shift of the TRF spectra of the final form is observed (Fig. 22.7, top).

-2 0 20

5

10

15

En

erg

y

[AU

]

[AU]

S1

S0

Abs.Abs.A

-2 0 20

5

10

En

erg

y [A

U]

[AU]

S'1

S'0

B

S0

S1

Abs.

1

1

2

3

C

exc

0

En

erg

y

[AU

]

XY

139

The presented results indicate that for the class of the molecules having donor and

acceptor D-A groups connected by a single bond the relaxation processes occurring in liquid

phase after excitation to the CT state are described in terms of the excited state geometry

transformation and solvent cage reorganization. There is no one universal model which can

describe the charge separation process, but rather the spectrum of the models that mirror

possible interaction between excited states, mono or multi dimensional intramolecular

geometry transformation and solvent relaxation. Any type of the D-A type molecules must be

treated separately.

140

9 Annex

For the study of the dynamics of the liquid phase in the femto and picosecond domain two

optical methods are used: time dependent Stokes shift detection and optical Kerr effect

spectroscopy.

Time dependent Stokes shift of fluorescence is the useful observable for the

determination of the energy stabilization of a lowest singlet state generated by the solvent

cage reorganization. The dipole moments of the ground ( 0) and the first excited state ( 1) of

the probe molecule should fulfill the relation 1 > 0 . Properly selected probe molecule

should not undergo the excited state geometry transformation. In accordance with the

literature, the time dependent Stokes shift can be understood in terms of the electrostatic

interactions between the solvent molecules and the excited solute [Reynolds et al., 1995],

[Castner and Maroncelli, 1998], [Barbara and Jarzęba, 1990]. After excitation the

redistribution of the electronic cloud occurs in the attosecond time domain. The relaxation of

the solvent cage occurs on the femto and picosecond scale. The initial fast large-amplitude

part of the spectral response function c(t) is, in many cases, well described by a Gaussian

function and is attributed to small, rapid rotational ”inertial” motions in the first solvation

shell. The slow small-amplitude component of c(t) can be satisfactorily described by mono or

multi exponential functions and is attributed to diffusive motions [Sudhold at al., 2000] and

references therein.

In order to study the molecular motions in pure solvents, optical Kerr effect

spectroscopy can be applied. In this method the dynamics of the response of the orientational

part of the total polarizability anisotropy response of the liquids is followed [Smith and

Meech, 2000], [Zhang et al., 2003].

The relaxation times and corresponding amplitudes evaluated from spectral solvation

response function c(t) and derived from the optical Kerr effect experiments for butyronitrile in

few temperatures are collected in Table 1.

Table 1

The relaxation times si [ps] and corresponding amplitudes Ai obtained from the analysis of

the spectral-solvation response function c(t) of 4-Dimethylamino-4’-cyanostilbene (DCS) in

butyronitrile compared with the data evaluated from optical Kerr effect in pure butyronitrile

[Druzhinin et al., 2006] and [Zhu et al.,2005].

T [ C] DCS Kerr

s1 s2 A1 A2 s1 s2 A1 A2

-60 8 34 0.9 0.10 5.8 20.7 0.2 0.8

-70 9 41 0.88 0.12 7.2 27.1 0.19 0.81

-80 9 47 0.76 0.24 9.7 38.8 0.18 0.82

-90 13 80 0.64 0.36 14 59.5 0.17 0.83

-100 20 138 0.66 0.34 21 95 0.16 0.84

The relaxation times evaluated from two different experimental methods correspond one to

the other. Contrary to that, the amplitudes show the inverted trend. To understand the nature

of these discrepancies the analysis of the events occurring after excitation of the probe

molecule immersed in the solvent cage was performed:

141

1) excitation, redistribution of the electron cloud of the probe and surrounded solvent

molecules – attosecond domain,

2) expansion / contraction of the molecule in the excited state in respect to the ground

state volume – tens-hundreds femtosecond domain,

3) solvent cage reorganization - femto-picoseconds domain.

In the case of optical Kerr spectroscopy the dynamics of the pure solvents is monitored.

Generally in the excited states the length of the bonds is larger than in the ground state. Let us

try estimate this effect in the case of DMABN’s CT excited state of the planar geometry of

DMABN (Fig. 1) [Cogan et al. 2006].

Fig. 1 Calculated structure and changes of the lengths of the bonds ( L) in the case of the CT

excited state of the planar geometry of DMABN molecule. The lengths of the C-H bonds

have not be shown in the source [Cogan et al., 2006].

To adopt new distances between the atoms, the DMABN molecule needs the time, which

corresponds to a half of the period of the stretching vibration. This time can be roughly

estimated as tens to hundreds femtoseconds.

The molecular dynamics simulations of the solvent effects in the intramolecular

charge transfer of DMABN were performed by Sobolewski, Domcke and coworkers [Sudhold

et al., 2005]. In accordance with their results the radial distribution of the acetonitrile around

DMABN depends on the assumed geometry and the nature of the state. The first maximum of

the calculated radial distribution function g(r), depending around which atom, is located at

about 3.7-5 Å, for example see figure 2A

Fig. 2 A- Radial distribution function of the center of mass of acetonitrile around nitrogen

atom of the cyano group; B- the Lennard-Jones potential (VL J) around the first maximum of

the g(r) function.

N

MeMe

CN

0.016A

0.05A

0.018A

-0.01A

0.037A

-0.002A

0.008A

L=L(SCT

)-L(S0)

0.096 A

0.052 A

o

o

2 4 6 8 10

0.0

0.5

1.0

1.5

2.0

N

M eM e

CN

g(r

)

Ao

A

3.7 Ao

3.0 3.5 4.0 4.5-2

0

2

4

6

VL

J [A

.U.]

A

r0=3.7 A

o

o

0.4

[A

.U.]

0.05 Ao

=1

[A

.U.]

B

142

The interaction between non bonded two atoms or molecules can be described by the potential

which must include two terms:

the long distance attracting term (in many cases coulomb interaction [Anner and Haas,

1986])

the short distance (r) repulsive term described by exponential, or r-n

function [[Piela, 2006].

To estimate the change of the intermolecular potential corresponding to the changes of the

molecular diameters in the S1 state, the commonly used Lennard-Jones potential was applied

VL J = {(ro/r)12

– 2(ro/r)6}

Where : ro corresponds to the minimum of the VL J potential; -the depth of the potential well.

Accepting that ro=3.7 Å corresponds to the maximum of the radial distribution

function g(r) (Fig. 2A), the zero value of the VL J potential is located at 3.3 Å. This value is

about 0.3 Å higher than the edge of the g(r ) function (Fig. 2A). The change of the VL J ( VL

J ) for the r=3.3-3.25 Å is equal to 0.4 , for r=3.3-3.2 Å, VL J=0.95 . Concluding - the

change of the r of 0.05-0.1 Å generate the change of the potential comparable with the depth

of the well.

The viscoelastic continuum model of nonpolar solvation [Berg , 1997] postulated that

the coupling of the probe molecule to the nonpolar solvent occurs through a change in the

probe’s size upon excitation. Having in mind the arguments of this annex it is reasonable to

include the excited state changes of the bonds lengths of the probe molecule into the polar

solvent relaxation model. It is worth to add that an expansion /collapse of the solute can

generate simultaneously the transfer of the mass center and rotational motions of the solvent

molecules.

143

10 References

[Andres and Moskovitz, 1989] – L.Andrews and M.Moskovitz ,ed., Chemistry and Physics

of Matrix-Isolated Species, North Holland, Amsterdam, 1989.

[Anner and Haas] – O.Anner and Y.Haas, J.Phys.Chem., 90(1986)4298.

[Arzhantsev, Maroncelli, 2006] – S. Arzhantsev, M. Maroncelli, Applied Spectroscopy

59(2005)2006.

[Assel et al., 1999] – M.Assel, R.Laenen, A.Laubereau, J. Chem.Phys., 111(1999)6869.

[Assel et al., 2000] – M.Assel, R.Laenen, A.Laubereau, Chem. Phys. Letters, 317(2000)13.

[Barbara and Jarzęba, 1990] – P.F.Barbara, W.Jarzęba, Advances in Photochemistry V.15,

John Wiley &Sons Inc., 1990.

[Been and Weller, 1968] – H.Beens, and A.Weller, Acta Phys. Pol., 93(1989)3439.

[Berg, 1997] – J.Phys.Chem.A, 102(1998)17.

[Bondybey et al., 1983] – V.E.Bondybey, S.V.Milton, J.H.English, P.M.Rentzepis, Chem.

Phys. Letters, 97(1983)130.

[Brand, 1966] – J.C.D.Brand, J. Mol. Spectr. 20(1966)359.

[Castner and Maroncelli, 1998] – E.W.Castner and M.Maroncelli, Journal of Molecular

Liquids 77(1998)1.

[Chergui, 2000] – M.Chergui, Low Temp. Phys., 26(2000)632.

[Cogan et al., 2006] – S.Cogan, S.Zilberg, Y.Haas, J.Am.Chem.Soc.,128(2006)3335.

[Cowley and Healy, 1977] – D.J.Cowley and P.J.Healy, Proc. Roy. Irish Acad.,

77B(1977)397.

[Dewey et al., 1982]- J.H.Dewey, R.D.Miller, J.Michl, J. Am. Chem. Soc., 104(1982)5298.

[Dicker et al., 1981] – A.I.M. Dicker, J.Dobkowski, S.Völker, Chem. Phys. Letters,

84(1981)23.

[Dicker, 1982] - A.I.M. Dicker D.Ph., Leiden, 1982.

[Dicker and Völker, 1982] – A.I.M. Dicker and , S.Völker, Chem. Phys. Letters,

87(1982)481.

[Dobkowski et al., 1982] – J.Dobkowski, E.Kirkor-Kamińska, J.Koput, A.Siemiarczuk,

J.Lumin., 27(1982)339.

[Dobkowski et al., 1989] – J.Dobkowski, J.Herbich, J.Waluk, J.Koput, W.Kühnle

J. Lumin. 44(1989)149.

[Dobkowski et al., 1995] – J.Dobkowski, Z.R.Grabowski, J.Jasny, Z.Zieliński, Acta Phys.

Pol. A 88(1995)455.

[Dobkowski et al. 1997] – J.Dobkowski, J.Waluk, W.Yang, C.Rullière, W.Rettig,

New.J.Chem., 21(1997)429.

[Dobkowski et al., 1998] – J.Dobkowski, J.Herbich, V.Galievsky, R.P.Thummel, F.Wu,

J.Waluk, Ber. Bunsenges. Phys. Chem., 102(1998)469.

[Dobkowski et al., 1999] – J.Dobkowski, V.Galievsky, A.Starukhin, J.Waluk,

Chem Phys. Letters, 318(2000)79.

[Dobkowski et al., 2002] – J.Dobkowski, J.Michl, J.Waluk, Phys. Chem. Chem. Phys.,

5(2003),1027.

[Dobkowski et al., 2002] – J.Dobkowski, J.Wójcik, W.Koźmiński, R.Kołos, J.Waluk,

J.Michl, J. Am. Chem. Soc., 124(2002)2406.

[Dobkowski et al., 2004] – J.Dobkowski, V.Galievsky, M.Gil, J.Waluk,

Chem. Phys. Letters, 394(2004)410.

144

[Dobkowski et al., 2004] – J.Dobkowski, V.A.Galievsky, J.Jasny, I.V.Sazanovich, Polish

J.Chem. 78(2004)961.

[Dobkowski et al., 2005] – J.Dobkowski, Y.Lobko, S.Gawinkowski, J.Waluk,

Chem. Phys. Letters, 416(2005)128.

[Druzhinin, et al., 2005] – S.I.Druzhinin, V.A.Galievsky, and K.A.Zachariasse,

J. Phys. Chem. A , 109(2005)11213.

[Druzhinin et al., 2006] - S.I.Druzhinin, V.Galievsky, T.Yoshihara, K.A.Zachariasse,

J.Phys.Chem. A, 110(2006)12760.

[Dubroca, 1972] – C.Dubroca, Chem.Phys.Letters, 15(1972)207.

[Dubraca and Lazano 1974] – Chem.Phys.Letters, 24(1974)49.

[Duguay, 1969] – M.A.Duguay, Progres in Optics (North-Holland, Amsterdam,1976), p.162.

[Ernstingand Kovalenko, 2004] – N.Ernsting and S.Kovalenko, private information.

[Fournier et al., 1983] – F.Fournier, J.Berthelof and Y.L.Pascal, Can. J. Chem.,

61(1983)2121.

[Furutsuka et al., 1974] – T.Furutsuka, T.Imura, T.Kojima, K.Kawabe, Technol. Rep. Osaka

Univ., 24(1974)1155-1190, 367.

[Galievsky et al. 2005] – V.A.Galievsky, S.I.Druzhinin, A.Demeter, Y-B Jiang,

S.A.Kovalenko, L.P.Lustres, K.Venugopal, N.P.Ernsting, X.Allonas, M.Noltemeyer,

R.Machinek, K.A.Zachariasse, CemPhysChem, 6(2005)2307.

[Gil and Waluk, 2007] – M. Gil, J. Waluk, J .Am. Chem. Soc.,129(2007)135.

[Gilabert et al., 1987] – E.Gilabert, A.Declèmy, C.Rullière, Rev. Sci. Instrum.58(1987)2049.

[Goodman and Koyanagi] – L.Goodman and M.Koyanagi, Mol.Photochem., 4(1972)369.

[Grabowski et al.,1978] - Z.R.Grabowski, K.Rotkiewicz, W.Rubaszewska and E.Kirkor-

Kamińska, Acta Phys. Pol. A, 54(1978)767.

[Grabowski et al.,1979] – Z.R.Grabowski, , K.Rotkiewicz, A.Siemiarczuk, D.J.Cowley and

W.Baumann, Nouv. J. Chim., 3(1979)443.

[Grabowski and Dobkowski, 1983] – Pure and Appl. Chem., 55(1983)245.

[Grabowski et al., 1984] – Z.R.Grabowski, J.Dobkowski, W.Kühnle, J. Mol. Struct.,

114(1984)93.

[Grabowski et al., 2003] – Z.R.Grabowski, K.Rotkiewicz, W.Rettig, Chemical Reviews,

103(2003)3899.

[Gurzadyan et al., 1998] – G.G.Gurzadyan, T.-H.Tran-Thi, T.Gustavsson, J.Chem. Phys.,

108(1998)385.

[Harris, 1977] – C.B. Harris, J. Chem. Phys., 67(1977)5607.

[Hashimoto and Hamaguchi, 1995] – M.Hashimoto, H.Hamaguchi, J. Phys. Chem.,

99(1995)7875.

[Hayashi et al., 1974] – H.Hayashi, S.Nagakura, Mol. Phys., 27(1974)969.

[Horng et al., 1995] – M.l.Horng, J.A.Gardecki, A.Papazyan, and M.Maroncelli,

J.Phys.Chem. 99(1995)17311.

[Herbich et al., 1989] – J.Herbich, Z.R.Grabowski, H.Wójtowicz, K.Golankiewicz,

J. Phys .Chem., 93(1989)3439.

[Herbich et al., 1989] – J.Herbich, K.Rotkiewicz,J.Waluk, B.Andersen, E.W.Thulstup,

Chem. Phys., 138(1989)105.

[Herbich and Kapturkiewicz, 1991] – J.Herbich and A.Kapturkiewicz, Chem.Phys.,

158(1991)143.

145

[Herbich et al.,1992] – J.Herbich, J.Karpiuk, Z.R.Grabowski, N.Tamai,and K.Yoshikara,

J.Lumin., 54(1992)165.

[Herbich and Kapturkiewicz, 1993] – J.Herbich and A.Kapturkiewicz, Chem. Phys.

170(1993)221.

[Herbich et al, 1993] – J.Herbich, Z.R.Grabowski, H.Wójtowicz and K.Golankiewicz,

J.Phys.Chem. 89(1993)3439.

[Herbich and Waluk,1994] – J.Herbich, J.Waluk, Chem.Phys., 188(1994)247.

[Hopkins and Rentzepis, 1985] – Chem. Phys. Letters 117(1985)414.

[Huppert et al., 1985] – D.Huppert, V.E.Bondybey, P.M.Rentzepis, J. Phys. Chem.,

89(1985)5811.

[Jasny, 1978] – J.Lumin., 17(1978)149.

[Jasny et al., 1994] - J.Jasny, J.Sepioł, J.Karpiuk, J.Gilewski, Rev. Sci., Instr. ,

65(1994)3646.

[Jodl, 1989] – H.J.Jodl, in L.Andrews, M.Moskowits (Eds.), Chemistry and Physics of

Matrix-isolated Species, Elsevier, North-Holland, 1989.

[Jonusauskas et al., 1994] – G.Jonusauskas, R.Gadonas, C.Rullière, Optics Communications,

112(1994)80

[Jonusauskas et al., 1997] – G.Jonusauskas, J.Oberlè, E.Abraham, C.Rullière, Optics

Communications 137(1997)199.

[Kahlow et al., 1988] – M.A.Kahlow, W.Jarzęba, T.P.DuBruil, P.F.Barbara, Rev. Sci.

Instrum. 59(1988)1098.

[Kanematsu, 2000] – Y.Kanematsu, H.Ozawa, I.Tanaka, S.Kinoshita, J.Lumin.87-

89(2000)917.

[Khalil et al. 1976] – O.S.Khalil, J.L.Meeks, and S.P.McGlynn, Chem. Phys. Lett.,

39(1976)457.

[Kirkor-Kamińska, 1978] – E.Kirkor-Kamińska, K.Rotkiewicz, A.Grabowska,

Chem.Phys.Lett., 58(1078)379.

[Kołos, 2003] – R.Kołos, Hab. Thesis, Institute of Physical Chemistry, Warsaw, 2003.

[Kwok et al., 2000] – M.W.Kwok, C..Ma, D.Phillips, P.Matousek, W.A.Parker, M.Towrie,

J. Phys. Chem. A, 104(2000)4188.

[Kwok et al., 2003] – M.W.Kwok, C..Ma, M.W.George, D.C.Grills, P.Matousek,

W.A.Parker, D.Phillips, W.Toner, M.Towrie, Phys. Chem. Chem. Phys. 5(2003)1043.

[Landolt-Bornstein, 1991] - Numerical Data and Functional Relationship in Science and

Technology, New Series, ed O. Madelung, Group IV, volume 6, Springer, Berlin, 1991.

[Lipkowski and Tabaszewska, 1994] - J.Lipkowski, A.Tabaszewska, private information

[Liptay,1962] – W.Liptay, Z. Naturforsch.,A. 17(1962)621.

[Liptay,1974] – W. Liptay in “Excited states”, ed. E.C. Lim, Academic Press, New York,

1974, vol. 1, p.129.

[Lippert et al., 1962] – E.Lippert, W.Lüder, H.Boos, in: Advances in Molecular

Spectroscopy, ed. A.Mangini (Pergamon, Oxford, 1962) p.443.

[Luzina et al.,2007] – E.Luzina, J.Sepioł, Y.N.Svartsov, A.Grabowska, J. Chem. Phys.

126(2007).

[Lower and El Sayed, 1966 ].-.S.K.Lower and M.A.El_Sayed Chem.. Rev., 66(1966)199.

[Maiman, 1960] – T.H.Maiman, Nature 187(1960)493.

[Maroncelli and Fleming, 1987] – M.Maroncelli and G.R.Fleming, J.Chem.Phys.,

86(1987)6221.

146

[Maus et al., 1996] – M.Maus, W.Rettig, W.Bonafoux, R.Lapouyade,

J. Phys. Chem. A, 103(1999)3388.

[Maus and Rettig, 1997] – M.Maus and W.Rettig, Chem. Phys. 218(1997)151.

[McGlynn et al., 1972] – S.P.McGlynn, T.Azumi, M.Kinoshita “Molecular Spectroscopy of

the Triplet State” Prentice-Hall, Inc, Englewood Cliffs, New Jersey, 1969.

[Meites and Zuman, 1974] – L.Meites, P.Zuman, Electrochemical Data, John Wiley& Sons,

1974.

[Michl, Bonačić-Koutecký, 1990] – J.Michl, V. Bonačić-Koutecký, Electronic Aspects of

Organic Photochemistry , John Wiley and Sons: New York, 1990.

[Nakamura and Kanematsu].-.R.Nakamura and Y.Kanematsu, Rev. Scient. Instruments

75(2004)636.

[Naskręcki, 2000] – R.Naskręcki, “Femtosekundowa spektroskopia absorpcji przejściowej”,

Wydawnictwo Naukowe UAM, 2000.

[Okada et al, 1987] – T.Okada, N.Mataga, W.Baumann, J.Phys.Chem. 91(1987)760.

[Okada et al., 1999] – T.Okada, M.Uesugi, G.Köhler, K.Rechthaler, K.Rotkiewicz,

W.Rettig,, G.Grabner, Chem. Phys., 241(1999)327.

[Okamoto et al., 2000] – H.Okamoto, H.Inishi, Y.Nakamura,S.Kohtani, R.Nakagaki, Chem.

Phys. 260(2000)193.

[Piela, 2006] – L.Piela, “Idee Chemii Kwantowej” PWN, Warszawa 2006.

[Porter, 1950] – G.Porter, Proc.R.Soc.Ser.A 200(1950)284; G.Porter Disc.Faraday Soc.

9(1950)60.

[Pryor et al., 1999] – B.A.Pryor, P.M.Palmer, Yu.Chen, M.R.Topp, Chem. Phys. Letters,

299(1999)536.

[Rentzepis et al., 1984] –P.M.Rentzepis, V.E.Bondybey, J. Chem. Phys. , 80(1984)4727.

[Reynolds etal., 1995] - L.Reynolds, J.A.Gardecki, S.J.V.Frankland, M.L.Horng, and

M.Maroncelli, J.Phys.Chem. 100(1996)10337.

[Rosenthal et al., 1994] – S.J.Rosenthal, R.Jimenez, G.R. Fleming, P.V.Kumar,

M.Maroncelli, J.Mol.Liq. 60(1994)25.

[Rotkiewicz et al., 1976] – K.Rotkiewicz, Z.R.Grabowski, A.Krówczyński, W.Kühnle, J.

Lumin., 12/13(1976)877.

[Rullière et al., 1983] – C.Rullière, A.Declèmy, Ph. Pèe, Rev. Phys. Appl. 18(1983)347.

[Rullière et al., 1987] – C.Rullière, Z.R.Grabowski, J.Dobkowski, Chem. Phys. Lett.

137(1987)408.

[Sassara et al., 1999] – A.Sassara, G.Zerza, M.Cherqui, V.Ciulin, J.-D.Ganiere, B.Devaud, J.

[Schuddeboom at al,1992] – W Schuddeboom, S.A.Jonker, J.M.Warman, U.Leinhos,

W.Kühnle and K.A.Zacharasse, J.Phys.Chem., 96(1992)10809.

[Sepioł et al., 1998] – J.Sepioł, Y.Stepanenko, A.Vdovin, A.Mordziński, E.Vogel, J.Waluk,

Chem. Phys. Letters, 296(1998)549.

[Shelby et al., 1979] – R.M. Shelby, C.B. Harris, P.A. Cornelius, J. Chem. Phys.,

70(1979)34.

[Shida, 1988] T.Shida, “Electronic Absorption Spectra of Radical Ions”, Physical Sciences

Data 34, Elsevier, 1988.

[Smith and Meech, 2000] – N.A.Smith and S.R.Meech, J.Phys.Chem. A, 104(2000)4223.

Chem. Phys., 111(1999)689.

147

[Sobolewski and Domcke, 2002] – A.L.Sobolewski and W.Domcke, J. Phys. Chem. A

106(2002)4158.

[Son et al., 2001] – D.H.Son, P.Kambhampati, T.W.Kee, P.F.Barbara, J. Phys. Chem. A

105(2001)8269.

[Starukhin et al., 1998] – A.Starukhin, E.Vogel, J.Waluk J. Phys. Chem., 102(1998)1999.

[Stepanow, 2002] – A.G.Stepanow, M.T.Portella-Oberli, A.Sassara, M.Chergui, Chem. Phys.

Letters, 358(2002)516.

[Sudhold et al., 2000] – W.Sudholt, A.Staib, A.L.Sobolewski, W.Domcke, Phys. Chem.

Chem Phys., 2(2000)4341

[Svarcov, 2006] – PhD Thesis, Institute of Physical Chemistry, Warsaw, 2006.

[Szydłowska et al., 2003] – I.Szydłowska, A.Kyrychenko, J.Nowacki and J.Herbich, Phys.

Chem. Chem. Phys. 5(2003)1032.

[Szydłowska et al., 2005] – I.Szydłowska, J.Kubicki and J.Herbich, Photochem. Photobiol.

Sci. 4(2005)106.

[Takeda et al., 2000] – J.Takeda, K.Nakajima, S.Kurita, S.Tomimoto, S.Saito, T.Suemoto,

J.Lumin., 87-89(2000)927.

[Thulstrup et al., 1970] – E.W.Thulstrup, J.Michl, J.H.Eggers J. Phys. Chem.,

74(1970)3868.

[Uiberacker et al., 2007] – M.Uiberacker, Th.Uphues, M.Schultze, A.J.Verhoef,

V.Yakovlev, M.F.Kling. J.Rauschenberger, N.M.Kabachnik, H.Schroder, M.Lezius,

K.L.Kompa, H.-G.Muller, M.J.J.Vrakking, S.Hendel, U.Kleineberg, U. Heinzmann,

M.Drescher, and F.Krausz, Nature, 446(2007)627.

[Waluk and Michl, 1981] – J.Waluk and J.Michl, Anl. Chem. 53(1981)236.

[Waluk et al., 1991] – J.Waluk, M.Müller, P.Świderek, M.Köcher. E.Vogel, G.Hohlneicher,

J.Michl, J.Am.Chem.Soc., 113(1991)5511.

[Whittle et al., 1954] – E.Whittle, D.A.Dows, G.C.Pimentel, J. Chem. Phys. 22(1954)1943.

[Wiosna-Sałyga et al., 2006] – G.Wiosna-sałyga, J.Dobkowski, M.S.Mudadu,

I.Sazanovich,R.P.Thummel, J.Waluk, Chem.Phys.Letters 423(2006)288.

[Zachariasse et al., 1993] – K.A.Zachariasse, T.von der Haar, A.Hebecker, U.Leinhos,

W.Kühnle, Pure and Appl. Chem., 65(1993)1745.

[Meities and Zuman, 1974] – L.Meites, P.Zuman, Electrochemical Data, John Wiley & Sons,

1974.

[Zachariasse et al.,1997] – K.A.Zachariasse, M.Grobys and E.Tauer, Chem.Phys.Lett.,

274(1997)372.

[Zewail, 2000] – A.H.Zewail, J. Phys. Chem. A, 104(2000)5660.

[Zhang et al., 2003] – Y.Zhang, J.Jiang and M.Berg, J.Chem.Phys., 118[2003]7534.

[Zhu et al., 2005] – X.Zhu, R.A.Farrer, Q.Zhong, J.T.Fourkas, J.Phys.:Condens. Matter

17(2005)S4105.

[Văsak et al., 1982] – M.Văsak, J.W.Downing, B.L.Townsend, J.Michl,

J. Tetrahedron, 38(1982)1571.

[Yoshihara et al., 2003] – T.Yoshihara, V.A.Galievsky, S.I.Druzhinin, S.Saha and

K.A.Zachariasse, Photochem. Photobiol. Sci. 2(2003)342.

148

11 Experimental: compounds and solvents

1. Porphycene (PC) and 2,7,12,17-tetra-t-butyloporpycene (PCTT) was synthesized and

purified according to procedures described in:

E. Vogel, M. Köcher, H. Schmickler, J. Lex, Angew. Chem. Int Ed. Engl., 25(1986)257.

2. The syntheses and purification of 2,7,12,17-tetra-butyl-3,6,13,16-

dibenzo[cde,mno]porphycene (TTPC) followed the procedure described in:

E. Vogel, Pure Appl. Chem. 65(1993)143.

D. Donnerstag, PhD Dissertation, University of Cologne, Germany, 1993.

U. Hübsch, PhD Dissertation, University of Cologne, Germany, 1994.

3. N,N-dialkylamino derivatives of m-cyanopyridine (PAC, PEC, PC) were synthesized

according to procedure described in:

J. Dobkowski, J. Wójcik, W. Koźmiński, R. Kołos, J. Waluk, J. Michl, J. Am. Chem. Soc.

124(2002)2406.

J. Dobkowski, J. Michl, J. Waluk, Phys. Chem. Chem. Phys. 5(2003)1027.

4. p-Dimethylamino-benzaldehyde (DMABA) and p-Dimethylamino-acetophenone

(DMAA), product of (P.O.Ch.,Gliwice) and (Chemipan, Warsaw), respectively, were

purified by repeated recrystalizations and sublimations.

5. p-Diethylamino-benzaldehyde (DEABA) product of Aldrich was purified by column

chromatography.

6. p-Diethylamino-acetophenone (DEAA) was obtained by acetylation of diethylaniline,

applying the method described in: G.M. Kosolapoff, J. Am. Chem. Soc., 69(1947)1651.

7. The synthesis of 5-N,N-dimethylamino-1-indanone (DMI) and 5-Formyl-1-

methylindoline (FMI) was described in: Z.R. Grabowski, J. Dobkowski, W. Kühnle, J.

Mol. Structure 114(1984)93.

8. The synthesis of 3-methyl-N,N-dimethylaminobenzaldehyde (mDMABA) was described

in: R. Gawinecki, S. Andrzejak, A. Puchała, Org. Prep. Proced. Int. 30(1998)455.

9. 4’-(1-pyrenyl) benzonitrile (Py-BN) and 4’-(1-pyrenyl) acetophenone (Py-AC) were

synthesized and purified according to procedures described in:

J. Dobkowski, W. Rettig, J. Waluk, Phys. Chem. Chem. Phys. 4(2002)4334.

J. Dobkowski, J. Waluk, W. Yang, C. Rullière, W. Rettig, New J. Chem. 21(1997)429.

10. 4-Acetyl-4’-N,N-dimethylaminobiphenyl (ADAB) was synthesized and purified as

described previously:

J. Dobkowski, Z.R. Grabowski, J. Waluk, W. Kühnle, W. Rettig, C. Rullière, W. Yang,

J.Adamus, J. Gębicki, Proc. Indian Acad. Sci. (Chem. Sci.), 104(1992)143.

11. 4-Acetyl-2,2’-dimethyl-4’-N,N-dimethylaminobiphenyl (I) was synthesized in a 4-step

procedure. (1) Commercial 3-methyl-4-bromo-acetophenone was heated in nitrobenzene with

2-iodotoluene and catalytic Cu. (2) The resulting 4-acetyl-2,2’-dimethyl-biphenyl was

nitrated by HNO3 to 4-acetyl-2,2’-dimethyl-4’-nitro-biphenyl; (3) the latter was catalytically

149

reduced (H2/PtO2) to 4-acetyl-2,2’-dimethyl-4’-amino-biphenyl, and methylated with CH3I to

the desired 4-acetyl-2,2’-dimethyl-4’-N,N-dimethylamino-biphenyl

12. The synthesis of 4-N,N-dimethylamino-4’-cyanobiphenyl (II) was described in:

M. Maus, W. Rettig, D. Bonafoux, R. Lapouyade, , J. Phys. Chem. A 103(1999)3388.

The solvents used for the absorption and emission studies (Aldrich or Merck spectral

grade) were used without further purification except butyronitrile (BuCN) and 2-methyl-

tetrahydrofuran (MTHF). BuCN (Merck for synthesis) was repeatedly distilled over CaCl2

and P2O5. MTHF (Merck for synthesis) was repeatedly distilled over CaCl2.

150

12 Glossary of the abbreviations

T temperature

h Planck constant

IC internal conversion

ISC intersystem crossing

CT charge transfer state

S0 ground singlet state/

S1/Sn lowest/higher excited singlet state

T1/Tn lowest/higher excited triplet state

TICT twisted intramolecular charge transfer state

SIF solvent induced flattening

ao radius of Onsager cavity

g ground state dipole moment

s excited state dipole moment

dielectric constant

( ) fluorescence quantum yield (in chapter IV)

kf radiative rate constant in fluorescence

knr nonradiative rate constant

M electronic transition moment related to fluorescence

lifetime R

risetime

s solvent relaxation time

c(t) solvation response function

NMR Nuclear Magnetic Resonance

p-JT pseudo-Jahn-Teller interaction

DA molecules with donor and acceptor groups linked by single bond

MC center of mass

MCD magnetic circular dichroism

CCD charge coupled device

BA-derivatives derivatives of benzaldehyde and acetophenone

LVD light velocity dispersion

FWHM full width at half maximum

ACN acetonitrile

BuCN butyronitrile

THF tetrahydrofuran

MTHF 2-methyltetrahydrofuran

EtOH ethanol

MeOH methanol

The abbreviations of the compounds are shown in the Experimental part.

151

13 Acknowledgements

I would like to express my gratitude to Professors Anna and Zbigniew R. Grabowski for

introducing me into the fascinating world of photophysics, where one second means eternity.

My special thanks go to Professor Jacek Waluk for his constant readiness for discussions and

for inspiring suggestions.

I wish to express my sincere appreciation to dr. Jan Jasny for the construction of the advanced

optical equipment and many thanks to Andrzej Ardasiewicz for the construction of the

electronic devices and development of the computer software.

My thanks go also to Professor Wolfgang Rettig (Humboldt Universität zu Berlin, Germany)

and to dr. Claudie Rullière (Centre de Physique Molèculaire Optique et Hertzienne, Bordeaux,

France) for efficient and fruitful cooperation.

I wish to thank to my collaborators dr. Victor Galievsky and dr. Igor Sazanovich from

Institute of Molecular and Atomic Physics, Academy of Sciences of Belarus for their

enthusiastic activity in the field of time-resolved spectroscopy.

The fascination for the microworld may generate some emptiness in the family world. I would

like therefore to apologize to my children: Piotr, Jan and Kasia and thanks them for their

support.