Rate constants of the thermal cis-trans isomerization of azobenzene dyes in solvents, acetone/water mixtures, and in microheterogeneous surfactant solutions

Article abstract of DOI:10.1002/(SICI)1097-4601(1999)31:5<337::AID-KIN3>3.0.CO;2-E

Rate constants k_iso of the thermal cis-trans isomerization of four 4,4′-nitro-aminoazobenzenes with different amino groups have been determined in homogeneous aprotic solvents and polyglykol oligomers, primarily by means of conventional flash photolysis. The rate constants have been correlated with polarity (according to λ_max from UV/Vis absorption spectra of the trans isomers) and bulk viscosity of the solvents. Qualitative conclusions about the influence of varying concentrations of water with respect to polarity and hydrogen bonding on k_iso- and λ_max-values in acetone/water mixtures were derived. Based on these results the data from microheterogeneous solutions have been interpreted. In microheterogeneous water/surfactant solutions k_iso-values of selected azo dyes were strongly dependent on the concentrations of SDS, Triton X-100, C₁₂EO₈ in water, and varied with the composition of bicontinuous microemulsions of Igepal CA-520/heptane/water. The large spread of isomerization rate constants is in part due to varying microviscosity. Replacement of H₂O by D₂O in aqueous surfactant solutions produced surprisingly large kinetic solvent isotope effects.

Full text of DOI:10.1002/(SICI)1097-4601(1999)31:5<337::AID-KIN3>3.0.CO;2-E

Rate Constants of the
Thermal Cis-Trans
Isomerization of
Azobenzene Dyes in
Solvents, Acetone/Water
Mixtures, and in
Microheterogeneous
Surfactant Solutions
KATHRIN GILLE, HELMUT KNOLL, KONRAD QUITZSCH
Wilhelm-Ostwald-Institut fu¨r Physikalische und Theoretische Chemie der Universita¨t Leipzig, Linne´str. 2, 04103 Leipzig,
Germany
Received 22 October 1997; accepted 7 December 1998
Dedicated to Prof. Dr. Klaus Scherzer on occasion of his 65th birthday
ABSTRACT: Rate constants kiso of the thermal cis-trans isomerization of four 4,4’-nitro-ami-
noazobenzenes with different amino groups have been determined in homogeneous aprotic
solvents and polyglykol oligomers, primarily by means of conventional flash photolysis. The
rate constants have been correlated with polarity (according to ␭ from UV/Vis absorption
max
spectra of the trans isomers) and bulk viscosity of the solvents. Qualitative conclusions about
the influence of varying concentrations of water with respect to polarity and hydrogen bonding
on kiso- and ␭max-values in acetone/water mixtures were derived. Based on these results the
data from microheterogeneous solutions have been interpreted.
In microheterogeneous water/surfactant solutions kiso-values of selected azo dyes were
strongly dependent on the concentrations of SDS, Triton᭨ X-100, C12EO8 in water, and varied
with the composition of bicontinuous microemulsions of Igepal᭨ CA-520/ heptane/water. The
large spread of isomerization rate constants is in part due to varying microviscosity.
Replacement of H2O by D2O in aqueous surfactant solutions produced surprisingly large
kinetic solvent isotope effects. ᭧ 1999 John Wiley & Sons, Inc. Int J Chem Kinet 31: 337–350, 1999
Correspondence to: H. Knoll
᭧ 1999 John Wiley & Sons, Inc.
CCC 0538-8066/99/050337-14

338
GILLE, KNOLL, AND QUITZSCH
INTRODUCTION
activity parameter kiso according to the following
scheme:
Microheterogeneous fluid reaction media (micellar
and vesicular solutions, microemulsions) have been
considered as simple models for biological reaction
media [1] and are of practical importance in technical
processes [2]. Their influence on the overall reaction
rate of complex reactions compared with simple ho-
mogeneous solvents is due to concentration, separa-
tion, and diffusion effects within and between the mi-
crocompartments and across their extended interfaces.
At the molecular level, the interface between the head-
groups of surfactant or lipid molecules and water is of
special interest. Reactivity at this interface, which is
in fact an interfacial layer, is determined by the solvent
structure around the probe molecule (in particular hy-
drogen bonding), polarity and viscosity.
Scheme I
There is a large number of spectroscopic studies in
which probe molecules are added to characterize par-
ticular interfacial properties, for example, hydrogen
bonding and polarity [3,4] and viscosity [5–7].
Reactivity parameters for the characterization of the
local environment in microheterogeneous reaction
media, however, have been studied less extensively
[8–12].
Interpretation of rate constants of elementary re-
action steps in microheterogeneous media, based on
properties of the microenvironment determined by
probe molecules, may be difficult if probe and reactant
molecules are different, and may have a different mi-
croenviroment. Therefore, reactivity parameters of el-
ementary reactions of molecules, which also exhibit
spectroscopic probe properties with respect to the mi-
croenvironment in water-surfactant interface layers,
are of interest.
Schanze, Whitten et al. [14–18] studied the rates
of thermal cis-trans isomerization of different 4,4’-ni-
tro-aminoazobenzenes in simple solvents and different
solutions of surfactants and lipids, in particular in re-
versed micelles of AOT, and mainly with 4,4’-nitro-
diethylaminoazobenzene. These authors obtained a
good correlation of ⌬G#-values of the cis-trans isom-
erization with the solvent polarity, and observed a rate
increasing effect of hydrogen bond donating solvents.
Steiner et al. [19] studied the thermal cis-trans isom-
erization of solvatochromic merocyanines and re-
ported similar linear correlations between solvent po-
larity and isomerization rate constants.
In this study we repeated some experiments with
4,4’-nitro-diethylaminoazobenzene to show that our
measurements are reliable, and we focused on the fol-
lowing new aspects:
The thermally stable trans isomers of donor-accep-
tor substituted azobenzene dyes show remarkable sol-
vatochromy, and they can be easily transformed into
the thermally unstable cis form by light. AM1 calcu-
lations show that the dipole moments of their cis iso-
mers are about 6 D, similar to 8 D of the trans isomers
[13]. Therefore, we assume that cis azobenzene dye
molecules formed by the flash will remain essentially
at the same sites as their trans form precursors. From
UV/Vis spectroscopy it follows that this is the water-
surfactant interface layer in most cases. So differences
of the average microenvironment can be monitored by
means of the UV/Vis absorption of the stable trans
isomers. Moreover, the rate of the thermal cis-trans
back isomerization is highly dependent on solvent
properties in simple homogeneous solvents, and there-
fore, an appropriate model reaction for probing the
water-surfactant interface layers by means of the re-
1. We have chosen three azo dyes which differ at
the amino group, to vary its electron donor ca-
pability, hydrogen bond formation ability, and
rotor volume (Scheme II). By comparing the
isomerization rate constants of different azo
dyes in one solvent and of one azo dye in dif-
ferent solvents, conclusions can be drawn about
the sensitivity of the isomerization rate constant
on different properties of the medium.
2. Based on these results, the data from microhet-
erogeneous solutions of mainly nonionic surfac-
tants were evaluated with respect to the micro-
environment of the probe molecules.
3. From the comparison of the rate constants of the
thermal isomerization in H2O/- and D2O/surfac-
tant solutions, respectively, conclusions are
drawn on the accessibility of water to the react-

ISOMERIZATION OF AZOBENZENE DYES IN SOLVENTS
339
Scheme II
ing probe molecules and the importance of hy-
drogen bonding.
(igepal CA-520, Aldrich 98%), dodecyl-octaethylen-
glycolether (C12EO8, Nikko). Water from a Milli-Q
Plus apparatus (Millipore) and D2O (Laborchemie
Berlin, 99.7%) freshly distilled over KMnO4 were
used.
EXPERIMENTAL
Apparatus
Materials
UV/Vis spectra and time-dependent measurements in
the minutes range were carried out with a Lambda 2-
4,4’-Nitro-aminoazobenzene (I, Aldrich), 4,4’-nitro-
(ethyl-2-cyanoethyl)aminoazobenzene (III, Aldrich),
have been used as received. 4,4’-Nitro-diethylamino-
azobenzene (II) has been prepared according to [20].
4,4’-Nitro-anilinoazobenzene (IV, Sigma) was dis-
solved in acetone and the solution decanted from the
salt residue.
The following solvents were used as received:
acetone (Merck, p.a. 0.01% water), acetonitrile, n-hep-
tane (heptane), N,N-dimethylformamide (DMF, all
Merck UVASOL), propionitrile and butyronitrile (Al-
drich 99%), cyclohexane (Fluka 99.5%), Poly(-ethyl-
englycol) (PEO 600, Aldrich, average molecular
weight 600), Poly(propylenglyol) (PPO 425, PPO
2000, Aldrich, average molecular weights 425 and
2000, respectively).
spectrometer (Perkin-Elmer, ⌬␭
was estimated to
max
Ϯ2 nm), irradiation for cis-trans isomerization with
reaction times of minutes was performed by means of
an argon ion laser Innova 305 (Coherent). For reaction
times of microseconds to seconds a home-made mi-
crosecond flash-photolysis apparatus (FWHM 25 ␮s,
up to 180 J/flash) equipped with a digital storage os-
cilloscope PM3533 (Fluke), and 5 or 10 cm cuvettes
(water jacket, bath thermostat) was used. Bulk viscos-
ities were taken from the literature [21] or measured
for the polyglycol oligomers with a LS100 rheometer
(Paar-Physica).
Procedures
3
The following surfactants were used as received:
sodiumdodecyl sulfate (SDS, Serva, p.a. recrystal.
twice), t-octylphenyl-nona/decaethyleneglycolether/
From stock solutions of the azo dyes (10Ϫ3 mol dmϪ )
in acetone, the desired aliquots were given to 10 ml
volumetric flasks and the solvent evaporated. The
samples were prepared from solvents or stock solu-
tions of the surfactants, bicontinuous microemulsions,
᭨
Triton X-100 (triton X-100, Merck), p-iso-
᭨
octylphenyl-pentaethylenglycolether/Igepal CA-520

340
GILLE, KNOLL, AND QUITZSCH
and water, respectively. The volumetric flasks were
agitated in an ultrasonic bath and were allowed to let
stand overnight. The concentrations of the azo dyes
were between 2 and 5 ϫ 10Ϫ6 mol dmϪ3 for the kinetic
measurements, respectively.
The irradiation of the dye solutions for seconds by
the argon-ion laser or for microseconds by the flash
resulted in a bleaching of the ␲␲*-CT band and trans-
formation of most of the stable trans form into the cis
isomer. The thermal back reaction into the trans form
occurred according to a first-order rate law in all pure
solvents, solvent/water mixtures, and surfactant solu-
tions, except for some cases with dyes I and IV in
bicontinuous microemulsions. Isomerization was fol-
lowed by means of the PMT signal and evaluated ac-
cording to equation (1) from the change in absorbance
during 2 to 3 half-lifes (Fig. 1).
pH Ͻ 5.2 when small amounts of HCl were added
stepwise. As water/ surfactant solutions were studied
at pH Ͼ 5.6, we assume that the rate constants refer
to unprotonated cis isomers of the azo dyes.
It was confirmed by NMR experiments that –NH,
–NH2, and –OH protons of the azo dyes and surfac-
tants, respectively, were immediately replaced by deu-
terium in D2O as a solvent.
RESULTS AND DISCUSSION
Isomerization Rate Constants in
Homogeneous Solvents and Acetone/Water
Mixtures
Understanding the influence of the microenvironment
in water surfactant interface layers on kiso requires an
understanding of the influence of the microenviron-
ment in solvents and solvent mixtures. Therefore, we
began with experiments in different aprotic solvents,
highly viscous polyglycols PEO(600), PPO(425),
PPO(2000), and mixtures of acetone with H2O. In Ta-
ble I ␭max-values of the trans form and the correspond-
ing kiso-values in pure solvents are given together with
available data from the literature. There is satisfactory
agreement between our data and literature values
where available.
ln {(A0 Ϫ Aoo)/(At Ϫ Aoo)} ϭ kisot
(1)
The sequence between photochemical trans : cis and
thermal cis : trans isomerization was completely re-
versible for several cycles which were carried out with
one sample.
For repeated experiments with the same solution
the standard deviation of kiso was found to be less than
Ϯ5%, and for experiments with independently pre-
pared solutions, Ϯ15%.
In acetone/water mixtures (1:1, v:v) an increase of
the isomerization rate constants was observed for
The electronic effects of the substituents at the
amino group on ␭max can be separated from other fac-
Figure 1a UV/VIS absorption spectra before and after irradiation (90 s with the 514.5 nm line) of
dye IV in cyclohexane in an 1 cm cuvette. 1 ϭ before irradiation and at infinite time after irradiation,
2 ϭ 1 min, 3 ϭ 4 min, 4 ϭ 9 min after irradiation.

ISOMERIZATION OF AZOBENZENE DYES IN SOLVENTS
341
Figure 1b Oscilloscope trace before and after the microsecond flash (observation wavelength 493
Ϫ
Ϫ3
nm) on dye IV in 4 ϫ 10 4 mol dm C12EO8; experimental data, and fit of the first-order rate law,
respectively.
tors by consideration of the data determined in the
unpolar aprotic solvents heptane and cyclohexane. The
order of ␭max-values for the four dyes in both solvents
is I Ͻ III Ͻ IV Ͻ II. This order reflects the increasing
ability of the substituents at the amino group to sta-
bilize the electronically excited ␲␲* charge transfer
state. This increasing electron donor ability of the sub-
stitutents at the amino nitrogen should also increas-
ingly stabilize the bipolar transition state of the ther-
mal cis-trans isomerization [15]. The isomerization
rate constants kiso increase in the order IV Ͻ III Ͻ
II Ͻ I, and there is a remarkable inversion in the order
of kiso between I and IV compared with the order of
␭max. The kiso order reflects the different steric require-
ments for rotation of the aminobenzene and the anili-
nobenzene group, respectively. Whereas, for the ami-
Table I UV/VIS Absorption Maxima ␭max and Cis-trans Isomerization Rate Constants kiso at 298 K of Azo Dyes in
Aprotic Solvents and Neat Polyglycol Oligomers of Viscosity ␩
I
II
III
IV
1
1
1
1
Solvent
Heptane
␩/mPa s
␭
max/nm
kiso/sϪ
␭
max/nm
kiso/sϪ
␭
max/nm
kiso/sϪ
␭
max/nm
kiso/sϪ
3
3
3
3
3
3
0.386
400
8.5 ϫ 10Ϫ
453
451
457
4.9 ϫ 10Ϫ
7.0 ϫ 10Ϫ
4.7 ϫ 10Ϫ
5.0 ϫ 10Ϫ
430
2.5 ϫ 10Ϫ
438
1.4 ϫ 10Ϫ
3a
3
3
Cyclohexane
Acetone
0.896
0.310
0.345
408
438
429
8.5 ϫ 10Ϫ
430
465
462
2.4 ϫ 10Ϫ
440
435
0.9 ϫ 10Ϫ
0.6 ϫ 10Ϫ
1.6 ϫ 10Ϫ
0.19
0.12b
0.14c
2.7 ϫ 10Ϫ
2.0 ϫ 10Ϫ
0.33
3a
3b
3c
1.35
1.56
490
493
17
18a
19d
52
58a
0.58
1.55
461
461
2
Acetonitrile
490
491
490
491
502
501
490
483
455
456
458
461
484
486
480
476
2b
Propionitrile
Butyronitrile
DMF
PEO 600
PPO 425
PPO 2000
0.413
0.565
0.924
141
76
335
430
432
457
455
446
443
1.00
1.14
78
29
21
461
465
479
480
462
456
1.00
2.44
8.1
2.21
0.29
0.12
0.20
4.9
1.23
0.17
1.3 ϫ 102
42
46
5.6
3.60
0.59
1.71
0.12
a ref. [18], b ref. [22], c ref. [23], d ref. [24]

342
GILLE, KNOLL, AND QUITZSCH
Figure 2 Plot of free activation energy of cis-trans isomerization ⌬G#298 K vs. Emax (from ␭max) in
aprotic solvents (open symbols), PEO, and PPO’s (solid symbols) for II (triangles) and IV (squares),
regression lines according to eq. (2) including only aprotic solvents.
nobenzene group (in I), a rotor volume of about 56 ml
parameter reflecting kiso at 298 K. Linear plots of
⌬G#298 K, vs. Emax, according to equation (2)
1
molϪ can be estimated, for the corresponding rotors
(in III, IV, and II) rotor volumes of about 100 ml
molϪ1 result [25]. The order of the kiso-values of I and
IV is a strong indication that viscosity influences the
cis-trans isomerization of azo compounds, as recently
confirmed for the isomerization at carbon-carbon dou-
ble bonds in bis-oxonols [26].
Before discussing the viscosity effect, the influence
of the solvent polarity on the isomerization rates of
particular azo dyes will be described. Several authors
[14–18] have shown that polarity around donor-ac-
ceptor-substituted trans azo dyes correlates with their
⌬G#
ϭ a ϩ b Emax
(2)
298 K
of azo dyes, without protic hydrogen atoms in aprotic
solvents, allowed separation from H-bonding effects.
Azo dyes II and III fulfill this requirement. The re-
gression parameters a and b were calculated also for I
and IV, neglecting possible hydrogen bonding be-
tween the hydrogen atoms at the amino groups of I
and IV and hydrogen bond acceptor groups in the
aprotic solvents. In Figure 2 the data and regression
lines of II and IV are shown as examples. Slopes b
between 0.6 and 1.1 were determined (Table II). A
slope b of about one suggests that there is a similar
degree of charge transfer in the transition state for the
isomerization reaction and in the Franck-Condon ex-
cited state after light absorption, as found earlier for
dye II [15].
Emax-values calculated from the maximum absorbance
wavelength ␭ (Emax ϭ h · c/␭max, where h is the
max
Planck constant and c is the speed of light). Therefore
Emax is a dye specific polarity parameter. We calcu-
lated the free activation energy of isomerization
⌬G#
ϭ 298 R{ln(kB298/h · kiso)}, where kB is the
298 K
Boltzmann constant and R the gas constant, as energy
Table II Parameters of Equations (2) and (3) Reflecting the Polarity and Viscosity Effects on Free Activation Energy
⌬G#298, and 1n kiso, Respectively, of Cis-trans Isomerization of the Azo Dyes
a
Azo Dye
a
b
C
␣
␤
I
II
III
IV
Ϫ101 Ϯ 16
Ϫ193 Ϯ 17
Ϫ129 Ϯ 12
Ϫ142 Ϯ 26
0.63 Ϯ 0.06
0.85 Ϯ 0.10
0.78 Ϯ 0.05
0.85 Ϯ 0.10
Ϫ5.60 Ϯ 0.51
Ϫ5.76 Ϯ 0.36
Ϫ6.06 Ϯ 0.35
Ϫ7.04 Ϯ 0.49
0.15 Ϯ 0.10
0.32 Ϯ 0.06
0.24 Ϯ 0.07
0.52 Ϯ 0.12
Ϫ0.26 Ϯ 0.02
Ϫ0.43 Ϯ 0.02
Ϫ0.30 Ϯ 0.02
Ϫ0.35 Ϯ 0.03
a Ϫ␤ ϫ RT corresponds to parameter b.

ISOMERIZATION OF AZOBENZENE DYES IN SOLVENTS
343
With increasing donor ability of the amino group,
bipolar transition state of cis-trans isomerization and
that is, increasing ␭ of the dyes in the unpolar sol-
the Frank-Condon state, and increases both kiso and
max
vents heptane and cyclohexane, an increasing slope b
was obtained. It corresponds to increasing sensitivity
of the isomerization rates of the various dyes to solvent
polarity. From these aprotic solvents with bulk vis-
cosities 0.3 Ͻ ␩ Ͻ 1 mPa · s the dependence of kiso
on viscosity cannot be derived.
␭
(and decreases Emax). Hydrogen bonding is like-
max
wise possible at the lone electron pair of the amino
group [27], which decreases both kiso and ␭max. Hydro-
gen bonding between hydrogen atoms at the amino
group in I and IV with hydrogen bond acceptor groups
of solvents, however, increases the donor capability of
the amino group. The balance of these effects results
in an increase of the isomerization rate [15] in solu-
tions containing water. We studied UV/Vis absorption
and cis-trans isomerization in acetone/water mixtures
in order to derive qualitatively the hydrogen bonding
To obtain kiso-values from solvents with high bulk
viscosity we studied the isomerization in pure
polyglycol oligomers PEO(425), PPO(600), and
PPO(2000). The isomerization rates are lower and the
⌬G#298 K higher than expected from the ⌬G#298 K/Emax
-
correlation with the aprotic solvents of low viscosity
(Fig. 2). The decrease in rate at higher viscosity is
obvious in these solvents, although the rate-increasing
hydrogen-bonding effect of the hydroxylic end groups
in PEO/PPOs cannot be considered separately. There-
fore, only a lower limit of the influence of viscosity
can be quantified by ␣, equation (3), which is similar
to an equation used by Benniston and Harriman [26].
It correlates ln kiso with bulk viscosity ␩ and Emax as a
measure of polarity. The Emax-value for each dye is
highest in heptane and it was taken as a reference for
zero polarity.
effect of water on ␭ and kiso. Differences between
the azo dyes in hydrogen bonding result from differ-
ences at their amino groups.
max
Acetone with added water (4:1 v:v) gave an in-
crease in ␭ and kiso compared to pure acetone due
max
to increasing solvent polarity (Table III). However, in-
creasing occupation of the amino nitrogen lone elec-
tron pair by hydrogen bonding with water leads to a
decrease of ␭ and kiso. As an overall effect, a max-
max
imum in ␭
(at acetone water 4:1 v:v) and kiso (at
max
acetone water 1:4 v:v) was observed for dye I. There-
fore, we conclude that an increase of polarity and hy-
drogen bonding may balance in ␭max and kiso. ␭max re-
flects the electronic effects of the environment of the
ln kiso ϭ C Ϫ ␣ ln ␩ ϩ ␤(Emax Ϫ Emax0
)
(3)
dye on a time average. Given the drop of ␭ of 19
max
The calculated intercept C has the meaning of C ϭ (ln
B Ϫ EA/RT), whereas B is the viscosity-independent
preexponential factor, and EA is the polarity-indepen-
dent Arrhenius activation energy. The ␣-values are
empirical parameters approximately dependent on vis-
cosity. They increase in the order I, III and II, IV
(Table II), and show the difference in the rotor vol-
umes between I and the other dyes. Despite similar
rotor volumes, the ␣-values show also a difference be-
tween dyes III, II, and IV, respectively, probably due
to the different molecular shapes of the amino groups.
All azo dyes of this study may accept hydrogen
bonding at the nitro group which stabilizes both the
nm on changing from acetone/water 1:4 to pure water,
a large decrease of kiso would be expected. However,
only a small decrease of kiso was observed indicating
an additional rate increasing factor. This might be a
favorable cooperative dynamics of hydrogen bond
forming (nitro group) and breaking (amino nitrogen
lone pair) with water molecules.
Only I is sufficiently soluble in pure water to cover
the entire range of mixtures with water. The other dyes
could only be studied in acetone/water mixtures Ն1:
4. They are less capable of hydrogen bonding at the
amino nitrogen for steric reasons but seem to behave
similarly, as shown for II in Table III. With acetone/
Table III UV/VIS Absorption Maxima ␭max and Cis-trans Isomerization Rate Constants kiso at 298 K in Acetone/Water
Mixtures (v:v)
I
II
Ϫ
1
Ϫ1
Solvent
Acetone
␭
max/nm
kiso/s
␭
max/nm
kiso/s
438
1.4
490
17
Acetone/water
4:1
1:1
1:4
1:9
Water
448
441
441
435
422
73
502
507
508
0.88 ϫ 103
2.4 ϫ 103
1.1 ϫ 103
0.64 ϫ 103
2.1 ϫ 103
1.7 ϫ 103
1.6 ϫ 103

344
GILLE, KNOLL, AND QUITZSCH
water of 1:1 a maximum of kiso and a plateau-value of
be due to viscosity changes around the reacting probe
molecule. An increase of the aggregation number of
SDS micelles between 63 and 89 occurs at concentra-
␭max was observed.
These results from homogeneous solvents and ac-
3
etone/water mixtures show that trends of the influence
of polarity, viscosity, and of water in the microenvi-
ronment of the isomerizing azo dye molecules in fluid
media can be interpreted.
tions between 0.02 and 0.2 mol dmϪ [28,30]. This
can be expected to lead to a denser packing of the
surfactant molecules and increasing microviscosity in
the micelles, and therefore to a decrease of the isom-
erization rate of IV. The decrease by factor 2 in kiso
3
between 0.04 and 0.05 mol dmϪ in H2O is in accord
with a result from the literature, where in the same
concentration range cis-trans isomerization of 4,4’-ni-
tro-dimethylaminoazobenzene showed the strongest
decrease of kiso on stepwise increasing SDS concen-
trations [31].
Changes of the Microenvironment for Azo
Dyes at Different Surfactant
Concentrations
Anionic (SDS). By increasing [SDS] from 0.01 mol
3
3
3
dmϪ (cmc ϭ 0.008 mol dmϪ [28]) to 0.5 mol dmϪ
For the solvent kinetic isotope effect, kiso(H2O)/
kiso(D2O) ϭ 2.1 Ϯ 0.1 was determined at [SDS] Ͻ
0.040 mol dmϪ3 and 1.0 Ϯ 0.2 at [SDS] Ͼ 0.040 mol
in H2O, and in D2O, respectively, a considerable de-
crease of the rate constant of IV was observed (Fig.
3). Unfortunately, kiso could not be determined in pure
water. Therefore, it cannot be excluded that the largest
values of kiso at the lowest SDS concentrations contain
some contribution of high rates from isomerization in
water due to incomplete binding of the dye to micelles
[29].
3
dmϪ . The large kinetic solvent isotope effect at
3
[SDS] Յ 0.04 mol dmϪ indicates a better access of
water to the microenvironment of the probe molecules
at lower SDS concentrations. A possible dynamic
H-bonding effect decreases with decreasing access of
water to the dye molecules. This might also contribute
3
␭max ϭ 492 Ϯ 1 nm was constant in the entire con-
to the decrease of kiso at [SDS] Ն 0.04 mol dmϪ
centration range. The constant ␭max-value indicates
constant electronic effects around the azo dye due to
constant or compensating “static” effects by micro-
polarity and H-bonding. Therefore, the large concen-
tration effect on the isomerization rate might, in part,
(Fig. 3).
Within the same concentration range of SDS, azo
dye I behaves similarly to IV. At constant ␭
445 Ϯ 1 nm, kiso decreases less for I than IV, from
1344 (0.01 mol dmϪ ) to 232 sϪ (0.2 mol dmϪ ),
ϭ
max
3
1
3
Figure 3 Plots of cis-trans isomerization rate constants kiso of IV at 298 K vs. SDS concentration
in H2O (full squares), and D2O (open squares), respectively. Lines are guides for the eye.

ISOMERIZATION OF AZOBENZENE DYES IN SOLVENTS
345
whereas kiso(H2O)/kiso(D2O) changes from 1.6 Ϯ 0.2
in the lower concentration range to 1.3 Ϯ 0.2 in the
higher concentration range. Unlike dye IV, dye I is
water soluble, and incomplete binding to the micelles
occurs in the lower concentration range. Therefore,
only data from the higher concentration range can be
compared with the results of dye IV. The ␭max-value
corresponds to an acetone/water mixture of about 2:1.
Moreover, the isotope effect and kiso-values indicate
the presence of water for I in all solutions and a so-
lubilization site of I at the water side of the water
surfactant interface.
concentration and increasing aggregation numbers, as
in the case of SDS, although no such data are available
for the low triton X-100 concentrations of this study.
Again, the most interesting fact is a clear break in the
values for the isotope effect, from about (5 Ϯ 1) at
[triton] Ͻ 2 cmc to about (2.6 Ϯ 0.7) at [triton] Ն 2
cmc. This might reflect a more rigid solvent structure/
higher viscosity of micellar aggregates in D2O and a
persisting solubilization site of IV within the water
containing polyethyleneoxide headgroup shell.
D2O is the usual solvent in studies of aggregation
behavior of amphiphiles by means of NMR [37] and
neutron scattering. However, experimental effects
caused by the exchange of H2O by D2O in surfactant/
water and lipid/water mixtures have been studied only
occasionally. Some general conclusions from such in-
vestigations are that hydrophobic and electrostatic in-
teractions between hosts (e.g., cyclodextrins) and
guests are stronger in D2O [38], and that the increase
of the aggregation number and decrease of cmc in D2O
[39] compared to H2O, phase diagrams of mixtures
from H2O and D2O, respectively, are not very different
[40]. Our results show that small differences in struc-
ture and equilibria properties may be amplified in ki-
netic parameters. Differences in H2O and D2O solu-
tions, respectively, are also evident in the slightly
better solubility of the azo dyes in D2O solutions com-
pared to H2O solutions at the same lowest triton X-
100 concentrations used.
Nonionic/triton X-100. Due to the better solubility of
the azo dyes by triton X-100 compared to SDS mi-
celles, solubilization and reduction of kiso begins below
the cmc for IV (Fig. 4). This decrease can only be
explained by assuming the existence of premicellar ag-
gregates. Recent papers give indications of premicellar
aggregates for nonionic surfactants [1,32]. It is an
open question if such premicellar aggregates exist
without the probe molecules or if these molecules in-
duce aggregations [33–36].
Although the ␭max-values of 493 Ϯ 1 nm were con-
stant between 0.5 and 20 times the cmc, indicating
constant (or balancing) micropolarity and extent of hy-
drogen bonding, there was a decrease of one order of
magnitude in kiso. Again, this might be due to an in-
crease of microviscosity with increasing surfactant
Figure 4 Plots of cis-trans isomerization rate constants kiso of IV at 298 K vs. triton X-100 con-
centration in H2O (full squares), and D2O (open squares), respectively. Lines are guides for the eye.

346
GILLE, KNOLL, AND QUITZSCH
Table IV UV/VIS Absorption Maxima ␭max and Cis-trans Isomerization Rate Constants kiso at 298 K of I in
Dependence on the Triton X-100 Concentration
H2O
D2O
[Triton X-100]/
Ϫ
Ϫ3
Ϫ1
Ϫ1
10 4 mol dm
␭max/nm
kiso/103 s
␭max/nm
kiso/103 s
kH2O/kD2O
1.0
1.2
2.0
5.0
10
426
427
436
456
460
460
462
3.0
2.9
3.0
2.4
2.1
1.7
1.3
426
428
433
453
453
1.2
1.2
1.5
1.1
1.2
2.5
2.4
2.0
2.2
1.7
25
50
Table IV gives the data for I in triton X-100, which
show quite different trends compared with those of IV.
centrations. The small change of kiso with [triton X-
100] is in accord with the lower sensitivity of kiso of I
to changes in microviscosity, so that the microviscos-
ity does not change as much at the solubilization site
of I in triton X-100 solutions compared to the micro-
␭
max
is 426 nm below the cmc, which indicates an
environment similar to acetone/water 1:9 (see Table
II), but kiso is twice as high as in pure water, obviously
due to interactions with premicellar triton X-100 ag-
environment of IV. The change of ␭ and kiso of I
max
4
3
gregates. ␭ increases with increasing triton X-100
between 2 and 5 ϫ 10Ϫ mol dmϪ of triton X-100,
however, reflects clearly the structure changes caused
by micelle formation.
max
concentration, in part probably because of more com-
plete binding to micelles. On the other hand, the isom-
erization rate constant decreases only by a factor of
two. The increase in ␭max might be due to a change in
the orientation of the amino group in I from the water
side with a high chance of hydrogen bonding (accep-
tor) onto the micelle interior at higher surfactant con-
C12EO8. The dependence of the isomerization rate
constant of IV on the concentration of C12EO8 above
the cmc is shown in Figure 5. There is a similar de-
creasing kinetic solvent isotope effect with increasing
Figure 5 Plots of cis-trans isomerization rate constants kiso of IV at 298 K in dependence of C12EO8
concentration in H2O (full squares), and D2O (open squares), respectively. Lines are guides for the
eye.

ISOMERIZATION OF AZOBENZENE DYES IN SOLVENTS
347
surfactant concentration as in triton X-100 solutions.
The constant ␭max-value of 493 nm of IV in both
C12EO8 and triton X-100, and the absence of a vanish-
ing kinetic solvent isotope effect at high surfactant
concentrations indicate a similar solubilization site in
the water-surfactant interface layers on increasing con-
centrations of both surfactants. The decrease in kiso in
the concentration range between the cmc (for triton X-
from igepal CA-520, heptane and water, based on a
published phase diagram [45]. Igepal CA-520 struc-
ture is similar to that of triton X-100, but has an un-
branched hydrocarbon tail. The experimental results
are summarized in Table V.
In all four microemulsions studied, azo dye I shows
␭
max
ϭ 437 nm, which indicates a less polar average
solubilization site than in micellar triton X-100 solu-
tions (Ϸ 460 nm). Considerable deviations from the
first-order rate law according to eq. (1) were obtained.
From absorbance/time curves (Aoo Ϫ At)/(Aoo Ϫ A0)
was calculated, which corresponds to the concentra-
tion ratio of cis isomers at times t and zero, respec-
tively. Its dependence on time is shown in Figure 6.
The curves could not be fitted to a biexponential rate
law, but reasonable fits according to eq. (1) were ob-
tained with estimated Aoo-values for the first part of
the curves. These kiso-values are given in Table V.
They are an order of magnitude smaller than those in
triton X-100 solutions (See Table IV).
4
100 at 2.5 ϫ 10Ϫ mol dmϪ3 [41], for C12EO8 at
4
3
1.0 ϫ 10Ϫ mol dmϪ [42]) and 20 cmc for both sur-
factants is about sixfold for triton X-100, and thirty-
fold for C12EO8 (Figs. 4 and 5). The phenyl group in
triton X-100 may anchor IV in the micelles fixing the
anilino group of the dye in a microenvironment of
higher viscosity at lower surfactant concentrations
compared to C12EO8.
Cis-trans Isomerization in Bicontinuous
Microemulsions
Bicontinuous microemulsions have a sponge-like mi-
crostructure between swollen micelles and reverse mi-
celles. They have been used as reaction media only
recently [2,43,44]. To the best of our knowledge, cis-
trans isomerization of azo dyes has not yet been stud-
ied in such surfactant media. Different from micellar
solutions, beside the extended water/surfactant inter-
face layer a large volume of unpolar oil domains may
dissolve the dye molecules and distributions between
different solubilization sites must be considered. We
prepared transparent bicontinuous microemulsions of
different compositions within the one-phase region
The (Aoo Ϫ At)/(Aoo Ϫ A0) ϭ f(t) curves change
with the wavelength range of exciting light ␭exc for the
trans-cis isomerization and observation wavelength
␭
obs
for the thermal cis-trans isomerization of I (Fig.
6). With a blue flash filter (350 Ͻ ␭ Ͻ 480 nm)
exc
more cis I can be prepared in heptane domains than
with a green flash filter (␭exc Ͼ 430 nm). With ␭obs ϭ
400 . . . 420 nm mainly low polar domains were
monitored. About 50% “slow” isomerization was ob-
served using the blue flash filter, but only about 15%
(␭ ϭ 420 nm) was observed with the green flash
obs
filter. Nethertheless, this “slow” isomerization is com-
Table V UV/VIS Absorption Maxima ␭max and Cis-trans Isomerization Rate Constants kiso at 298 K of Azo Dyes in
Bicontinuous Microemulsions A, B, C, and D from Igepal CA-520/Heptane/Watera
H2O
D2O
Azo Dye/
Microemulsion
Ϫ1
Ϫ1
␭max/nm
kiso/s
␭max/nm
kiso/s
kH2O/kD2O
I
A
B
C
D
437
437
437
437
139
141
148
206
434
109
1.3
II
IV
A
B
C
D
463
462
460
480
172
165
171
505
464
482
105
1.6
1.4
A
B
C
D
485
484
484
484
2.13
2.01
2.45
5.5
1.55
a Microemulsion compositions: igepal CA-520/heptane/water in weight percent: A (20 : 40 : 40), B (20 : 35 : 45), C (20: 50: 30), D
(25 : 30 : 45) with H2O, and A (19.2: 38.3 : 42.5) with D2O, to have equal molar ratios in A with both H2O and D2O. Bulk viscosities at 298
K in mPas were A, 34; B, 34; C, 18; and D, 48. Small amounts of 0.1 mol dmϪ3 NaOH were added to adjust pH Ͼ 6.0.

348
GILLE, KNOLL, AND QUITZSCH
Figure 6 Plots of (Aoo Ϫ At)/(Aoo Ϫ A0) vs. time for three dyes in microemulsion A. Dyes II and
IV (␭exc Ͼ 430 nm) show curves independent on ␭obs and ␭exc. Data determined at ␭obs ϭ 400, 435,
and 470 nm for I were obtained with a blue flash filter (350 Ͻ ␭ Ͻ 480 nm).
exc
plete within a few seconds and therefore much faster
than would be expected for cis-trans isomerization in
pure heptane (see Table I). Therefore, we assume that
the second branch of the (Aoo Ϫ At)/ (Aoo Ϫ- A0) ϭ
f(t) curve originates from the isomerization rate deter-
mining redistibution of cis isomers into the polar sur-
There is an initial increase of cis IV at 450 nm Ͻ
obs Ͻ 525 nm until about 30 ms after the flash, which
␭
shows that a small part of the cis isomer was redisti-
buted before isomerization (Fig. 6). Decreasing values
(Aoo Ϫ At)/(Aoo Ϫ A0) Յ 1 at reaction times Ͼ 0.1 s
were used for calculation of first-order rate constants
(Table V). A dependence of rate constants on different
flash filters and on observation wavelength was not
factant water interface. By increasing ␭ to 435 and
obs
475 nm (blue flash filter), more polar domains were
monitored and the relative contribution of the slow
component decreased.
II is the most hydrophobic dye studied and mainly
solubilized in a low polar environment as ␭max Ϸ 460
nm in microemulsions A, B, and C shows, compared
found. ␭ Ϸ 484 nm reflects a moderate polar mi-
max
croenvironment and rate constants between 2.0 and 5.5
1
sϪ in the microemulsions with H2O are comparable
1
3
to kiso ϭ 8 sϪ in the 5 mmol dmϪ solution of triton
X-100. Dye IV has the lowest isomerization rate con-
stants of all dyes studied in bicontinuous microemul-
sions because it has the strongest sensitivity to mi-
croviscosity and because of possible fixation by the
anilino group at the phenyl group of igepal CA-520.
However, the rate constants for IV in microheteroge-
neous bicontinuous microemulsions do not correlate
with their bulk viscosities (Table V). In the case of IV,
redistribution occurs before most of the cis isomer can
isomerize due to the slow isomerization rate of IV.
to ␭
ϭ 505 nm in micellar triton X-100 solution
max
1
(kiso Ϸ 4 ⅐ 103 sϪ ). Only microemulsion D, with the
highest content of surfactant ␭max ϭ 480 nm, indicates
a more polar microenvironment. All isomerization rate
constants are larger than expected from the ␭max-values
in Table I. In contrast to I the kinetic curves of II are
monoexponential up to about 95% conversion and in-
dependent of excitation and observation wavelength.
A fast redistribution of the cis isomer into polar water
containing domains with faster isomerization rates can
be assumed. The same conclusion had been drawn
from isomerization experiments with II in reverse mi-
cellar solutions of AOT [18]. However, the difference
in redistribution behavior between I (slow) and II
(fast) cannot be explained from our results.
Kiso of all three dyes is largest in microemulsion D
with the highest content of surfactant and correspond-
ingly the largest volume of the water/surfactant inter-
face layer. The kinetic solvent isotope effect kiso, H2O
/
kiso, D2O Ϸ 1.4 (fast component) proves that water has
access to all three isomerizing dyes in microemulsion

ISOMERIZATION OF AZOBENZENE DYES IN SOLVENTS
349
A and that isomerization occurs in the water/surfactant
interface layer.
BIBLIOGRAPHY
1. Cho, C. H.; Chung, M.; Lee, J.; Nguyen, T.; Singhh, S.;
Vedamurthu, M.; Yao, S.; Zhu, J.- B.; Robinson, G. W.
J Phys Chem 1995, 99, 7806.
2. Schwuger, M.-J.; Stickdorn, K.; Schoma¨ker, R. Chem
Rev 1995, 95, 849.
3. Tsukamoto, I.; Ogli, K. J Colloid Interface Sci 1994,
168, 323.
4. Saroja, G.; Samanta, A. Chem Phys Lett 1995, 246, 506.
5. Miyagishi, S.; Asakawa, T.; Nishida, M. J Colloid In-
terface Sci 1995, 115, 199.
6. Miyagishi, S.; Kurimoto, H.; Ishihara, Y.; Asakawa, T.
Bull Chem Soc Jpn 1994, 67, 2398.
7. Miyagishi, S.; Kurimoto, H.; Asakawa, T. Bull Chem
Soc Jpn 1995, 68, 135.
8. Yao, J.; Romsted, L. S. J Am Chem Soc 1993, 115, 8351.
9. Romsted, L. S.; Yao, J. Langmuir 1996, 12, 2425.
10. Garcia-Ris, L.; Leis, J.-R.; Iglesias, E. J Phys Chem
1995, 99, 12318.
SUMMARY AND CONCLUSIONS
The influence of polarity and bulk viscosity on the rate
constants of the thermal cis-trans isomerization of
4,4’-nitroaminoazobenzenes were estimated quantita-
tively from ␭
(␲␲* CT-transition of the trans
max
forms) in homogeneous solvents and in solvents of
high bulk viscosity.
The influence of increasing concentration of water
in the microenvironment of the azo dyes on ␭ and
kiso, respectively, is threefold and can be described
qualitatively: increase of polarity (positive for ␭max and
kiso), decrease of the electron donating ability of the
amino group due to H-bonding at the nitrogen electron
max
lone pair (negative for ␭ and kiso), and increasing
11. Szadjdzinska-Pietek, E.; Gebicki, J. L. J Phys Chem
1995, 99, 13500.
12. Mancini, G.; Schiavo, C.; Cerichelli, G. Langmuir 1996,
12, 3567.
13. Kluge, T. AM1 calculations, unpublished.
14. Schanze, K. S.; Mattox, T. F.; Whitten, D. G. J Am
Chem Soc 1982, 104, 1733.
15. Schanze, K. S.; Mattox, T. F.; Whitten, D. G. J Org
Chem 1983, 48, 2808.
16. Schanze, K. S.; Whitten, D. G. J Am Chem Soc 1983,
105, 6734.
17. Shin, D.-M.; Whitten, D. G. J Am Chem Soc 1988, 110,
5206.
18. Shin, D.-M.; Schanze, K. S.; Whitten, D. G. J Am Chem
Soc 1989, 111, 8494.
19. Abdel-Halim, S. T.; Abdel-Kader, M. H.; Steiner, E. U.
J Phys Chem 1988, 92, 4324, and private comm. by
Prof. Steiner.
max
influence of another factor, possibly dynamics of hy-
drogen bond forming and breaking steps on kiso (pos-
itive, by favorable charge fluctuations).
According to the ␭max-values, the trans azo dyes
have a constant, or with respect to polarity and hydro-
gen bonding, balancing average microenvironment in
the water/surfactant interface layer in a wide range of
micelle concentrations. The isomerization rates of
their cis isomers decrease up to an order of magnitude
with increasing surfactant concentration, in part due to
increasing microviscosity.
The large kinetic solvent isotope effects at equal
surfactant concentrations in H2O, and D2O solutions,
respectively, are caused by differences in structures
dominated by hydrogen bonding. They give a clear
indication about the presence of water in the micro-
environment of the reacting azo dyes.
20. Beilsteins Handbuch der Organischien Chemie, Vol 16,
314, Springer-Verlag Berlin, 1933.
In bicontinuous microemulsions of igepal CA-520/
water/heptane the azo dyes studied may be solubilized
at different sites mainly in the extended pseudophases
oil and water/surfactant interface. Deviations from
first-order behavior indicate redistribution between
different solubilization sites.
The kiso increasing effects of hydrogen bonding and
structural fluctuations of the water/surfactant interface
layer could not be separated.
21. Landoldt-Bo¨rnstein, Vol II/5a, 6th ed., 1969, Springer-
Verlag, Berlin.
22. Kobayashi, S.; Yokohama, H.; Kamei, H. Chem Phys
Lett 1987, 138, 333.
23. Hair, S. R.; Taylor, G. A.; Schultz, L. W. J Chem Educ
1990, 67, 709.
24. Marcandalli, B.; Pelliciari-Di Liddo, L.; Di Fede, C.;
Bellobono, J. R. J Chem Soc, Perkin Trans 1994, 2,
589.
25. Bondi, A. J Phys Chem 1964, 67, 441.
26. Benniston, A. C.; Harriman, A. J Chem Soc, Faraday
Trans 1994, 90, 2627.
27. Vinogradov, S. N.; Linnell, R. H. Hydrogen Bonding;
Van Norstrand Reinhold Comp.: New York 1971.
28. Bales, B. L.; Almgren, M. J Phys Chem 1995, 99, 15153.
29. A reviewer called our attention to the possibly incom-
plete binding at our lowest surfactant concentrations
with reference to results of ref. [31].
K. G. gratefully acknowledges a grant and further support
by the Graduiertenkolleg “Physikalische Chemie der Grenz-
fla¨chen” at the University of Leipzig, sponsored by Deutsche
Forschungsgemeinschaft and Sa¨chsisches Staatsministerium
fu¨r Wissenschaft und Kunst. We thank R. Wang for some
measurements and the Fonds der Chemischen Industrie for
financial support.
30. Quina, F. H.; Nasser, P. M.; Bonilha, J. B. S.; Bales, B.
L. J Phys Chem 1995, 99, 17028.

350
GILLE, KNOLL, AND QUITZSCH
31. Miyagishi, S.; Matsumura, S.; Asahara, T.; Nishida, M.
J Colloid Interface Sci 1988, 125, 237.
32. Uppu, R. M. Langmuir 1995, 11, 1038.
33. Kato, T.; Kanada, M.; Seimiya, T. Langmuir 1995, 11,
1867.
39. Chang, N. J.; Kaler, E. W. J Phys Chem 1985, 89, 2996.
40. Baglioni, P.; Dei, L.; Gambi, C. M. C. J Phys Chem
1995, 99, 5035.
41. Mean value of the data given in K. Kalyanasundaram,
Photochemistry in Microheterogeneous Systems; Aca-
demic Press, 1987 and Catalog Sigma Chemie GmbH,
1996.
34. Kusumoto, Y.; Shizuke, M.; Ssataku, I.; Kagoshima, K.
Chem Lett 1986, 529.
35. Seret, A.; Gandio, E.; Van de Vorst, A. Chem Phys Lett
1987, 135, 427.
42. Heerklotz, H.; Binder, H.; Lantzsch, G.; Klose, G.
Biochem Biophys Acta 1994, 1196, 114.
36. Loran, C. P.; Von Wandruszka, R. Talanta 1991, 38, 497.
37. El Seoud, O. A.; Blasko, A.; Bunton, C. A. Langmuir
1994, 10, 653.
38. Wang, A. S.; Matsui, Y. Bull Chem Soc Jpn 1994, 67,
2917.
43. Schoma¨ker, R.; Stickdorn, K.; Knoche, W. J Chem Soc,
Faraday Trans 1991, 87, 847.
44. Jo´hannson, R.; Almgren, M. Langmuir 1993, 9, 1259.
45. Gra¨tz, K.; Helmstedt, M.; Meyer, H. W.; Quitzsch, K.
Colloid Polym Sci 1998, 276, 131.