3. SYNTHESIS AND THERMAL STUDIES OF
PHOTORESPONSNE AZO POLYMERS
Polymers containing azobenzene moieties have recently attracted
considerable interest in view of their potential use as photoresponsive
systems 102'128. In particular, the - _ trans - cis photo-isomerisation of the azo groups is accompanied by reversible changes in the macromolecular
conformation and eventually this would lead to a change in physical
properties of the polymers. Therefore, the basic principle involved
in the molecular design is to construct a polymer system having azo
chromophores that can transfer light energy into a change in conformation.
The photo-isomerisation which is responsible for the photochromism
of certain polyamides with pendant azo groups prompted to explore
the possibility of designing much more easily accessible photochromic
polyesters based on polyvinyl alcohol and polyesters with azobenzene
residues in the backbone. These photoresponsive polyesters can serve
as models for energy converting polymers. In the present study, the
synthesis and characterization of the hitherto unreported photosensitive
polyesters with azobenzene residues in the pendant groups and in the
backou~ii. ure discussed. A systematic investigation of the thermal
stability of polyesters with azobenzene residues in the backbone are
also described.
Results and Discussion
3.1 Synthesis of Polyesters with Azobenzene Residues in the Side Chain
(AZO PVE)
The first step towards the synthesis of polymers was the preparation
of different azobenzenedicarboxylic acids. Azobenzene-4,4'-dicarboxylic
acid (13), azobenzene-3,3'-dicarboxylic acid (14) and azobenzene-2,2'-
dicarboxylic acid (15) were prepared by the reduction of para-, e- or - ortho-nitrobenzoic acids respectively in ethanol with zinc dust and
129 sodium hydroxide . All the three dicarboxylic acids were converted to the corresponding dicarbonyl chlorides (azobenzene-4,4'-dicarbonyl
chloride (16), azobenzene 3,3'-dicarbonylchloride (17) and azobenzene-2,2'-
dicarbonyl chloride (18)) by refluxing with molar excess of thionyl
chloride in presence of a few drops of N,N-dimethylformamide as
catalyst l3''l3l. The excess thionyl chloride was distilled off and
the resultant solid residue was recrystallised from petroleum ether
or cyclohexane. They were characterize3 by different analytical and
-1 . spectral methods. The IR peaks around 1680 cm in the azobenzene-
dicarboxylic acids are shifted to peaks around 1730 em-' showing the
replacement of the hydroxyl group of carboxylic acid by chlorine.
The introduction of azo group is confirmed by the peak at 1580 em-'
in the 1R spectrum, corresponding to the N = N stretching frequency.
The characterization data of different azobenzenedicarboxylic acids
and dicarbonyl chlorides are presented in Table 3.1.
Table 3.1. Characterization data of azobenzenedicarboxylic acids and azobenzenedicarbonyl chlorides
Compound h max (nm)
Azobenzene-4.4'-dicarboxylic acid ( 13)
Azobenzene-3,3'-dicarboxylic acid (14)
Azobenzene-2,2'-dicarboxylic acid (15)
Azobenzene-4,4'-dicarbonyl chloride (16)
Azobenzene-3.3'-dicarbonyl chloride (17)
323 330, 430 2900 (OH), 1680 iC=Oi (lit 324) 1580 (N=N)
340 319, 426 2890 (OH), 1680 (C=O) (lit 342) 1588 (N=N)
242 330, 433 2900 (OH), 1682 (C=O) (lit 243) 1.588 (N=N)
163 332, 470 1752 (C=O), 1590 (N=N) (lit 164)
97 316, 443 1758 (C=O), 1586 (N=N) (lit 97)
Azobenzene-2,2'-dicarbonyl chloride (18) 100 334, 451 1758 (C=O), 1589 (N=N)
Photoresponsive polyesters of polyvinyl alcohol (PVA) (MW 10,000)
with pendant azobenzene groups were synthesised by t he interfacial
polycondensation method 132i133. This method was adopted t o obtain
the polymers in good yield. Here, t o t he PVA (19) s t i r red with aqueous
sodium hydroxide in presence of sodium sulphate under nitrogen atmosphere
a t room temperature , an equimolar amount of azobenzene-4,4'-dicarbonyl
chloride (16) in a minimum volume of freshly distilled chloroform was
quickly added. The emulsified reaction mixture was s t i r red for 40-50 mln
and then poured into acetone in order t o coagulate the polymer.
Filtration followed by purification yielded red fibrous polymer
AZO PVEla (20). 'The condensation reaction between the dicarbonyl
chloride and the hydroxyl group of PVA t o yield the polymer AZO PVEla
is shown in the Scheme 3.1. The s t ruc ture of t he polymer was confirmed
by analytical and spectra l data . The polymer AZO PVEla is composed
of pendant 8zobenzene groups through ester linkages.
The polymer was found t o be insoluble in most of the solvents
except in N,N-dimethylacetamide and concentra ted sulphuric acid.
Elemental analysis of t he polymer showed C, 38.66; H , 3.96 and N , 4.28%.
Fairly high value of nitrogen indicated t ha t most of the hydroxyl groups
of PVA were funclionalised. The appearance of the character is t ic
stretching frequencies of the carbon-oxygen linkage confirmed the
formation of the polyester. The carbonvl (C=O) stretching frequency
of the polymer appeared at 1 7 3 0 ~ cm-I and t he C-0 stretching frequency
a t 1100 em-'.
Scheme 3.1. Pwmation of polymer AZO PVEl, (20)
The UV-visible spectrum of the AZO PVEla was recorded in
N,N-dimethylacetamide (Fig. 3.1). A strong band was observed in
the UV region (323 n m ) and a very weak band at 438 nm. Absorptions
of trans azobenzene in the U V region are known to be due to 7'7-T*
transition (320 nm) and those in the visible region (440 nm)
are due to n - T * transition1''. The slight shift of the transitions
in the polymer when compared to unsubstituted azobenzene can be
attributed to the substituent effect.
Fig. 3.1. UV-visible spectrum of AZO PVEl, (20) in DMA
The interfacial condensation method was also extended to the
synthesis of other polyvinyl esters having azo linkages a t meta or
ortho positions. Thus, the condensation of PVA (19) with azobenzene-3.3'- - dicabronyl chloride (17) and 19 with azobenzene-2.2'-dicarbonyl chloride (18) ,.
yielded AZO P V E l b (21) and AZO PVElc (22) respectively. The details of
the different azo polyvinyl esters synthesised and their' characterization
data are presented in Table 3.2.
-CH2- CH- I
3.2 Synthesis of Polyesters with Azobenzene Crosslinks
The formation of a zo polyvinyl es te r s from PVA and azobenzene-
dicarbonyl chlorides, and t he presence of another . functional group
on thc side chains of thc rcsultcd polymers prompted t o c8rry out
the synthesis of azobenzene crosslinked polymers of PVA. Thus t he
polycondensation of PVA (19) dissolved in aqueous sodium hydroxide,
with the chloroform solution of azobenzene-4,4*-dicarbonyl chloride (16)
in a 2:l molar ra t io in the presence of sodium sulphate under nitrogen
afforded a highly insoluble red fibrous polymer AZO PVEZa (23). The
crossinking reaction is shown in Scheme 3.2. IR spectrum of the polymer
showed character is t ic ester carbonyl s t re tching frequencies a t 1730 em-'
and C-0 s t re tching frequencies at 1075 cm-',
Scheme 3.2. Formationof polymer AZO PVEZa (23)
Similar condensation of PVA (19) with azobenzene-3,3'-dicarbonyl
chloride (17) o r azobenzene-2,2'-dicarbonyl chloride (18) yielded crosslinked
polymers AZO PVEZb (24) o r AZO PVEZc (25) respectively. The
. characterization d a t a of these crosslinked polymers a r e presented in Table 3.2.
- - Y S
?
11 U
U-
- V)
om
m 0
t- - 7 -
2%
0 z
-- (D
O
am
e V
) m
- I I P 2 - 0 al P - w m N r. I I - P) N - m N W > a 0 N -x
3.3 Synthesis of Polyesters (PE) with Azobenzene Residues in the Backbone
Poly.esters having azobenzene residues in the backbone were synthesised
in view of their possible photo-induced property changes. These polyesters
contain a number of azo groups in the backbone. lnterfacial polycondensa-
tion of bis(pheno1s) with azobenzenedicarbonyl chlorides was found to
be very effective in the synthesis of polyesters. Different bidphenols)
(Scheme 3.3) were prepared by the diazotizution of p-uminophenol
or tetrazotization of benzidine or tolidine followed by coupling with
phenolic ~ o m ~ o u n d s ! ~ ~ The details of the bi$henols) prepared and their
chnrncterization d ~ t n are presented in Table 3.3.
Table 3.3. Synthesis and analytical d a t a of bis(pheno1.s) used for t h e synthesis of polyesters (PE)
Synthetic Recryst Colour Yield rn.p. Analytical da ta
reaction solvent ( % ) ('"2) Found (Calcd.)
4,41-Dihydroxyazobenzene (26! diazotisation toluene black 3 2 215 C, 65.4 (67.26); coupling (lit 218) H, 4.2 (4.71);
N, 12.52 (13.08)
4-[(4-Hydroxypheny1)azol-1- diazotisation benzene black green 80 179 (dec) C, 63.5 (65.52); naphthol (27) coupling H, 4.2 (4.6);
N , 15.6 (16.09)
Bis(pheno1-4-azok4,4'-biphenyl (28) tetrazotisation toluene green 7 2 182 C 72.0 (73.1); coupling H, 4.1 (4.57);
N , 14.01 (14.21)
Bis(pheno1-4-azo-2-carboxy!- tetrazotisation toluene yellow green 80 190 (dec) C, 62.93 (64.73); 4,4'-biphenyl (29) coupling H, 3.2 (3.73);
N , 10.32 (11.62)
Bis(pheno1-4-azo)-4,4'-(3.3'- tetrazotisation toluene brown 80 248 (dec) C, 73.2 (74.25); dimethylbiphenyl) (30) coupling (l i t 250) H, 4.9 (5.28),
N, 12.06 (13.33)
\ COOH
Scheme 3.3. Bis(pheno1s) used for the synthesis o f polymers PE
For the synthesis of polymer PEla (31), equimolar amounts of
4,4'-dihydroxyazobenzene (26) in aqueous sodium hydroxide solution
and azobenzene-4,4'-dicarbonyl chloride (16) in freshly distilled chloroform
were st irred well for 20 min under nitrogen atmosphere and in t he
presence of sodium sulphate a s the emulsifying agen t a t room temperature.
The resulting product was allowed t o coagulate by adding into acetone.
I t was f i l tered, washed with water and dried t o g e t reddish spongy
mass of the polymer PEla (31) in 45% yield. The s t ruc tu re of the
polymer PEla (Scheme 3.4) was confirmed by lhermogravimetry and
spectra l methods. The IR spectrum showed character is t ic e s t e r carbonyl
(C=O) s t re tching frequency at 1730 cm-' and C - 0 s t re tching frequency
a t 1110 cm-'. UV-visible spec t rum was recorded in N,N-dimethylacetamide
( 'max 448, 346 nm).
Scheme 3.4. Formation of polymer PE,, (31)
Extension of this in terfacia l polycondensation approach t o other
azobenzenedicarbonyl chlorides and bi$henols) yielded similar products.
Thus the condensation of azobenzene-3,3'-dicarbonyl chloride (17) o r
azobenzene-2,Z'-dicarbonyl chloride -(I81 with 4,4'-dihydroxyazobenzene (26)
lead t o the formation of the polymer P E l b (32) o r PElc (33) respectively.
Similarly the different bis(pheno1s) like 4-[(4-hydroxypheny1)azol-1-naphthol
(27), bis(pheno1-4-azo)-4,4'-biphenyl (28), bidphenol-4-azo-2-carboxy)-4,4'-
biphenyl (29) and bis(pheno1-4-azo)-4,'-(3,3'-dimethylbiphenyl) (30),
each of which on interfacial polycondensation with azobenzene-4,4'-
dicarbonyl chloride (16) afforded the polymers PE2, (341, PE3a (37),
PE4a (40) and PE5, (43), with azobenzene-3,3'-dicarbonyl chloride (I?'),
yielded PE2b (35), PE3b (38), PE4b (41) and PESb (44) and with
azobenzene-2,2'-dicarbonylchloride (18) yielded PEZc (36), PE3c (39),
PE4c (42) and PESc (45) respectively. The details of the different
photosensitive polyesters synthesised are given in Table 3.4.
3.4 Thermal Stability of Polyesters with A m Groups in the Polymer
Backbone
Though much attention has been invested on the synthesis and
photochromic properties of photoresponsive polymers, only very little
work has been done on their thermal stability. Thermal stability of
a substance indicates its ability to maintain the properties as nearly
unchanged as possible on heating 134-136. On heating, a solid may
undergo thermal decomposition in addition to other physical and chemical
changes. Thermal analysis is a general term to represent a group
of related techniques in which some physical properties of a sample
is continuously measured as a function of temperature, while the sample
is subjected to a controlled temperature change. In the present investi-
gation, the techniques used for the thermal analysis are non-isothermal
thermogravimetry (TG) and differential thermal analysis (DTA).
co-
-0- 0 0 N=N ~ N = N @ o - c o 0 CO-
- o ~ N = N ~ N = N Q o - c o ~ N = N @ ~ ~
HOOC COOH
HOO c COOH
+N=NmN=NBOQN=NQ co-
H3C CH3
Table 3.4 Synthesis end analytical data of different polyesters (PE)
Azobenzene- solvent m.p. Bis (phenol) dicarbonyl Polyester ("C)
Colour chloride
26 16 PEla (31) DMA 270 (dee) pink 3490 (OH), 1730 (C=O), 1590 (N=N), 1110 (C-0-C)
26 17 PElb (32) DMA 280 (dec) pink 3460 (OH), 1730 (C=O), 1600 (N=N), 1104 (C-0-C)
26 18 PElc (33) DMA 290 (dec) pink 3480 (OH), 1732 (C=O), I600 (N=N), 1098 (C-0-C)
27 16 PE2a (34) NMP 252 (dec) red 3490 (OH), 1730 (C=O), 1596 (N=N), 1060 (C-0-C)
PE2b (35) NMP 263 (dec) red 3460 (OH), 1720 (C=O), 1598 (N=N), 1070 (C-0-C)
27 18 PEZc (36) NMP 270 (dec) orange 3492 (OH), 1729 (C=O), red 1588 (N=N), 1100 (C-0-C)
PE3a (37 ) Morpholine 277 (dec) red 3490 (OH), 1732 ( G O ) , 1590 (N=N), 1100 (C-0-C)
28 17 PE3b (38) Morpholine 286 (dec) ' orange 3488 (OH), 1730 ( G O ) 1598 (N=N), 1099 (C-0-C)
Table 3.4 contd .......
28 1 8 PE3c (39) hlorpholine 272 (dec) red 3470 (OH). 1740 (C=O),
1600 (N=N), 1102 (C-0-C)
29 16 . PE4a (40) NMP 292 (dec) red 3410 (OH). 1726 !C=0) , 1590 ( N = N ) . 1110 (C-0-C)
29 17 PE4b (41) NMP 280 (dec) orange 3480 (OH), 1732 (C=O), red 1596 (N=N), 1090 (C-0-C)
29 18 PEfc (42) N MP 290 (dec) pink 3482 (OH), 1730 (C=O), 1602 (N=N), 1108 (C-0-C)
30 16 PE5a (43) Morpholine 252 (dec) red 3452 (OH), 1739 (C=O), 1592 (N=N) , 1060 (C-0-C)
30 17 PESb (44) Morpholine 263 (dec) red 3450 (OH), 1738 (C=O), 1590 (N=N), 1150 (C-0-C)
30 18 PEjc (45) Norpholine 260 (dec) orange 3440 (OH), 1739 (C=O), red 1598 (N=N), 1050 (C-0-C)
The change in mass of the materials on heating forms the basis
of TG. DTA provides information regarding the exothermic or endothermic
nature of the reaction in addition to the enthalpy changes. Detailed
informations of the nature of a particular sample can be obtained
only by using more than one thermal analysis technique. Since all
chemical reactions are accompanied by either a change in energy and/or
mass, the techniques which enable us to measure the moss of energy
changes during a reaction can be used to study the solid state reactions.
In non-isothermal TG, the changes in mass of a sample is recorded
as a function of temperature. In DTA the temperature of the sample
is compared with that of an inert reference material while both are
subjected to a programmed change of temperature. If the sample
undergoes any transition which results in an absorption or evolution
of heat energy, a corresponding deviation will occur in its temperature
from that of the reference. This difference in temperature between
the sample and reference i.e., A T = Tsample - Treference is recorded
as a function of temperature in a DTA curve. When the temperature
difference between the sample and reference is zero, the sample does
not undergo any physical or chemical change. Using this simple but
widely used thermal analysis technique, phase transitions or chemical
reactions can be followed by observing the heat evolved or absorbed.
Physical changes usually result in endothermic curves whereas chemical
reactions, particularly those of an oxidative nature are exothermic.
DTA is especially suitable for studying the structural changes within
a solid a t elevated temperatures 137'138. Thus, in the present study,
a t t emp t has been made t o investigate the thermal behaviour of the
newly synthesised photoresponsive polyesters having azobenzene groups
in the backbone. The thermal studies were carried out only t o obtain
qualitative results on the decomposition behaviour of different polyesters.
Evaluation of kinetic parameters from non-isothermal TG curves and
the phenomenological aspects from both TG and DTA curves have
also been done.
3.4.1 Phenomenological Aspects
The thermal studies were carried out on a Shimadzu DT-40 Thermal
Analyser in an atmosphere of nitrogen a t a flow r a t e of 50 ml/min
with a heating r a t e of 10°C/min. Sample mass of 5-8 mg was used
in the studies. On the cha r t , percentage mass losses of the sample
were plotted against the furnace temperature . The thermal decomposi-
tion of polyesters having azo groups in the polymer backbone was
studied by using non-isothermal TG and DTA techniques. The da ta
obtained from non-isothermal TG and DTA curves have been utilised
for ascertaining the phenomenological aspects such a s temperatures
of initiation (Ti) , maximum decomposition (Ts) and the completion
of decomposition (Tf), since this could give information regarding the
regions of stability and regions of thermal decomposition.
The TG and DTA curves of t he polyesters PEla- PEsa a r e given
Temperature (OCI
Pig. 3.2. TG and DTG curves of PEl, (31)
in the figures (Fig. 3.2 - Fig. 3.6). The thermal decomposition pat tern
was found t o be similar in a l l cases. The stabil i ty differences observed
among the polyesters can be a t t r ibuted t o the difference in their
s t ructures . The substi tuents can favourably or unCavourubly a f fec t
the dissociation of the bonds within the polymers. The I'G curves
in all the cases show a smal l mass loss (2-3%) around 373-393 K due
to the loss of water and other solvents absorbed by the polymers.
This is also evidenced by t he first endothermic peak observed in DTA
in this region.
In the case o~f PEla (31) a fall in the TG curve was noticed
around 540 K with a mass loss of 5 % (Fig. 3.2). After 640 K another
fall in the TG curve was seen with a mass loss of 14%. This is supported
by the exothermic peaks observed in DTA curves. For PE2, (34)
a deep fall in t he TG curve was observed with a mass loss of 47%
a t 520-580 K region corresponds to the first s t age of thermal decompo-
sition (Fig. 3.3). The second s tage of decomposition with a mass
loss of 2 6 % was in t he region 656-765 K. The different s t ages of
decomposition were accompanied by exotherms in the DTA curve.
The TG and DTA curves of PE3a (37) a r e represented in Fig. 3.4.
After 550 K a fall in the TG curve was noted with a mass loss of 30%.
A deep fall in the 'TG curve was resulted at 898 K . These decompositions 1
were appeared in the DTA curve a s sharp and well-defined exothermic '
peaks in the DTA curve. The polymer system PE4, (40) was found
to be the most s table among the polyesters (Fig. 3.5). A first fall
in the TG curve occurred a t 565 K assisted with a mass loss of 3%.
The second s tage of decomposition appeared in the region 673-920 K
with a mass loss of 20% For PE5, (43) the first s tage of thermal
decomposition was in the temperature region 525-600 K with a
corresponding mas:; loss of 7% (Fig. 3.6). Interestingly, the second
100 100 300 LO0 500 600 '700
Temperature i°Cl
Pig. 3.5. TG and DTA curves of PE4, (40)
s t e p of degradation occurred ra the r very rapidly and the maximum
weight loss (75%) was the result over the temperature range of 650-810 K.
The corresponding exotherm in t h e DTA curve observed in this region
is well-defined with maximum peak area . The decomposition was
almost complete by 980 K . The thermal behaviour of polymers PE l b
and PElc was foiund t o be comparable t o t h a t of PEla and almost
identical thermal pat terns were observed in t h e TG curves of the
100 200 300 600 500 600 700
Temperature I OC)
Fig. 3.6. TG and DTA curves of PE5, (43)
polymers. T h e phenomenologica l d a t a of t h e t h e r m a l decomposi t ion
o l po lymers PEIaL- I?Esa a r e s u m m a r i z e d in Tab le 3.5.
In gene ra l , t h e azo po lymers are found to exhib i t m o d e r a t e t h e r m a l
s tnbi l i ty , t h e r igicl i t :~ be ing enhanccd by t h e a z o groups p re sen t . T h e
decomposi t ion t e m p e r a t u r e is a measu re of t h e t he rma l s t ab i l i t y of
t h e polymers. T h e pa thways through which a polyamide and a polyes te r
Table 3.5. Phenomenological data of the thermal decomposition of polymers PEla- PE5,
Decomposition Temp Nlass Decomposition Temp Peak Temp Thermal Polymer Decomposition Range in 'rG Range in DTA
LOSS ( % ) T (K) Nature system stage s l i ( K ) Tf (K) Ti ( K ) T f ( K )
I 540 619 6 556 63 i 590 E m
P E l a If 640 958 14 635 973 673 Exo
I 516 58 1 47 5 19 588 558 Exo
PE2a If 656 765 2 6 604 811 704 Exo
! 548 840 30 494 840 727 Exo
PE3a 11 898 1056 15 87 3 990 906 Exo
1 565 612 3 58 1 635 596 Exo
PE4a I1 673 927 20 650 988 773 Exo
I 524 598 7 512 573 550 Exo
PE5a 11 658 812 75 623 812 765 Ex0
degrades thermally has been investigated by pyrolysis-mass spectral
studies. It has been reported 13' that in a~oaromatic polyamides
the facile cleavage of the C-N bond of the amide group is the initial
degradation step. The next step of decomposition is the expulsion
of the azo group as nitrogen gas. The major primary process involves
the loss of carbon dioxide. In polyesters, pohl140 suggested the ether
links as the main point of thermal instability and carbon oxides as
the principal products. According to Marshall and Todd 141 the first
stage of thermal decomposition in polyesters is due to the scission
occurring a t the ether links to give one carboxyl group per chain scission.
Considering all these aspects it is possible to explain the endotherms
and exotherms occurring in the DTA curves of the synthesised polyesters
and the corresponding therniogravimetric curves.
In all the polyesters under investigation (PEla- PESa), the first
stage of thermal decomposition was in the temperature region above
520 K . In this region an exothermic peak of less peak area was seen
in the DTA curves with a corresponding small mass loss in the TG
curves. This can be attributed to the scission of C-0 bond of the
ester links which is the point of least thermal stability, leading to
the formation of -OH and -COOH as the end-capped units, as interpreted
by Marshall and 'To'dd in certain polyesters'41. The next step of degrada-
tion in all cases is the expulsion of azo group as molecular nitrogen.
The sharp and well-defined exotherms of the DTA curves in the
temperature region 620-720 K points to such elimination, as explained
in the thermal degradation of some azo polymers. 'The last stage
of decomposition includes the complete thermal breakdown of the
polymers leaving behind a small percentage of residual mass.
3.4.2 Evaluation o f Kinetic Parameters
The kinetic parameters for the second stage of decomposition
of the different polyesters were evaluated using the Coats-Redfern
equation which can be written in the form:
~ n [ ~ ( o c ) / T ~ l = Inl(AR./$E) (1-ZRT/E)I - where g(&) = -In(l-oC)
The d values were determined from the TG curves for various stages
2 of decomposition a t different temperatures. The log[g(oC)/T ] values were
plotted against 103/T. The slope and the intercept vicre determined
by the method of least squares and hence the activation energy (E)
and the correlation coefficient ( y ) were evaluated (Table 3.6).
The slight stability differences observed among the different
polyesters PEln- PE5a' as shown by the energy differences for the
second stage of thermal decomposition can be attributed to their
structural differences. The naphthyl and biphenyl systems can induce
effective steric hindrance and hence more energy may be needed for
decomposition to occur. 'I'he substituents can also affect the stability
towards decomposition by assisting or preventing the bond scissions
to some extent.
Table 3.6. YAnd E values of the second stage of thermal decomposition of polymers PEl, - PE5,
Polymer E ( k ~ rnolil) A (s-') .(
PHOTOSTIMULATED PROPERTY CHANGES
OF AZO POLYMERS