Spectroscopic properties and photosensitivity of epoxy resins

Spectroscopic properties and photosensitivity of epoxy resins

PolymerPhotoche~ 2 (1982) 97-107 SPECTROSCOPIC PROPERTIES AND PHOTOSENSITIVITY OF EPOXY RESINS NORMAN S. ALLEN,t JOHN P. Bn~KLE¥ Department of Chem...

407KB Sizes 15 Downloads 70 Views

PolymerPhotoche~ 2 (1982) 97-107

SPECTROSCOPIC PROPERTIES AND PHOTOSENSITIVITY OF EPOXY RESINS

NORMAN S. ALLEN,t JOHN P. Bn~KLE¥

Department of Chemistry, John Dalton Faculty of Technology, Manchester Polytechnic, Chester Street, Manchester M1 5DG, Great Britain BARRY J. PARSONS,GLYN O. PHILLIPS

School of Natural Sciences, Kelstenon College, NE Wales Institute of Higher Education, Connah's Quay, Clwyd, N. Wales, Great Britain and NORMAN H. TENNENT

Glasgow Art Gallery and Museum, Kelvingrove, Glasgow G38AG, Scotland, Great Britain (Received: 21 April, 1981)

ABSTRACT

The spectroscopic properties of various commercial epoxy resins are examined using infra-red, fluorescence~phosphorescence and ultra-violet-visible derivative techniques. Infra-red studies on the uncured resins and hardeners show the presence of impurity carbonyl chromophores but their chemical nature appears to be very dependent upon the resin/hardener type. All the hardeners examined exhibit very weak fluorescence, the origin of which is unknown and those hardeners containing an alkyl phenol accelerator give strong phosphorescence. All the uncured resins examined exhibit very intense phosphorescence, due to the Bisphenol A component, while the cured resins gave emission due to a triplet eximer. Ultra-violet-visible derivative absorption spectroscopy showed the presence of impurity chromophores, probably quinone in type, which absorb light wavelengths > 300 nm. The results are discussed in relation to the known light stabilities of the resins and their ability to sensitise the photofading of dyestuffs. t To whom correspondence should be sent.

97 Polymer Photochemistry 0144-2880/82/0002-0097/$2.75©Applied England, 1982 Printed in Northern Ireland

Science Publishers Ltd,

98

N O R M A N S. A L L E N E T AL.

INTRODUCTION

Epoxy resins are extensively used as adhesives and fillers in a variety of applications. In many of these applications, for example, stained glass restoration, a,= stability to prolonged sunlight exposure is essentfal. Unfortunately, epoxy resins deteriorate on exposure to sunlight and undergo extensive chain scission, crosslinking and yellowing. ~-~ The processes involved are quite complex and the nature of the yellowing process is little understood. Since sunlight-induced oxidation involves light wavelengths greater than 300 nm (ref. 3) the exact nature of the chromophore responsible for light absorption is not too dear. For example, while many early studies considered light absorption by the Bisphenol A component of the polymer, a recent study by George et al. 4 suggests that light absorption in the near ultra-violet region of the spectrum occurs through traces of aromatic carbonyl impurities formed by oxidation of the methylene group in Bisphenol A during high temperature curing. In this paper we have examined the spectroscopic properties of a number of commercial epoxy resin systems to elucidate the nature of the species responsible for their sensitivity to sunlight exposure. The results are related to the known light stabilities of the resins and also their ability to sensitise the photofading of dyestuffs. The latter property is extremely important since many commercial dyestuffs which normally have high light stability in say polyesters 3 have poor light stability in epoxy resins, z This can be a problem in a number of applications, for example, museum conservation, where colour matching is important. 2 EXPERIMENTAL

Materials

All the resins used in this study are based on the diglycidyl ether of Bisphenol A (diphenylolpropane) which has the general formula (I): O

F

CH3

OH

]

CH3

O

(1) T he resin type is dependent on the value of n. The hardeners used are all aliphatic amines of v ~ o u s types. Both the resins and hardeners used in this

work together with the names of their manufacturers are listed in Table 1. Some of the hardeners examined also contained an alkyl substituted phenol as an accelerator.

99

SPECTROSCOPIC PROPERTIES OF EPOXY RESINS TABLE 1 EPOXY RESIN SYSTEMS USED

Epoxy systems Araldite A Y 103 resin a H Y 9 5 1 hardener s T h e r m o s e t 600 resin c No. 64 hardener b Araldite M Y 7 9 0 resin a X D 7 1 6 hardener A b l e b o n d 342-1 resin a A b l e b o n d 342-1 h a r d e n e r b

Constituent

Mixing proportions

Bisphenol A diglycidyl ether + dibutylphthalate plasticiser 100 Triethylene tetramine 10 Bisphenol A diglycidyl ether 100 N H 2 C H ( C H a ) C H 2 - - [ O C H 2 C H ( C H 3 ) ] x - - N H 2 35 Bisphenol A diglycidyl ether 100

[---CH(NH2)C(CH3)(NH2)CH(NH2)--O--]x Bisphenol A diglycidyl ether Polyoxypropylene diamine

37 100 32

a Ciba-Geigy (Plastics and Additives) C o m p a n y , Cambridge. b Contained an alkyl phenol accelerator. c T h e r m o s e t Plastics Inc, Indianna, USA. a Ablestik Laboratories, California, U S A .

The solvents, propan-2-ol and chloroform, were of an 'Analar' quality.

Spectroscopic measurements Infra-red spectra were recorded using a Perkin-Elmer Model 457 spectrometer. The resins and hardeners were examined directly as thin liquid films between two salt flats. Fluorescence spectra were recorded at room temperature (300 K) using a Baird-Atomic Model 100 E Fluorispec. Phosphorescence spectra were recorded at 77 K (-liquid nitrogen) using an Amino-Bowman Phosphorimeter. Ultra-violet-visible absorption spectra were recorded using a Perkin-Elmer Model 554 spectrometer equipped with a microprocessor for recording first (dA/dh) and second (daA/dA 2) order spectra.

RESULTS AND DISCUSSION

Infra-red spectra A close examination of the infra-red spectra of the resins and hardeners showed the presence of weak absorptions in the region 2000 to 1500 cm -1 which could not be associated with their basic structures. The spectra of two resins, Thermoset 600 and Ablebond 342-1, and their corresponding hardeners are shown in Fig. 1. Both resins exhibit an absorption at 1760 cm -1 which is due to a carbonyl group. The exact nature of the carbonyl group is uncertain although early work on the oxidation of polyolefins assigned absorption in this

100

NORMAN S. ALLEN ~ AL. ' THERMOSET 600

RESIN

ABLEBOND

342-1 RESIN

90

80 ta 70

u

z 60 II'HERMOSET 64

A B L E B O N D 342-1

~ so o

70 6O 5O !

2000

15OO WAVENUMBER

I

18OO

!

1500

C M "1

Fig. 1. Infra-redspectraof some uncuredresins and hardeners in the region 2000 to 1500cm-1. region to peresters of the structure R(C------O)OOR.7 The influence of these species on polymer photo-oxidation, although uncertain, is not expected to contribute to the light sensitivity of the epoxy resin system since their absorption would be well below 300 nm. The corresponding hardeners, on the other hand, were variable and only the Thermoset 64 exhibited an absorption. In this case an absorption band was observed at 1705 cm-* due to a carboxylic acid. Again these chromophores are unimportant in photo-oxidation. 3 The presence of these impurity carbonyl absorptions, although unimportant in photo-oxidation, dearly show that both the resins and certain hardeners have undergone some oxidation during their preparation.

Luminescence spectra The luminescence spectra of the resins and hardeners showed several important features relevant to the photosensitivity of the cured systems. An examination of all the resins and hardeners by fluorescence spectroscopy showed that while the former are non-fluorescent, the latter are weakly fluorescent with an emission wavelength maximum at 424 nm and an excitation wavelength at 350 nm. Typical spectra are shown in Fig. 2, where it is also seen that the

SPECTROSCOPIC PROPERTIES OF EPOXY RESINS

EX

101

EM

>i-z w Iz IJJ

.J w

=

2 o W A V E L E N G T H , nm

Fig. 2. Fluorescence excitation (EX) and emission (EM) spectra of Thermoset 64 (--), T h e r m ~ e t 600+ Thermoset 64 cured film ( - - - - ) and Ablebond 342-1 hardener ( - - . - - . --). The resins were 1% w/v solutions in propan-2-ol.

emission is not affected by the curing process. The emission was very weak and is almost certainly due to an impurity chomophore that could be important in photo-oxidation since it is capable of absorbing sunlight in the near ultra-violet region, 300-400 nm. The phosphorescence spectra of the resins and hardeners are shown in Figs. 3 and 4, respectively. All the resins exhibited phosphorescence emission at '~

EX

EM

z w p.

_z

l

_J uJ

=

3b 0

4 06

5b 0

WAVE LENGTH, nm Fig. 3. Phosphorescence excitation (EX) and emission (EM) spectra of 1% w/v solutions of Bisphenol A (--), Araldite MY790 ( - - - - - - ) and Ablebond 342-1 ( - - . - - . - - ) resins in isopropanol.

102

NORMANs. ALLENwr ~ . EX

EM A

/ / " \ E x'X= 33 0 n m

\

I ,i-,, / e

>.

Ex'X=2

,,,

'-z

L

-

~

',!1 k'.,\', ~!/ // t

I1

I

//

~,\\ \\

DI,,/

._ I-.J

,.,

n,

\ \

\

2~io

WAVELENGTHjnm Fig. 4. Phosphorescence excitation (EX) and emission (EM) spectra of 1% w/v solutions of p-cresol (--), Ablebond 342-1 (--x--x--) and HY951 ( - - - - - ) hardeners in isopropanol.

390 nm with an excitation maximum at 290 nm, except Araldite AY103 which gave a strong emission due to the plasticiser. Two typical spectra are shown in Fig. 3 for Araldite MY790 and Ablebond 342-1. Also shown in the Figure are the phosphorescence excitation and emission spectra of Bisphenol A which are seen to match closely the spectra of the resins. It is obvious from these results that the phosphorescence from the resins is due to the Bisphenol A unit present in the molecular structure. The phosphorescence excitation and emission spectra of the hardeners showed a number of interesting features different from the resins. In this case, only the hardeners containing an alkyl substituted phenol as an accelerator exhibited phosphorescence emission. It is seen from Fig. 4 that the phosphorescence spectra of the hardener Ablebond 342-1 matches closely that of a model phenol, p-cresol. Both have excitation and emission wavelength maxima at 290 and 390 nm, respectively. Also shown in the Figure are the excitation and emission spectra of the hardener HY951. This hardener exhibits an anomalous emission at 450 nm with an excitation wavelength at 330 um and this is almost certainly due to an impurity chromophore. This particular hardener in combination with the resin Araldite AY103 leads to a highly photo-unstable cured system.2 Figure 5 compares the phosphorescence excitation and emission spectra of the cured resins. Here two interesting features are noted. First, while the excitation wavelength is unaffected at 290 nm the emission wavelength has

SPECTROSCOPIC PROPERTIES OF EPOXY RESINS

103

EM

EX

l 30(

400

500

WAVELENGTH~ nm Fig. 5. Phosphorescence excitation (F_~) and emission (EM) spectra of cured films of Araldite

AY103/HYg51 (--), Thermoset 600fllaermoset 64 (------), Araldite MY790/XD716 (. . . . . . . ) and Ablebond342-1 resin/hardener(--x--x--).

shifted from 390 to 425 nm. The only exception to this effect is with the Araldite AY103 resin which contained a plasticiser. In this case the anomalous emission at 450 nm deafly dominated the spectrum, together with that from the plasticiser. The shift in the emission could be due to a triplet excimer formed by either an intra or intermolecular interaction between the Bisphenol A units in the molecular backbone of the polymeric resin. This phenomenon is well established in many aromatic containing polymers. 8'9 The exact nature of the interaction in epoxy resins would be difficult to establish due to their intractable properties. The second important feature is the marked difference in the intensity of the phosphorescence emission from the resins. The emission intensities decreased in the order Araldite AY103>Thermoset TS600>Araldite MY790> Ablebond 342-1 which correlates well with the photostabilities of the resins reported by Tennent. 2 Using the degree of yellowing as a means of assessment the light stabilities of the resins increased in the order Araldite AY103 < Thermoset 600 < Araldite MY790 < Ablebond 342-1. It would appear that the Bisphenol A unit is clearly responsible for light absorption and subsequent yellowing of the cured resin.

Ultra-violet-visible derivative spectra Ultra-violet-visible derivative spectroscopy has been recently shown to be a valuable and sensitive method for analysing the presence of light absorbing

104

NORMAN S. At,LF.N Ear At,.

I

2-1

lil~

2

~'t 0'01 .

.

.

.

_.aO.O I

I-2 u z.< m nO

U) m .<

\ \

J I

200 0 ~ WAVELENGTH, nm Fig. 6.

4~)0

Zero (A) and second order (d2A/d), 2) ultra-violet-visible derivative absorption spectra of 1% w/v solutions of XD716 (--) and HY951 ( - - - - - - ) hardeners in chloroform.

chromophores in polymer films, x° For a second order spectrum all minima are absorption maxima or vibrational structure on a main absorption band. The zero and second order derivative absorption spectra of the 'impure' hardener, HY951, are compared with those of a relatively pure hardener, XD716, in Fig. 6. It is seen that while the XD716 hardener possesses no absorption above 250 nm the HY951 shows three absorptions at 270, 280 and 310nm. The first two are evidently due to the alkyl phenol accelerator, whereas that at 310 nm is an impurity and could be the chromophore giving rise to the anomalous phosphorescence at 450nm. The difference in the excitation wavelength at 330 nm is simply due to the fact that the phosphorimeter records 'uncorrected' spectra. The second order derivative spectra of all the cured resins shows a further interesting feature and this is demonstrated for two cured resins in Fig. 7. Here it is seen that the films exhibit 'impurity' absorption bands above 300 nm. Neither of the uncured resin solutions exhibited absorption in this region of the spectrum. These absorptions are evidently formed as a result of oxidation

105

SPECTROSCOPIC PROPERTIES OF EPOXY RESINS

1'5

1"2

uJ u ZO.6 m nO Io

I

o. o.oo,

' V

~=" "~JO001

A

I 300

46o

4bO

sbo

WAVELENGTH,nm

Fig. 7. Zero (A) and second order (d2A/d~. 2) ultra-violet-visible absorption spectra of cured films of Thermoset 600/Thermoset 64 and Ablebond 342-1 resin/hardener. ( - - - - - - ) second order derivative spectra of 1% w/v solutions of Thermoset 600 and Ablebond 342-1 resins in chloroform.

during the curing process or manufacture of the resins. Early work by George e t al. 4 indicated that curing resulted in the formation of aromatic carbonyl groups that could sensitise photo-oxidation of the resins. The absorptions observed here could be due to aromatic carbonyls of the quinone type. These chromophores possess extended absorbance up to 4 0 0 n m and are wellestablished products of the thermal oxidation/degradation of phenols, u Thus, terminal phenolic groups on the polymer or the alkyl phenol accelerators could be easily oxidised during curing to give quinones. The quinones are highly photoactive and will almost certainly enhance the photosensitivity of the resin. Furthermore, they could also sensitise the photofading of dyestuffs by a hydrogen abstraction process or energy transfer. 3

CONCLUSIONS

Epoxy resins are capable of absorbing light through two different types of chromophoric units. The first type of chromophore is the Bisphenol A unit in the molecular structure of the resin. This will absorb light at the high energy

106

NORMA~ S. ALLESwr At..

end of the sunlight spectrum and undergo a scission process to give CH3 ~ ~ O - - - C H 2 C H 2 ~

hv

/H3L~/

~ H3 ~ H ~

O" +" C H 2 C H 2 ~

(1) as an in-chain process or, __

CH 3

~H3

CH3 as a possible end-chain process. These processes are well-established3'12 for phenolic-based polymers. The other chromophore is an oxidation product, probably quinone in type due to its extended absorption into the visible region of the spectrum. Again the photoreactions of these species are well-established. TM Finally, the photosensitised fading of dyestuffs in epoxy resins could be due to either of the above chromophores. The Bisphenol A unit may induce dye fading through photoreduction or energy transfer, whereas the aromatic carbonyls may accelerate fading by hydrogen-atom abstraction, a These processes are now currently under study and will be reported on later.

ACKNOWLEDGEMENT

The authors thank SRC for financial support for one of them (J.P.B.).

REFERENCES 1. LOWE,W, Studies in Conservation, $ (1960) 139. 2. TENNENT,N. H., Studies in Conservation, 24 (1979) 153. 3. McKmIAR, J. F. and ALL~, N. S., Photochemistry of man-made polymers, Applied Science Publishers Ltd, London, 1979. 4. GFORGE,G. A., SACHER, R. E. and SPROUSE,J. F., J. Appl. Polym. Sci., 21 (1977) 2241.

SPECTROSCOPIC PROPERTIES OF EPOXY RESINS

107

5. KEt~tmR, P. G. and G E s ~ a , B. D., Jr. Appl. Polym. Sci., 13 (1969) 9. 6. GF_S~__~a,B. D. and KELLEtmR, P. G., I. Appl. Polym. Sci., 13 (1969) 2183. 7. LUONGO,J. P., I. Polym. Sci., XLll (1960) 139. 8. SOMERSALL,A. C. and GUILLET, J. E., J. Macromol, Sci., Re,vs. Macromol., Chem., C (B) (1975) 135. 9. Fox, R. B., Pure Appl. Chem., 30 (1972) 87. 10. ALI~N, N. S., Polym. Photochem., 1 (1981) 43. 11. SEDLAR,J., KOVAROVA,J. and POSPISIL, J., Polym. Photoehem., I (1981) 25. 12. ALTON, N. S. and McKEU.AR J. F., Macromol, Chem., 180 (1979) 2875.