Thermal stability of radiation curable materials

Thermal stability of radiation curable materials

Radiat. Phys. Chem. Vol. 25, Nos. L-3, pp. 323-332, L985 01-16--5724/85 $3.00 + .00 Pergamon Press Ltd Printed in Great Britain. THERMAL STABILITYO...

443KB Sizes 5 Downloads 53 Views

Radiat. Phys. Chem. Vol. 25, Nos. L-3, pp. 323-332, L985

01-16--5724/85 $3.00 + .00 Pergamon Press Ltd

Printed in Great Britain.

THERMAL STABILITYOF RADIATIONCURABLEMATERIALS

V. P. THALACKERAND T. E. BOETTCHER 3M, CENTRALRESEARCHLABORATORIES PROCESS TECHNOLOGY LABORATORY ST. PAUL, MINNESOTA55144

ABSTRACT The thermal s t a b i l i t y of radiation curable materials was determined via TGA measurements and related to functionality, chemical type, and cure conditions. The results indicate that Tri-, Tetra-, and Penta- functional acrylates are more thermally stable than lower functional acrylates and that acrylates have better thermal s t a b i l i t y than methacrylates. A comparison of EB, UV, and thermal curing conditions for TMPTAshowed that EB-cured material was more thermally stable than UV and that thermally cured TMPTAwas the least stable. The thermal stabilii~y:of acrylated epoxy oligomers is again in the order of increasing functionality The '~C carbonyl region and oxygen permeability measurements were used to determine the extent of cure and then related to thermal s t a b i l i t y . KEYWORDS Radiation; thermal; cure; s t a b i l i t y ; acrylate; E-beam; UV INTRODUCTION The attraction of radiation curing to industry as an alternative drying method for coatings was o r i g i n a l l y spurred by the escalating costs of energy, and solvent emission standards. The use of radiation curing and the development of materials which can be used in UV and EB processing has lead to several successful processes which provide these advantages. Rapid cure leading to fast work turnover, application v e r s a t i l i t y , and low energy requirements are a few of these documented attributes.(1) More recently, however, the cost effectiveness of radiation curing compared to thermal processes and the development of new products made possible by this emerging technology have become driving forces. Attention is being focused on the use of radiation curable materials to provide unique properties such as solvent, abrasion, and stain resistance, gloss and outdoor durability.(2) Another attribute that has only recently been given attention is the property of thermal s t a b i l i t y . Thermal s t a b i l i t y becomes important when considering applications such as printed c i r c u i t boards, solder masks for photoresists, gasket coatings, wire/cable coatings, tapes, and solar collectors or reflectors. An investigation of the thermal s t a b i l i t y of radiation curable coatings was recently reported(3) in which the authors focused primarily on coatings cured by UV, although a small comparison was given to EB. We wish to expand on the knowledge of the thermal s t a b i l i t y of EB-cured coatings because of the important differences with those cured via UV. Our objective in this work was to determine the relationships, i f any, between chemical type, process conditions, and thermal s t a b i l i t y of EB-cured materials. RESULTS AND DISCUSSION The i n i t i a l focus was on the relationship between thermal s t a b i l i t y , the monomer type, and monomer functionality. Figure l i l l u s t r a t e s the TGA curves for acrylates which have 2 to 5 acrylate functional groups. As can be seen from the figure, the diacrylate, HDODA, exhibits significantly less thermal s t a b i l i t y than do the T r i - , Tetra-, or Penta- functional materials especially below 300°C. The higher crosslink density of the 3 to 5 multi-functionals leads to lower oxygen permeability and thus less oxidative degradation. The thermal weight loss for HDODAbegins at a much lower temperature, lO0°C, than do the other multi-functionals (> 250 to 300°C). The difference could be due to the lower reactivity of HDODAat low dosage levels. The 323

)24

V.P. THALACKER ~ D T. E. BOEITCHER

monomers which begin to show significant degradation in the 250 to 300°C region are TEGDA, TPGDA, and TMPPTA. Table I gives complete data including other diacrylates. Figure 2 and Table 2 i l l u s t r a t e what happens to the thermal s t a b i l i t y of HDODAwhen cured at different dosage levels. As can be seen, when HDODAis exposed to high irradiation levels i t s thermal s t a b i l i t y is improved so that at 5 and 10 Mrads i t s s t a b i l i t y is approaching that of the higher functional materials. The difference in thermal s t a b i l i t y at the different cure conditions can also be explained in terms of crosslink density. I t might be expected that materials that are decomposed via oxidative degradation as the major pathway would show reduced degradation i f more highly crosslinked and this is the case here with HDODA. We also sought to link oxygen permeability to thermal s t a b i l i t y since i t is expected that more highly crosslinked structures of the same type would have less oxygen permeability. OXYGENPERMEABILITY An investigation of oxygen permeability was undertaken with EB-cured samples. Oxygen permeability can be used as a measure of crosslink density and "tightness" of crosslink structure in cured systems. Table 6 shows some of the typical results of oxygen permeability in several different chemistries. Oxygenpermeability for the monomer, HDODA, varies d i r e c t l y with the dose, i.e., the "tighter" crosslink network achieved at 10 Mrad has less oxygen permeability than does the f i l m cured at 3 Mrad. The oxygen permeability results may be a more sensitive measure of crosslink density because even through the thermal s t a b i l i t y data for HDODAis similar at 5 and 10 Mrad, the oxygen permeability is less at 10 Mrad than at 5 Mrad. The 3 Mrad cure is i n f e r i o r to either 5 or 10 Mrad. Further work on oxygen permeability with oligomers is discussed below. 0LIGOMERS Several thermal s t a b i l i t y studies were made of acrylated epoxy oligomers which had been EB-cured. The results are shown in Table 3. The observed order of s t a b i l i t y at 300°C for the epoxy-type oligomers which varied in type and functionality was: CELRAD 3700 > CELRAD 3600 > CELRAD 3500 > CELRAD 3200 Table 3 and Fig. 5 show the results. Celrad 3700, 3600, and 3500 are diacrylate esters of a bisphenol A type epoxy resin, while Celrad 3200 is an acrylate ester of an aromatic/aliphatic epoxy blend. The functionality is 2 for both 3700 and 3600, 1.85 for 3500 and 1.6 for the 3200 oligomer. Thus the s t a b i l i t y order can be related to functionality much as the work above showed higher functional monomers to be more stable than those of lower functionality. Oxygen permeability measurements were made on Celrad 3500 (epoxy-based acrylate) at different cure conditions. The results are less clear, however, than those discussed above for the acrylate monomer HDODA. The order of permeability for Celrad 3500 is from high irradiation to low irradiation (i.e., low permeability to high permeability) but the range is narrow and may not be s t a t i s t i c a l l y significant. The differing chemical types also demonstrated the u t i l i t y of using oxygen permeability to relate to thermal s t a b i l i t y . ACRYLATE VS. METHACRYLATE The thermal degradation pathways for acrylate vs. methacrylate polymers has been determined by comparing the degradation of poly(methyl acrylate) with that of poly(methyl methacrylate).(4) The authors have pointed out that poly(methyl acrylate) undergoes a complex degradation and yields only a trace of monomer, whereas poly(methyl methacrylate) undergoes predominantly a depol~nnerization reaction to monomer. A comparison was made here of the thermal s t a b i l i t y of an acrylate and methacrylate monomers cured under EB irradiation. Figure 3 and Table 4 i l l u s t r a t e the results using TMPTAand TMPTMAmonomers. The methacrylate shows s i g n i f i c a n t l y lower thermal s t a b i l i t y than does the acrylate, especially below 450°C. The acrylate vs. methacrylate thermal s t a b i l i t y shown here correlates with the previous detailed work.

Thermal Stability of Radiation Curable Materials

CURE CONDITIONS The thermal s t a b i l i t y of TMPTAwhich had been cured at several conditions including EB at 1,3,5,10 Mrad, and thermally with and without a peroxide catalyst were compared with the results reported previously via UV curing.(3) The results are listed in Table 5 and i l l u s t r a t e d in Fig. 4. The order of thermal s t a b i l i t y for TMPTAat 350°C and below is lO Mrad = 5 Mrad = 3 > UV-cured > l Mrad > thermally catalyzed > thermally uncatalyzed. TMPTA cured at 3,5,10 Mrad exhibits identical thermal s t a b i l i t y and compares quite favorably to the UV-cured data reported (no results available on UV above 350°C). The l Mrad sample, however= shows less thermal s t a b i l i t y below 4OO°Cthan those cured at higher dosages. An advantage of radiation over thermal curing is shown when comparing samples thermally cured with or without a catalyst. The non-catalyst sample shows very significant degradation at 3500C while the sample containing a catalyst does not show significant weight loss u n t i l 400°C. In both cases, however, the radiation cured materials are more stable at 400 to 4500C than are the thermally cured ones. At 400 and 450°C the l Mrad sample is, surprisingly, the most stable sample, while the thermally cured samples remain the least stable. Apparently the speed of cure via radiation allows less chain scission or oxidation to occur during the curing process than is the case with thermally cured materials. The higher oxidation or degradation is then observed in TGA measurements as much faster weight loss for thermally cured materials. NMR EB-cur~ poly(trimethylolpropane triacrylate) (poly(TMPTA))was examined by solidstate "~C NMR for degree of cure. Of particular interest was the development of information on the degree of polymerization of highly crosslinked networks which are generally insoluble and thus limited in the kinds of analysis that can be performed. A second, more general, question was to develop information on the nature of the crosslinking function and of the extent of cure. A third interest was in correlating the chemical and physical description of the cured network with thermal s t a b i l i t y properties. Solid-state NMR relaxational experiments have shown that motional information on the molecular level can be obtained. I t is expected that the extent of cure influences molecularmotion in polymers, so these relaxation studies could help to determine the nature of crosslinking in these systems. EXTENT OF CURE The resonances due to carbonyl carbo~ in unreacted and reacted acrylate polymers are generally resolvable by solid-state / a C NMR. The ImunreactedU nuclei appearing s l i g h t l y upfield (165 PPM versus 176 PPM). The ratio of reacted to (reacted + unreacted) is the degree of polymerization. The main d i f f i c u l t y in making this determination by solid-state NMR is obtaining spectra with quantitatively accurate peak areas. Generally, CPMASspectra at long contact times yield semi-quantitative results, but recent work on similar polyacrylates has demonstrated that the intensities of the °'unreacted" carbonyl peaks may be strongly dependent on the "goodness" of the Hartman-Hahn match. The reasons for this are not clear at this time, but the consequences are obvious - CP spectra should not be used for such experiments. 90-degree pulse excitation does provide quantitative results i f a l l the nuclei are given sufficient time to relax f u l l y between pulses. This can be an appreciable time for solids, possibly several minutes or more for carbons remote from protons, such as carbonyls. In the case at hand, this constraint may be relaxed somewhat because both peaks of interest are carbonyls. Thus while their intensities compared to the other nuclei may be somewhat lower, carbonyls compared to other carbonyls may yield quantitative ratios at much shorter times. This cannot be assumed, however, the experiments must be done to demonstrate this to be true. A comparison of the carbonyl region for samples of TMPTAcured at 1,5, and lO Mrad is shown in Fig. 6A. Analysis of the peak areas shows that degree of cure changes from 66 to 96%. See Fig. 6B. The degree of cure is defined here as the percentage of total acrylate functionality which has polymerized. Assuming no cure at zero Mrad, Fig. 6B shows that the majority of curing occurs for the lower radiation exposures. This, of course, is expected because as the material cures, the growing polymer network inhibits the mobility necessary for further reaction. The degree of cure change observed for this acrylate via NMR corresponds to the changes observed in thermal s t a b i l i t y as the dosage is increased.

33

326

v. P THALACKER~'~DT E BOETFCH ' ER EXPERIMENTAL

Radiation Curable Films The films were prepared for curing by knotched bar draw down at 4 mil on polyester film. The materials were used neat as delivered by the manufacturer. When the viscosity of the material was too high at room temperature for the preparation of good films the monomeror oligomer was warmed s l i g h t l y prior to draw down in order to produce good films. The films were then stripped from the polyester prior to further analysis. Electron-Beam Curing The wet films on polyester were exposed to electron-beam irradiation delivered by a 250kev "electrocurtain" (ESI) accelerator at various dosages. The acceleration voltage was 200kev with variable beamcurrent and l i n e speed of 25 feet/minute. The oxygen l e v e l was maintained at less than 200 ppm. Thermogravimetric Anal~sis (TGA1 Thermogravimetric analysis was conducted on the cured materials after stripping them from the polyester substrate. The TC~Ameasurements were conducted under an a i r atmosphere at a flow rate of 100 cc/minute with a constant heating rate of 20°C/minute to a maximumof 55°C on a Perkin Elmer mode] TGS-2 thermal analyzer. The TGA data was obtained as a plot of weight percent versus temperature. Weight loss percentages were then read from the graph. Nuclear Magnetic Resonance The cured films were powdered in a liquid-nitrogen m i l l . The samples were spun at the magic-angle in alumina rotors with KEL-F endcaps, generally, ~ a rate of 3 KHz. A l l spectra were acquired with about 50KHz of proton decoupling; "~C RF f i e l d was about 50KHz. Recycle times were 1.0 s for cross-polarization experiments and 60 s for 90 degree excitation. Oxygen PermeabilitX The oxygen permeability was determined using a Mocon 0X-tran Twin via ASTM test method D-3985-81. Two films were conditioned simultaneously and the data read out on the computerized data logger. CONCLUSION The study of thermal s t a b i l i t y of both acrylate and methacrylate functional materials showed that acrylates were the more stable when cured under l i k e conditions via EB. The higher functional acrylates are more thermally stable than t h e i r lower functional ~mologs and this thermal s t a b i l i t y was a function of crosslink density as shown by ~C NMR and by oxygen permeability. A comparison of the thermal s t a b i l t y of l i k e materials cured via EB, UV and thermal conditions showed that EB curing resulted in greater thermal s t a b i l i t y than UV which in turn gave more thermal s t a b i l i t y to the cured samples than did either catalyzed uncatalyzed thermal curing.

ABBREVIATIONS

HDODA DPMHP PETA-tri PETA-tetra TEGDA TMPPTA TMPTA TPGDA TDI MDI Celrad

= = = = = = = = =

Hexanediol diacrylate (Sartomer) Pentaerythritol monohydroxypentaacrylate (Sartomer) Pentaerythritol t r i a c r y l a t e (Sartomer) Pentaerythritol tetraacrylate (Sartomer) Tetraethylene glycol diacrylate (Celanese) Trimethyol propane propoxylate t r i a c r y l a t e (Diamond Shamrock) Trimethyol propane t r i a c r y l a t e (Sartomer) Tripropylene glycol diacrylate (Celanese) Toluene diisocyanate = Methylene diisocyanate = registered trademark - Celanese Chemical Co.

Thermal Stability of Radiation Curable Materials

ACKNOWLEDGEMENTS The authors would like to thank W.l~onway and R. Skarjune of 3M's Central Research Analytical Department for TGA and ~C NMR results as well as interpretation of those results, ~ Hwang for use of the Mocon oxygen permeability equipment, and S. Van Meerten for typing this final draft. REFERENCES l-Lacey, J. and Keough, A. H., Conference Proceedings, RadCure V, September, 1980, "Radiation Curing: A Discussion of Advantages, Features, and Applications" 2-Burn, I. and Skelhore, G., Rad'Cure 'B3, Laussane, Switzerland, May 9 - I I , 1983, "Product Development in the '80's - Extending the Field of Radiation Curing Applications" 3-Armbruster, D. C. and Molina, J. F, Radiation Curing 8(4), November 1981, P4-28 4-Conley, R. T., editor, "Thermal Stability of Polymers", Volume l, Marcel Dekker, Inc., New York, New York (1970) (Chap. 8, Part If, Acrylate Based Polymers P225ff) 5-Schaefer, J., Stejskal, E. 0., Steger, T. R., Sefcik, M. D., and McKay, R. A., Macromolecules I__3_3(Ig8O), l 121 and references therein.

327

'.S

V.P.

TH.-XL.-',.CKgR .-',:',,D T, E. BOETrCHER

20

40.

60. L 6-HDOOA T~TA 80-~ PETA-T[T~

t00 i 0

I 50

I 100

I

I

~50

I

' I

200 ~0 300 TENO~L~TURE ~EG. CJ

Fig. t . E f f e c t of f u n c t i o n a l i t y on thersa! s t a b i l i t y

I

I

350

400

i 450

I 500

at a close of 3 ldrad.

0

\ 20.

I

40.

§0 3 ;red. 5 14"ed.

80 tO Iq'ed.

100

! 5O

I 100

I 150

'I 200

~TUI:E Fig. 2

Thera.iZ o t e b l l ~ t y

I 250

I 300

(I]FJ;. C) o f i. f r-HIX]OA

! 350

I 400

I 450

Thermal Stability of Radiation Curable Materials

TAIIE1

329

MON{I~R"rIPEA~. FUNCTI~I~ %wn"LOSSAT °C

MON~BR

TRACE

100

150

200

250

300

14

iS

400

450

22

50

9'2

500

[F.SIG~TION

~]OUA

SR238

7

12

14

TEQ)A

SR258

0

0

0

3

12

30

67

84

0,5

0,7

i

7

17

37

78

88

91

IPGDA TMPTA

SR351

0

0

0

0

i

2

10

30

85

II'IgTA

P4072

0

0

0,5

i

6

23

60

72

88

PETA-TRX

SR444

0

0

0

0,5

i

3,5

14

38

81

PETA-~,~A

SR2g5

0

0

0

0

0

2

10

33

80

HYI3ROXY--~.NTA SP,399

0

0

0

0,5

I

3

8

35

85

DIPETAMONO

X WTLO~ AT °C

DOSE(FRAD)

1~

200

250

300

400

450

3

7

14

14

15

22

50

92

5

0

0

0

0

1

7

33

"82

0

0

0,5

0,5

i

5

31

82

TAB.E 3

~IC..<]M~WTLOSS_Z

OLIGQ'ER ACRYLAIE EPOXY

TRAIl

Z ~r LOSS AT C

IISI(3'~ATIO~

100

(ZLRAD3200

1

1,5

3

4,5

1

1,5

2

2,5

3.5

-

150

200

250

300

350

10

400

450

5(30

23

45

61

74

10

45

66

81

H

"

3600

1

1,5

2

3.5

6

9

24

50

69

n

"

3700

0,5

1

1

1,5

3

5

20

40

63

TABLE4_ ACRYLAIEvs I,EII~OIYLATE WTLOSS~ AT "C

100

150

EB 3 YRAD

0

5"

0

MONC~

CURE

I~TA #

#

200

250

300

350

400

~0

0

0

0

1

2

10

30

0

0

0

1

3

II

32

85

0

0

0

0,

1

3

13

35

83

EB 3 VRAD

0

0,5

1

2

6

16

31

43

91

#

5"

0

0

0

0

6

21

45

59

83

n

10"

0

0

0

0

5

16

36

50

88

T~TMA

10"

330

V.P.

THALACKER AND T . E. BOETTCHER

20-

40-

60114~rA P

80-

tO0 0

I

I

I

I

I

I

I

I

I

!

50

~oo

~so

2o0

2so

3o0

3so

40o

4so

500

TBfB~nJR[ Fig. 3. Ther~ml s t a b i l i t y

(m[S. CJ

of a c r y l a t e vs. ~ U l ~ y l e t e

\\

20

I

at • dose of 3 14red.

\

40

\

\

\,

\ \

60 i i4~O.

4HRS. AT SSO ¢. NO CATALYST

\

3 14RAO.

80'

iO0 0

5 NRAO.

4 HRS. AT SO0 C. CATALYST

lO NRaD.

IN CATALYST

I 50

I tO0

I t50

I 200

I 250

TB4q~Tm[

I 300

I 350

I 400

(DES. C)

Fig. 4. Theraal I t l d S i l i t y of TI~TA at vm'lous Cure conditions

i 450

I 500

Thermal Stability of Radiation Curable Materials

TAI!I_E~_ CURIN6IECHNIQUE_~ ~

331

STABILITY

F_~ ~I}TA CII~ I I l l ~

WTLOSS%AT °C

100

3OO

~JO

40O

45O

5OO

0.5

1

1.5

2.5

3

4

8

22

89

-3 "

0

0

0

0

1

2

10

30

85

-5"

0

0

0

0

1

3

11

32

84

EB- 1 MRAD

-10 "

150

2O0

250

0

0

0

0

1

3

13

35

83

150"C/4 I'~

0.5

0.5

0.5

2.5

6

31

60

72

80

"(+.25% @P)100 C/41~

0.5

0.5

1

2.5

3.5

10

28

47

84

0

0.2

0.4

0.9

1.2

NA

HA

NA

UV(+.5%BEE)(3PA,[email protected] 20 F~)

TABLE6

2.5

(]K'Y{I}IPERIT:ABILITY CCZM/I~Y

B~IERI~I. EB31~

1.35

51~

n

#

124

10 MRA~

110

EB3 MRAD

65

u

#

5MRAD

59

#

n

10 MRAD

45

C~tAD 5500

20

~t

40 -

60c~J~O 3a90 oS_q~ 35oo

% %

60c~.&tO 36oo

CELRkO 3 7 ~ 100 0

I 50

I 100

! 150

I 200

~lul~ Ftg. S. ~ 1

! 250

! 300

I 350

I 400

450

rues. c)

s ~ M t l l t y of W 4 r / o c r y l e t t ollgoeor a t • cloq of 3 14rl4.

I 50Q

332

V . P . THALACKER .~ND T. E. BOET'FCHER

A

f

/-

F

L/L 1

5

10

B 100 -

C U R E

5O-

0

I

0

Figure

5 MRad

6. A. S p e c t r a of C a r b o n y l B. % C u r e v e r s u s M R a d

i

10

Region

for

10, 5, a n d

I MRad