Effects of temperature on the weathering of engineering thermoplastics

Effects of temperature on the weathering of engineering thermoplastics

Available online at www.sciencedirect.com Polymer Degradation and Stability 93 (2008) 684e691 www.elsevier.com/locate/polydegstab Effects of tempera...

446KB Sizes 5 Downloads 31 Views

Available online at www.sciencedirect.com

Polymer Degradation and Stability 93 (2008) 684e691 www.elsevier.com/locate/polydegstab

Effects of temperature on the weathering of engineering thermoplastics J.E. Pickett*, D.A. Gibson, S.T. Rice, M.M. Gardner GE Global Research, 1 Research Circle, Niskayuna, NY 12309, USA Received 6 December 2007; accepted 20 December 2007 Available online 6 January 2008

Abstract We have determined the activation energies (Ea) of yellowing and gloss loss for a large number of engineering thermoplastics and blends under accelerated weathering conditions. The Ea often depend on the property measured and exposure conditions, although they vary over a fairly small range. Under the CIRA/sodalime-filtered xenon arc conditions most likely to be representative of outdoor exposure, the Ea for gloss loss was 5 kcal/mol for all samples tested. The Ea for yellowing was also 5 kcal/mol except for SAN and ABS. Evidently the color bodies formed from photo-oxidation of SAN are more sensitive to temperature. A reaction with an Ea of 5 kcal/mol will increase its rate by about 33% for each 10  C increase in temperature near room temperature. Temperature is an important, but not overwhelming, variable in the weathering of most engineering thermoplastics. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Accelerated weathering; Activation energy; Thermoplastics; Polycarbonate

1. Introduction Several variables come into play during the weathering of materials including sunlight, heat, moisture, mechanical stresses, and biological growth. Over the past few years we have investigated the first three of these in some detail for aromatic engineering thermoplastics in an effort to improve accelerated weathering techniques [1]. We have reported on the effects of irradiation conditions in an accompanying paper [2], and this report will focus on the effects of temperature e the activation energies for discoloration and gloss loss. Most accelerated weathering protocols involve increasing the temperature of the samples over ambient. In xenon arc testers, this is a nearly unavoidable consequence of the high intensity of visible and infrared radiation from the lamps. Fluorescent testers can run at near-ambient temperature

Abbreviations: PC, BPA polycarbonate; PBT, poly(butylene terephthalate); SAN, styrene acrylonitrile copolymer; ABS, acrylonitrile/butadiene/styrene copolymer; ASA, acrylonitrile/styrene/butyl acrylate copolymer; IM, acrylic impact modifier; HALS, hindered amine light stabilizer: 1% SanduvorÒ 3058 N-acylated HALS; UVA, UV absorber: CyasorbÒ 5411 benzotriazole. * Corresponding author. Tel.: þ1 518 387 6629; fax: þ1 518 387 7403. E-mail address: [email protected] (J.E. Pickett). 0141-3910/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymdegradstab.2007.12.013

because the lamps emit primarily UV radiation and do not appreciably heat the samples. Nevertheless, samples are deliberately heated in many protocols. There has been very little discussion in the literature on the activation energies for the weathering of plastics and at what temperature the tests should be run. Fischer and Ketola have reported that the average activation energy for weathering was about 7 kcal/mol for a group of 50 unspecified reference materials [3], but little quantitative data have been published. The goal of much testing is predicting the performance of a material at a standard location such as Miami or Phoenix. Analysis of data supplied by Atlas Weathering Services gives the averages shown in Table 1. On sunny days, the temperature rises quickly and is within a few degrees of maximum throughout the sunny part of the day. If temperature affects primarily the photochemical reactions, then the average daily high temperature is probably important. However, if temperature affects both photochemical and dark reactions, then the daily average temperature is probably more important. Variability in Miami and Phoenix temperatures has been discussed in another report [4]. We are currently attempting to determine an irradiance-weighted ‘‘effective’’ temperature in Arizona. Preliminary estimates suggest that this is approximately 30  C for ambient and 40  C for a black panel.

J.E. Pickett et al. / Polymer Degradation and Stability 93 (2008) 684e691 Table 1 Average daily temperatures ( C) at standard outdoor weathering sites [4]

Air, daily high Air, daily average Black panel, daily high Black panel, daily average

Miami site (1994e2006) ( C)

Phoenix site (1997e2006) ( C)

29.4 23.7 46.6 25.9

29.5 22.3 47.3 25.5

685

The relative reaction rate resulting from increased temperature is given by Eq. (2). The calculated rate changes for small temperature increases occurring around room temperature are shown in Table 3 for various activation energies. Note that the old rule of thumb of doubling the reaction rate for each 10  C increase in temperature is not generally valid and can be quite misleading. 2. Experimental

The temperature that a material reaches is dependent on a number of factors: color, absorptivity in the infrared, light intensity, orientation, wind speed, thermal conductivity, and backing. Our plastic samples sent to outdoor sites typically are mounted at a 45 angle facing south and are either unbacked or mounted on another plastic panel. In full sun, white samples are usually about 5  C warmer than the air. Fischer and Ketola of 3M Company have measured the temperature of vinyl panels of various colors exposed to the summer sun [5]. Table 2 shows the temperatures of panels when the temperature of the black panel is maintained at 40, 50, or 60  C. Reaction rates are expected to increase with increasing temperature according to the Eyring equation or the simpler Arrhenius equation (Eq. (1)) where k is the rate of reaction, A is a pre-exponential factor, Ea is the activation energy (cal/ mol or J/mol), R is the gas constant, and T is the absolute temperature (K). Eq. (2) shows the relative rate of a reaction at two temperatures where Ea has units of calories (not kilocalories) and T is expressed in kelvins (K ¼ 273.12 þ  C). k ¼ A expð  Ea =RTÞ

ð1Þ

k2 =k1 ¼ exp½ðEa =1:987Þð1=T1  1=T2 Þ

ð2Þ

Purely photochemical reactions should have very low, perhaps even negative, activation energies. However, photodegradation processes are rarely simple, one-step reactions. The measured activation energy is an aggregate of many steps, some of which may have thermally driven steps not involving light. As a result, the net activation energy is strictly phenomenological and has little mechanistic significance. Degradation due to multiple reactions can result in non-linear Arrhenius plots [6], but this is unlikely to be a problem over the relatively narrow temperature range of outdoor or artificial weathering experiments. Table 2 Temperatures of colored vinyl panels exposed to sunshine when the black panel temperature is set [5] Black panel set temperature ( C)

Black Blue Green Red Orange Yellow White

40.0

50.0

60.0

40.0 39.6 40.0 38.8 35.6 34.8 32.4

50.0 47.6 48.4 46.0 42.0 40.4 36.4

60.0 55.2 56.4 54.0 47.6 45.2 40.4

2.1. Xenon arc exposures 2.1.1. Samples General compositions are shown in Table 9. In each case, UVA ¼ 1% CyasorbÒ 5411, HALS ¼ 1% SanduvorÒ 3058, and TiO2 ¼ a coated, passivated, rutile grade. Compositions were extruded and injection molded into plaques approximately 3 mm thick. The plaques were cut down to 0.500  1.2500 (1.3  3.2 cm) and attached to 300  6.700 (7.6  17.0 cm) aluminum panels using silicone RTV adhesive to make arrays of 24 samples. A space of about 1/400 (6 mm) was left at the top and bottom of the panels for attachment to the heater sections as shown below. 2.1.2. Exposure conditions Weathering was performed in a modified Atlas Ci4000 Weather-ometerÒ using conditions shown in Table 4. Readings were taken three times per week for the first few weeks then twice per week for a few weeks and then weekly as rates of change decreased. The misting spray was disabled. Instead, the samples received a weekly rain-like spray (approximately 20 min) with tap water from an oscillating lawn sprinkler mounted downward in a sink followed by a rinse with deionized water. Additional 1 mm holes were drilled between the existing holes of the sprinkler to give more uniform coverage. Samples were blown dry using an air jet before reading gloss and color. 2.1.3. Temperature control The Ci4000 Weather-ometer was modified to allow multiple temperature control on sample arrays. Three aluminum Table 3 Rates of reaction relative to 25  C for reactions having activation energies of 3e25 kcal/mol calculated using Eq. (2) T ( C)

25 26 30 35 40 45 50 55 65 75 85 100

DT

0 1 5 10 15 20 25 30 40 50 60 75

Activation energies (kcal/mol) 3

5

7

10

15

20

25

1.00 1.02 1.09 1.18 1.27 1.38 1.48 1.59 1.82 2.07 2.34 2.77

1.00 1.03 1.15 1.32 1.50 1.70 1.92 2.16 2.72 3.36 4.12 5.46

1.00 1.04 1.22 1.47 1.76 2.10 2.50 2.95 4.05 5.47 7.25 10.77

1.00 1.06 1.32 1.73 2.25 2.89 3.70 4.69 7.38 11.3 17.0 29.8

1.00 1.09 1.52 2.28 3.37 4.92 7.10 10.2 20.0 38.1 69.8 163

1.00 1.12 1.75 2.99 5.05 8.37 13.7 22.0 54.5 128 287 890

1.00 1.15 2.01 3.94 7.56 14.2 26.3 47.5 148 431 1183 4863

J.E. Pickett et al. / Polymer Degradation and Stability 93 (2008) 684e691

686 Table 4 Conditions for xenon arc exposure

Instrument Inner filter Outer filter Wavelength cutoff (nm) Irradiance (W/m2nm @ 340 nm) Air temp. ( C) Black panel temp. ( C) Relative humidity (%) Light period External ‘‘garden’’ spray

A

B

Ci4000 Type S boro Type S boro 290 0.75 35 55 30 Continuous 20 min weekly

Ci4000 CIRA Soda lime 300 0.75 35 55 30 Continuous 20 min weekly

panels, each 6.7500  1800  3/3200 (17.1  45.7  0.24 cm) were bent at five places to the curvature of the middle ring of the Weather-ometer sample rack to make a six-sided portion of a circle as shown in Fig. 1. A rectangular silicone rubber heater (600  1800 (15.2  45.7 cm); Omega SRFG 618/2) was held tightly to the back by attaching to each segment a 300  6.7500  1/800 (7.6  17.1  0.32 cm) polycarbonate panel using four small screws on the top and bottom. The bottom screws were run through the panel and held on by nuts. The top screws were threaded into tapped holes and filed flush on the inside. A small aluminum strip was attached to the inside bottom of each segment as shown in Fig. 1 and bent outward slightly at the top. The sample arrays were slid between the strip and the aluminum sheet to hold the bottom securely, then attached at the top using a spring clip. In this way, the samples were held firmly against the aluminum sheet for good thermal contact but could be easily removed for data collection. The heater sections were then firmly attached to the inside of the middle ring of the sample rack using aluminum strips screwed onto the back. The leads for the heater and a thermocouple were directed to small electrical boxes attached to the rack support arms above the top sample ring. Three Watlow Series 935A temperature controllers were equipped with solid state relays and mounted on the disk located in the top section of the Weather-ometer that carries the black panel temperature sensor box and rotates with the sample rack as shown in Fig. 2. The hollow shaft of the sample rack was equipped from the manufacturer with a Mercotac 340

slip ring (4 conductors; 2 rated at 30 A (240 V) and 2 at 4 A) to get 24 V to the temperature sensor and the signal out. We used the higher rated conductors to pass 120 VAC to the temperature controllers and added a 24 V power supply to the disk to provide voltage for the temperature sensor. Because of problematic grounding and current leakage through the solid state relays, the 120 V line was equipped with an external switch that could be turned off before opening the sample chamber. Thermocouple wires and wires to power the heaters were run through the hollow shaft, out ports already present on the shaft inside the chamber, and into the small electrical boxes mounted on the rack support arms. Thermocouples were embedded into the surface of white polycarbonate plaques, the plaques were adhered to aluminum plates using silicone RTV, and the plates were mounted onto the heater panels. The thermocouples were routed to the junction boxes and onto the temperature controllers. The surface temperatures of the samples could be independently measured using an infrared thermocouple (Omega OS36-K-140F) temporarily mounted inside the sample chamber. The set and measured temperatures are shown in Table 5. In addition to the three heated panels, samples were also mounted on aluminum panels without heaters. Thus, samples could be maintained at four temperatures each for white and black, offset by about 15  C. 2.2. Fluorescent UV tester (QUV) exposures 2.2.1. Exposure conditions Experiments were carried out in a modified Q-Panel QUVÒ weathering device equipped with UVA-340 lamps. A partition of polycarbonate panels was placed down the centerline of the instrument and sealed with silicone RTV to isolate the two sides. The water chamber was sealed on one side to eliminate humidity. Copper coils cooled with flowing tap water were placed near the center wall of both sides. Two 400 muffin fans were mounted inside each half to circulate air. (Under Section 2.2.2.2 (Series #2) conditions described below, the fans were mounted in unused sample positions to bring in outside air and more cooling.) The wiring was modified to allow heating of the water chamber independently from the light

Heater leads 1/8” polycarbonate panel

Heater 3/32” Aluminum sheet Top view

Back view

Fig. 1. Top view of heater section and a back and front view of one segment.

Front view

J.E. Pickett et al. / Polymer Degradation and Stability 93 (2008) 684e691

687

Table 6 Set and measured temperatures (24 locations) for samples on the two sides of the modified QUV tester Set temperature

No heat 50  C 60  C 70  C

Side A (high humidity)

Side B (low humidity)

Mean ( C)

Std. dev. ( C)

Mean ( C)

Std. dev. ( C)

43.5 51.1 61.4 70.1

0.4 0.9 1.0 2.6

43.5 50.2 60.1 68.6

0.4 0.8 1.7 2.5

of the arrays was used. In addition, arrays were rotated from left to right (and side to side, if appropriate) regularly to ensure uniform exposure. Temperature uniformity is shown in Table 6. 2.2.2. Samples

Fig. 2. Schematic of sample basket in a Ci4000 Weather-ometer and mounting of modifications. Temperature controllers are mounted on the disk in the upper part. Wires lead through the hollow shaft to junction boxes on the support arms (three total). Power leads go from the boxes to the heater panels. A thermocouple embedded in a white PC sample mounted on the inside of the heater panel acts as the control sensor.

cycle timer. By passing dry compressed air into the ‘‘dry’’ side, the relative humidity could be maintained at nearly 0%. By bubbling air though the water bath (maintained at 45  C), the relative humidity could be maintained at about 45% (air temperature 44  C). Heating panels were made by sandwiching a 600  1200 (15.2  30.5 cm) heater pad (Omega SRFG 612/2) between a backing of 600  1200  1/800 (15.2  30.5  0.32 cm) polycarbonate and 600  1200  1/1600 (15.2  30.5  0.16 cm) aluminum. Samples were mounted on thin 600  1200 (15.2  30.5 cm) aluminum panels using silicone RTV, and the entire assembly of sample array/aluminum panel/heater/ PC backing was held to a standard QUV 600  1200 sample holder using large spring clips. The heaters were regulated using Omega CN76000 PID controllers and a thermocouple imbedded into the surface of a sample on each array. Because the irradiance is not uniform over the QUV sample area, only the center 800 (20.3 cm) of the space (at maximum)

Table 5 Set (on white) and measured temperatures ( C) for white, black (0.2% carbon black) and green samples (Ds are relative to the white samples) Unheated

Set temperature 50  C

60  C

65  C

White

44

48

60

65

Black D

60 15

65 15

74 14

80 15

Green D

52 7

57 7

69 9

72 7

2.2.2.1. Series #1. Samples with formulations shown in Table 7 were cut into 1.2500 (3.2 cm) squares and adhered to 600  1200 aluminum panels using silicone RTV in two 2  6 arrays per panel as shown in Table 7. The control thermocouple was imbedded into sample 3 near the center. The second series (italics) on each array received a misting spray from Weather-ometer nozzles in the sink before data was taken. Samples were exposed at 40, 50, 60, and 65  C. 2.2.2.2. Series #2. Samples with formulations shown in Table 8 were cut to 0.500  1.2500 (1.3  3.2 cm) and adhered to the center portion of a 600  1200 aluminum panel using silicone RTV in a single 3  9 array. Row #5 (center row) consisted of dummy black samples in which control and monitoring thermocouples were mounted. In this series, one set of samples received an oscillating ‘‘lawn sprinkler’’ spray for 20 min once per week while a second set was sprayed 3  per week. The humidity was not controlled for this experiment, but it ran through the fall and winter months with a dry water chamber, so the humidity was very low (w15%). 2.3. Data analysis The data at all temperatures were plotted on a single graph for each sample. Shift factors were applied to the abscissa Table 7 Samples in QUV Series #1 #

Formulation

1 2 3 4 5 6 7 8 9 10 11 12

PC black dye + UVA ASA white PC/PBT 3% TiO2 PC/ABS 3% TiO2 PC/PBT blend 2 black PC 2% TiO2 PC 2% TiO2 PC/ASA/PMMA blend black ABS 3% TiO2 PC/PBT 1 3% TiO2 PC/PBT blend 3 black Polypropylene 2% TiO2

Array layout 1 7 1 7 2 8 2 8 3 9 3 9 4 10 4 10 5 11 5 11 6 12 6 12

J.E. Pickett et al. / Polymer Degradation and Stability 93 (2008) 684e691

688 Table 8 Samples in QUV Series #2

1 2

PC 0.6% carbon black PC/ASA/PMMA blend 1 black PC/ASA/PMMA blend 2 black PC/PBT blend 1 black PC/PBT blend 2 black PC/ABS blend black ABS black PC/PBT blend 3 black PC/PBT blend 4 black PC/PBT blend 5 black Nylon 6 0.5% carbon black PC 0.6% carbon black PC/BPA polyarylate gray 1 PC/BPA polyarylate gray 2 PC/BPA polyarylate black SAN 0.5% carbon black PBT 0.6% carbon black PC/PBT 0.6% carbon black PC/PBT/IM 0.6% carbon black PC-b-resorcinol polyarylate black PC-b-resorcinol polyarylate green PC-b-resorcinol polyarylate red PC-b-resorcinol polyarylate 2% TiO2 PC 0.6% carbon black

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

25 20 15 45 C 50 C 60 C 65 C

10

Array layout 1 9 17 2 10 18 3 11 19 4 12 20 D* D D* 5 13 21 6 14 22 7 15 23 8 16 24

5 0 0

1000

2000

3000

4000

Exposure (kJ/m2) 35

b

30 25 20 15 10

45 C shift = 1.0 50 C shift = 1.2 60 C shift = 1.4

5

65 C shift = 1.65

0 0

1000

2000

3000

4000

Exposure (kJ/m2) x shift Fig. 3. (a) Raw data for white PC exposed to CIRA/sodalime xenon arc. (b) Data for white PC after shift factors have been applied to X axis.

Thermocouples embedded into surface of these black dummy samples.

values (X axis) for the higher temperatures until the points were made to superpose onto the data for the lowest temperature as shown in Fig. 3. The shift factors are the relative rates [7e9]. The activation energy is determined by plotting rate data according to Eq. (3). The natural logarithm of the relative rates is plotted against the inverse of the absolute temperature. The slope is multiplied by the negative of the gas constant, (R ¼ 1.987) to give the activation energy, Ea, in cal/mol. Because temperature control is relatively poor and the temperature range is relatively narrow, the resulting activation energies have uncertainties of at least 1 kcal/mol. The intercept, A, has no significance in this analysis. The relative rates from Fig. 3b are plotted in Fig. 4 as an example of typical data.   Ea 1 þA ð3Þ lnðkÞ ¼  1:987 T

Samples 3 and 4 gave insignificant property change during the course of exposure, so relative rates could not be determined. Examination of repeats (PC þ carbon black: #1, 30, 48 and PBT þ stabilizers: #25, 43) shows uncertainties of at least 1 kcal/mol for both gloss loss and color shift. In general, the activation energies are 5 kcal/mol except for the yellowing of SAN, ABS and a PC/ABS blend under CIRA/sodalime conditions (#39, 40, 41, 42).

0.6

0.4

Ln(krel)

*

a

30

Delta YI

Formulation

Delta YI

#

35

y = -2.49x + 7.84 R2 = 0.97

0.2

0.0

3. Results and discussion 3.1. Xenon arc exposures A summary of the activation energies from boro/boro and CIRA/sodalime xenon arc exposure is given in Table 9.

-0.2 2.90

2.95

3.00

3.05

3.10

3.15

3.20

1000/T Fig. 4. Arrhenius plot using rates from Fig. 3b. In this case, the activation energy is 4.9 kcal/mol.

J.E. Pickett et al. / Polymer Degradation and Stability 93 (2008) 684e691

The activation energies under boro/boro conditions average about 1 kcal/mol lower than under CIRA/sodalime conditions. The shorter wavelength light may drive purely photochemical reactions faster and diminish the importance of some thermally driven steps in the reaction sequence. Also, on average, the activation energies for yellowing are about 2 kcal/mol higher than the activation energies for gloss loss. Although

689

the activation energies for yellowing of SAN and ABS were higher than average, the activation energies for gloss loss were within the range of the other polymers. We know that aromatic thermoplastics are subject to reversible post-actinic yellowing [10] that could be due to cisetrans isomerism of unsaturated species. The equilibrium will be affected both by light intensity and temperature. As a result, at higher

Table 9 Activation energies (kcal/mol) and R2 for xenon arc exposure #

Formulation

Boro/Boro

CIRA/sodalime

DGloss

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

PC 0.6% carbon black PC/ASA/PMMA blend 1 black 1 PC/ASA/PMMA blend 2 black 1 PC/ASA/PMMA blend 1 black 2 PC/ASA/PMMA blend 2 black 2 PC/BPT blend 1 black PC/BPT blend 2 black. PC/BPT blend 3 black PC/ABS blend 1 black PC/ABS blend 2 black ABS 1 black ABS 2 black PC/PBT blend 4 black PC/PBT blend 5 black PC/BPA polyarylate blend gray 1 PC/BPA polyarylate blend gray 2 PC/BPA polyarylate blend black PC-b-resorcinol polyarylate black PC-b-resorcinol polyarylate green PC-b-resorcinol polyarylate blue PC-b-resorcinol polyarylate red PC-b-resorcinol polyarylate 2% TiO2 PBT 3% TiO2 PBT 3% TiO2 þ UVA PBT 3% TiO2 þ UVA þ HALS PBT 0.6% carbon black PC 3% TiO2 PC 3% TiO2 þ UVA PC 3% TiO2 þ UVA þ HALS PC 0.6% carbon black PC/PBT 3% TiO2 PC/PBT 3% TiO2 þ HALS PC/PBT 0.6% carbon black PC/PBT/IM 3% TiO2 PC/PBT/IM 3% TiO2 þ UVA PC/PBT/IM 3% TiO2 þ UVA þ HALS PC/PBT/IM 0.6% carbon black SAN 3% TiO2 SAN 3% TiO2 þ HALS ABS 3% TiO2 ABS 3% TiO2 þ HALS PC/ABS 3% TiO2 PBT 3% TiO2 þ UVA þ HALS ASA 3% TiO2 ASA 3% TiO2/HALS PC/ASA/PMMA 3 TiO2 PC/ASA/PMMA 3 TiO2 þ HALS PC 0.6% carbon black

na ¼ Not applicable. a Indicates insufficient property change for analysis.

DYI

Ea

R

2

Ea

1

0.89

na na na na na na na na na na na na na na na na na na na na na 1 3 4 2 na 4 4 3 na 4 5 na 6 5 6

a a

4 3 0 2 1 0 1 4 4 2 0 1 0 3 1 2 0 0 0 3 3 0 0 3 3 3 2 1 0 0 1 0 1 0 3 3 3 2 3 2 2 2 3 3 3

0.92 0.99 0.83 0.97 0.89 0.92 0.97 0.99 0.91 0.99 0.91 0.79

0.97 0.99

0.95 0.99 0.99 0.92 0.91

0.95 0.90 0.97 0.99 0.99 0.99 0.99 0.99 0.95 0.96 0.98 0.97 0.95 0.79

4 4 5 4 5 2 3 3 5 5 na

DGloss R

2

DYI

Ea

2

R

Ea

3

0.97

na na na na na na na na na na na na na na na na na na na na na 2 4 5 5 na 5 5 4 na 5 5 na 5 5 5 na 5 6 7 7 7 4 4 3 5 5 na

a a

0.80 0.99 0.98 0.94 0.97 0.99 0.99 0.99 0.99 0.89 0.92 0.89 na 0.99 0.99 0.99 0.97 0.96 0.93 0.97 0.98 0.90 0.94

4 4 4 3 3 0 3 5 4 2 1 0 0 0 0 0 0 0 0 2 2 2 3 4 4 3 4 3 3 2 3 3 3 3 4 5 4 4 4 3 5 3 4 4 4

0.98 0.91 0.91 0.71 0.88 0.88 0.96 0.97 0.93 0.91

0.94 0.94 0.89 0.99 0.96 0.95 0.96 0.97 0.91 0.89 0.92 0.93 0.93 0.95 0.81 0.91 0.97 0.96 0.96 0.90 0.99 0.98 0.85 0.93 0.96 0.90

R2

0.98 0.98 0.99 0.98 0.97 0.96 0.98 0.96 0.96 0.96 0.94 0.94 0.97 0.96 0.94 0.94 0.97 0.93 0.96 0.87 0.94 0.94

J.E. Pickett et al. / Polymer Degradation and Stability 93 (2008) 684e691

690

temperatures a greater proportion of the products might be in the more thermodynamically stable, and presumably more highly colored, trans form. Fig. 3 shows that the plateau yellowing of PC increases with higher temperature. This was observed for all of the white formulations in this study. Interestingly, ABS and SAN had the highest activation energies for yellowing and also did not obey reciprocity in light intensity experiments [2]. 3.2. Fluorescent UV exposure The results for samples exposed in QUV Series #1 are shown in Table 10. Samples 1 and 8 had insufficient changes in properties to evaluate. The moisture effects will be discussed in detail in another report. The color shift data were independent of moisture and spray conditions and had R2 of at least 0.98 throughout. The Ea for color shifts were consistently higher than for either xenon arc condition. This undoubtedly has to do with the absence of visible light in the QUV that should allow accumulation of the more highly colored products. The rates of gloss loss were highly affected by the spraying although not much by humidity. However, the activation energies for gloss loss were relatively independent of spray. The activation energies for gloss loss are much lower than for the color shifts, although greater than seen for samples exposed to xenon arc. The data were generally noisier with R2 ranging from 0.91 to 0.99. This series contained a sample of white polypropylene, and a very high activation energy of 24 kcal/mol was observed for its color shift. The data for gloss loss were too scattered to evaluate. The activation energy photo-oxidation of low density polyethylene also has been reported to be relatively high: 14e15 kcal/mol [11]. The results for samples exposed in QUV Series #2 are shown in Table 11. Since most of these samples were black, only gloss loss was measured. Again, we found the Ea to be small, sometimes zero or even negative, for all samples. Frequency of water spray did affect the results because carbon black remains adhered to the surface as the polymer erodes resulting in very low gloss unless it is removed by natural wind

Table 11 Activation energies (kcal/mol) for gloss loss from QUV Series #2 #

Formulation

1 spray

3 spray

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

PC 0.6% carbon black PC/ASA/PMMA blend 1 black PC/ASA/PMMA blend 2 black PC/PBT blend 1 black PC/PBT blend 2 black PC/ABS blend black ABS black PC/PBT blend 3 black PC/PBT blend 4 black PC/PBT blend 5 black Nylon 6 0.5% carbon black PC 0.6% carbon black PC/BPA polyarylate gray 1 PC/BPA polyarylate gray 2 PC/BPA polyarylate black SAN 0.5% carbon black PBT 0.6% carbon black PC/PBT 0.6% carbon black PC/PBT/IM 0.6% carbon black PC-b-resorcinol polyarylate black PC-b-resorcinol polyarylate green PC-b-resorcinol polyarylate red PC-b-resorcinol polyarylate 2% TiO2 PC 0.6% carbon black

3 5 4 2 3 3 4 4

3 4 5 3 3 4 4 3 0

5

6

0 0

0 0

4 2 0 2 0 0 0 0 6

4 2 3 3 0 0 3 0 4

Underlined indicates poor superposition of curves. R2 > 0.90 unless in italics. Blanks indicate insufficient property change to establish relative rates.

and rain, forceful water spray, or sponge wiping. This series included a sample of nylon 6, which had an Ea of approximately 5e6 kcal/mol. Samples containing polyarylates based on bisphenol-A (BPA) or resorcinol usually had zero or small negative Ea in both xenon arc and fluorescent exposures. This probably is because the major photochemistry is a simple photo-Fries reaction on the backbone of the polymer and not the multi-step photo-oxidation of most polymers. Polypropylene is just the opposite e all oxidation and little intrinsic photochemistry e and seems to have an Ea > 20 kcal/mol, at least for yellowing. 4. Conclusions

Table 10 Activation energies (kcal/mol) for samples from QUV Series #1 #

Formulation

DGloss No spray

1 2 3 4 5 6 7 8 9 10 11 12

PC black dye þ UVA ASA white PC/PBT 3% TiO2 PC/ABS 3% TiO2 PC/PBT blend 2 black PC 2% TiO2 PC 2% TiO2 PC/ASA/PMMA blend black ABS 3% TiO2 PC/PBT 1 3% TiO2 PC/PBT blend 3 black Polypropylene 2% TiO2

DYI Sprayed

4 3 6 4 5 5

5 3 6 3 5 4

7 4 4

7 3 4

Dry

Humid

na

na

7 9 na 6 7 na 10 7 na 24

7 9 na 6 6 na 10 6 na 24

Blanks indicate insufficient property change to establish relative rates.

Overall we find that the activation energies for gloss loss and color shift of many aromatic engineering thermoplastics depend on the property measured and exposure conditions, although they vary over a fairly small range. The conditions most likely to be representative of outdoor exposure are those of the CIRA/sodalime-filtered xenon arc experiment. In this case, the Ea for gloss loss was 5 kcal/mol for all samples tested while the Ea for yellowing was also 5 kcal/mol except for SAN and ABS. Materials such as aromatic esters that undergo direct photochemical reactions seem to have very small or zero activation energies while those with a greater oxidation component have higher ones. A reaction with an Ea of 5 kcal/ mol will increase its rate by about 33% for each 10  C increase in temperature near room temperature. This makes temperature an important variable in polymer degradation and must be taken into account for accurate lifetime predictions.

J.E. Pickett et al. / Polymer Degradation and Stability 93 (2008) 684e691

Acknowledgements We thank Al Berzinis, Frank Hoefflin, and Patty Hubbard of GE Plastics (now SABIC Innovative Plastics) and Randy Carter, Pratima Rangaragan, and Vicki Watkins of GE Global Research for contributing samples and SABIC Innovative Plastics for permission to publish this work. Special thanks to Don Sorensen for modifications to the Ci4000 Weather-ometer. References

[4] [5]

[6]

[7] [8]

[1] Pickett JE. Highly predictive accelerated weathering of engineering thermoplastics. In: Martin JW, Ryntz RA, Dickie RA, editors. Service life prediction: challenging the status quo. Federation of societies for coatings technology; 2005. p. 93e106. [2] Pickett JE, Gibson DA, Gardner MM. Effects of irradiation conditions on the weathering of engineering thermoplastics. Polym Degrad Stab, submitted for publication. [3] Fischer RM, Ketola WD. Error analysis and associated risks for accelerated weathering results. In: Martin JW, Ryntz RA, Dickie RA, editors.

[9] [10] [11]

691

Service life prediction: challenging the status quo. Federation of societies for coatings technology; 2005. p. 79e92. Pickett JE, Gardner MM. Reproducibility of Florida weathering data. Polym Degrad Stab 2005;90:418e30. Fischer RM, Ketola WD. Surface temperatures of materials in exterior exposures and artificial accelerated tests. In: Ketola WD, Grossman D, editors. Accelerated and outdoor testing of organic materials. ASTM; 1994. p. 88e111. STP 1202. Gillen KT, Bernstein R, Celina M. Non-Arrhenius behavior for oxidative degradation of chlorosulfonated polyethylene materials. Polym Degrad Stab 2005;87:335e46. Simms JA. Acceleration shift factor and its use in evaluating weathering data. J Coat Technol 1987;59(748):45. Gillen KT, Clough RL. Timeetemperatureedose rate superposition: a methodology for extrapolating accelerated radiation aging data to low dose rate conditions. Polym Degrad Stab 1989;24:137. Gillen KT, Celina M. The wear-out approach for predicting the remaining lifetime of materials. Polym Degrad Stab 2001;71:15e30. Pickett JE. Reversible post-exposure yellowing of engineering thermoplastics. Polym Degrad Stab 2004;85:681e7. Cruz-Pinto JJC, Carvalho MES, Ferreira JFA. The kinetics and mechanism of polyethylene photo-oxidation. Angew Makromol Chem 1994;216:113e33.