- Email: [email protected]
Two new test methods for assessing environmental stress cracking of amorphous thermoplastics
Polymer Testing 15 (19%) 407-421 0 1996 Elsevier Science Ltd 0142-9418(95)00048-5
Printed in Great Britain. All rights reserved 0142-9418/!96/$15.00
ELSEVIER
MATERIAL PROPERTIES Two New Test Methods for Assessing Environmental Stress Cracking of Amorphous Thermoplastics M.C. Hough & D.C. Wright Rapra Technology
Ltd, Shawbury,
(Received
23 October
Shrewsbury,
SY4 4NE, UK
1995; accepted 23 November
1995)
ABSTRACT Two new test methods, monotonic creep and micro-hardness,
are used to study the
environmental stress crack (ESC) resistance of polymethyl methacrylate (PMMA) and unplasticised polyvinyl chloride (UPVC). Monotonic creep is shown to discriminate, to a high resolution and in the short term, the ESC resistance of polymer/fluid pairs,
including those polymer/Jluid pairs which exhibit mild/weak interactions.
Micro-hadness
is shown to offer
a cost-effective
polymer/fluid pairs for compatibility. Copyright 01996
method
of mass screening
Elsevier Science Ltd
1. INTRODUCTION Environmental stress cracking (ESC) has been the subject of extensive investigation for almost 50 years. It has deserved this attention because it has been responsible for an estimated 20% of all plastics product failures in service. In addition the phenomenon of ESC is very interesting to both chemists and physicists as it involves stress enhanced absorption, permeation, the thermodynamics of mixtures, local yielding, cavitation, fibrillation, and fracture. As a result of this interest and activity, many test methods have been developed. The most direct method compares the static creep rupture characteristic under tensile stress with the plastic in air and a fluid environment. As shown diagrammatically in Fig. 1, the influence of the fluid is only apparent after an ‘induction period’. For less aggressive ESC agents the induction period may be so long as to invite acceleration by testing at elevated temperatures.
408
M.C. Hough, D.C. Wright
1 DOE+00
l.OOE+Ol
1 .ooE+oz
1 .OOE+03
l.OOE+W
TIME TO FAILURE
1 .OOE+05
1 OOE+O6
1 .M)E+07
(seconds)
Fig. 1. Typi :al static creep rupture characteristic of a polymer in air and in contact with an environmental stress cracking fluid.
The monitoring of tensile creep strain is at least as expensive but in that it detects the apparent initiation of the phenomenon, rather than the final event (fracture), it has the advantage of reducing testing time scales by as much as a factor of 10. Although less direct than creep rupture it also provides a more rational safe design stress and strain. Typical results are shown in Fig. 2. A major problem with the constant stress creep and creep rupture methods, in addition to expense, is that without prior knowledge of the severity of attack, it is difficult to choose a stress level that will provide a positive result within a reasonable time scale. This difficulty, and the expense, increases as the severity of attack decreases and to a degree it is a problem shared by all test methods. As a consequence mild to moderate fluid polymer interactions 2.2 2 1.6 1.6 1.4 z z g
1.2 1 0.6 0.6 0.4 0.2 0 1 .OoE+OO
l.OOE+Ol
1.00E+02
1 .OOE+03
1 ce+O4
1 .OOE+05
1 OOE+ce
1.3OE+07
TIME (seconds)
Fig. 2. Typical tensile creep curves of a polymer in air and in contact with an environmental stress cracking fluid.
409
Two new test methods
are less well characterised than severe interactions. Paradoxically, it is the less severe interactions that tend to produce the most costly failures in that they occur after the point of sale and within a few years of service. For reasons of economy, the vast majority of archival data has been generated using relatively short-term tests involving bending beams under constant strain. Such tests frequently require subjective assessment and always suffer from the problem of unknown levels of stress relaxation. However even after 50 years of low cost testing the amount of archival data in the public domain is small compared with the potential number of plastic fluid/fluid mixture pairs. There is a need for lower cost test methods and/or an effective means of predicting the severity of the phenomenon for a given pair. These were the principle objectives of one project in a suite with the theme of ‘Polymer Degradation’ awarded by the UK’s Department of Trade and Industry. This paper concentrates on two new test methods emanating from that project.
2. MONOTONIC CREEP This is similar to the slow strain rate testing technique used for many years by the metals industry to assess stress corrosion cracking and hydrogen embrittlement (Ref. 1). However, here the strain response to a constant stressing rate is monitored. The departure of the stress-strain characteristic in air and in the fluid of interest as shown in Fig. 3 is taken to be the initiation of ESC and in this respect is similar to the assumption made in the static creep testing method (Fig. 2). For transparent plastics the departure coincided with the appearance of visible crazing. 40
2 25 SE g 20 L ti 15 10
0 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.6
0.9
1
1.1
1.2
1.3
1.4
1.5
16
STRAIN%
Fig. 3. Typical monotonic stress creep curves for a polymer in air and an environmental stress cracking fluid.
MC
410
Hough, D.C. Wright
The method as shown in Fig. 4 employs a tensile creep machine with the weight pan replaced by a blow moulded vessel of 75 litres capacity. Water is pulse fed into the container via a peristaltic pump. The rotational speed of the pump is continuously variable and this provides a means of applying a known stressing rate to the plastic specimen. The effective range of stressing rate was 0.1-10 MPa/h. Specimen strain is monitored via a Moire fringe extensometer employing ELECTRONICS
BASIC CREEP MACHINE
SPECIMEN AND EXTENSOMETER
Fig. 4. Monotonic creep testing machine.
Two new test methods
411
diffraction gratings of 250 lines per mm. With four photodiodes, 90” phase resolution and A/D conversion, this provides a digital electronic pulse for each micrometer of extensometer displacement. With a gauge length of 66 mm this translates to a recorded tensile strain increment of 0.000015. Within the gauge length as shown in Fig. 5, is incorporated a glass tube sealed with a grommet at each end. This contains the ESC fluid. 2.1 Results for a stressing rate of 4 MPa/h This rate was chosen for convenience as the slowest rate that ensures a result for all fluids, apart from the least aggressive within an 8 h working day (32 MPa). The test temperature was 20°C. Figures 6 and 7 show the stress-strain characteristics of PMMA and UPVC, respectively in air and 41 selected fluids. The resolution of the departure point is typically as shown in Fig. 8. The initiation time and strain and computed stress are listed in Tables 1 and 2. Various approaches were employed to correlate these results with known fluid parameters, ranging from total solubility parameter (Ref. 2) to combinations of partial solubility parameters and molar volumes (Ref. 3). Figures 9 and 10 display the hydrogen bonding and polar solubility parameter co-ordinate positions of each of the fluids as listed by Ref. 4. The number adjacent to each point is termed here the ‘severity index’ and is the time to initiate
DIFFRACTION
ENVIRONMENTAL CHAMBER
Fig. 5. Rapra Moird fringe extensometer
with environmental
chamber attached to specimen.
M.C. Hough, D.C. Wright
412 40
AIR CURVE
-I 0
0.2
0.4
06
0.6
1
1.2
1.4
1.6
STRAIN %
Fig. 6. Monotonic 4 MPalh stress-strain characteristics of PMMA in air and 41 fluids at 20°C.
10 -.“..--
___--_-0 0
0.2
0.4
0.6
0.6
1
1.2
1.4
16
1.6
2
2.2
2.4
STRAIN %
Fig. 7. Monotonic 4 MPa/h stress-strain characteristics of UPVC in air and 41 fluids at 2OT.
ESC (in seconds) divided by the molar volume raised to the power 213 (which approximates to the molar surface area). Contours of the severity index were found to provide the best correlation between fluid parameters and the severity of ESC. The predictive power of these contours will no doubt be assessed independently by others. Choosing Aniline, this according to Ref. 5 has polar and hydrogen bonding parameters of 5.1 and 10.2 (MPa)“.5. With reference to Figs 9 and 10, Aniline would be expected to have a severity index (Sindex)of approximately 200 for both PMMA and UPVC. Aniline has a molar volume
413
Two nezu test methods
I A
8.5 032
033
0.34
0.35 STRAIN
Fig. 8. Typical environmental
0.36
0.37
0.38
%
monotonic stress creep test departure point resolution.
of 91.5 cc, therefore the predicted time to the initiation of ESC at 4 MPa/h would be: Si,d,,200*(91 .5)“.66= 4060 s 2.2 Rate and temperature effects The stress and strain to initiate ESC should decrease with increasing temperature and decreasing stressing rate. This indeed was found to be the case. Figure 11 shows the influence of rate and temperature and Fig. 12 shows the data normalised at ambient temperature by time temperature superposition. The quality of the data would indicate (as with the contour plots) that the method has utility. Analysis of the results and in particular the development of a viscoelastic model for predicting long-term ESC behaviour under static stress or strain via short term results obtained via this method will be published at a later date.
3. MICRO-HARDNESS
MEASUREMENT
Interest in micro-hardness as a means of detecting degradation of ,polymer surfaces has increased over the last few years with reports of its use to assess chemical compatibility (Ref. 6) and UV resistance (Ref. 7). Its fundamental attraction is the shallow nominal depth of penetration, which can be as low as a few micrometers. The principle is simple; a pyramidal diamond indentor with a facing angle of 136” is impressed on the surface of the plastic at a known load and for a known time. An example of the impression is shown
414
M.C. Hough, D.C. Wright TABLE 1 PMMA initiation data at 4 MPaJh and 20”
Environment
Initiation time (s)
Initiation stress (MPa)
Initiation strain (%)
MEK CHLOROFORM ACETONE ETHYL ACETATE CHLOROBENZENE BENZENE BENZYL ALCOHOL 2-METHOXY ETHANOL TOLUENE 2-ETHOXY ETHANOL DI ACETONE ALCOHOL N-BUTYL LACTATE BUTYL ACETATE METHYL SALICYLATE BUTAN-2-OL DI-ETHYL ETHER 0-XYLENE 2-BUTOXY ETHANOL PROPAN-ZOL 1, 1,l -TRICHLORO ETHANE AMYL ACETATE ETHANOL PROPAN- 1-0L N-BUTANOL DBE TETRA CHLORO ETHYLENE 2-METHYL PROPAN- l-OL TRI-N-BUTYL-O-PHOSPHATE CARBON TETRA CHLORIDE DOP DNBP DECAN- l-OL DOS N-DECANE PROPYLENEGLYCOL CYCLOHEXANOL METHANOL ETHANE DIOL CYCLOHEXANE N-DODECANE WATER
9.00E+Ol 1.80EtO2 4SOE+02 4SOE+02 7.2OE+O2 1.62EtO3
0.10 0.20 0.50 0.50 0.80 I .80 2.10 2.40 2.50 3.00 4.20 4.20 4.30 4.50 4.80 5.00 5.50 5.90 6.00 6.00 6.30 6.50 7.00 8.50 IO.00 10.00 10.50 11.80 12.00 12.30 17.00 18.00 18.00 18.50 20.00 22.50 23.00 26.00 27.50 29.50 36.00
0.001 0.010 0.015 0.015 0.028 0.050 0.070 0.080 0.080 0.100 0.140 0.140 0.143 0.150 0.170 0. I70 0.190 0.200 0.210 0.210 0.220 0.220 0.240 0.300 0.350 0.350 0.370 0.420 0.430 0.440 0.640 0.660 0.660 0.680 0.760 0.860 0.880 1.020 1.100 1.200 1.600
1.89E+O3 2.16E+03 2.25E+O3 2.70E+03 3.78E+O3 3.78E+O3 3.87E+O3 4.05E+03 4.32E+03 4.50E+03 4.95E+O3 5.31E+O3 5.40E+03 5.40E+03 5.67E+O3 5.8.5E+03 6.30E+O3 7.65E+03 9.00E+03 9.00E+03 9.45E+O3 l.O6E+04 l.O8E+04 l.llE+O4 1.53E+O4
1.62E+04 1.62E+04 1.67E+O4 1.80E+O4 2.03E+04 2.07E+04 2.34E+O4 2.48E+O4 2.66E+04 3.24E+O4
415
Two new test methods TABLE 2 UPVC initiation data at 4 MPa/h and 20°C Environment
MEK CHLOROFORM ACETONE ETHYL ACETATE CHLOROBENZENE BENZENE BUTYL ACETATE DBE AMYL ACETATE DI-ETHYL ETHER TOLUENE 2-ETHOXY ETHANOL METHYL SALICYLATE N-BUTYL LACTATE TRI-N-BUTYL-O-PHOSPHATE 2-METHOXY ETHANOL 1,l, 1-TRICHLORO ETHANE DI ACETONE ALCOHOL DNBP TETRA CHLORO ETHYLENE 2-BUTOXY ETHANOL BENZYL ALCOHOL 0-XYLENE CARBON TETRA CHLORIDE DECAN- l-OL DOP PROPAN-ZOL DOS PROPAN- 1-0L CYCLOHEXANE N-BUTANOL BUTAN-2-OL ETHANOL 2-METHYL PROPAN- I-OL METHANOL N-DECANE CYCLOHEXANOL N-DODECANE PROPYLENEGLYCOL ETHANE DIOL WATER
Initiation
Initiation
time
stress
strain
(s)
fMPa)
@)
Initiation
3.6OE+O2 4SOE+O2 9.OOE+O2 9.00E+02 9.90E+O2 1.35E+O3 1.3.5E+03 1.80E+03 2.25E+O3 2.25E+O3 2.25E+03 2.61E+O3 2.70E+O3 2.70E+O3 2.70E+O3 4.05E+O3 4.50E+03 5.40E+03 5.40E+O3 6.3OE+O3 6.48E+O3 7.20E+03 7.56E+O3 9.18E+O3 1.31E+04 1.35E+04 1.35E+O4 lS3E+O4 1.80E+04 2.07E+O4 2.07E+O4 2.16E+04 2.16E+04 2.25E+O4 2.34E+O4 2.39E+O4 2.43E+Q4 2.61E+O4 3.338+04 3.38E+04 4.59E+O4
0.40 0.50 1.oo 1.00 1.10 1.50 1so 2.00 2.50 2.50 2.50 2.90 3.00 3.00 3.00 4.50 5.00 6.00 6.00 7.00 7.20 8.00 8.40 10.20 14.50 15.00 15.00 17.00 20.00 23.00 23.00 24.00 24.00 25.00 26.00 26.50 27.00 29.00 37.00 37.50 51.00
0.02 0.02 0.03 0.03 0.04 0.05 0.05 0.06 0.08 0.08 0.08 0.09 0.09 0.09 0.09 0.13 0.15 0.17 0.17 0.20 0.21 0.23 0.24 0.30 0.41 0.44 0.44 0.50 0.60 0.70 0.70 0.74 0.74 0.78 0.80 0.83 0.85 0.92 1.26 I .30 2.50
M.C. Hough, D.C. Wright
416
0 0494 1 710
2
3
4
5
7
6
6
9
10
11
12
13
14
15
16
POLAR BONDING PARAMETER Wa”O.5
Fig. 9. PMMA ‘Severity index’ for various fluids on the hydrogen bonding versus polar bond-
ing plot.
1 0% 701
2
3
4
5
6
7
6
9
10
11
12
13
14
15
16
POLAR BONMNG PARAMETER W’a”0.5
Fig. 10. PVC ‘Severity index’ for various fluids on the hydrogen bonding versus polar bonding
plot.
in Fig. 13. The area of the impression after removal of the indentor is calculated from the measurement of its mean diagonal length (d in mm) using an in-built microscope and XY micrometer adjustment of the specimen table. The micro-hardness number is given by: MH=
1.854xF d2
where F is the indentor load expressed in Kgf. The instrument used was a Vickers model MXT30 supplied by Matsuzawa Seiko Co. Ltd.
Two new test methods 30
0 0
0.2
0.4
0.8
0.6 % STRAIN TO INMATE
Fig. 11. PMMA monotonic
1
1.2
1.4
DAMAGE
creep stress versus strain to initiate damage in cyclohexane different stress rates and temperatures.
at
1 WE+02
1.OOE+OO 1 WE+03
100E+O4
1.xmO5
1.00E+O6
DAMAGE INITIATIGN
TIM
1 .mE+o7
1.ooE+o8
1 .00E+OQ
(seconds)
Fig. 12. PMMAkyclohexane
monotonic creep damage initiation stress versus damage initiation time at various temperatures and time temperature shift of monotonic creep data to 2OT.
A most important test variable with regards to the detection of surface effects was as expected, the indentor load. The lower the load, the shallower the impression and the more sensitive the test with regards to the detection of surface effects. For this reason the lowest available load (0.01 Kgf) was adopted. The MH number was derived in air and after increasing periods of time with the tested surface exposed to a wide range of fluids. 3.1 ABS and di-propylene
glycol mono-methyl-ether
(DPM)
Figure 14 shows two repeat test runs for an ABS in contact with DPM. This fluid plastic pair was of particular interest because it was known to produce very rapid stress cracking (within 60 s at 0.5% surface strain) yet it required
418
M.C. Hough, D.C. Wright
Fig. 13. Impression of a microhardness indentor on the surface of a piece of plastic.
14 days of stress free immersion before any significant weight change could be detected. As shown in Fig. 15, the reduction in micro-hardness with time of fluid contact would appear to be continuous and proportional to time raised to the power 1.5. On this basis after 60 s of contact the micro-hardness number
TWOnew fesf methods
1 WE+01
l.OOE+00
1 .DOE+02
100E+O3
[email protected]+O4
1 .M)E+O5
1.00E+M
1 WE+07
Exposure Time bl
Fig. 14. Environmental
OL-
zero stress microhardness,
,
I
oODE+@O MOE+05
l.OOE+O6
1.5OE+C6 Z.M)E+O6 Ewo.slJRE
Fig. 15. Reduction
of micro-hardness
ABS in DPM.
! .J
I
2.5OE+C6 3.OOE+O6 35OE+C6
4.OOE+06 4.50E+f)6
5.COE+O6
TIME “1.6 (seconbs)
of ABS with time of fluid contact 1.5.
raised to the power
would have decreased by only 0.0006 and the penetration depth increased by only 0.014 micrometers. This of course does not indicate the degree of interaction at a craze site under stress (stress enhanced absorption). The first detectable reduction in micro-hardness occurs after ca. 2000 s, compared with 14 days for weight pick-up. 3.2 UPVC and PMMA in a range of fluids The 41 fluids used to investigate the monotonic creep method were also seiected here to investigate the general utility of the micro-hardness test method.
M.C. Hough, D.C. Wright
420
l.ODE+OZ Ll.OOE+OO
l.COE+Ol
l.CCE+OZ
l.WEtO3
TIME TO INmATE
SOFlENlNG
l.OOE+O4
l.OOE+OS
i.OOE+OB
USING MlCRO-HMDNE§S
f.OOE+O7
l.OOE+OE
(-da)
Fig. 16. PVC time to initiate damage using monotonic creep versus time to soften using micro-
hardness in a range of fluids at 20°C.
Figures 16 and 17 show the relationship between time to initiate damage using monotonic creep and micro-hardness for UPVC and PMMA respectively. For UPVC the correlation is quite good with monotonic initiation taking approximately 2 decades longer to detect as changes in surface hardness for severe fluids. The moderate and mild fluids, however, are detected approximately 2.5 decades earlier by monotonic testing. For PMMA the correlation is comparatively poor. This is probably due to the many fluids that induce weight loss in PMMA, via oligomer extraction, resulting in a loss of surface material which would intefere with the softening initiation as assessed via micro-harness. Monotonic initiation for PMMA is l.OOE+O5
4
l.CQE+Ol 1.wE+oz
i.MJE+O3
l.CCiXM
llME TO INKLATE WFlENlNG
l.ooE+o5
USING MCRO-HARDNESS
Fig. 17. PMMA time to initiate damage using monotonic
micro-hardness
l.WE+06
1.Yx+o7
1.M)E+O8
(-d5)
creep versus time to soften using in a range of fluids at 20°C.
Two new test methods
421
approximately 0.5 of a decade earlier to detect as changes in surface hardness for severe fluids. The moderate and mild fluids, as with the UPVC, are detected approximately 2.5 decades earlier by monotonic testing.
4. CONCLUSIONS The monotonic creep method is capable of high resolution and discrimination as to the relative stress cracking severity of fluids. It is of particular value for the accelerated testing of fluids which have a mild interaction with a plastic. These include, for example, oils and greases which can induce fracture in a few years at ambient temperature and recommended design stresses. The fact that the method generates critical time, critical stress and critical strain invites the use of the method for investigating viscoelastic criterion for initiation of the ESC phenomenon. The microhardness method has the potential for mass screening of plastic/fluid compatibility, including extraction as well as absorption and should be of interest to polymer suppliers.
REFERENCES 1. Kane R. D., ASTM Standardization News, 35, May 1993. 2. Kambour, R. P. & Gruner, C. L., Journal of Polymer Science, Pol. Phys. Ed., 16 (1978) 703 3. Jacques, b. H. M. & Wyzgoski, M. G., Journal of Applied Polymer Science, 23 (1979) 1153 4. Brandrup, J. & Immergut, E. H., Polymer Handbook 3rd Edition, Wiley Interscience, 1989. 5. Huang, J. C. & Wang, M. S., Adv. Polym. Tech., 239 (1989). 6. Pandy, S., Bajpai, R. & Datt, S. C., Polymer Testing, 10 (1991) 116 7. Gonzales, A., Pastor, J. M. & De Seja, J. A., Journal of Applied Polymer Science, 38 (1989)
1879
Printed in Great Britain. All rights reserved 0142-9418/!96/$15.00
ELSEVIER
MATERIAL PROPERTIES Two New Test Methods for Assessing Environmental Stress Cracking of Amorphous Thermoplastics M.C. Hough & D.C. Wright Rapra Technology
Ltd, Shawbury,
(Received
23 October
Shrewsbury,
SY4 4NE, UK
1995; accepted 23 November
1995)
ABSTRACT Two new test methods, monotonic creep and micro-hardness,
are used to study the
environmental stress crack (ESC) resistance of polymethyl methacrylate (PMMA) and unplasticised polyvinyl chloride (UPVC). Monotonic creep is shown to discriminate, to a high resolution and in the short term, the ESC resistance of polymer/fluid pairs,
including those polymer/Jluid pairs which exhibit mild/weak interactions.
Micro-hadness
is shown to offer
a cost-effective
polymer/fluid pairs for compatibility. Copyright 01996
method
of mass screening
Elsevier Science Ltd
1. INTRODUCTION Environmental stress cracking (ESC) has been the subject of extensive investigation for almost 50 years. It has deserved this attention because it has been responsible for an estimated 20% of all plastics product failures in service. In addition the phenomenon of ESC is very interesting to both chemists and physicists as it involves stress enhanced absorption, permeation, the thermodynamics of mixtures, local yielding, cavitation, fibrillation, and fracture. As a result of this interest and activity, many test methods have been developed. The most direct method compares the static creep rupture characteristic under tensile stress with the plastic in air and a fluid environment. As shown diagrammatically in Fig. 1, the influence of the fluid is only apparent after an ‘induction period’. For less aggressive ESC agents the induction period may be so long as to invite acceleration by testing at elevated temperatures.
408
M.C. Hough, D.C. Wright
1 DOE+00
l.OOE+Ol
1 .ooE+oz
1 .OOE+03
l.OOE+W
TIME TO FAILURE
1 .OOE+05
1 OOE+O6
1 .M)E+07
(seconds)
Fig. 1. Typi :al static creep rupture characteristic of a polymer in air and in contact with an environmental stress cracking fluid.
The monitoring of tensile creep strain is at least as expensive but in that it detects the apparent initiation of the phenomenon, rather than the final event (fracture), it has the advantage of reducing testing time scales by as much as a factor of 10. Although less direct than creep rupture it also provides a more rational safe design stress and strain. Typical results are shown in Fig. 2. A major problem with the constant stress creep and creep rupture methods, in addition to expense, is that without prior knowledge of the severity of attack, it is difficult to choose a stress level that will provide a positive result within a reasonable time scale. This difficulty, and the expense, increases as the severity of attack decreases and to a degree it is a problem shared by all test methods. As a consequence mild to moderate fluid polymer interactions 2.2 2 1.6 1.6 1.4 z z g
1.2 1 0.6 0.6 0.4 0.2 0 1 .OoE+OO
l.OOE+Ol
1.00E+02
1 .OOE+03
1 ce+O4
1 .OOE+05
1 OOE+ce
1.3OE+07
TIME (seconds)
Fig. 2. Typical tensile creep curves of a polymer in air and in contact with an environmental stress cracking fluid.
409
Two new test methods
are less well characterised than severe interactions. Paradoxically, it is the less severe interactions that tend to produce the most costly failures in that they occur after the point of sale and within a few years of service. For reasons of economy, the vast majority of archival data has been generated using relatively short-term tests involving bending beams under constant strain. Such tests frequently require subjective assessment and always suffer from the problem of unknown levels of stress relaxation. However even after 50 years of low cost testing the amount of archival data in the public domain is small compared with the potential number of plastic fluid/fluid mixture pairs. There is a need for lower cost test methods and/or an effective means of predicting the severity of the phenomenon for a given pair. These were the principle objectives of one project in a suite with the theme of ‘Polymer Degradation’ awarded by the UK’s Department of Trade and Industry. This paper concentrates on two new test methods emanating from that project.
2. MONOTONIC CREEP This is similar to the slow strain rate testing technique used for many years by the metals industry to assess stress corrosion cracking and hydrogen embrittlement (Ref. 1). However, here the strain response to a constant stressing rate is monitored. The departure of the stress-strain characteristic in air and in the fluid of interest as shown in Fig. 3 is taken to be the initiation of ESC and in this respect is similar to the assumption made in the static creep testing method (Fig. 2). For transparent plastics the departure coincided with the appearance of visible crazing. 40
2 25 SE g 20 L ti 15 10
0 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.6
0.9
1
1.1
1.2
1.3
1.4
1.5
16
STRAIN%
Fig. 3. Typical monotonic stress creep curves for a polymer in air and an environmental stress cracking fluid.
MC
410
Hough, D.C. Wright
The method as shown in Fig. 4 employs a tensile creep machine with the weight pan replaced by a blow moulded vessel of 75 litres capacity. Water is pulse fed into the container via a peristaltic pump. The rotational speed of the pump is continuously variable and this provides a means of applying a known stressing rate to the plastic specimen. The effective range of stressing rate was 0.1-10 MPa/h. Specimen strain is monitored via a Moire fringe extensometer employing ELECTRONICS
BASIC CREEP MACHINE
SPECIMEN AND EXTENSOMETER
Fig. 4. Monotonic creep testing machine.
Two new test methods
411
diffraction gratings of 250 lines per mm. With four photodiodes, 90” phase resolution and A/D conversion, this provides a digital electronic pulse for each micrometer of extensometer displacement. With a gauge length of 66 mm this translates to a recorded tensile strain increment of 0.000015. Within the gauge length as shown in Fig. 5, is incorporated a glass tube sealed with a grommet at each end. This contains the ESC fluid. 2.1 Results for a stressing rate of 4 MPa/h This rate was chosen for convenience as the slowest rate that ensures a result for all fluids, apart from the least aggressive within an 8 h working day (32 MPa). The test temperature was 20°C. Figures 6 and 7 show the stress-strain characteristics of PMMA and UPVC, respectively in air and 41 selected fluids. The resolution of the departure point is typically as shown in Fig. 8. The initiation time and strain and computed stress are listed in Tables 1 and 2. Various approaches were employed to correlate these results with known fluid parameters, ranging from total solubility parameter (Ref. 2) to combinations of partial solubility parameters and molar volumes (Ref. 3). Figures 9 and 10 display the hydrogen bonding and polar solubility parameter co-ordinate positions of each of the fluids as listed by Ref. 4. The number adjacent to each point is termed here the ‘severity index’ and is the time to initiate
DIFFRACTION
ENVIRONMENTAL CHAMBER
Fig. 5. Rapra Moird fringe extensometer
with environmental
chamber attached to specimen.
M.C. Hough, D.C. Wright
412 40
AIR CURVE
-I 0
0.2
0.4
06
0.6
1
1.2
1.4
1.6
STRAIN %
Fig. 6. Monotonic 4 MPalh stress-strain characteristics of PMMA in air and 41 fluids at 20°C.
10 -.“..--
___--_-0 0
0.2
0.4
0.6
0.6
1
1.2
1.4
16
1.6
2
2.2
2.4
STRAIN %
Fig. 7. Monotonic 4 MPa/h stress-strain characteristics of UPVC in air and 41 fluids at 2OT.
ESC (in seconds) divided by the molar volume raised to the power 213 (which approximates to the molar surface area). Contours of the severity index were found to provide the best correlation between fluid parameters and the severity of ESC. The predictive power of these contours will no doubt be assessed independently by others. Choosing Aniline, this according to Ref. 5 has polar and hydrogen bonding parameters of 5.1 and 10.2 (MPa)“.5. With reference to Figs 9 and 10, Aniline would be expected to have a severity index (Sindex)of approximately 200 for both PMMA and UPVC. Aniline has a molar volume
413
Two nezu test methods
I A
8.5 032
033
0.34
0.35 STRAIN
Fig. 8. Typical environmental
0.36
0.37
0.38
%
monotonic stress creep test departure point resolution.
of 91.5 cc, therefore the predicted time to the initiation of ESC at 4 MPa/h would be: Si,d,,200*(91 .5)“.66= 4060 s 2.2 Rate and temperature effects The stress and strain to initiate ESC should decrease with increasing temperature and decreasing stressing rate. This indeed was found to be the case. Figure 11 shows the influence of rate and temperature and Fig. 12 shows the data normalised at ambient temperature by time temperature superposition. The quality of the data would indicate (as with the contour plots) that the method has utility. Analysis of the results and in particular the development of a viscoelastic model for predicting long-term ESC behaviour under static stress or strain via short term results obtained via this method will be published at a later date.
3. MICRO-HARDNESS
MEASUREMENT
Interest in micro-hardness as a means of detecting degradation of ,polymer surfaces has increased over the last few years with reports of its use to assess chemical compatibility (Ref. 6) and UV resistance (Ref. 7). Its fundamental attraction is the shallow nominal depth of penetration, which can be as low as a few micrometers. The principle is simple; a pyramidal diamond indentor with a facing angle of 136” is impressed on the surface of the plastic at a known load and for a known time. An example of the impression is shown
414
M.C. Hough, D.C. Wright TABLE 1 PMMA initiation data at 4 MPaJh and 20”
Environment
Initiation time (s)
Initiation stress (MPa)
Initiation strain (%)
MEK CHLOROFORM ACETONE ETHYL ACETATE CHLOROBENZENE BENZENE BENZYL ALCOHOL 2-METHOXY ETHANOL TOLUENE 2-ETHOXY ETHANOL DI ACETONE ALCOHOL N-BUTYL LACTATE BUTYL ACETATE METHYL SALICYLATE BUTAN-2-OL DI-ETHYL ETHER 0-XYLENE 2-BUTOXY ETHANOL PROPAN-ZOL 1, 1,l -TRICHLORO ETHANE AMYL ACETATE ETHANOL PROPAN- 1-0L N-BUTANOL DBE TETRA CHLORO ETHYLENE 2-METHYL PROPAN- l-OL TRI-N-BUTYL-O-PHOSPHATE CARBON TETRA CHLORIDE DOP DNBP DECAN- l-OL DOS N-DECANE PROPYLENEGLYCOL CYCLOHEXANOL METHANOL ETHANE DIOL CYCLOHEXANE N-DODECANE WATER
9.00E+Ol 1.80EtO2 4SOE+02 4SOE+02 7.2OE+O2 1.62EtO3
0.10 0.20 0.50 0.50 0.80 I .80 2.10 2.40 2.50 3.00 4.20 4.20 4.30 4.50 4.80 5.00 5.50 5.90 6.00 6.00 6.30 6.50 7.00 8.50 IO.00 10.00 10.50 11.80 12.00 12.30 17.00 18.00 18.00 18.50 20.00 22.50 23.00 26.00 27.50 29.50 36.00
0.001 0.010 0.015 0.015 0.028 0.050 0.070 0.080 0.080 0.100 0.140 0.140 0.143 0.150 0.170 0. I70 0.190 0.200 0.210 0.210 0.220 0.220 0.240 0.300 0.350 0.350 0.370 0.420 0.430 0.440 0.640 0.660 0.660 0.680 0.760 0.860 0.880 1.020 1.100 1.200 1.600
1.89E+O3 2.16E+03 2.25E+O3 2.70E+03 3.78E+O3 3.78E+O3 3.87E+O3 4.05E+03 4.32E+03 4.50E+03 4.95E+O3 5.31E+O3 5.40E+03 5.40E+03 5.67E+O3 5.8.5E+03 6.30E+O3 7.65E+03 9.00E+03 9.00E+03 9.45E+O3 l.O6E+04 l.O8E+04 l.llE+O4 1.53E+O4
1.62E+04 1.62E+04 1.67E+O4 1.80E+O4 2.03E+04 2.07E+04 2.34E+O4 2.48E+O4 2.66E+04 3.24E+O4
415
Two new test methods TABLE 2 UPVC initiation data at 4 MPa/h and 20°C Environment
MEK CHLOROFORM ACETONE ETHYL ACETATE CHLOROBENZENE BENZENE BUTYL ACETATE DBE AMYL ACETATE DI-ETHYL ETHER TOLUENE 2-ETHOXY ETHANOL METHYL SALICYLATE N-BUTYL LACTATE TRI-N-BUTYL-O-PHOSPHATE 2-METHOXY ETHANOL 1,l, 1-TRICHLORO ETHANE DI ACETONE ALCOHOL DNBP TETRA CHLORO ETHYLENE 2-BUTOXY ETHANOL BENZYL ALCOHOL 0-XYLENE CARBON TETRA CHLORIDE DECAN- l-OL DOP PROPAN-ZOL DOS PROPAN- 1-0L CYCLOHEXANE N-BUTANOL BUTAN-2-OL ETHANOL 2-METHYL PROPAN- I-OL METHANOL N-DECANE CYCLOHEXANOL N-DODECANE PROPYLENEGLYCOL ETHANE DIOL WATER
Initiation
Initiation
time
stress
strain
(s)
fMPa)
@)
Initiation
3.6OE+O2 4SOE+O2 9.OOE+O2 9.00E+02 9.90E+O2 1.35E+O3 1.3.5E+03 1.80E+03 2.25E+O3 2.25E+O3 2.25E+03 2.61E+O3 2.70E+O3 2.70E+O3 2.70E+O3 4.05E+O3 4.50E+03 5.40E+03 5.40E+O3 6.3OE+O3 6.48E+O3 7.20E+03 7.56E+O3 9.18E+O3 1.31E+04 1.35E+04 1.35E+O4 lS3E+O4 1.80E+04 2.07E+O4 2.07E+O4 2.16E+04 2.16E+04 2.25E+O4 2.34E+O4 2.39E+O4 2.43E+Q4 2.61E+O4 3.338+04 3.38E+04 4.59E+O4
0.40 0.50 1.oo 1.00 1.10 1.50 1so 2.00 2.50 2.50 2.50 2.90 3.00 3.00 3.00 4.50 5.00 6.00 6.00 7.00 7.20 8.00 8.40 10.20 14.50 15.00 15.00 17.00 20.00 23.00 23.00 24.00 24.00 25.00 26.00 26.50 27.00 29.00 37.00 37.50 51.00
0.02 0.02 0.03 0.03 0.04 0.05 0.05 0.06 0.08 0.08 0.08 0.09 0.09 0.09 0.09 0.13 0.15 0.17 0.17 0.20 0.21 0.23 0.24 0.30 0.41 0.44 0.44 0.50 0.60 0.70 0.70 0.74 0.74 0.78 0.80 0.83 0.85 0.92 1.26 I .30 2.50
M.C. Hough, D.C. Wright
416
0 0494 1 710
2
3
4
5
7
6
6
9
10
11
12
13
14
15
16
POLAR BONDING PARAMETER Wa”O.5
Fig. 9. PMMA ‘Severity index’ for various fluids on the hydrogen bonding versus polar bond-
ing plot.
1 0% 701
2
3
4
5
6
7
6
9
10
11
12
13
14
15
16
POLAR BONMNG PARAMETER W’a”0.5
Fig. 10. PVC ‘Severity index’ for various fluids on the hydrogen bonding versus polar bonding
plot.
in Fig. 13. The area of the impression after removal of the indentor is calculated from the measurement of its mean diagonal length (d in mm) using an in-built microscope and XY micrometer adjustment of the specimen table. The micro-hardness number is given by: MH=
1.854xF d2
where F is the indentor load expressed in Kgf. The instrument used was a Vickers model MXT30 supplied by Matsuzawa Seiko Co. Ltd.
Two new test methods 30
0 0
0.2
0.4
0.8
0.6 % STRAIN TO INMATE
Fig. 11. PMMA monotonic
1
1.2
1.4
DAMAGE
creep stress versus strain to initiate damage in cyclohexane different stress rates and temperatures.
at
1 WE+02
1.OOE+OO 1 WE+03
100E+O4
1.xmO5
1.00E+O6
DAMAGE INITIATIGN
TIM
1 .mE+o7
1.ooE+o8
1 .00E+OQ
(seconds)
Fig. 12. PMMAkyclohexane
monotonic creep damage initiation stress versus damage initiation time at various temperatures and time temperature shift of monotonic creep data to 2OT.
A most important test variable with regards to the detection of surface effects was as expected, the indentor load. The lower the load, the shallower the impression and the more sensitive the test with regards to the detection of surface effects. For this reason the lowest available load (0.01 Kgf) was adopted. The MH number was derived in air and after increasing periods of time with the tested surface exposed to a wide range of fluids. 3.1 ABS and di-propylene
glycol mono-methyl-ether
(DPM)
Figure 14 shows two repeat test runs for an ABS in contact with DPM. This fluid plastic pair was of particular interest because it was known to produce very rapid stress cracking (within 60 s at 0.5% surface strain) yet it required
418
M.C. Hough, D.C. Wright
Fig. 13. Impression of a microhardness indentor on the surface of a piece of plastic.
14 days of stress free immersion before any significant weight change could be detected. As shown in Fig. 15, the reduction in micro-hardness with time of fluid contact would appear to be continuous and proportional to time raised to the power 1.5. On this basis after 60 s of contact the micro-hardness number
TWOnew fesf methods
1 WE+01
l.OOE+00
1 .DOE+02
100E+O3
[email protected]+O4
1 .M)E+O5
1.00E+M
1 WE+07
Exposure Time bl
Fig. 14. Environmental
OL-
zero stress microhardness,
,
I
oODE+@O MOE+05
l.OOE+O6
1.5OE+C6 Z.M)E+O6 Ewo.slJRE
Fig. 15. Reduction
of micro-hardness
ABS in DPM.
! .J
I
2.5OE+C6 3.OOE+O6 35OE+C6
4.OOE+06 4.50E+f)6
5.COE+O6
TIME “1.6 (seconbs)
of ABS with time of fluid contact 1.5.
raised to the power
would have decreased by only 0.0006 and the penetration depth increased by only 0.014 micrometers. This of course does not indicate the degree of interaction at a craze site under stress (stress enhanced absorption). The first detectable reduction in micro-hardness occurs after ca. 2000 s, compared with 14 days for weight pick-up. 3.2 UPVC and PMMA in a range of fluids The 41 fluids used to investigate the monotonic creep method were also seiected here to investigate the general utility of the micro-hardness test method.
M.C. Hough, D.C. Wright
420
l.ODE+OZ Ll.OOE+OO
l.COE+Ol
l.CCE+OZ
l.WEtO3
TIME TO INmATE
SOFlENlNG
l.OOE+O4
l.OOE+OS
i.OOE+OB
USING MlCRO-HMDNE§S
f.OOE+O7
l.OOE+OE
(-da)
Fig. 16. PVC time to initiate damage using monotonic creep versus time to soften using micro-
hardness in a range of fluids at 20°C.
Figures 16 and 17 show the relationship between time to initiate damage using monotonic creep and micro-hardness for UPVC and PMMA respectively. For UPVC the correlation is quite good with monotonic initiation taking approximately 2 decades longer to detect as changes in surface hardness for severe fluids. The moderate and mild fluids, however, are detected approximately 2.5 decades earlier by monotonic testing. For PMMA the correlation is comparatively poor. This is probably due to the many fluids that induce weight loss in PMMA, via oligomer extraction, resulting in a loss of surface material which would intefere with the softening initiation as assessed via micro-harness. Monotonic initiation for PMMA is l.OOE+O5
4
l.CQE+Ol 1.wE+oz
i.MJE+O3
l.CCiXM
llME TO INKLATE WFlENlNG
l.ooE+o5
USING MCRO-HARDNESS
Fig. 17. PMMA time to initiate damage using monotonic
micro-hardness
l.WE+06
1.Yx+o7
1.M)E+O8
(-d5)
creep versus time to soften using in a range of fluids at 20°C.
Two new test methods
421
approximately 0.5 of a decade earlier to detect as changes in surface hardness for severe fluids. The moderate and mild fluids, as with the UPVC, are detected approximately 2.5 decades earlier by monotonic testing.
4. CONCLUSIONS The monotonic creep method is capable of high resolution and discrimination as to the relative stress cracking severity of fluids. It is of particular value for the accelerated testing of fluids which have a mild interaction with a plastic. These include, for example, oils and greases which can induce fracture in a few years at ambient temperature and recommended design stresses. The fact that the method generates critical time, critical stress and critical strain invites the use of the method for investigating viscoelastic criterion for initiation of the ESC phenomenon. The microhardness method has the potential for mass screening of plastic/fluid compatibility, including extraction as well as absorption and should be of interest to polymer suppliers.
REFERENCES 1. Kane R. D., ASTM Standardization News, 35, May 1993. 2. Kambour, R. P. & Gruner, C. L., Journal of Polymer Science, Pol. Phys. Ed., 16 (1978) 703 3. Jacques, b. H. M. & Wyzgoski, M. G., Journal of Applied Polymer Science, 23 (1979) 1153 4. Brandrup, J. & Immergut, E. H., Polymer Handbook 3rd Edition, Wiley Interscience, 1989. 5. Huang, J. C. & Wang, M. S., Adv. Polym. Tech., 239 (1989). 6. Pandy, S., Bajpai, R. & Datt, S. C., Polymer Testing, 10 (1991) 116 7. Gonzales, A., Pastor, J. M. & De Seja, J. A., Journal of Applied Polymer Science, 38 (1989)
1879