Polymers/Organic Materials Aging Overview
Robert Bernstein
Sandia National Laboratories
Organic Materials Department
Albuquerque, NM
rbernst@sandia.gov
505-284-3690
179th Technical Meeting Rubber Division
April 18-20, 2011
Akron, Ohio
Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company,
for the United States Department of Energy’s National Nuclear Security Administration under
contract DE-AC04-94AL85000.
1
Organic Materials Problems;
Organic Materials Aging and Degradation
Nuclear Power Plant Cable Insulation
O-rings
Shorting Plugs
13
CH2
CH
CH2
CH
CH3
CH3
n
n
13
CH2
CH
CH2
CH
13
CH3
CH3
n
Labeled Polymers
n
Textiles
2
Temperature
‘Accelerated Aging’
Reaction Coordinate
3
General Approach/Goals
Macroscopic level
Molecular Level
Physical property
Sealing Force
Chemical Property
Compression Set
Tensile Strength
Permeation
Elongation
Additives
Dimensional changes
Goals
• Prediction of physical properties vs. time
• Predict remaining lifetime of field materials
• Develop condition monitoring method
4
Deception!
Conclusions derived from initial high temperature,
short duration (even out to 1 year) accelerated
aging can be misleading.
Chemistry / mechanisms must be understood.
Results must be critically analyzed to identify and
understand mechanism changes
5
Thermal-oxidative Aging: Nylon
Average % tensile strength remaining
100
90
80
70
60
138 °C
138 °C 2nd Spool
124 °C
124 °C 2nd Run
124 °C 2nd Spool
109 °C
99 °C
95° C
95 °C 2nd Run
80 °C
64 °C
48 °C
37 °C
50
40
30
20
10
0
1
10
100
1000
Time in days at temperature
Bernstein, R.; Gillen, K. T. Polymer Degradation and Stability, Nylon 6.6 accelerating aging studies: II. Longterm thermal-oxidative and hydrolysis results 2010, 95, 1471-1479.
2000 days ~ 5.5 years
6
7
Arrhenius Equation
Arrhenius equation:
k =Ae-Ea/RT
Old Chemist expression:
increase rate by 10 °C will double the rate
8
Time Warp
Back to High School….
…..but only briefly….
9
Equation of a Line
y=mx + b
what you want
y-intercept
what you know
slope
10
Function of a Line
y intercept =b
y=mx + b
slope =m
y
x
11
Arrhenius Equation
Activation Energy
k =Ae-Ea/RT
rate
Gas constant
Temperature (Kelvin)
Pre-exponential factor
e
Empirical equation
12
Arrhenius Equation
k =Ae-Ea/RT
ln(k) = ln(A) – Ea/RT
ln(k) =– Ea/RT + ln(A)
ln(k) =–( Ea/R)(1/T) + ln(A)
13
Function of a Line
y intercept =b
y=mx + b
slope =m
y
x
14
Function of a Line
y intercept =ln(A)
ln(k) =–( Ea/R)(1/T) + ln(A)
slope = -Ea/R
ln(k)
1/T
15
Arrhenius Equation
Arrhenius equation:
k =Ae-Ea/RT
ln(k) =–( Ea/R)(1/T) + ln(A)
k = anything
Plot log(aT) vs 1/T linear if Arrhenius
What is Ea?
16
Ea
Energy
reactants
products
Reaction coordinate
---Imagine a marble---
17
Ea
Energy
reactants
products
Reaction coordinate
18
Ea
Intermediates/Transition states
Ea
Energy
reactants
products
Reaction coordinate
19
Are Diamonds forever?
Kinetics vs. Thermodynamics
Energy
Diamond
Graphite
Reaction coordinate
20
(really the same thing)
21
Ea
Intermediates/Transition states
Ea
Energy
Diamond
Graphite
Reaction coordinate
22
Arrhenius Equation
k =Ae-Ea/RT
Critical assumption is that Ea is CONSTANT
Assume
Ass-u-me
23
Time-Temperature Superposition
Does mechanism change as a function of temperature?
If same mechanism:
• same shape (log graph)
• should be constant acceleration (multiple)
1. Pick a reference temperature
2. Multiply the time at each temperature by the
constant that gives the best overlap with the
reference temperature data
3. Define that multiple as ‘aT’ (aT = 1 for ref. temp.)
4. Find aT for each temperature
Plot log(aT) vs 1/T linear if Arrhenius
Gillen, K. T.; Celina, M.; Clough, R. L.; Wise, J. Trends in Polymer Science, Extrapolation of Accelerated Aging Data -Arrhenius or Erroneous? 1997, 5, 250-257.
Arrhenius equation: Empirical equation
k =Ae-Ea/RT
ln(k) = ln(A) – Ea/RT
24
Thermal-oxidative Aging: Nylon
Average % tensile strength remaining
100
90
80
70
60
138 °C
138 °C 2nd Spool
124 °C
124 °C 2nd Run
124 °C 2nd Spool
109 °C
99 °C
95° C
95 °C 2nd Run
80 °C
64 °C
48 °C
37 °C
50
40
30
20
10
0
1
10
100
1000
Time in days at temperature
Bernstein, R.; Gillen, K. T. Polymer Degradation and Stability, Nylon 6.6 accelerating aging studies: II. Longterm thermal-oxidative and hydrolysis results 2010, 95, 1471-1479.
2000 days ~ 5.5 years
25
Thermal-oxidative Aging: Nylon Shifted Data
Average % tensile strength remaining
100
90
80
70
60
138 °C
138 °C 2nd Spool
124 °C
124 °C 2nd Run
124 °C 2nd Spool
109 °C
99 °C
95 °C
95 °C 2nd Run
80 °C
64 °C
48 °C
37 °C
50
40
30
20
10
Shift Factor
8.5
9.0
3.25
2.75
3.0
1.0
0.45
0.75
0.65
0.25
0.20
0.12
0.08
0
1
10
100
1000
Shifted time to 109 °C, days
Bernstein, R.; Gillen, K. T. Polymer Degradation and Stability, Nylon 6.6 accelerating aging studies: II. Long-term thermal-oxidative and hydrolysis results 2010, 95, 1471-1479.
26
Shift factor, aT
Thermal-oxidative Aging: Nylon Shift Factor Graph
10
10
1
1
0.1
0.1
0.01
0.01
~23 °C
0.001
0.001
0.0001
0.0001
2.4
2.5
2.6
2.7
2.8
2.9
3.0
3.1
3.2
3.3
3.4
3.5
1000/T, 1/K
Bernstein, R.; Gillen, K. T. Polymer Degradation and Stability, Nylon 6.6 accelerating aging studies: II. Long-term thermal-oxidative and hydrolysis results 2010, 95, 1471-1479.
27
Thermal Exposure
Thermal-Oxidation
Polymer + O2
Oxidized Polymer
Quantify amount of oxygen consumed
•Simple in theory
•Difficult in practice
•Amazingly sensitive
28
Schematic of Oxuptake
Initial Pressure of O2
O2
O2
O2
O2
O2
O2
O2
O2
Final Pressure of O2
D
O2
O2
Polymer
O2
+ Time
Oxidized Polymer
O2
O2
29
Oxygen Consumption
Shift Factor, aT
Normalized Measured Property
Enhanced Extrapolation ‘Good’
O
XX O
X XO O
X
X
X
Measured Property
O
Oxygen Consumption
X O
O
O
O
1/Temperature, K-1
High Temp
Low Temp
30
31
Shift Factor, aT
Normalized Measured Property
Enhanced Extrapolation: ‘Bad’
O
XX O
X XO O
X
X
X
Measured Property
O
Oxygen Consumption
O
X
O
O
O
O
1/Temperature, K-1
High Temp
Low Temp
32
DLO, Need to Know
Diffusion Limited Oxidation (DLO) effects if oxygen dissolved in material
used up faster by reaction than it can be replenished by diffusion from
surrounding air atmosphere
Race between:
the oxygen consumption rate versus the oxygen diffusion rate
Therefore we need estimates of:
1. O2 permeability versus aging temperature
2. O2 consumption versus aging temperature
33
Diffusion-Limited Oxidation (DLO)
O2
O2
O2
O2
O2
O2
O2
rxn rate > diffusion rate
Heterogeneous
O2
O2
O2
O2
O2
rxn rate < diffusion rate
Homogeneous
34
Modulus Profiling
Measure of Inverse tensile
compliance
Closely related to tensile
modulus
Modulus vs. Shore A
100
M o d u lu s, M P a
Indentation technique
ca. 50mm resolution
10
1
20
30
40
Excellent to examine ‘geneity’ of aging
(heteo- or homo-) (DLO issues)
50
60
70
S h o re A h a rd n e ss
80
90
100
35
Schematic of Modulus Profile Experiment
Probe tip, sample and mass
Mass is applied in two steps
Gillen, K. T.; Clough, R. L.; Quintana, C. A. Polym. Degrad. Stab., Modulus profiling of polymers 1987, 17, 31-47
36
Modulus Profiler
37
Modulus Profiler Sample
38
Homogeneous Aging
Aging of a nitrile rubber at temperatures
ranging from 65°C to 125°C
Modulus profiles of samples aged at
65°C indicate the presence of
homogeneous aging
735 days
497 days
350 days
226 days
0 days
o
65 C
0
20
40
60
P, %
Wise, J.; Gillen, K. T.; Clough, R. L. Polymer Degradation and Stability, An ultrasensitive technique for testing the Arrhenius extrapolation assumption for thermally aged elastomers 1995, 49, 403-418.
80
100
BNM O D65
39
Heterogeneous Aging
Modulus profiles for samples aged at 95°C show that diffusion-limited oxidation
(DLO) is becoming important; at 125°C, DLO effects are very significant
176 days
1000
127 days
1000
84 days
18 days
36 days
14 days
0 days
7 days
4 days
100
M o d u lu s, M P a
100
-1
D , MPa
0 days
10
10
o
o
95 C
125 C
1
1
0
20
40
60
P, %
80
100
BNM O D95
0
20
40
60
P, %
Wise, J.; Gillen, K. T.; Clough, R. L. Polymer Degradation and Stability, An ultrasensitive technique for testing the Arrhenius extrapolation assumption for thermally
aged elastomers 1995, 49, 403-418.
80
100
BNM O D125
40
Nylon: Tensile versus Oxygen Consumption
10
Shift factor, aT
10
10
10
10
10
10
1
10
0
10
-1
10
-2
10
~23°C
-3
-4
10
10
Tensile Strength Shift Factor
Oxygen Consumption Shift Factor
-5
10
2.4
2.6
2.8
3.0
1000/T, 1/K
3.2
3.4
1
0
-1
-2
-3
-4
-5
Thermal-oxidative tensile:
Prediction vs. Experimental
Arrhenius Predictions
64 °C Thermal-oxidative
41
Arrhenius predictions severely off target
suggest change in mechanism/nonArrhenius behavior
Initial data Predicted: 92% at ca. 3700 days
Observe: 92% at ca. 835 days
Oxygen consumption suggests no change
in thermal-oxidative mechanism
Possible explanation involving mechanism change?
42
Nylon Structure
H 2N
NH2
1,6 -Hexanediamine
H
N
+
O
O
N
H
H
N
O
n
O
OH
HO
O
Adipic Acid
H2O
Nylon 6.6
43
Humidity Aging Schematic
H2O
H2O
O2
H2O
O2
H2O
O2
H2O O2
O2
H2O
O2
O2
H2O
H2O
O2
O2
O2
H2O
O2
H 2O
H2O
Time,
D
O2
H2O O2
H2O
O2
H2O
O2
44
Humidity Aging Hardware
45
Organic Materials Aging and Degradation
Specifics -o-rings
General path –most organic materials
This talk –details not important (all published)
46
O-ring Published Documentation
Bernstein, R.; Gillen, K. T. Polymer Degradation and Stability, Predicting the
Lifetime of Fluorosilicone O-rings 2009, 94, 2107-2133.
Bernstein, R.; Gillen, K. T. "Fluorosilicone and Silicone O-Ring Aging Study,"
SAND2007-6781, Sandia National Laboratories, 2007.
Chavez, S. L.; Domeier, L. A. "Laboratory Component Test
Program (LCTP), Stockpile O-Rings," BB1A3964, 2004.
Gillen, K. T.; Bernstein, R.; Wilson, M. H. Polymer Degradation and Stability,
Predicting and Confirming the Lifetime of O-rings 2005, 87, 257-270.
Gillen, K. T.; Celina, M.; Bernstein, R. In Polymer Degradation and Stability
Validation of Improved Methods for Predicting Long-Term Elastomeric Seal
Lifetimes from Compression Stress-Relaxation and Oxygen Consumption
Techniques, 2003; Vol. 82, pp 25-35.
47
O-rings Background
Used as environmental seals or other seals
O-RING CROSS-SECTIONS
Most systems filled with inert gas
to protect interior components
from oxidation & hydrolysis
UNAGED
15 yr in field
Previously:
No technique to measure equilibrium sealing force
No technique to rapidly achieve equilibrium compression set
No correlation
48
CSR Jigs
Gap of jig can be adjusted to any desired
size
O-ring pieces cut to allow air circulation
Measurement of force involves very slow
and slight compression until electrical
contact is broken between the top and
bottom plates
Jigs can be placed in ovens, thus
providing isothermal measurements
49
Compression Stress Relaxation (CSR)
Shawbury-Wallace Compression Stress Relaxometer (CSR) MK II
Commercial Instrument
Measure of Force
-O-ring sealing force
Can Adjust Gap Size to Approximate Actual
Compression in System
(Wallace Test Equipment, Cryodon, England)
50
Accelerated aging
1) Physical force decay
-Equilibrium values achieved –starting point
-Ability to get field returned o-ring force –ending point
2) Chemical force decay
Prediction of force changes as a function of aging
Why we do isothermal measurements…
9
8
7
~2 hrs
F/L, N/cm
6
5
4
3
2
1
0
-4
10
-3
10
-2
10
-1
10
0
10
1
10
2
10
o
Time after removal from 110 C oven, days
Sealing force per unit length versus time out of a 110 °C oven for two CSR jigs containing Butyl-A o-ring segments that had aged under
25% compression until the force degraded by ~42% (top curve) and ~72% (bottom curve), respectively.
Gillen, K. T.; Celina, M.; Bernstein, R. In Polymer Degradation and Stability Validation of Improved Methods for Predicting Long-Term Elastomeric Seal Lifetimes from Compression StressRelaxation and Oxygen Consumption Techniques, 2003; Vol. 82, pp 25-35.
51
52
All Jigs at Temperatures -Fluorosilicone
100
90
80
70
% F/Fo
60
50
Jig #3 138 °C
Jig #4 138 °C
Jig #5 138 °C
Jig #1 138 °C
Jig #5 124 °C
Jig #6 124 °C
Jig #8 124 °C
Jig #14 109 °C
Jig #15 109 °C
Jig #4 80 °C
Jig #7 80 °C
40
30
20
10
0
0.1
1
10
100
1000
Time, Days
Bernstein, R.; Gillen, K. T. "Fluorosilicone and Silicone O-Ring Aging Study," SAND2007-6781, Sandia National Laboratories, 2007.
Bernstein, R.; Gillen, K. T. Polymer Degradation and Stability, Predicting the Lifetime of Fluorosilicone O-rings 2009, 94, 2107-2133.
53
Time-Temperature Superposition
Does mechanism change as a function of temperature?
If same mechanism:
• same shape (log graph)
• should be constant acceleration (multiple)
1. Pick a reference temperature
2. Multiply the time at each temperature by the
constant that gives the best overlap with the
reference temperature data
3. Define that multiple as ‘aT’ (aT = 1 for ref. temp.)
4. Find aT for each temperature
Plot log(aT) vs 1/T linear if Arrhenius
Gillen, K. T.; Celina, M.; Clough, R. L.; Wise, J. Trends in Polymer Science, Extrapolation of Accelerated Aging Data -Arrhenius or Erroneous? 1997, 5, 250-257.
Arrhenius equation: Empirical equation
k =Ae-Ea/RT
ln(k) = ln(A) – Ea/RT
54
Time-Temperature Superposition
100
90
80
70
% F/Fo
60
50
40
30
20
10
Shift Factor
7
5
3.5
6
2
3
2.5
1
0.9
0.5
0.4
S Jig #3 138 °C
S Jig #4 138 °C
S Jig #5 138 °C
S Jig #1 138 °C
S Jig #5 124 °C
S Jig #6 124 °C
S Jig #8 124 °C
S Jig #14 109 °C
S Jig #15 109 °C
S Jig #4 80 °C
S Jig #7 80 °C
0
0.1
1
10
100
Shifted time to 109 °C, Days
1000
55
Shift factor, aT
Shift Factor Plot
10
10
1
1
0.1
0.1
0.01
0.01
~80 °C
0.001
0.001
~23 °C
0.0001
2.4
2.5
2.6
2.7
2.8
2.9
3.0
1000/T, 1/K
3.1
3.2
3.3
3.4
0.0001
3.5
Temperature
‘Accelerated Aging’
Reaction Coordinate
56
57
Shift factor, aT
Shift Factor Plot
10
10
1
1
23 °C aT =0.065
0.1
0.1
0.01
0.01
23 °C aT = 0.011
~80 °C
0.001
0.001
23 °C aT = 0.000926
~23 °C
0.0001
2.4
2.5
2.6
2.7
2.8
2.9
3.0
1000/T, 1/K
3.1
3.2
3.3
3.4
0.0001
3.5
58
Shifted Data with RT ‘Prediction’ w/o 80 C data
Predicted Time at 23 °C, Years
1
10
100
1000
100
90
80
70
% F/Fo
60
50
40
30
20
10
0
0.1
1
10
100
Shifted time to 109 °C, Days
1000
59
Shift factor, aT
Shift Factor Plot
10
10
1
1
23 °C aT =0.065
0.1
0.1
0.01
0.01
23 °C aT = 0.011
~80 °C
0.001
0.001
23 °C aT = 0.000926
~23 °C
0.0001
2.4
2.5
2.6
2.7
2.8
2.9
3.0
1000/T, 1/K
3.1
3.2
3.3
3.4
0.0001
3.5
Shifted Data with RT ‘Prediction’ All data
Predicted Time at 23 °C, Years
0.1
1
10
100
100
90
80
70
% F/Fo
60
50
40
30
20
10
0
0.1
1
10
100
Shifted time to 109 °C, Days
1000
60
61
Shift factor, aT
Shift Factor Plot
10
10
1
1
23 °C aT =0.065
0.1
0.1
0.01
0.01
23 °C aT = 0.011
~80 °C
0.001
0.001
23 °C aT = 0.000926
~23 °C
0.0001
2.4
2.5
2.6
2.7
2.8
2.9
3.0
1000/T, 1/K
3.1
3.2
3.3
3.4
0.0001
3.5
62
Shifted Data with RT ‘Prediction’ 109 and 80 only
Predicted Time at 23 °C, Years
0.01
0.1
1
10
100
100
90
80
70
% F/Fo
60
50
40
30
20
10
0
0.1
1
10
100
Shifted time to 109 °C, Days
1000
63
Shift Factor Plot
10
10
~50% in ~ 20 years
Shift factor, aT
1
1
23 °C aT =0.065
0.1
0.1
~50% in ~ 100 years
0.01
0.01
23 °C aT = 0.011
~80 °C
0.001
0.001
23 °C aT = 0.000926
~50% in ~ 1000 years
~23 °C
0.0001
2.4
2.5
2.6
2.7
2.8
2.9
3.0
1000/T, 1/K
3.1
3.2
3.3
3.4
0.0001
3.5
64
O-ring
Sealing force arguably most important parameter
Correlation between equilibrium values
sealing force
compression set
Difficult to measure
slow and laborious
Easy to measure
quick and simple
O-RING CROSS-SECTIONS
UNAGED
15 yr in field
65
Force versus Compression Set Data -Fluorosilicone
100
% Force Remaining at Temperature
90
80
70
60
50
40
30
20
10
0
0
20
40
60
80
100
% Equilibrium Compression set -Corrected for Temperature
Field Data*
66
* not quite the whole story, but good enough for this conversation!
100
J1
Compression set (~1 day), %
90
J2
J8
80
best prediction
70
hi-T prediction
60
50
40
30
20
10
0
0
2
4
6
8 10 12 14 16 18 20 22 24 26 28
Aging
Aging time,
time, years
years
W69set2
Compression set measurements of three fluorosilicone o-rings taken on surveillance units approximately 1 day
after removal from the unit. The solid curve and the dashed curve assume a linear relationship between set and
force decay.
Bernstein, R.; Gillen, K. T. Polymer Degradation and Stability, Predicting the Lifetime of Fluorosilicone O-rings 2009, 94, 2107-2133
67
Other Compression Set Data
XE-5601 SILICONE (E. A. Salazar data)
100
Different Silicone
Different Program
Different PI
10,000-hour Compression Set, %
40% initial strain at
80
25C
66C
100C
60
121C
149C
40
20
0
1
10
100
Aging time, hrs
1000
10000
Sil-set2
Gillen, K. T. "Silicone seal analysis," Internal Memo, SNL, 2001.
Compression Set
68
100
o
10,000-hour Compression Set, %
T, C,
80
60
a
Filled circles show
40 yr British data for"temperate conditions"
T
25
1
66
3.1
100
21
121
55
149
320
Data Analyzed by Gillen
Added in 40 year RT
data from another
source
40
20
0
-3
10
-2
10
-1
10
0
10
1
10
Aging time, years at 25C
2
10
Silsh2
Gillen, K. T. "Silicone seal analysis," Internal Memo, SNL, 2001.
69
Arrhenius Plot for Compression Set
10
3
10,000 hour recovery data
2
Shift factor a
T
10
10
10
1
0
2.2
2.4
2.6
2.8
1000/T, K
3.0
3.2
3.4
-1
Sil-Arr2
Arrhenius plot of the shift factors for silicone compression set which leads to an
aging room temperature prediction for compression set
Gillen, K. T. "Silicone seal analysis," Internal Memo, SNL, 2001.
70
Force versus Compression Set Data
Predicted Time at Room Temperature, Years
0.01
100
0.1
1
10
100
0
80
20
70
% F/Fo
60
40
50
40
60
30
20
10
0
80
Force
Compression Set
100
Compression set, %
90
71
Force versus Compression Set Data
Predicted Time at Room Temperature, Years
0.01
100
0.1
1
10
100
0
80
20
70
% F/Fo
60
40
50
40
60
30
20
10
0
80
Compression set, %
90
Force
Compression Set
100
Correlation between current Silicone Force data and
Compression set data obtained from three different
sources (and different sizes!)
Fluorosilicone versus Silicone!!
Displays confidence in generalized predictions about
silicone o-rings state of health (CS easy to measure)
under oxidative environments*
72
10
0 .9
9
0 .8
8
0 .7
7
0 .6
6
0 .5
5
0 .4
4
F /F
0
1 .0
la b -a g e d B u ty l-A
0 .3
F , N /cm
Butyl Force vs. Compression set; lab and field aged
3
la b -a g e d B u ty l-B
2
0 .2
la b -a g e d B u ty l-C
0 .1
1
fie ld -a g e d B u tyl-B
0
0 .0
0
20
40
60
80
100
C o m p re ssio n se t, %
Equilibrium values of compression set plotted versus F/F0 for laboratoryaged o-rings for three butyl materials plus field results for Butyl-B plotted
assuming that F0 = 10 N/cm.
Gillen, K. T.; Bernstein, R.; Wilson, M. H. Polymer Degradation and Stability, Predicting and Confirming the Lifetime of O-Rings 2005, 87, 257-270.
73
Heavily Filled Silicone
100
% Force Remaining at Temperature
90
80
70
60
50
40
30
20
10
0
0
20
40
60
80
100
% Equilibrium Compression set -Corrected for Temperature
Conducting o-rings compression set versus force remaining
74
Progression of Stress Relaxation due to Chemical Aging
Time = 0
Time = .5sec
Time =.8sec
Time = 1sec
SIGYY
-191 psi=133 N/cm2
-89 psi= 61 N/cm2
Time = 30 years
Time = 56 year
Time 90 years
Slide Courtesy of David Lo
13 psi= 9 N/cm2
Take home messages…
1) Be aware of mechanism changes
2) Understand chemistry/be careful with the details
–DLO, O2 vs. H2O etc
1) Find something to measure
2) Do things at many temp (as far apart as possible)
3) Do things for very very long time
4) Validate against real world
5) It would be nice to know your performance
requirements
75
Lots of help…
Dora Derzon, Brad Hance, Don Bradley, Roger Assink, James
Hochrein, Steven Thornberg, David Lo, Kathy Alam, Laura Martin,
John Schroeder, Patti Sawyer, Mark Stavig, and Ken Gillen
76
Questions…
77