Lab Test Method Validation Protocol for Laminated Films

Date: Test Method: 01/06/12 This test method study includes 4 different tests summarized in the table below: Puncture Instron (100lb) FTMS-101C Tensile Instron (100lb) D-882 Impact (Spencer) Elmendorf (6400g) D-3420 Tear Elmendorf (200g) D-1922

Test Type Test Instrument National/Fed STD #

Instrument Model and Make: a) Instron Model 4411 MFMD007 b) Thwing Albert Pendulum/Elmendorf MMIS052 c) Dumb Bell Die Tear Sample Cutter RK-N7003CA-91W c) JDC Precision Sample Cutter Model 1-10 (For Tensile samples) d) Mitutoyo Digital Micrometer IDC-112DEB (To determine film gauge) * Test equipments are regularly calibrated; calibration log reflects dates and calibration periods.

Detail of Study:
Material: Sample size: Polyester/Polyethylene Laminated Film 3 Trials of 10 samples per tester for each test. n = 30 (90 total per test -3 testers x 30 samples each)

Purpose: The purpose of this study is to validate puncture, tensile, impact and tear test methods used to determine mechanical properties of packaging products at the R&D Lab. Test Methods: The 4 tests involved in this study are: Puncture, Impact, Tensile, and Tear strength. Due to the nature of the tests (destructive) meaning each sample will be destroyed to obtain the measurement reading. Therefore it is important to note that an important assumption is made. The assumption is that the samples will be very similar in physical characteristics (homogenous) since they are from the same work order (same film roll). Since it is not possible for the testers to measure exact parts, a nested gage R&R study will be conducted. Also Anova analysis of means will be conducted to further investigate mean variability between the different testers for each of the tests. Puncture Test:

Page 1/19

Test will be conducted in accordance to Federal Test Method STD. NO. 101C and Company ABCs internal work instruction RWI-0242 (Instron 4411). Note that the 100lb Load will be used to conduct this test. Samples are cut into 2 inch strips using metal template and razor blade. Data for this test will be recorded automatically by Instrons software. Tensile Strength Test: Test will be conducted in accordance to ASTM Designation D-882 and Company ABCs internal work instruction RWI-0242 (Instron 4411). Samples will be taken in MD direction only for the purpose of this study. a 11.5 x 9 Metal template will be used to cut sheets from film then use JDC sample cutter to cut samples measuring 1 inch in width and about 6 inches in length. It is critical to this test that the samples edges are free of any nicks. Any nicks on the edge of the sample can cause an early tear and potentially skew the results of the tensile test. One hundred pound Instron load will be used for this test. The drawing on the right represents the sampling landscape cut from a sheet of RLA-069. The drawing only shows the samples used for trial 1, subsequent trials will follow the same cutting pattern utilizing additional film sheets. Data for this test will be recorded automatically by Instrons software. Note: Instron method 26 will be used (compression mode 00 - 12in/m). Impact Testing (Spencer): Impact testing method will be carried out in accordance with ASTM Designation D-3420 and Company ABCs internal work instruction RWI-0239 (Spencer Impact Test). The Thwing Albert Pendulum/Elmendorf apparatus (using 6400g Pendulum) will be used to conduct this test. A 4x4 metal template and a sharp cutter will be used to cut the samples for this test. Test data will be recorded on form RFCD-0419. Tear Testing: Tear testing method will be carried out in accordance with ASTM Designation D-1922 and Company ABCs internal work instruction RWI-0991 (Elmendorf Tear Propagation). The Thwing Albert Pendulum/Elmendorf apparatus (using 200g Pendulum) will be used to conduct this test. A spring loaded Dumb Bell die Cutter will be utilized to cut the samples. This cutter produces samples with dimensions adhering to the requirements found in the ASTM D-1922. Test data will be recorded on form RFCD-0419. Data Analysis: As mentioned above the data produced in this study will be analyzed by Anova analysis of means and a nested gage R&R study. The data analysis will be conducted using Minitab and the results will be interpreted in the summary and conclusion. The aim of the study is to determine validity of the test methods as interpreted by the selected measurement systems.

Page 2 of 19

Lab Test Method Validation Data and Results Analysis

Introduction: Test method validation was carried out according to above protocol for Puncture, Tensile, Impact, and tear. Test method results were analyzed in Minitab software accross 3 testers testing 30 samples for each test (3 trials of 10 samples per tester per test). Two types of statistical data analysis has been used for each of the tests, Anova analysis of means as well as Gage R&R were selected to investigate the validatity of the test methods. Puncture Testing Results: Selected puncture test for this study is Company ABCs internal work instruction RWI0242 (Instron 4411) which is based on Federal Test Method STD. NO. 101C). Three Testers tested 30 samples each across 3 trials using Instron. The Instron automatically logs test values, the recorded values represents maximum load force in pounds the sample withstands up to the point of puncturing. Instron Crosshead speed was set to 20 in/min using a 100 lbf load unit. The Puncture probe selected for this testing is 1/8 radius which is listed in the Federal Test Method Std 101C. Test Data:
Sample # 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 Mean Std. Dev Tester 1 13.92 14.22 11.81 14.45 14.53 13.74 14.86 14.29 14.61 13.33 14.35 13.95 14.01 13.7 14.32 11.75 14.52 13.05 14.05 13.57 13.45 14.67 13.69 13.77 13.35 13.21 11.88 14.37 14.56 13.61 13.79 0.82 Tester 2 14.23 12.93 14.09 14.38 12.59 12.64 12.91 14.84 13.32 13.22 13.94 14.14 14.44 12.46 13.71 14.55 13.26 13.6 14.1 14.09 14.33 14.15 13.23 12.64 12.71 13.86 11.98 13.85 13.52 10.59 13.48 0.91 Tester 3 14.88 14.71 13.37 14.23 11.73 14.3 14.15 12.69 13.49 14.25 14.15 14.76 13.25 14.79 13.81 14.95 12.72 13.59 14.18 13.55 13.99 13.45 14.37 13.41 12.98 13.15 13.93 13.19 13.01 13.43 13.75 0.75

Page 3 of 19

Puncture Resistance Test Gage R&R:

G age R&R (N ested) for Puncture lb/ f
Components of Variation
100 Percent
% Contribution % Study Var

Puncture lb/ f By part set ( op )

14 12 10
1 2 3 4 5 6 7 8 9 10 1 1 2 3 4 5 6 7 8 9 10 2 1 2 3 4 5 6 7 8 9 10 3

50

Gage R&R

Repeat

Reprod

Part-to-Part

part set

op

R Chart by op
Sample Range 4 2 0
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 1 0

3 UCL=3.462 _ R=1.345 LCL=0 part set 14 12 10

Puncture lb/ f by op

Xbar Chart by op
Sample Mean 15 14 13 LCL=12.295
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 1 0

2 op

UCL=15.046 _ _ X=13.670

part set

Gage R&R Study - Nested ANOVA Gage R&R (Nested) for Puncture lb/f Source op part set (op) Repeatability Total DF 2 27 60 89 SS 1.7106 21.1141 38.6192 61.4440 MS 0.855324 0.782005 0.643653 F 1.09376 1.21495 P 0.349 0.261

Gage R&R %Contribution Source Total Gage R&R Repeatability Reproducibility Part-To-Part Total Variation VarComp 0.646097 0.643653 0.002444 0.046117 0.692215 (of VarComp) 93.34 92.98 0.35 6.66 100.00

Study Var Source Total Gage R&R Repeatability Reproducibility Part-To-Part StdDev (SD) 0.803802 0.802280 0.049437 0.214749 (5.15 * SD) 4.13958 4.13174 0.25460 1.10596

%Study Var (%SV) 96.61 96.43 5.94 25.81

Page 4 of 19

Total Variation

0.831994

4.28477

100.00

According to minitab gage R&R results, repeatablity was the most significant source of variation. This is understandable considering the destrucutive nature of this test and performance variation inherent in the laminated film.

Puncture Resistance, Anova analysis of means: Before conducting anova analysis of the means, a normality test must take place.

Probability Plot of Puncture lb/ f

Normal
99.9 99 95 90 80 70 60 50 40 30 20 10 5 1 0.1 Mean StDev N KS P-Value 13.67 0.8309 90 0.085 0.110

Percent

10

11

12

13 14 Puncture lb/ f

15

16

17

Hence the P value is greater than 0.05 then one must conclude that the data is normal and can proceed with anova anlysis of the means. Below are the results from Minitab for executing a One way Anova anlysis on the puncture data accross 3 testers to determine if there is a statistical difference amongst the means.

Page 5 of 19

The following is a graphical representation the results of Anova analysis of the means amongsts all 3 testers.

One-Way Normal ANOM for Puncture lb/ f

Alpha = 0.05
14.1 14.0 13.9 13.8 Mean 13.7 13.6 13.5 13.4 13.3 1 2 op 3 13.3754 13.6704 13.9655

One-way ANOVA: Puncture lb/f versus op

Source op Error Total DF 2 87 89 SS 1.711 59.733 61.444 MS 0.855 0.687 F 1.25 P 0.293

S = 0.8286

R-Sq = 2.78%

R-Sq(adj) = 0.55%

The results above show a P value of 0.293. Since P value is greater than 0.05 then there is a 95% chance that there is no statistically significant difference amongst the means between all 3 testers. This result validates the test method results among all 3 testers and concludes that the method is reliable and the variation is inherted from the material.

Page 6 of 19

Tensile Strength Test Results: Tensile testing was carried out according to Company ABCs internal work instruction RWI-0242 (Instron 4411) , which is based on ASTM method D882. Three Testers tested 30 samples each across 3 trials using the Instron. The Instron automatically logs test values that represent the load force in pounds the sample can withstand up to break. Instron Crosshead speed was set to 12 in/min using a 100 lbf load unit. Raw Test Data:
Tensile Results Sample # 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 Mean Std. Dev Tester 1 16.84 17.65 18.74 18.15 18.9 17.82 14.25* 11.73* 17.56 19.02 15.2* 18.89 16.19 14.22* 18.59 18.2 14.79* 18.62 18.43 15 14.23* 14.23* 15.57 15.75 13.69 18.83 14.59* 18 14.75 14.32* 16.43 2.07 Tester 2 16.26 14.57* 14.02* 17.48 11.51* 13.59* 13.7 17.62 17.12 13.63* 16.79 18.11 17.27 17.77 18.95 13.01* 14.42 18.22 18.76 19 18.97 18.17 17.06 18.08 18.15 17.47 17.44 17.78 17.68 15.73 16.61 2.06 Tester 3 16.66 18.07 19.19 17.71 17.61 14.92* 19.46 18.99 18.75 18.29 14.05* 17.95 16.28 18.58 19.01 10.9* 18.66 17.35 14.86* 18.42 19.45 18.76 17.3 16.56 17.99 17.22 13.67* 19.33 18.39 14.92* 17.31 2.03

Tensile Strength Gage R&R:

Page 7 of 19

G age R&R (N ested) for Tensile lb/ f

Components of Variation
100 Percent
% Contribution % Study Var

Tensile lb/ f By part set ( op )

20

50

15

Gage R&R

Repeat

Reprod

Part-to-Part

part set

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

op

R Chart by op
Sample Range 10 5 0
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 1 0

3 UCL=8.91 _ R=3.46 LCL=0 part set

Tensile lb/ f by op
20

15

10

Xbar Chart by op
1 Sample Mean 20.0 17.5 15.0 LCL=13.241
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 1 0

2 op

UCL=20.323 _ _ X=16.782

part set

Gage R&R Study - Nested ANOVA Gage R&R (Nested) for Tensile lb/f Source op part set (op) Repeatability Total DF 2 27 60 89 SS 13.064 100.430 266.323 379.817 MS 6.53211 3.71962 4.43872 F 1.75612 0.83799 P 0.192 0.687

Gage R&R %Contribution Source Total Gage R&R Repeatability Reproducibility Part-To-Part Total Variation VarComp 4.53247 4.43872 0.09375 0.00000 4.53247 (of VarComp) 100.00 97.93 2.07 0.00 100.00

Study Var Source Total Gage R&R Repeatability Reproducibility Part-To-Part Total Variation StdDev (SD) 2.12896 2.10683 0.30619 0.00000 2.12896 (5.15 * SD) 10.9641 10.8502 1.5769 0.0000 10.9641

%Study Var (%SV) 100.00 98.96 14.38 0.00 100.00

Page 8 of 19

Gage R&R report indicates that the main source of variation is repeatability. Similar to puncture test method validation, the data was further analysed using Anova analysis of the means. Tensile Strenth Anova analysis of means: First a normality test is conducted. Following ASTMs D882 section 4.5s direction, abnormal data points were removed from the analysis. Section 4.5 states Materials that fail by tearing give anomalous data which cannot be compared with those from normal failure.

Probability Plot of R - Tensile (lb/ f)

Normal
99.9 99 95 90 80 70 60 50 40 30 20 10 5 1 0.1 Mean StDev N KS P-Value 17.85 1.060 66 0.085 >0.150

Percent

14

15

16

17 18 19 R - Tensile (lb/ f)

20

21

22

Since the P value is greater than 0.05 therefore the data is normal and anova anlysis of the means is pursued.

Page 9 of 19

One-Way Normal ANOM for R - Tensile lb/ f

Alpha = 0.05

18.4 18.2 18.0 Mean 17.8 17.6 17.4 17.2 1 2 op 3 17.428 17.846 18.263

One-way ANOVA: R - Tensile lb/f versus op

Source op Error Total DF 2 63 65 SS 4.03 68.96 72.99 MS 2.01 1.09 F 1.84 P 0.167

S = 1.046

R-Sq = 5.52%

R-Sq(adj) = 2.52%

The results above show a P value of 0.167. Since P value is greater than 0.05 then there is a 95% chance that there is no statistically significant difference. amongst the means between all 3 testers.

Page 10 of 19

Spencer Impact Test Results: Selected Impact test methos is Company ABCs internal work instruction RWI-0239, which is based on ASTM method D-3420. Three Testers tested 30 samples each across 3 trials using the 6400g Thwing Albert Pendulum unit. Test data results were divided by 100 and multiplied by a factor of 5.4. This is in accordance with ASTMs D-3420 Impact testing procedure, the data results are in Jewel units. Test Data:
Spencer Impact Results Sample # 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 Mean Std. Dev Tester 1 1.836 2.268 2.052 1.512 1.998 2.322 2.808 2.754 2.214 2.43 2.376 2.322 2.214 2.214 1.836 2.052 2.43 2.538 1.836 1.674 1.62 2.916 2.646 2.484 2.322 2.052 3.132 1.944 1.674 1.998 2.22 0.40 Tester 2 2.376 2.106 2.592 2.268 1.836 2.322 2.106 2.646 1.944 1.89 2.106 1.998 2.322 1.944 2.322 2.16 1.944 2.43 2.268 2.322 2.484 2.16 2.16 2.214 1.89 2.43 2.16 2.106 2.322 2.376 2.21 0.21 Tester 3 2.7 2.43 2.16 1.89 2.052 2.97 2.052 2.16 2.322 2.214 2.268 2.268 2.322 2.538 2.538 1.998 2.106 2.322 1.89 2.484 2.052 2.43 1.998 2.268 2.43 2.484 1.782 2.16 1.89 2.16 2.24 0.26

Spencer Impact Gage R&R test Results:

Page 11 of 19

Gage R&R (N ested) for Spencer J

Components of Variation
100 Percent
% Contribution % Study Var

Spencer J By part set ( op )

3.2

50

2.4

1.6 Gage R&R Repeat Reprod Part-to-Part

part set op
1 2 3 4 5 6 7 8 9 10 1 1 2 3 4 5 6 7 8 9 10 2 1 2 3 4 5 6 7 8 9 10 3

R Chart by op
Sample Range 1 1.0 0.5 0.0
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 1 0

Spencer J by op
UCL=1.260 _ R=0.490 LCL=0 1.6 1 2 op 3 3.2

2.4

part set

Xbar Chart by op
Sample Mean 2.8 2.4 2.0 LCL=1.722
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 1 0

3 UCL=2.723 _ _ X=2.222

part set

Gage R&R Study - Nested ANOVA Gage R&R (Nested) for Spencer J Source op part set (op) Repeatability Total Gage R&R Source Total Gage R&R Repeatability Reproducibility Part-To-Part Total Variation VarComp 0.0775332 0.0775332 0.0000000 0.0143484 0.0918816 %Contribution (of VarComp) 84.38 84.38 0.00 15.62 100.00 Study Var (5.15 * SD) 1.43401 1.43401 0.00000 0.61689 1.56107 %Study Var (%SV) 91.86 91.86 0.00 39.52 100.00 DF 2 27 60 89 SS 0.02339 3.25562 4.65199 7.93100 MS 0.011696 0.120578 0.077533 F 0.09700 1.55518 P 0.908 0.079

Source Total Gage R&R Repeatability Reproducibility Part-To-Part Total Variation

StdDev (SD) 0.278448 0.278448 0.000000 0.119785 0.303120

Page 12 of 19

Looking at the Gage R&R results, the main source of variability is repeatability. Therefore, additional analysis will be conducted to investigate variability amongst the testers via Anova analysis of means. Spencer Impact, Anova analysis of means: First a normality test will be conducted.

Probability Plot of Spencer (J )

Normal
99.9 99 95 90 80 70 60 50 40 30 20 10 5 1 0.1 Mean StDev N AD P-Value 2.222 0.2985 90 0.389 0.379

Percent

1.5

2.0 2.5 Spencer (J)

3.0

Based on a P value of 0.379, which is greater than P=0.05, the data is normal and therefore Anova Analysis of Means is pursued.

Page 13 of 19

The following is a graphical representation of the Anova analysis of means:

One-Way Normal ANOM for Spencer (J )

Alpha = 0.05

2.35 2.3298 2.30

2.25 Mean 2.2224 2.20

2.15 2.1150

2.10 1 2 op 3

One-way ANOVA: Spencer J versus op

Source op Error Total DF 2 87 89 SS 0.0234 7.9076 7.9310 MS 0.0117 0.0909 F 0.13 P 0.879

S = 0.3015

R-Sq = 0.29%

R-Sq(adj) = 0.00%

Since P value is 0.879, which is greater than P=0.05, it is concluded that there is a 95% chance that there is no statistical difference between all 3 testers and that the test method is valid and consistent.

Elmendorf Tear Test Results:

Page 14 of 19

Test method used for Elmendorf tear test validation is Company ABCs internal work instruction RWI-0991, which is based on ASTM method D-1922. Three Testers tested 30 samples each across 3 trials using the 200g Thwing Albert Pendulum unit. The 200 gram unit was picked because it provides the most accurate reading for the film tested. According to ASTM, test values are multiplied by 2 since the 200gram arm was used. Test data:
Tear Results Sample # 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 Tester 1 40 38 38 38 30 30 30 30 30 32 28 32 32 30 28 34 34 34 32 32 32 38 38 32 36 38 38
% Contribution % Study Var

Tester 2 32 34 34 32 36 32 30 38 32 30 30 34 36 36 32 32 32 38 30 32 36 32 32 34 32 32 30
40 35

Tester 3 36 36 36 34 38 34 34 40 34 34 32 30 28 32 40 34 36 32 36 32 36 36 28 30 30 32
Tear G By part set ( op )

G age R&R (N ested) for Tear G

Components of Variation
100 Percent

26 27 28 29 30

28 36 34 30

34 32 36

32 32 38

50

Mean
0 Gage R&R Repeat Reprod

33.53 3.47

33.07 30 2.45 part set

op
1

33.60 3.25
1 2 3 4 5 6 7 8 9 10 2 1 2 3 4 5 6 7 8 9 10 3

Std. Dev

Part-to-Part

1 2 3 4 5 6 7 8 9 10

R Chart by op
Sample Range 1 10 5 0
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 1 0

Tear G by op
UCL=13.73 _ R=5.33 LCL=0 40 35 30 1 2 op 3

part set R&R: Tear Strength Gauge

Xbar Chart by op
Sample Mean 40 35 30 LCL=27.90
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 1 0

3 UCL=38.81 _ _ X=33.36

Page 15 of 19

part set

Gage R&R Study - Nested ANOVA Gage R&R (Nested) for Tear (g)
Source op part set (op) Repeatability Total DF 2 27 60 89 SS 3.822 218.800 608.000 830.622 MS 1.9111 8.1037 10.1333 F 0.235832 0.799708 P 0.792 0.734

Gage R&R
Source Total Gage R&R Repeatability Reproducibility Part-To-Part Total Variation VarComp 10.1333 10.1333 0.0000 0.0000 10.1333 %Contribution (of VarComp) 100.00 100.00 0.00 0.00 100.00 Study Var (5.15 * SD) 16.3939 16.3939 0.0000 0.0000 %Study Var (%SV) 100.00 100.00 0.00 0.00

Source Total Gage R&R Repeatability Reproducibility Part-To-Part

StdDev (SD) 3.18329 3.18329 0.00000 0.00000

Looking at the Gage R&R results, repeatability is the only contributor to the variation. Next step is to conduct normality test on the tear data to investigate further differences among the means of the testers.

Page 16 of 19

Tear Strength, Anova analysis of means:

Probability Plot of Tear (g)

Normal
99.9 99 95 90 80 70 60 50 40 30 20 10 5 1 0.1 Mean StDev N RJ P-Value 33.36 3.055 90 0.993 >0.100

Percent

25

30

35 Tear (g)

40

45

The P value is equal to 0.100, which is greater than 0.05, this is an indication that the data assumes a normal distribution with a confidence interval of 95%. The next step is to conduct an Anova anlysis of the means to determine if there is a statistical difference between the means.

Page 17 of 19

One-Way Normal ANOM for Tear (g)

Alpha = 0.05
35.0 34.5 34.0 Mean 33.5 33.0 32.5 32.258 32.0 1 2 op 3

34.453

33.356

The graphical representation demonstrates clearly the closeness of the means. Below is Anova analysis of the means across all 3 testers.

One-way ANOVA: Tear (g) versus op

Source op Error Total DF 2 87 89 SS 3.82 826.80 830.62 MS 1.91 9.50 F 0.20 P 0.818

S = 3.083

R-Sq = 0.46%

R-Sq(adj) = 0.00%

With a P value of 0.818, which is greater than P=0.05, it is concluded that there is a 95% chance of no statistical difference between all 3 testers and therefore the test method is acceptable.

Page 18 of 19

Final Conclusion: Test methods selected for validation study are utilized to determine physical properties of laminated film; Puncture, Tensile, Impact, and Tear. The test methods are based on ASTM and Federal standards interpreted into Company ABC internal procedures as referenced above. Output data from all four test runs was analyzed in Minitab using Anova means analysis and gage R&R. 1. Due to material variability and the destructive nature of the tests it was not expected to attain acceptable Gage R&R results. 2. Test run data was further analyzed to evaluate statistical variation by comparing testers data distribution using Anova Analysis of Means. In all 4 test runs the results pointed to the same conclusion; the means of all the testers results were statistically similar. Therefore it is concluded that the testers and testing equipment are capable of producing reliable results and that the main source of variation is inherent in product and applied manufacturing processes.

Page 19 of 19