6 INTERPRETATION OF TEST DATA

THE PREVIOUS DISCUSSIONS are intended to show the composite structure of paper, the many variables affecting the testing of that structure, the sensitivity of the variables to their prior history and testing conditions, and the possible effects of this history on test data. The following discussion addresses the interpretation of test results with emphasis on the fold endurance test, which is commonly used in testing for conservation research.

Valid paper conservation research must be based on the relationship of the inquiry to the research design. The sample material, sample population size, conditioning of the samples, testing methodology, and appropriate instrumentation for testing must be carefully chosen. The data must be analyzed and interpreted with appropriate consideration given to possible ambiguity caused by the samples' manufacturing techniques and history, and the testing methodology.

Test results that demonstrate either an increase or decrease in strength properties, as measured by fold endurance for example, could be interpreted in several ways:

If hydrolysis occurs, one could record a low fold endurance; if cross-linking occurs, one could record a high fold endurance; if both processes are occurring the results are unpredictable.

If a paper has a large number of aldehyde groups, data may show an increase in strength, but this could be due to cross-linking.

In a paper with high internal bonding that is aged, it may cross-link more than another sheet with less initial bonding affecting the fold endurance values differently.29

Understanding the choice of instrumentation used in a project is important to overall evaluation of the results. Each test method is useful for obtaining a particular type of information. For example, fold endurance testing has been used because of its ability to detect changes earlier than other physical methods. This information can be useful in conjunction with data from other physical methods, or as a general indication of trends useful in designing a research project. However, by itself, fold endurance is not a reliable research tool.30

As early as 1939, the Institute of Paper Chemistry made a comparison of three fold endurance instruments and concluded that in all cases the percent deviation was very high.31

Casey discusses the wide variation in individual test results which “has caused some people to consider the fold endurance test useless.”32 He discusses the number of tests needed to obtain significant data and the magnitude of differences in data necessary for significant data and the magnitude of differences in data necessary for significant results:

It is important to keep the variability of the folding-endurance test in mind when evaluating test results. An average of 5 to 10 measurements should be used and a difference between two samples of only 10 to 20 percent is seldom significant. The precision of any test method can be increased by using more tests to obtain the average, but for one paper sample with a folding endurance of 1000, it was calculated that over 6000 tests would have to be made to produce an average reproducible within 10 folds, at the 3 δ assurance level (99.7% of the time).33

In 1956 Hoffman found that there is a greater deviation in high fold than in low fold values.34

In 1964 and 1966 Kahlson and Lindholm investigated the cause of wide test variation in fold endurance data. They determined that motor warm up and mechanical frictional heat caused a local rise in temperature at the point of fold. The resulting drop in local relative humidity at the sample test site greatly reduced fold endurance. They discovered that using a fan to maintain constant air circulation around the specimen could reduce variability.35 Current TAPPI standards require a fan as part of the instrumental control. Even with a fan, variability is still high. TAPPI standards state that the degree of precision for “repeatability” (same lab, same instrument, same operator, same sample) is 20% and for “reproducibility” (different lab, different operator, same instrument, same sample) is 30% at the 95% probability level for the MIT instrument. The standard further states:

This method is very susceptible to small errors in the adjustment and calibration of the instrument and in the relative humidity of the test room. Limits as much as two times those shown [20% and 30%] may be expected if the instructions are not followed meticulously.36

Fold endurance tests are very sensitive to moisture. In their summary of all tests of physical properties, Crook and Bennett state in reference to fold endurance:

Changes of relative humidity, when the temperature is held constant, influence paper properties to different degrees. Of the properties studied, folding endurance was the most affected … The outstanding fact about folding endurance is its extreme dependence upon relative humidity. The maximum rate of increase of folding endurance was greater than 10% for each 1% R.H. increase of humidity. The influence of temperature change, at a constant relative humidity, also varies from paper to paper. The most marked effect is on folding endurance …37

Controversy exists over the type and amount of loading during the fold test and the magnitude of discrepancy caused by the weight. The MIT tester allows for variable weights, the Schopper is set for one load setting. Luner discusses the problem of using a standard load on sheets of various basis weights and tensile strength.38 Cardwell et al. also discusses the effect of equally loading differing papers during aging tests.39 Casey states that using the same 1 kg. weight on all samples “is unfair to papers of low tensile strength; it is not a true test of flexibility is such cases.”40 Kohler recommends not using a standard load, but one that is a percentage of the initial tensile strength of the paper.41 Crane comments on the use of standard loads on samples before and after aging tests:

Due to the strong dependence of folding values on load, one questions the value of comparing the very low values encountered after aging to the unaged sample under the same load.42

Fold endurance results are influenced by formation and furnish variables. Different values are obtained depending on how the paper was manufactured. Fold endurance rises in the early stages of beating, but drops with further beating and wet pressing.43 It increases with basis weight up to a point, but decreases with higher weights.44 Crane found that fiber orientation, as a result of fiber composition and drying tension, influences fold values.45 He also reported that the presence of fines affects the relationship of fiber length to fold endurance.46 Given this information, it is questionable that two different papers can be compared accurately using this test.

In the conclusion of his study of fold endurance testing, Crane comments:

Fold endurance is related to certain network properties. It does not correlate as clearly as do other tests such as tear and tensile with specific characteristics of the web. The relationship between folding endurance and viscoelastic properties is complex. Again, viscoelastic behavior is not as easily related to folding endurance as other strength properties.

While there is little argument that folding endurance is important in paper and papermaking, the fold test will likely remain an unpopular one. It is generally felt that fold will roughly parallel changes in the more easily measured strength values. Consequently, it seems unlikely that fold endurance will see any surge in popularity in describing paper strength.47

The following general statements can be made regarding the interpretations of data and evaluation of research based on fold endurance values:

The test method was designed for checking continuity of production within the paper industry.

By itself it is not a research tool, although it may be useful in initial planning of a project.

Careful preconditioning of samples and careful control of temperature and humidity during testing is necessary for reliable data.

Data may be questionable if the same load is used before and after artificial aging.

A large sample population is necessary due to large variability in test results.

The same type of investigation that has been used in the discussion on fold endurance tests can be applied to other instruments and research designs in order to better evaluate paper research projects and interpret the data.