Author
Dr. Samuele Giani
Softening point (SP) is an important thermal parameter for the characterization of materials in many industries. The two standard analytical methods used across a range of samples from bitumen to greases, waxes, and resins, are the ring-and-ball and the METTLER TOLEDO cup-and-ball method. The question that often arises is how the SP results obtained with the two methods compare. Statistical practice can provide a definitive answer.
Methods for softening point determinations
The softening point is the temperature at which, under defined testing conditions, the test substance escapes from a sample holder with an aperture and flows through a stated distance. A ball can be used as a weight to promote the flow. The determination of the softening point has been standardized with international norms and standards to ensure comparable results. Historically, the ring-and-ball (RB) setups came first [1]. It involves the use of a thermostatted liquid bath, a mercury thermometer, and a gauge for the distance. The specified sample holder is in form of a ring, giving this method its name.
Although the RB method has a simple setup, it has several disadvantages. Depending on the softening temperature of the test substance, different liquids have to be used in the liquid bath. As the substance under investigation is in direct contact with the liquid, there must be no reactivity between the test specimen and the medium. It is also important that the liquid should display uniform viscosity throughout the experimental temperature window. Once the ball has flowed through the ring, the setup must cool down and be cleaned thoroughly, making the RB method time-consuming. A large volume of liquid needs to be replaced with fresh liquid after a few experiments.
METTLER TOLEDO DP Excellence Systems for softening point determination operate according to the cup-and-ball method [2]. This setup differs in various respects. The temperature control is ensured by a metal block heating principle and the cup-and-ball temperature is recorded by a digital thermometer.
The sample is placed in a cup and can flow freely downwards through an aperture in the cup. As with the RB setup, a ball also promotes the flow of the sample; however, here it is blocked by the smaller diameter of the cup and does not flow through with the sample. The analysis takes place in a glass container that is disposed of after the experiment, thus avoiding furnace contamination. With a DP70 or a DP90 instrument, the SP detection is fully automatic and performed by video analysis.
Agreement between ring-and-ball and cup-and-ball methods
A question that typically arises is how the softening point results obtained by the two methods compare. Do the two techniques deliver the same results? In the same way that SP can be measured with different methods, many types of statistical analysis can in principle be used to assess the quality of the results.
We decided to proceed according to ASTM D6708 “Standard Practice for Statistical Assessment and Improvement of Expected Agreement Between Two Test Methods that Purport to Measure the Same Property of a Material”. In this standard test, the methodology and criteria are fixed, described, and unequivocally agreed upon.
ASTM D6708 aims to assess the equivalence of the results. This equivalence is measured to statistically relevant criteria, using a confidence limit of 95%. It also identifies if and with which constant, proportional, or linear bias correction can improve the degree of agreement.
Interlaboratory study of adhesives and ink resins
Previous studies already addressed the SP obtained via RB and CB methods [3]. In this work, an analysis of an interlaboratory study (ILS) on 7 substances was carried out according to ASTM D6708. The measurements were conducted in 11 different laboratories and the SP measurements were carried out both with RB and CB methods. Six-fold measurements per technique and per sample were performed. The substances were chosen to cover a large temperature range. These substances are summarized in Table 1. Adhesive resins display a lower temperature SP compared with ink resins.
 |
Table 1. Sample substances used for the ILS softening point study. |
Calculations and results
We followed ASTM D6708 using statistics to quantify the degree of agreement between two test methods after bias correction. Glycerol ester is taken as a representative example and Figure 1 summarizes all results obtained from the ILS for this sample. The overall mean, standard deviation of the results for intra- and interlaboratory results, and standard error are plotted for RB and CB tests.
Figure 1 demonstrates how close the SP results of RB and CB are, based on the example of glycerol ester. Both CB standard deviations are smaller than the corresponding RB, which indicates better precision, and thus better repeatability and reproducibility of the CB method. This is also the case for the other substances. The outcome of the scientific study according to ASTM D6708 is shown in Figure 2.
It can be summarized as follows:
- The two methods, RB and CB, can be considered equivalent.
- There is no bias correction that could improve the degree of agreement between CB and RB results. This means that there is no better correction to be applied to a single SP result obtained from CB method to obtain a result for the RB method.
 |
Figure 1. SP results for glycerol esther. |
 |
Figure 2. ASTM D6708 states that no correction for bias between CB and RB results exists. As is represented in this figure, there is a close correlation between the RB and CB methods and the results can be considered equivalent. |
Conclusion
The results of all the samples measured by RB and CB methods were assessed in accordance with the procedure outlined in practice ASTM D6708. They showed that the two softening point techniques, CB and RB, may be considered to be practically equivalent for sample types and property ranges studied.
No sample-specific bias was observed for the materials studied. Differences between results from the two test methods are expected to exceed 1.3 °C (between methods reproducibility) about 5% of the time.
In other words, no bias correction has to be applied to the results obtained from CB method to obtain a result for the RB method. The two methods, cup-and-ball and ring-and-ball are equivalent: the two measurement methods can be interchanged.
References
[1] ASTM E28, ISO 4625-1
[2] ASTM D6090, ISO 4625-2
[3] AnaChem UserCom 19, 18
