Standardization of EPDM characterization tests for QC and specification purposes.

(This is the conclusion of a two-part series. The first installment appeared in May, 1997).

Unsaturation content of EPDM: Ethylidene norbornene and dicyclopentadiene

The exercise carried out in this context was restricted to EPDMs containing ethylidene norbornene (ENB), dicyclopentadiene (DCPD) or both as third monomers for unsaturation.

Because of the more readily analytically accessible unsaturated -C = C- carbon-carbon double bond, the the amount of unsaturation of EPDM has always been somewhat easier than that of the ethylene or propylene content. For this reason, a variety of analytical test methods has based developed (refs. 15 and 16), of which the most important are:

* iodometric titration, resulting in a so-called iodine number, or expressed as C = C double bonds per 1,000 C-atoms;

* refractive index measurement of the pure polymer;

* pyrolysis gas chromatography;

* infrared spectroscopy;

* [sup.1H]-NMR spectroscopy.

Infrared determination of the unsaturation content of EPDM

In the same way as in the exercise concerning ethylene determination, the five participating laboratories measured the ethylidene norbornene content of the samples 1-3, using their own in-house test method. The results are shown in table 6. Again, quite unsatisfactory results were obtained between the different laboratories as regards differences in average numbers as well as regards repeatability. Part of the reason again was the lack of a unified set of calibration standards. Based on an analysis of the coefficients of variation, the infrared spectroscopic measurement seemed to be the best method. Therefore, this method was further pursued within the working group by the exchange of information.

Table 6 - ENB content of the three selected EPDM samples 1-3, measured by own in-house

                   Sample 1               Sample 2

                      x         s     v      x      s     v
Lab 1
Refractive index   4.61.     .148   3.2   2.20   .149   6.8

Lab 2
Infrared spectr.                          2.56   .048   1.8

Lab 3
Infrared spectr.   3.80(*)   .034   0.9   2.45   .025   1.0

Lab 4
Refractive index   4.7        .03   0.6    2.4    .07   2.9

Lab 5
Iodine number      4.8        .13   2.7   2.45    .08   3.3

                   Sample 3

                      x      s     v
Lab 1
Refractive index   9.10   .126   1.4

Lab 2
Infrared spectr.   6.73   .043   0.6

Lab 3
Infrared spectr.   9.60   .077   0.8

Lab 4
Refractive index    8.9    .05   0.6

Lab 5
Iodine number       9.0   .165   1.8

x = average (mass%) s = standard deviation (mass%) v = coefficient of varation (%) * Sample 1 contained some extender oil; result not corrected for oil content.

Further study showed that the determination of the film thickness of the sample used for the infrared technique was the main cause of the variation in the results. It is rather difficult to press a film with a homogeneous thickness, particularly for polymers with a high ethylene content. The quantitative interpretation of an infrared spectrum is therefore hampered by the inaccuracy in the sample thickness at the point where the actual spectrum is measured. Different methods were used to determine the sample thickness:

* mechanical means, such as thickness measurement using callipers;

* amount of absorption of beta radiation;

* measurement of the thickness included in the infrared technique by measurement of the infrared absorbance at a so-called isotopic point, where no specific absorption of chemical groups takes place.

One of the members of the working group (Exxon Chemical) has recently developed the latter method, apparently with very positive results. This method provides automatic film thickness determination at the same point of the film at which the spectrum is taken. Depending on the particular sample used, the thickness gauge is the net absorbance difference between:

* Group 1: 2,708 [cm.sup.-1] (isotopic point) and 2,450 [cm.sup.-1] (anchor point), or

* Group 2: 2,668 [cm.sup.-1] (isotopic point) and 2,450 [cm.sup.-1] (anchor point).

The contents of unsaturation are measured by using the following response peaks:

* ENB: 1,690 [cm.sup.-1], or rather the second derivative in the region 1,681 - 1,688 [cm.sup.-1];

* DCPD: 1,611 [cm.sup.-1], or rather the second derivative in the region 1,612 - 1,619 [cm.sup.-1].

This method has been published (ref. 17), and the reader is referred to this article for further details.

As infrared spectroscopy is only a relative technique, this method also needs standard calibrants. See the section on calibration standard for ASTM D3900 for a detailed description. For ENB-containing polymers, which generally cover a broader range of contents than DCPD-containing polymers, four standard calibrants are available, as opposed to two for DCPD. In fitting the infrared responses to these standard calibrants, the experience of the working group was that the use of the least square fit of these ENB standard calibrants gave slightly better results overall than a linear extrapolation for EPDMs with intermediate ENB contents. The working group therefore recommends the least square fit.

In addition, this method for determining unsaturation content sets certain requirements for the Fourier transform infrared instrument; it should have spectral accumulation, averaging and subtracting capabilities and should be fitted with a sample shuttle to permit alternating and repetitive collection of single beam sample and "empty sample compartment" spectra. If a shuttle is not available, this method cannot be implemented exactly as described. In such cases, the method may be modified. but only at the expense of some development work. A preferred (but not mandatory) instrument for the procedure is the Perkin Elmer 1760 series spectrophotometer with its dedicated 7500 or 7600 computer.

In the case of oil-extended EPDM polymers, the oil must be extracted before the unsaturation is determined. A suitable procedure is described later. Further caution has to be taken with stearic acid or stearates contained in the EPDM polymer, for which the spectrum needs to be "cleaned."

Based on this Fourier transform infrared method and corresponding standard calibrants, three laboratories (one laboratory did not have the right instrument at that time) compared their ability to achieve the same results and sufficient repeatability on the four standard calibrants belonging to this technique as well as on the samples 1-3 (tables 7 and 8).

Table 7 - Fourier transform infrared determination of ENB content of four ENB standard calibrants

                   ENB standard 2      ENB standard 3

                     x     s      v     x       s     v
Assigned ENB
content (mass %)   2.26                5.11

Lab 1              2.15   .028   1.3   5.29   .057   1.1

Lab 2              2.18   .061   2.8   5.09   .080   1.6

Lab 3              2.25   .058   2.6   5.02   .082   1.6

                   ENB standard 4

                    x      s      v
Assigned ENB
content (mass %)   9.99

Lab 1              9.91   .264   2.7

Lab 2              9.80   .088   0.9

Lab 3              9.80   .118   1.2

ENB standard 1 contains 0.0 mass % ENB and was not measured, but used only for calibration purposes.

Table 8 - ENB content of three selected EPDM samples 1-3, determined using the new Fourier transform infrared method

        Sample 1            Sample 2            Sample 3

         x       s     v     x      s      v     x      s      v

Lab 1                       2.45   .025   1.0   9.60   .077   0.8

Lab 2   4.53   .042   0.9   2.42   .034   1.4   9.62   .046   0.5

Lab 3   4.64   .045   1.0   2.36   .008   0.4   9.68   .171   1.8

The correspondence between laboratories is remarkably good. The conclusion can be drawn that the participating laboratories managed repeatability levels using this method which were on average three times smaller (as expressed by the coefficient of variation) than those achieved using their own in-house method.

On this basis, the HSRP working group decided to recommend the Exxon test method for the determination of mass % unsaturation in EPDM.

Calibration standards for unsaturation

It is commonly accepted that [sup.1H]-NMR is the best technique for obtaining quantitative data on the content of unsaturation in EPDM, irrespective of whether the unsaturation is based on ENB or DCPD. Therefore, a set of standard calibrants were certified using [sup.1H]-NMR, using samples dissolved in deuterated o-dichlorobenzene at 120 [degrees] C. The following standard calibrants were developed:

Ethylidene norbornene

* Standard 1: 0 mass %

* Standard 2: 2.26 mass %

* Standard 3: 5.11 mass %

* Standard 4: 9.99 mass %

Dicyclopentadiene

* Standard 1: 0 mass %

* Standard 5: 4.80 mass %

All contents were based on the recommended compositional definition given in the section on the description of molecular composition of EPDM. These standard calibrants cover the range of commonly available commercial EPDM rubbers, where slight extrapolation to somewhat higher numbers may be necessary in exceptional cases.

Determination of the stabilizer content (phenolic antioxidants)

During the course of the study. the question was raised as to whether harmonization of the determination of stabilizer content of EP(D)M would be desirable. An overview of the different methods in use at the different suppliers revealed the following:

* infrared spectroscopy: three suppliers; typical coefficient of variation 4-13%;

* HPLC (high pressure liquid chromatography): one supplier; typical coefficient of variation 3%;

* photometric detection: one supplier; no further details.

The fact that the determination of stabilizer content can be included in the infrared determination of ethylene content and unsaturation content without much extra effort is a strong point. The HPLC method involves a great deal of work. In addition, the amount of stabilizer does not normally need to be determined with such great accuracy, as stabilizer content is usually specified as a one-sided minimum requirement. It was agreed, therefore, that the possibility of a slight gain in repeatability with HPLC compared with infrared spectroscopy was not really worth the effort.

Infrared determination of the stabilizer content EP(D)M

In the case of commonly used stabilizers of the hindered phenol type, the determination can be included in the infrared determination of ethylene content, making use of the infrared absorbance at 1,740 [cm.sup.-1].

The height of the peak is measured and is quantified by calibrating the instrument with known standards. In the same way as for the ethylene content determination, the film thickness gauge may be determined as the net absorbance difference between the isotopic points and the anchor point for groups 1 and 2.

The test method as described above has to be the only preferred test method, due to certain restrictions:

* it is useful only for sterically hindered phenolic antioxidants;

* it may show an interference between ester additives and the EP(D)M (stearates);

* in the case of oil-extended products, it can be used only with paraffinic oil.

The IISRP working group recommends infrared spectroscopy for the determination of the stabilizer content of EP(D)M.

Calibration standards for stabilizer content

Calibration standards are best achieved by mill-mixing known amounts of the actual stabilizer in pure EP(D)M samples. Because of the limited stability of the stabilizer content during the shelf-life of such calibration standards, it was not considered useful to develop a unified set of standard calibrants.

Determination of the oil content of oil-extended EPDM

Determination of the extender oil content of EPDM is another important analytical technique. In addition, many of the aforementioned analytical techniques (ethylene content and unsaturation content) require the oil to be removed before the measurement can be carried out.

Methods commonly in use can be divided into two categories - extraction methods and precipitation methods.

The first category makes use of a solvent which dissolves the extender oil but which does not dissolve the EPDM. Depending on the experimental configuration, extraction is achieved either by step-by-step extractions in flasks with regular renewal of the extraction medium or using Soxhlet apparatus. The second category makes use of suitable solvent/non-solvent combinations, where first of all the oil-extended polymer is dissolved completely, then the EPDM polymer is precipitated and separated from the liquid phase containing the extender oil. For both categories either the EPDM moiety or the oil moiety after evaporation of the extraction medium can be used for the calculation of the oil content.

There was no reason beforehand to prefer either of the two categories, as can be seen from the first attempt to standardize this analytical method. Each participating laboratory measured the oil content of sample 1, using its own in-house method (table 9).

Table 9 - oil content of sample 1, measured by own in-house method

    Method            Solvent      Non-solvent        Appr. duration

Lab 1 Extraction                   2-propanol            4 hours

                                                        16 hours

Lab 2 Extraction                       MEK               2 hours

Lab 3 Extraction                 Ethanol/toluene
                                     azeotrope

Lab 4 Precipitation   Toluene    Methanol/acetone        3 hours
                                     50/50

Lab 5 Precipitation   Toluene         Acetone           2.5 hours

      Extraction                Acetone/cyclohexane       1 hour
                                        2/1

    Method                 x         s     v

Lab 1 Extraction         23.6       .31   1.3
                      via extract
                         24.1       .03   .14
                      via extract
Lab 2 Extraction         24.5       .25   1.0

Lab 3 Extraction         23.9       .17   3.4

Lab 4 Precipitation      22.3       .77   3.4

Lab 5 Precipitation      23.6       .32   1.4

      Extraction         23.8       .10   .42

x = average (mass %); s = standard deviation (mass %); v = coefficient variation (%)

An important criterion in this test is the duration, particularly as this test is to be used for quality-control purposes. On die other hand, as can be seen from table 9, the repeatability of the extraction method seems to improve when longer extraction times are employed. This is of particular importance for EPDMs with a relatively high ethylene content, for which it is well known that the extraction of extender oil can be a time-consuming exercise: the diffusion of the extender oil from the EPDM into the solvent is apparently retarded by crystallinity or by a lower degree of swelling of the EPDM in the non-solvent.

After thorough consideration the decision was taken to pursue standardization by improving the extraction methods employed by laboratories 2 and 5. Laboratory 2 made use of Soxhlet apparatus with boiling MEK, laboratory 5 of a one-off extraction of [+ or -] 200 mg of EPDM rubber in a boiling azeotrope of 40 ml acetone and 20 ml cyclohexane using a conical flask. The main difference was the speed with which the Soxhlet method reached equilibrium compared to the conical flask method for a high-ethylene-containing EPDM (sample 4) extended with 100 phr of paraffinic oil (see figure 2). The higher speed of equilibrium achieved with the Soxhlet method gives an extra guarantee for better repeatability within as short a time-span as possible for the analysis. Why the Soxhlet method, using MEK as the extraction agent, systematically tends to give higher extraction numbers than the conical flask method, using the 2/1 mix of acetone/ cyclohexane, could not be explained. It may well be that the different extraction agents tend to differ a little, depending on their solvent power for EPDM, as regards the amount of very-low-molecular-weight, oligomeric EPDM they extract from the rubber along with the oil. This is a systematic problem with all oil-determination tests, irrespective of whether they are extraction or precipitation methods. To the committee's knowledge there is no fundamental solution to this problem, according to the present state of the art. As long as everybody uses the same method, the same systematic error (bias) will be made by everybody and this will not lead to major differences in comparison of the data. Although it was felt that the conical flask method had not been fully exploited, it was generally felt that overall all the laboratories achieved the best reproducibility and the same average results with the Soxhlet extraction method. The duration of the extraction, either 30 or 45 minutes, had only a small influence on the results. The IISRP working group decided to recommend the Soxhlet extraction method, using MEK as the solvent, as the preferred method for the determination of the oil content of oil-extended EPDM.

This method has been published (ref. 18), and for a detailed description of the method the reader is referred to the pertinent article.

Mooney viscosity

Sample preparation

An important factor affecting the precision of the Mooney measurement has over the last few years proved to be the sample preparation (refs. 6-8). Two types of sample preparations in particular have become more or less officially accepted:

* massing on a mill;

* no mill-massing: by cutting a sample from a bale or, in the case of the rubber's being in crumb or pellet form, by compacting a pre-dimensioned sample in a press.

The obvious question was whether either of these two sample preparations could also be recommended for EP(D)M. A review of existing standards for Mooney measurement and/or sample preparation shows the following:

* ISO 4097 (ref. 19) and ISO 1795 (ref. 20): no mill-massing preferred, if massed: 35 [+ or -] 5 [degrees] C;

* DIN 53 523 (ref. 21): no mill-massing, vacuum-compacting or mill-massing at 35 [+ or -] 5 [degrees] C;

* ASTM D1646-90 (ref. 22): massed or unmassed, if massed: 50 [+ or -] 5 [degrees] C.

A review of the actual conditions used by the different laboratories revealed the following:

* Lab 1: mill-massing at 105 [degrees] C;

* Lab 2: no massing: compacting;

* Lab 3: mill-massing at approx. 125 [degrees] C;

* Lab 4: mill-massing at 140 [degrees] C;

* Lab 5: mill-massing at 50 [degrees] C.

In any case, on the basis of their own experience of reducing the test variability, most members do not adopt the lower temperatures recommended in the standard test methods.

A review of the experiences of the participants of this study concerning the influences of mill-massing and a lack of mill-massing on the actual Mooney values measured and on the repeatability of the Mooney measurement gives rise to conflicting evidence. Depending on the type of EPDM rubber being tested and on the particular participant carrying out the test, several cases were observed:

* average Mooney value is the same; massing at high temperature gives lower standard deviation;

* average Mooney value is a few units lower with massing; massing at high temperature gives lower standard deviation;

* average Mooney value is a few units lower with massing; standard deviation the same;

* average Mooney value is somewhat higher with massing, particularly for free-flowing polymers.

It is quite clear that no consistent picture can be derived from these observations. In any case, the advantage of the use of unmassed samples does not show up clearly if compared to the high temperature mill-massing.

The IISRP working group decided to support the unmassed sample preparation as the preferred method, simply because of ease of handling. The choice of whether or not to use mill-massing is left to the discretion of the various producers. It is recommended that the sample preparation be specified in reporting data.

Mooney test conditions

In the course of time, a wide variety of test conditions for Mooney viscosity has come to be used by the various suppliers. A good overview can be gained from the Synthetic Rubber Manual (ref. 23), wherein eight different conditions are quoted as being in common use for measuring Mooney viscosities of EP(D)M:

* ML(1+4) 100 [degrees] C;

* ML(1+8) 100 [degrees] C;

* ML(1+8) 120 [degrees] C;

* ML(1+4) 121 [degrees] C;

* ML(2+10) 121 [degrees] C;

* ML(1+4) 125 [degrees] C;

* ML(1+8) 125 [degrees] C;

* ML(1+8) 150 [degrees] C.

A summary of the preferred conditions recommended in the appropriate standard test methods either for EPDM or for the Mooney measurement itself shows tile following:

* IS0 4097-1980 (ref. 19): ML(1+4) 125 [degrees] C

* DIN 53 523 (ref. 21): ML(1+4) 100 [degrees] C

* ASTM D1646-90 (ref. 22): ML(1+4) 125 [degrees] C

The origin of all these conditions is impossible to trace but they are to a large extent the result of difficulties encountered in accurately measuring the Mooney viscosities for the range of EP(D)M polymers on the market (refs. 5, 24 and 25). In particular, the fact is often overlooked that all pertinent standard test methods for Mooney viscosity call for caution if measurements are carried out on high-molecular-weight polymers. They put the provisional restrictions on the Mooney value that it should not exceed 80, due to mechanical equipment limitations which may result in erratic results. It is quite clear that one condition is by no means sufficient to cover the whole range of EP(D)Ms available and therefore different conditions have originated. The question may be posed though whether the amount as cited above is practical.

The main point of discrimination between die various test conditions is the temperature at which the test is carried out. This is the main tool for bringing the Mooney viscosity within the necessary experimental range. And because there are various EP(D)M types being marketed with extremely low Mooney viscosities ([is less than] 10 at 100 [degrees] C), there is sometimes a tendency to lower the temperature conditions to such a temperature and even to standardize the test at that temperature. However, a complicating factor at a temperature of 100 [degrees] C is that for polymers with a high ethylene content remnants of crystallinity may persist up to temperatures above 100 [degrees] C, which will have a rather significant but irreproducible influence on the Mooney viscosity measured at that temperature. There is basically no way out of this dilemma.

In order to standardize the Mooney measurement for EP(D)M rubber, the IISRP working group based its considerations on the experimental work already carried out on earlier occasions (refs. 5, 23 and 24) and reached agreement on a recommendation to reduce the various possible conditions to mainly two:

* ML(1+4) 125 [degrees] C;

* ML(1+8) 150 [degrees] C, leaving the possibility open to use other conditions if the need should arise.

The first condition is the one which is most widely used and which is generally recommended in the appropriate standard test methods. The second condition is recommended for high Mooney and/or heavily branched grades. A clear guideline as to when there should be a switch to the second condition cannot be given because it depends on equipment factors in combination with the particularities of the various grades. This is left to the discretion of the particular supplier.

Availability of standard calibrants

The standard calibrants for the ethylene and unsaturation determination were produced in commercial plants, and sizeable quantities have been set aside. Samples of these calibrants are available from:

Exxon Chemical Technology Center Brussels

Hermeslaan 2

B-1831 Machelen

Belgium and

Exxon Chemical Co., Baytown Polymers Center

Vistalon Technology Manager

5200 Bayway Drive

P.O. Box 5200

Baytown, TX 77522 USA

Conclusions

In a joint project between European EPDM producers under the auspices of the IISRP a set of recommended characterization tests was developed for quality-control and specification purposes for EPDM rubber. These test methods include procedures for:

* ethylene content;

* unsaturation content (ENB and/or DCPD level);

* stabilizer content (phenolic type);

* oil content of oil-extended EPDM types;

* Mooney viscosity.

For ethylene content, a new set of standard calibrants was developed in cooperation with the ASTM D11.11 rubber-testing sub-committee. These are partially based on the older calibrants of ASTM Standard D3900, and involve a shift in mass % ethylene up to 4.5 mass %, relative to the formerly assigned numbers. Standard calibrants were also developed for unsaturation content. All standard calibrants are available upon request.

The decision to implement such test methods, as well as the timing thereof, are left to the discretion of the individual suppliers and consumers.

References

[1.] Worldwide rubber statistics, International Institute of Synthetic Rubber Producers Inc., Houston, TX.

[2.] J.W.M. Noordermeer, Kirk-Othmer Encyclopedia of Chemical Technology - Fourth Ed., 8, 978 (1993), John Wiley & Sons, Inc.

[3.] ASTM D3568-90. Standard Test Methods for Rubber - Evaluation of EPDM (ethylene propylene diene terpolymers) including mixtures with oil.

[4.] Prufung von Kautschuk und Elastomeren, DIN 53 670 Teil 10: Prufung von Kautschuk in Standard Testmischungen; ethylen-propylen-dien kautschuk EPDM, 1983 issue.

[5.] International Standard, ISO 4097 rubber, ethylene-propylene-diene (EPDM), non-oil extended raw general-purpose rubber - evaluation procedures, 1991 issue.

[6.] J. Markert, Gummi Asbest Kunststoffe, 9, 568 (1976).

[7.] R. Koopmann, Kautschuk + Gummi, Kunststoffe, 38, 281 (1985).

[8.] H. Kramer, Kautschuk + Gummi, Kunststoffe, 43, 912 (1990); Rubber World, 204, 35 1991).

[9.] W. Breemhaar, R. Koopmann, J. Markert and J. Noordermeer, Kautschuk Gummi Kunststoffe, 46, 957 (1993).

[10.] International Standard, ISO 5725, Precision of test methods - determination of repeatability and reproducibilily for a standard test method by inter-laboratory, tests.

[11.] I.J. Gardner, C. Cozewith and G. Verstrate, Rubber Chemistry and Technology, 44, 1015 (1971).

[12.] Standard Test Methods for Rubber, ASTM D3900, determination of ethylene units in EPM and EPDM, 1980 issue.

[13.] S. DiMartino and M. Kelchtermans, Determination of the composition of ethylene-propylene-rubbers using [sup.13C]-NMR spectroscopy, J. Appl. Poly. Sci. in print.

[14.] Standard Test Methods for Rubber, ASTM D3900, determination of ethylene units in EPM and EPDM, 1994 issue.

[15.] J. Gardner and G. Verstrate, Rubb. Chem. and Techn., 46, 1019-1034 (1973).

[16.] J. van Schooten and J.K. Evenhuis, Polymer 6 (11), 561-577 (1965).

[17.] Determination of 5-ethylenenorbornene (ENB) by Fourier transform infrared spectroscopy, Materiaux et Techniques December (1991), 69-72.

[18.] J. Fourreau, Oil determination by MEK extraction, Materiaux et Techniques, 4-5, 80 (1992).

[19.] ASTM D1646-93, Standard Test Method for Rubber - Viscosity and Vulcanization Characteristics (Mooney viscometer).

[20.] International Standard ISO 1795, Rubber, raw, natural and synthetic; sampling and further preparative procedures, 1992 issue.

[21.] Prufung von Kautschuk und Elastomeren, DIN 53 523, Prufung mit den Scherscheiben-Viscosimeter nach Mooney, 1992 issue.

[22.] International Standard, ISO 289-1, Determination of viscosity of natural and synthetic rubbers by the shearing disk viscometer, 1994 issue.

[23.] The Synthetic Rubber Manual, International Institute of Synthetic Rubber Producers, Inc., Houston 1992, 12th Ed.

[24.] R.J.H. America, G.W. Visser and W. Breemhaar, PRI Conference, Belgian Section, Leuven, Belgium; 16-17 April 1991.

[25.] E.T. Italiaander, EPR raw polymer Mooney and its importance of using the correct test temperature to ensure reproducibility of test data; Polysar, technical memorandum, Antwerp, 8 November 1988.

Dr. Noordermeer authored this on behalf of the International Institute of Synthetic Rubber Procedures. The objective is to achieve standardization among EPDM producers of testing procedures for the basic properties. This article was originally published in Kautschuk + Gummi, Kunststoffe. The first part will ran in May.