New technology for the devulcanization of sulfur-cured scrap elastomers.

by Rudi Kohler, Praxair, Inc., and John O'Neill, Custom Cryogenic Grinding Corp.

The properties of two different types of rubber containing up to 30% devulcanized scrap rubber were compared to those made of virgin material. There was no significant degradation of properties caused by the incorporation of the devulcanized product. The blends provided technical performance near that of the virgin compound.

New cryogenic grinding developments are producing high-value fine crumb at lower costs as compared to the older conventional ambient grinding techniques.

The problem

Most readers are probably aware of the problem - we are producing tons of scrap rubber every year that must be disposed of in some manner. In the United States alone, 170,000 tons of factory scrap are produced annually. That's 5% of total production. In addition, Americans discard 250 million used tires every year and Canada adds another 25 million. What do we do with this factory scrap and all these worn-out tires?

The current solution

A number of technologies have been tried to solve this major waste problem. One of the most common is to burn the scrap rubber as fuel. This certainly works, but it presents new problems of air-pollution control. And it certainly is a low-value use of scrap rubber.

A related technology involves pyrolysis of the rubber to create gas and oil that can be used as fuel and possibly for other applications. This, however, is still a laboratory technology; there are no commercial installations. The economics of the process seem to be a problem.

Civil engineering uses some scrap rubber for such things as playground surfaces, parking lots, bank stabilization, fill under road surfaces and asphalt modifier. While many of these uses make excellent use of old tires, the recent fires have set back the use of ground up scrap rubber for many of these applications.

The future solution

We believe that the best prospect for recycling scrap rubber is to reuse it in new rubber products. This places the highest value on the scrap. Instead of being a disposal problem, the scrap becomes a raw material. There are several approaches to this.

The simplest approach is to use finely granulated rubber as a filler in new rubber products. The finer the crumb the larger the percentage that can be added without degrading the properties of the new product. Generally, however, the addition of unmodified crumb filler is limited to fairly low percentages before the properties of the blend start to degrade.

A number of technologies modify the surface of the crumb rubber, giving it new properties. With surface-modified crumb, a larger percentage of scrap can be incorporated into the compound. There are several technologies used for surface modification. Some depolymerize (break the rubber chains) the surface, and some coat the surface to help it blend into and adhere to the fresh compound. Some technologies coat the rubber particles with plastics and others treat the surface with strong oxidizers to make it sticky. The amount of surface-modified crumb that can be added to a fresh batch is generally higher than that possible with unmodified crumb but still limited to fairly low percentages. Moreover, there are limits on the product application and type of compounds that can employ surface-modified crumb.

Finally, there is true devulcanization of the crumb rubber, which is the technology we will discuss. Devulcanized scrap can be added to fresh compound in much greater percentages than can surface-modified or unmodified crumb. In general terms, for every part of unmodified crumb that can be added to virgin material without degrading the properties, about three parts of surface-modified or seven parts of devulcanized crumb can be incorporated. Because devulcanized rubber can be revulcanized without additional vulcanizing agents, some products can be made with 100% De-Vulc (DV) (ref. 1) devulcanized rubber. Obviously, to recycle large amounts of scrap rubber, devulcanization is the technology of choice. We will compare the properties of rubber made with additions of 30% DV.

Cryogenic grinding

Regardless of whether you are incorporating untreated, surface-modified or devulcanized scrap into virgin compound, the process starts with grinding the scrap to crumb. This can be done at ambient temperatures or cryogenically. Most processors prefer cryogenic grinding, however, because the equipment costs less, operating costs are lower, productivity is increased and the product has better flow characteristics than ambiently ground rubber.

In each case, producing very fine crumb is essential. To produce a satisfactory product, untreated filler crumb must be at least 10 mesh, but 40 or 60 mesh, or even finer, is better. Most surface modification technologies work best with 40 mesh or finer crumb, as does the De-Link (DL) (ref. 2) devulcanization process.

Typically, the first step involves coarse grinding, or granulating, the scrap rubber at cryogenic or ambient temperatures. If tires are being ground, wire and fiber reinforcements must be removed. The end product of this granulation is usually one quarter inch or less.

The coarse product is then put through a precooler where it is sprayed with liquid nitrogen at -320 [degrees] or 195 [degrees] C. This embrittles the rubber making it easier to grind to fine crumb 40 to 60 mesh is readily produced. Grinding is done in a high velocity impact type mill.

Future grinding technologies

Companies that grind scrap rubber want to produce high-value fine crumb economically. As a result, a number of new cryogenic grinding technologies are being investigated. High speed mills that produce finer crumb with less energy and ultra-high-energy mills that grind in a vacuum are being developed. The goal of these new mills is to lower operating costs and increase productivity.

The De-Link devulcanization process

Once the crumb is produced, it can be devulcanized using common rubber machinery - either a two-roll mill or an internal mixer. With a two-roll mill, 100 parts of 40 mesh or finer crumb are mixed with between two and six parts of DL reactant. The amount required depends on the type of rubber being devulcanized and the final form of the devulcanized product desired. The DL process will work with any sulfur-cured compound, and in eight minutes or less, the process will be completed.

Once the devulcanization process is complete, you may elect to keep it as devulcanized crumb or to sheet the DV for shipment or later use. To sheet the DV, it is typically necessary to add about six parts of virgin rubber and two parts of processing aid to improve the handling characteristics of the product and run the mixture through a sheeting mill. An alternative is to add the devulcanized crumb directly to the final rubber compound prior to sheeting.

Devulcanization can also be done in an internal mixer. Two to six parts of DL reactant per 100 parts of 40 mesh or finer crumb are added to the mixer. In five to six minutes, the devulcanization will be complete. If the DV is to be added to a fresh batch of compound, it is easiest just to leave the product as a slightly sticky powder, rather than producing sheets. Cool the crumb and bag it until you are ready to use it. This is the procedure most of our customers follow, if the DV is to be used in-house. Depending on the compound, customers generally blend between 10 and 75% DV with the virgin material without experiencing any significant loss of properties. The blend of DV with virgin material can be molded, extruded or calendered.

The parameters studied

Blending studies were done at Custom Cryogenic Grinding Corp. Additional testing was conducted by the independent BFGoodrich testing laboratories under the direction of Dr. Aggarwal of Global Polymer Technology Associates.

We investigated seven parameters that we felt might affect the properties of the finished product. Two elastomers were tested - natural rubber and SBR compounds. From past experience we knew that different types of rubber required different amounts of reactant for the devulcanization, and we assumed that they might behave differently when devulcanized scrap was incorporated into the compound. The compounds tested are listed in table 1.

Table 1 - NR, SBR compounds

NR compound              Parts by weight
Natural rubber                100
Carbon black N220              45
Zinc oxide                      5
Stearic acid                    2
Santoflex 13                  1.5
Santocure NS                  0.7
Sulfur                        2.5

SBR compound
SBR 1502                      100
Carbon black N220              50
Zinc oxide                      3
Stearic acid                  0.5
Santoflex 13                    1
OBTS accelerator              2.2
Sulfur                        1.8

We then compared the properties of the vulcanized product with various amounts of DV product, made from the same compound, added to the mixture. Concentrations ranged from 0% DV (virgin compound) to 100% DV. The DV was made by vulcanizing samples of the compound, cryogenically grinding them and devulcanizing the crumb before blending it with virgin material. We found that, for the specific types of rubber tested, adding up to 30% devulcanized material to the batch resulted in a final product which performed nearly as well as that made from virgin materials. We chose this as our standard and compared a product made with the addition of 30% recycled rubber to that made from virgin compound.

We also investigated the effects of particle size. We tested crumb from 10 to 80 mesh. We found that crumb size had very little effect on tensile strength, but that elongation at break and Mooney viscosity improved with finer crumb. We chose 40 mesh as our standard, because it seemed to give satisfactory results and is easily and economically obtained via cryogenic grinding technology.

Next we looked at the effect of varying the amount of devulcanizing agent. The DV compound used in the tests was prepared using between three and six parts of DL reactant for every 100 parts of rubber crumb. For the final natural rubber and SBR test samples, the DV was prepared using five parts of DL per 100 parts of rubber crumb.

We tested both ambiently and cryogenically ground crumb. Crumb produced at ambient temperatures has a rough surface, while cryogenically ground crumb is more crystalline in nature with flat planes and straight-line fractures. We wanted to determine if surface morphology of the devulcanized crumb had any effect on the final product. We found that for devulcanizing, both types of crumb behaved very similarly. The choice of method depends mostly on economics. For our tests, cryogenically ground crumb rubber was used.

We also wanted to see what effect mill shear might have on the devulcanizing process, so the nip opening of the two-roll mill was varied from 0.010 to 0.050 inches. As the nip opening increased, the tensile strength and elongation at break of the pure DV decreased and Mooney viscosity increased. For the most effective devulcanization, a nip opening of 0.010 inches was chosen as standard. This test clearly illustrated the importance of shear in the devulcanization process.

We also wanted to determine the effect of reaction time on the devulcanization in the two-roll mill. Using a nip of 0.010 inches, the number of passes was varied from 20 to 40. While the tensile strength and elongation at break improved with the number of passes and the Mooney viscosity decreased, the advantage gained appeared to decline after 30 passes.

Therefore, 30 passes through the two-roll mill was chosen as the standard for subsequent testing.

Two types of rubber were tested - NR and SBR. For the blends, the DV was made using five parts of DL reactant per 100 parts of 40-mesh cryogenically ground crumb that was passed through the two-roll mill 30 times with the nip set at 0.0 10 inches.

The tests

We then vulcanized the four rubber samples to 95% optimum cure and subjected them to different tests given in tables 4-14. In each case rubber made from virgin materials was compared to that containing 30% devulcanized product. Natural rubber and SBR were the two types tested. No adjustments to the cure package were made. However, DV users commonly find minor adjustments, such as withdrawal of expensive accelerators from the virgin compound, will improve the properties even more.

Table 2 - effect of nip opening

Opening,    Tensile strength,    Elongation at     Mooney
inches           MPa               break, %       viscosity

0.010           6.99                311.5             104
0.020           5.69                239.8             125
0.030           3.90                199.2             141
0.040           2.42                184.2          Off scale
Table 3 - effect of reaction time

Number   Tensile strength,   Elongation at    Mooney
passes     MPa                 break, %      viscosity
20        3.38                 254.7            125
25        6.04                 282.4            116
30        6.99                 311.5            104
35        7.67                 285.1            104
40        7.08                 288.2             99
Table 4 - Mooney viscosity

Sample   ML-4@212 [degrees] F   Relative value
NR            61.9                  100
NR-D          72.3                  117
SBR           96.6                  100
SBR-D        109.2                  113
Table 5 - dynamic mechanical properties

Sample        Tan Delta @ maximum torque
NR                      0.056
NR-D                    0.062
SBR                     0.091
SBR-D                   0.097
Table 6 - elongation to break

Sample   Elongation to    Standard    Relative
           break, %       deviation     value

NR          534              13          100
NR-D        489              17           91.6
SBR         358              11          100
SBR-D       345              10           96.4
Table 7 - tensile strength

Sample    Tensile strength,    Standard      Relative
             lb./sq. in.       deviation       value
NR             4,270               110          100
NR-D           4,020               130           94.1
SBR            3,880               120          100
SBR-D          3,580                90           92.3
Table 8 - 300% modulus

Sample   300% modulus,    Standard    Relative
            lb./sq. in.   deviation     value

NR          1,987            37        100
NR-D        2,151            31        108
SBR         3,059            53        100
SBR-D       2,923            26         95.5
Table 9 - tear resistance

Sample   Tear strength,   Standard   Relative
           lb./in.        deviation    value

NR           511             127         100
NR-D         368              10          72.6
SBR          235              16         100
SBR-D        225               4          95.7
Table 10 - compression set

Sample  Compression    Standard    Relative
            set, %     deviation     value
NR          19.5          0.2         100
NR-D        22.4          6.9         115
SBR         15.0          0.1         100
SBR-D       13.6          1.5          90.7
Table 11 - heat build-up

Sample      Temperature       Relative   Compression   Relative
           rise, [degrees]F     value       set, %       value

NR              21               100          9.7         100
NR-D            23               110          7.1          73.2
SBR             37               100          3.9         100
SBR-D           34                91.9        3.5          89.7
Table 12 - blow-out test

Sample      Temperature       Relative         Blow-out
           rise, degrees]F      value     value time, min.
NR             127              100               12.5
NR-D           114               89.8             14.7
SBR            114              100               15.2
SBR-D          104               91.2             16.6

Sample    Relative
           value
NR          100
NR-D        118
SBR         100
SBR-D       109
Table 13 - crack growth

Sample   Kilocycles to 8X
          crack growth

NR             20
               20
NR-D           15
               17.5
SBR             1.0
                1.5
SBR-D           1.0
                1.0
Table 14 - glass transition

Sample  Glass transition, [degrees] C  Transition range, [degrees] C

NR                   -58.4                           5.7
NR-D                 -59.7                           5.5
SBR                  -47.7                           7.6
SBR-D                -48.3                           8.4

In the tables, NR indicates natural rubber made from virgin materials, NR-D indicates natural rubber with 30% devulcanized rubber added to the blend, SBR is that made from virgin materials. SBR-D contains 30% devulcanized SBR in the mix.

* Mooney viscosity (table 4): When devulcanized product is added to the mixture, the Mooney viscosity increases somewhat. This, of course. is a condition easily corrected.

* Dynamic mechanical properties (table 5): Dynamic mechanical properties were measured using a rotorless cure meter at 300 [degrees] F. Only tan delta at maximum torque was tabulated. The tan delta values are not significantly affected by the addition of devulcanized product to the mix. Product made from the blend performed nearly as well as that made from virgin materials.

* Tensile properties (tables 6-8): The elongation to break and tensile from virgin materials and those with 30% devulcanized content are nearly equal. Likewise, the 300% modulus, or the tension needed to stretch the sample to three times its original length, is nearly the same for the blends and the 100% virgin samples.

* Tear resistance (table 9): The presence of devulcanized scrap in the mixture lowers the tear resistance of natural rubber, but has little effect on that of SBR.

* Compression set (table 10): Compression set was measured after 22 hours at 70[degrees]F There was no serious degradation of properties in either rubber.

* Heat build-up (table 11): The heat build-up was measured with a Goodrich Flexometer at 212 [degrees] F. The heat buildup in the natural rubber with DV was very close to that in the natural rubber sample made with virgin materials. SBR with DV showed less heat build-up than did the sample made with virgin compound. Heat build-up correlates with the percent compression.

* Blow out (table 12): The blow-out test, which measures durability under severe use conditions, was also done on the Goodrich Flexometer and is complementary to the heat build-up test. In this test, the sample, at a base temperature of 207 [degrees] F, is flexed to failure. The temperature rise and time to failure are recorded. With both the natural rubber and SBR samples, the temperature rise is significantly lower for the DV blend than for the virgin batch and the time to blow-out is longer. The DV blends performed better than the virgin samples.

* Crack growth (table 13): The De-Mattia test is conventionally used to measure crack growth, an important functional property. It measures the number of kilocycles of tensile deformation for a specific crack to grow to 800%. While this test lacks reproducibility and precision, it is noteworthy that for both SBR and NR, the DV blends are essentially as crack resistant as the corresponding samples made of virgin materials.

* Transition behavior and glass transition (table 14): There is no appreciable difference in the glass transition temperature between the virgin batches and those with DV. There are also no significant differences in the transition behavior and breadth of the transition peaks. These glass transition results indicate that there is no marked difference in the cross-link density of the matrix and the dispersed phase in the DV blends. There was no identifiable phase separation and no multiple transitions.

Conclusions

The results of these tests indicate that using the De-Link process to devulcanize scrap rubber is effective. Blow-out times and the associated temperature rises for the De-Vulc blends are better than for the virgin compounds. Tensile properties, tear resistance, compression set, heat build-up and tan delta values of the DV blends are very similar to those for the samples made with virgin materials. All all, it's an excellent way to turn scrap from a disposal expense into a profit-making raw material.

References

[1.] Kok-Kee Hon and Frederic Siesseger, STI-K Polymers America, Inc. and B.C. Sekhar and S.W. Sin, STI-K Polymers SDN BHD, Malaysia, "De-Link recycling system: A revolutionary process for devulcanization of post consumer and factory waste rubber," presented at the meeting of the Rubber Division of the American Chemical Society, Cleveland, Ohio, October 17-20, 1995, Paper #55.

[2.] European patent application EP 0 690 091 Al, Tan Sri Dr. Balachandra Chakkinggal Sekhar, MY-59100 Kuala Lumpur, Malaysia, and Vitaly Abramovich Kormer, 199057 St. Petersburg, Russia, "Improvements in and relating to the reclaiming of natural and synthetic rubbers," Date of filing 03.03.1995.