Mold fouling during rubber vulcanization.

Mold fouling is a commonly seen phenomenon in the rubber processing industry. During vulcanization, a layer is deposited on the wall of the mold that builds up gradually with each subsequent production cycle. A previous article (ref. 1) discussed the investigation of different factors that can cause mold fouling. Here it was found that for the different sulfur (and zinc oxide) containing polymeric rubber compounds, zinc sulfide (ZnS) was the most disturbing reaction by-product of curingvulcanization that causes mold fouling. No semi-permanent release agents or permanent (metal) coatings can avoid this deposition. The conclusion was that mold fouling is initially caused by ZnS (inorganic deposition) which adheres to the mold and causes a gray (building up) layer of deposit. As a function of temperature, low molecular weight components from the compound adhere to the ZnS crystallites and cause a second stage of (organic) deposition, and this forms, in due time, oxidation products and causes carbon deposition.

By understanding the cause of mold fouling and the underlying mechanism of the formation of ZnS crystallites on the metal surface (of the mold) it should be possible to generate adaptations to the current process to reduce this ZnS formation and thus prevent mold fouling. By investigation of the cause of the mold fouling, it should be possible to reduce this phenomenon.

There are two possible solutions to prevent or minimize the formation of mold fouling:

* change compounding ingredients; or

* modification of the surface of the mold.

Reduction by compounding ingredients

Mold fouling by compounding zinc oxide or sulfur has to be reduced or eliminated. Most deposition products are related to high sulfur compounds and zinc oxide. These ingredients are commonly used in rubber compounds for tires. Tire production is worldwide the largest by volume (up to 75%) of total rubber products, and therefore, most experiments are carried out on blends of NR/BR compounds and SBR compounds commonly used in the manufacturing of tires. In the case of reduction of mold fouling by compounding ingredients, investigations were carried out on the determination of zinc sulfide, short run vulcanization experiments and the influence of compounding ingredients.

Determination of zinc sulfide (ZnS)

This study started with the investigation of the formation of ZnS, which is the initial cause of mold fouling. Vulcanization experiments showed that ZnS was formed on the metal surface. The deposits on the inserts were examined microscopically at a magnitude of 1,000x to determine the first 'visible' crystallites, and afterwards by means of Rontgen micro analysis (RMA), as shown in figure 1. The RMA elementary analysis detected primarily zinc and sulfur. Based on the detected zinc/sulfur ratio, it was concluded that the crystallites were primarily made of insoluble ZnS (figure 2). A physical analytical method (AP-TPR) was used to analyze the H2S content in the vulcanized compound in order to determine the presence of formed ZnS (indirect method). A model vulcanization experiment was carried out at 200oC under oxygen-free conditions in a closed tube containing squalane, zinc oxide, sulfur and high-surface elemental iron to determine the formation of ZnS in the presence of iron. In this experiment also, ZnS was detected by means of RMA. Both experiments led to the formation of ZnS, as was expected (ref. 1). However, there was no evidence that ZnS was formed at the interface of the compound and mold, or was formed during vulcanization as a reaction by-product of zinc and sulfur.

To determine the ZnS content in the rubber compound, another method was used. The molded (vulcanized) rubber was cryogenically ground into small granules, extracted with acetone and treated with a mixture of hydrochloric acid and acetic acid; whereby the metallic sulfides are decomposed. The resulting hydrogen sulfide was absorbed in a buffered cadmium acetate solution and the cadmium sulfide that was formed was determined iodometrically. Furthermore, the extracted rubber was digested with sulfuric acid/nitric acid in a microwave oven. In the digested solution, an element scan was performed by ICP-ES.

From these results, it was concluded that ZnS was formed as a reaction product from zinc oxide and sulfur. In the vulcanized product, this reaction product is available for the formation of ZnS crystallites at the interface of the robber product and the mold.

The most acceptable hypothesis is a chemical formation of ZnS as a reaction product from zinc oxide and sulfur. This general reaction is described in different rubber handbooks. A simplification of the reaction mechanism is: 2 RH + [S.sub.x] + ZnO + (accelerator) [right arrow] R-S[(.sub.x-1)] -R + ZnS + [H.sub.2]O

Most tire compounds contain 5 phr zinc oxide and about 2 phr sulfur. For a tire compound, it can be calculated that a formulation based on 100 phr rubber (total of about 175 phr) contains 2.8 wt. % zinc oxide and 1.1 wt. % sulfur. It also can be calculated that 1 gram of ZnO gives about 0.6 gram ZnS. As can be seen from these calculations, a substantial amount of ZnS could be generated. In practice, only the ZnS present in the upper layer of the tire gives ZnS crystallites (possibly caused by the metal surface). About 500 moldings are performed before a mold cleaning operation has to be carried out.

Experiments

It was already known that inserts (small metal plates) used to act as a mold wall and containing zinc sulfide crystallites can, in principle, easily be analyzed by means of RMA, but this is a rather expensive test. Therefore, a simple test was developed to determine the first (visible) crystallites on the inserts. By means of a light microscope at a magnitude of 500x, individual crystallites of 0.5 to 1 micron are visible. To generate crystallites, vulcanization experiments were carried out on different compounds, different temperatures and times. Two compounds were selected and are shown in table 1, including a tread compound based on s-SBR and an intermediate compound based on a blend of NR/BR. Both compounds were used as a master-batch for the different vulcanization experiments.

To carry out vulcanization experiments, a simple compression mold, applicable for up to eight inserts, was manufactured for short run experiments up to 20 cycles. After each five subsequent cycles, the inserts were visually examined. In this way, parameters such as compounds with different additives or, for the inserts, the choice of metal, roughness, or coatings could be examined. The preliminary experiments were carried out on the two basic tire compounds cured for 20 minutes at 160[degrees]C and two minutes at 200[degrees]C (data calculated from rheometer curves). The results in table 2 showed that no difference in the quantity of ZnS crystallites can be found visually. To reduce time, all further experiments were carried out at a temperature of 200[degrees]C. After short run experiments up to 20 vulcanizations, further experiments have been carried out by means of injection molding up to 500 vulcanizations.

Influence of compounding ingredients

Zinc alternatives

Experiments were carried out on compounding ingredients. As already shown, ZnS is formed as a reaction product of zinc oxide (or zinc containing ingredients) and sulfur. It will not be easy to eliminate sulfur or zinc sulfide from the formulation, because both ingredients are essential in rubber compounding. Elemental sulfur gives high mechanical strength and bonding behavior, and zinc oxide activates the vulcanization system.

Rheometer experiments showed that the zinc oxide level may be reduced from 5 phr to 3 phr without change in the rheometer maximum torque, almost a reduction by a factor of two. However, even with this level of reduction, no change in the deposit of ZnS was visible (table 3). Also, by replacing the zinc oxide by a smaller particle size, no difference in deposit is visible compared to using ZnO (RS).

But, when the zinc oxide was replaced by 0.25 phr of nano zinc oxide with 40 nm particle size (with the same rheometer maximum), and thus reducing the zinc oxide level by a factor of 20, a difference in deposit was evident. Short run vulcanization experiments, up to 20 cycles in a compression mold with nano zinc oxide, showed no deposition of ZnS.

Replacement of zinc oxide

The replacement of zinc oxide (ZnO) by other metal oxides (table 4) such as calcium oxide (CaO) or magnesium oxide (MgO) was not a solution, because in that case CaS or MgS was formed.

An amine was also investigated as an activator in comparison to zinc oxide. For this purpose, a multifunctional amine (MFA = Duomeen TDO, Akzo) was added to the compound and, in addition, the vulcanization system was optimized by adding 0.2 phr of ZBEC. No crystallites or other deposits were visible on the inserts. The use of this low concentration of ZBEC gave no deposit of ZnS in short run experiments. However, the use of amines is not advisable because of the unpleasant smell and toxicity (depends on type). No experiments on these tire compounds were conducted in tire molds.

Some additional experiments were carried out on inhibitors. It is claimed that inhibitors are materials that prevent or reduce mold fouling. Different kinds of inhibitors have been tested, such as ethylene diamine tetra-acetic acid (EDTA), lauryl pyridine chloride (LPC), 2-amino-2 methyl-1propanol (AMP) and benzotriazole (BTZ). None of those investigated chemicals reduced the deposition of ZnS.

Environmental benefits of the reduction of zinc oxide

Reduction of zinc oxide or replacement by nano-zinc oxide as a part of the activator system or by alternative vulcanization systems reduces the deposits onto the mold. This leads to longer stand times (presuming up to at least 10 times) of molds, and therefore a cost reduction in mold cleaning operations. The reduction of zinc oxide is also beneficial to the environment. Pollution of rubber by tire abrasion on roads also causes an environmental problem. Leaching of zinc components from the tire rubber into the surface water disturbs or destroys the microbiological balance in the aquatic environment.

However, tire manufacturers are not willing to reduce the zinc oxide levels in robber. The reason is that those compounds give reduced properties such as reversion, rolling resistance and heat build-up. Those properties are (strongly) related to the zinc oxide level, but until now no comparable investigation has been carried out yet. From the point of view of the (European) governments, a reduction of zinc oxide level in rubber compounds is strongly recommended because of the eco-toxic behavior of zinc oxide in an aquatic environment.

Reduction by mold modification

Introduction

Reduction of mold fouling by mold modification was carried out by the investigation of coatings. It has been shown that the most permanent coatings, such as chromium or vanadium, are metal-based with a thickness of 5 to l0 microns. In the earlier study (ref. 1), it was shown that zinc sulfide was responsible for the larger part of mold fouling. No fouling could be found on a thick polymeric coating based on polytetrafluoroethylene (PTFE). However, the coating became weak at high processing temperatures and was partly destroyed after several cycles due to the high shear stress during the injection of the rubber. Therefore, an investigation and a closer look at coatings could give new insights into mold fouling behavior. In this study, the influence of mold parameters, non-metal coatings and magnetite coatings were investigated.

Influence of mold parameters

Different parameters were investigated in short run experiments. Stainless steel and varying roughness from 0.1 Ra to 2.0 Ra gave significantly no reduction in mold fouling. A closer look at different metals could give new insights. A series of inserts manufactured from different metals was investigated in short run experiments. Different metals were chosen with different electro-potentials. In the series of electro-negative potentials, magnesium (Mg), aluminium (AI), zinc (Zn), chromium (Cr), iron (Fe) and nickel (Ni) were chosen. From the electro-positive series, copper (Cu) and silver (Ag) were chosen. The compound NR/BR was vulcanized in a compression mold for 20 times at 2 min./200[degrees]C. The insert samples were visually investigated after each five vulcanization cycles, and afterwards an elemental analysis on the inserts was done with RMA. The results are shown in table 5. Non-precious metals based on the elements Mg, A1, Zn, Fe and Cr showed zinc sulfide crystallites on the inserts, and the particle size increased going from Mg to Fe. On the electro-positive side, AgS was formed. The rubber compound was partly bonded to the metal surface of nickel (Ni) and copper (Cu). This phenomenon is known as the sulfidizing effect of the surface to bond compounds to the metal (ref. 3). The conclusion is that changing iron or stainless steel molds to other metal molds did not solve the mold fouling problem.

An interesting investigation has been carried out to influence the electro-chemical potential of the mold by applying a voltage over the mold. The deposit of zinc sulfide could possibly be reduced. It was stated by TNO Coatings that zinc sulfide could be a corrosion reaction of different mold construction materials at the interface. By applying a current, it was possible to measure the polarization reaction at the interface.

As presented in figure 3, polarization measurements were carried out from -5 V to + 5 V. During the vulcanization reaction, currents of one ampere were measured, however, no significant difference in deposits (such as zinc sulfide) was visible in comparison to standard vulcanization experiments. Based on these results, it is not probable that electro-chemical reactions took place at the interlace.

Non metal coatings

As already shown in table 5, metal coatings could not be used to reduce mold fouling. Therefore, other coatings were investigated. Coatings could be divided into different groups of materials, such as hybrid coatings, PVD- or CVD coatings (coatings prepared from the physical- or chemical vapor damping process), diamond coating (DLC), ceramic coating (enamel), and plastic coatings such as polyphenylenesulfide (PPS) and PTFE's. One or two coatings from each group were selected for short run vulcanization experiments. After 5, 10, 15 and 20 vulcanization cycles, the inserts were visually inspected by using a light microscope with a magnitude of 500x, and afterwards, if possible, an RMA analysis was performed. Results, including a reference of a stainless steel insert, are shown in table 6. As can be seen from the table, a DLC coating failed due to subsequent bonding of the rubber. After these short run experiments, tests were carried out on an injection molding machine for large run experiments.

For injection molding experiments, a multi-functional mold was constructed. Eight different inserts could be evaluated with this mold, each easily replaceable. The inserts were coated by special mold makers or job coaters. The mold was self-releasing, and the cavities were injected centrally. The injection molding temperature was 210[degrees]C, and the cycle time was about 45 seconds. In this way, about 500 cycles were possible in one shift. The vulcanized samples were inspected for surface damages, and afterwards the inserts were analyzed by RMA for mold fouling. The results are shown in table 7. It can be seen that the hybrid coatings A and F (fluor), and the PVD coatings CrN (chromic nitride) and Cr/CrN (multi-layer), showed crystallites of zinc sulfide (ZnS). Also, a PPS (polyphenylene sulfide) coating and a thin PTFE (polytetrafluoroethylene) coating showed crystallites. In this case, the coating was porous, and the crystallites were formed in the holes (figure 4). Only enamel (ceramic) and a thick PTFE coating showed no crystallites and formed a closed barrier on the metal surface.

All of the thin coatings showed crystallites. For tire molds, thick ceramic (enamel) coatings and thick PTFE coatings could not be applied because of the design of the profile and venting involved. Another problem was the poorer heat transfer from the mold (through the coating) to the rubber compound. For tire molds, a thin coating is advisable. A look at thin coatings with a non-porous closed surface

Magnetite coating

As was investigated, the electro-chemical mechanism that was studied cannot account for the formation of zinc sulfide (ZnS). This implies that the formation of ZnS at the interface of mold and rubber compound would be caused by a certain physicochemical reaction at the mold surface (ZnS crystallizes at the interface) due to a reaction of zinc oxide and sulfur during vulcanization.

Vulcanization experiments at high temperature confirmed that the ZnS was formed as crystals on the mold surface. It was already shown that nano-sized ZnS crystallites were formed (determination of ZnS) in the rubber and diffused from the rubber compound to the mold surface where they interact on a molar basis with iron (Fe203) present at the surface and could form a marmetite (ZnSFe) lattice. This could then act as a grafting point for the build-up of the ZnS crystal. In order to prove or disprove this hypothesis, it is expected that this mechanism of formation depends upon the oxidized or sulfurized iron atoms present on the surface of the mold.

Therefore, the objective would be to change the iron, and at the same time still allow that surface to participate in a physico-chemical reaction, it should be possible to oxidize the surface of hematite (F[e.sub.2][O.sub.3]) to magnetite (F[e.sub.3][O.sub.4]). It is known that in central heating systems or boilers, a magnetite coating is applied to prevent corrosion. In this case, a passivated iron mold (insert) was prepared. The passivation can be carried out chemically (alkali) or physically (steam). Initially, short run vulcanization experiments were carried out, and afterwards large scale runs were investigated by means of an injection molding machine. It was shown that no ZnS crystallites were visible at the mold surface. Not only the formation of ZnS (mold fouling) was improved, but also the mold release behavior. In table 8, the mold fouling and release behavior were studied for the few most interesting coatings by using an injection molding machine. Two different magnetite coatings were investigated, both based on an alkali surface preparation method, and two PTFE-coatings. The stainless steel coating is a reference material. As can be seen from the table, after about 500 cycles, no mold fouling was visible on the magnetite coatings and the PTFE coatings. On the PPS coating and the multi-layer of chromic/chromic nitride coating, some crystallites were visible. During vulcanization experiments, all coatings showed good release properties.

No practical experiments have been carried out yet on tire compounds or tire molds.

Conclusions

Mechanism of mold fouling

It was concluded that ZnS crystals are formed on the surface of iron molds through:

* In-situ formation of ZnS in the compound as a reaction product during vulcanization;

* the diffusion of nano-sized ZnS crystals to the hematite-type oxidized surface of the iron mold, to form a marmetite-type layer; and

* further growth of this crystallite to form micro-sized ZnS crystals by the deposition of additional nano-sized ZnS crystallites, either during the same vulcanization run or as a result of successive vulcanization runs.

Compound

* Mold fouling may be reduced by reduction of the zinc oxide level in rubber compounds;

* mold fouling can be eliminated or reduced by using nano-zinc oxide, however, for optimal mechanical and dynamic properties, more investigation has to be carried out;

* other metal oxide activators gave no solution to the mold fouling problem; and

* alternative activators such as amines or MFAs also reduced mold fouling, however, those activators/accelerators can give an unpleasant smell and some are toxic. Plus, an additional zinc accelerator (ZBEC) has to be used for comparable cure rate.

Coating

* Electro-chemical reactions are not involved in the solution of the mold fouling problem;

* metal coatings gave no solution to reduce mold fouling. It is essential that a closed barrier on the surface of the mold has to be applied;

* plastic coatings such as PTFE types could solve mold fouling, but these coatings are not applicable on tire molds due to hand lay-up problems and dimensional instability; and

* magnetite coatings are a very likely candidate. Results were very promising, but experiments in tire molds are necessary to verify the potential of magnetite coatings in preventing mold fouling. However, no practical investigation on tire molds has been carried out.

Recommendation

Future work on tire molds is necessary to complete this investigation.

Table 1 - tire components

Components              s-SBR (phr)     NR/BR (phr)

S-SBR                       100              -
NR                           -              70
BR                           -              30
Zinc oxide (RS)              5               5
Stearic acid                 2               2
Carbon black (N375)         50              50
Oil (high aromatic)          5               5
Accelerator (TBBS)          1.5            1.25
Sulfur                      1.5            2.25

Table 2 - deposition of ZnS crystallites at different temperatures

                    s-SBR              s-SBR
                    (20 min./          (2 min./
                    160[degrees]C)     200[degrees]C)

As such             0                  0
After 5 cycles      2                  2
After 10 cycles     2                  2
After 15 cycles     4                  4
After 20 cycles     4                  4
RMA                 ZnS                ZnS

                    NR/BR              NR/BR
                    (20 min./          (2 min./
                    160[degrees]C)     200[degrees]C)

As such                                0
After 5 cycles      0                  2
After 10 cycles     2                  2
After 15 cycles     2                  4
After 20 cycles     4                  4
RMA                 4                  ZnS

0 - no deposition; 1 - few small crystallites; 2 - many crystallites
covering <50%; 3 - small and large crystallites; 4 - crystallites
covering >50%

Table 3 - deposition of ZnS crystallites by using zinc alternatives
in an NR/BR blend

                    ZnO       ZnO       ZnO         Nano
                    (RS)      (RS)      (Silox)     ZnO
                    5 phr     3 phr     3 phr       0.25 phr

As such             0         0         0           0
After 5 cycles      2         2         2           0
After 10 cycles     2         2         2           0
After 15 cycles     4         4         4           0
After 20 cycles     4         4         4           0
RMA                 ZnS       ZnS       ZnS         -

0 - no deposition; 1 - few small crystallites; 2 - many crystallites
covering <50%; 3 - small and large crystallites; 4 - crystallites
covering >50%

Table 4 - deposition of crystallites with different additives
in an NR/BR blend

                    ZnO       CaO       MgO       MFA 3 phr
                    3 phr     3 phr     3 phr     ZBEC 0.2 phr

As such             0         0         0         0
After 5 cycles      2         2         2         0
After 10 cycles     2         3         3         0
After 15 cycles     4         3         3         0
After 20 cycles     4         3         3         0
RMA                 ZnS       CaS+S     MgS+S     -

MFA = n-tallow-1,3 propanediamine distearate

0 - no deposition; 1 - few small crystallites; 2 - many crystallites
covering <50%; 3 - small and large crystallites; 4 - crystallites
covering >50%

Table 5 - influence of the metal electro-potential

Element          Density       Electro     Component   Crystallites
                 (gram/        potential   (RMA)       (micron)
                 [cm.sup.3])   (mV)

Magnesium (Mg)   1.74          -2.4        ZnS         <0.2
Aluminum (Al)    2.70          -1.76       ZnS         <0.2
Zinc (Zn)        7.14          -0.76       ZnS         0.5 - 1
Chromium (Cr)    7.20          -0.60       ZnS         1 - 2
Iron (Fe)        7.87          -0.44       ZnS         1 - 2
Nickel (Ni)      8.55          -0.22       #           -
Hydrogen (H)     1             0           -           -
Copper (Cu)      8.96          +0.34       #           -
Silver (Ag)      10.50         +1.38       AgS         -

# (partly) bonded to the insert

Table 6 - results of short run vulcanizations on different coatings
with NR/BR blend

Coatings          Thickness   5x   10x   15x   20x    RMA
                  (micron)

Stainless steel       -       2     2     4     4     Zns
Hybrid A              2       0     0     0     0      -
Hybrid F              3       0     0     0     0      -
DLC                   4       B     B     B     B    rubber
CrN                  10       0     0     0     0
Cr/CrN               11       0     0     0     0      -
Enamel               80       0     0     0     0      -
PPS                  26       0     0     0     0      -
PTFE (1)              6       0     0     0     0      -
PTFE (2)             25       0     0     0     0      -

B = DLC coating failed (rubber bonding)

0 - no deposition; 1 - few small crystallites; 2 - many crystallites
covering <50%; 3 - small and large crystallites; 4 - crystallites
covering >50%

Table 7 - results of injection molding on different coatings
with NR/BR blend

Coatings   Thickness     50      100      200      400     RMA
           (micron)    cycles   cycles   cycles   cycles

Hybrid A       2         2        4        4        4      ZnS
Hybrid F       3         2        4        4        4      ZnS
CrN           10         4        4        4        4      ZnS
Cr/CrN        11         0        2        2        2      ZnS
Enamel        80         0        0        0        0       -
PPS           26         2        2        2        2      ZnS
PTFE (1)       6         0        2        2        2      ZnS
PTFE (2)      25         0        0        0        0       -

0 - no deposition; 1 - few small crystallites; 2 - many crystallites
covering <50%; 3 - small and large crystallites; 4 - crystallites
covering >50%

Table 8--evaluation of the most interesting coatings on a blend of
NR/BR

                  Thickness    Mold      Mold
Coating           (micron)    fouling   release

Stainless steel      na          -        +/-
Magnetite (1)        2.5         +         +
Magnetite (2)         4          +         +
PTFE (1)             25          +         +
PTFE (2)             18          +         +
PIPS                 26         +/-        +
Cr/CrN               11         +/-        +

+ = no fouling or very good release

+/- = little fouling or good release

- = fouling or bad release

References

1. B. van Baarle, "Mold fouling during vulcanization," Rubber World, December 2001.

2. Peter J. Nieuwenhuizen and Ben van Baarle, "The mechanism of mold fouling during sulfur vulcanization," unpublished article.

3. W.J. Van Ooij, "Mechanism and theories of rubber adhesion to steel tire cords--an overview," Rubber Chemistry and Technology, vol. 57-3, p. 421-456 (1984).

4. Reports of the different investigations, TNO Industrial Technology.