Removing fouling residue from molds in-the-press with solid C[O.sub.2] pellet blasting.

The use of C[O.sub.2] solid particle blasting for non-abrasive cleaning of rubber molds without removing molds from the press or creating a secondary waste stream has intrigued molded rubber products manufacturers since the mid-1980s. Today, many molded rubber product manufacturers are aware of C[O.sub.2] particle blasting technology and its potential to reduced production downtime and maintenance labor costs and enhance product quality and appearance. In fact, molded rubber products manufacturers make up a major segment of the C[O.sub.2] particle blast cleaning equipment market worldwide.

While working to develop applications solutions for the molded rubber products industry, C[O.sub.2] blast cleaning technology suppliers began to find solutions for the general problems of noise, ergonomics, operator safety, work area accessibility, system reliability and operating costs. Within the molded rubber products industry, the economic impact of implementing C[O.sub.2] particle blast mold cleaning is overwhelmingly significant, and the cost to benefit analysis usually indicates a pay back in terms of months, not years.

The following discussion will provide an understanding of the state-of-the-art solid C[O.sub.2] particle blast rubber mold cleaning technology. From a cost-to-benefit standpoint, solid C[O.sub.2] particle blast mold cleaning is, more often than not, shown to be the best choice among the many methods and technologies currently available for rubber mold cleaning. Within the C[O.sub.2] particle blasting industry, however, there is a variety of technologies that offer different levels of mold cleaning cost and performance. Similar to abrasive blasting equipment, there are two types of C[O.sub.2] particle blast systems. One type is the direct feed or "single-hose" system. The other is the inductive feed, or "two-hose'" system. In abrasive blast systems, there may not be significant cleaning performance differences between the two types of systems. However, for C[O.sub.2] particle blast systems, there is a dramatic performance differential between the types; and the differences must be fully understood and considered before selecting a mold cleaning system. There are also two basic forms of the solid C[O.sub.2] blasting media that must be understood. There are discrete "pelletized" dry ice particles, and shaved dry ice "flakes" produced from a block of dry ice.

The mold cleaning performance levels of the different types of C[O.sub.2] particle blast systems and C[O.sub.2] media are very significant. Failing to understand these fundamental differences may lead to the selection of an inappropriate system and a significant reduction in potential productivity increases and cost savings.

Why rubber curing molds need to be cleaned

A major problem faced by all molded rubber products manufacturers is that of mold fouling, a residue build-up on the curing surfaces of rubber molds caused primarily by chemical reactions of the release agent, or the adhesive used for rubber-to-metal bonded parts, with the base polymer under heat and pressure. For injection molding systems, a major area of concern is keeping mold surfaces that mate or come into intimate contact during cure, free of residue build-up to minimize flash. For multi-section rubber molds, these are the surfaces between the mold sections that come in contact when the mold is closed and produce parting lines in the part. If too much residue is allowed to build up in this area of the mold, the mold sections will not mate together completely, even under the extreme press squeeze pressure. The result is a noticeable flash at the parting line on the surface of the product. An excessive amount results in additional labor costs to remove the flash. For molds that are designed to be flashless/ trimless, keeping the precision mating surfaces free of fouling is even more critically important. Excessive fouling residue build-up in intricate details and in sharp edges or corners of molds can cause the parts to lose details (lettering, logos, etc.) and critical cross sectional shapes (seal lips, o-ring cross section, etc.), all resulting in scrapped parts.

Excessive build up of release agent in the cavities can cause high releasant transfer to the parts, resulting in surface gloss level deficiencies like gluing, etc. As much as the release agent can benefit the rubber molding process, an excess of it in the cavities can be detrimental to the process.

Another problem area, especially in injection and injection transfer molds, is the clogging of sprues, runners, gates and vents with fouling residue, overbuilt-up mold release residue, and semi-cured rubber. Clogged vents can cause non-fills and other serious part defects. Finally, the overbuild up of mold release agent residue or adhesive residue used in rubber-to-metal bonded parts can cause defects like knit lines and surface gloss loss, or even cause the parts to stick in the mold and tear upon removal.

Comparing costs to benefits

As an example, a current user of C[O.sub.2] pellet blast mold cleaning technology in the robber industry conducted an in depth cost to-benefit analysis to compare the value of C[O.sub.2] blasting with the abrasive blast method already in use at his facility. This study was conducted after completion of a trial usage period with a C[O.sub.2] pellet blasting system, and before the final purchase of the C[O.sub.2] system. This user took into consideration all aspects of capital and operating costs, including purchase price of the C[O.sub.2] blasting equipment, C[O.sub.2] pellet usage and cost vs. abrasive media, compressed air and electricity costs for both types of systems over time, mold and press downtime for cleaning with both types of systems, and other aspects of mold maintenance costs relative to his operation. C[O.sub.2] pellet blasting can be accomplished in-the-press in one-tenth to one-fourth the time required for traditional off-line methods. Interestingly, the cost to benefit analysis was presented to the company management in terms of increased product sales dollars per day by using C[O.sub.2] pellet blasting instead of their existing off-line abrasive mold cleaning method. The study indicated that the lost sales dollars due to mold cleaning would be decreased by 75% every day. In other words, whatever total parts sales dollars were being lost each day due to press downtime were cut to only 25% of that amount by adopting C[O.sub.2] pellet blast cleaning technology for all of their molds. Based on this increased productivity ratio, the user determined complete payback on a portable C[O.sub.2] pellet blast mold cleaning system would be realized in under 60 days. Other cost to benefit studies have been conducted by robber product manufacturers with comparable results, and given the similarities in the manufacturing processes of most rubber molders, this level of increased productivity and payback is typically experienced when abrasive mold cleaning is replaced with C[O.sub.2] particle blasting technology.

Mold condition factors that affect cleaning performance

Temperature

Data gathered over the past decade, since the emergence of C[O.sub.2] particle blast cleaning in the molded rubber products industry, supports the tact that rubber molds between 300[degrees]F and 350[degrees]F can be cleaned three to four times faster than the same molds at ambient temperature (cold molds). Although the reasons and mechanisms which give rise to this phenomenon are not completely understood, mold cleaning experience in curing departments at many different robber products manufacturers have proven this to be the case. In case studies involving rubber mold fouling, particularly with EPDM, FKM, NR, NBR, HNBR, butyl compounds and flouroelastomers, it has been determined that the reactive chemicals present in the base polymer, the chemicals in cure accelerators and inhibitors, and the chemicals in many mold release agents combine at the curing temperatures to form an almost glass-like material at the product-mold interface. This glass-like material is different from the polymer material of the cured product. The glass-like property of this fouling residue at elevated temperatures allows it to be easily removed from the mold surface and fractured into small particles by inducing high levels of thermal stress, or thermal shock with C[O.sub.2] pellets. Since the temperature of solid C[O.sub.2] is -109[degrees]F, the C[O.sub.2] pellet blast stream is an ideal source for inducing thermal shock in the residue layer. At lower temperatures (below 150[degrees]F), the fouling residue can become much more difficult to remove from the mold surface because it resembles a very dense viscoelastic material which absorbs the impact energy of the C[O.sub.2] pellets. The thermal shock mechanism becomes less effective because there is very little temperature differential between the material and the mold surface. The overall result is that it may become very difficult to remove the fouling residue from room temperature, or cold, rubber molds, and sometimes the residue will not respond to the C[O.sub.2] particle blasting at all.

Adhesive fouling of molds producing rubber-to-metal bonded parts

Many rubber products, especially in the automotive and vibration mount markets, consist of a robber isolation shape molded around and bonded to a steel mounting plate or a cylindrical center steel tube. These rubber-to-metal bonded parts require a heat activated adhesive applied to the metal portion which creates an incredibly strong bond between the metal and robber during the curing cycle. The concern with this process is that the heated adhesives become almost liquefied at curing temperatures and the excess flows out of the part during the high temperature curing cycle, creating a large amount of rapidly built-up fouling on the mold surfaces. Figure 1 shows typical rubber-to-metal bonding adhesive fouling on a mold surface.

[FIGURE 1 OMITTED]

Fortunately, this type of baked-on adhesive fouling also resembles a glass-like residue that responds exceptionally well to the temperature gradient induced thermal shock removal mechanism of C[O.sub.2] pellet blasting. Rubber-to-metal bonded part molds can be cleaned in the press as often as once a shift if necessary, so C[O.sub.2] pellet blasting offers the robber-to-metal bonded part manufacturer a safe, easy to use and extremely cost effective way to keep these molds in continuous production.

Mold metallurgy and temperature effects

Early on, valid concerns were expressed by rubber product molders about the possibility of mold sub-surface damage resulting from thermally induced changes in the metallurgy of the mold material. This concern was strong where martensitic and other heat-treated and hardened alloy tool steels are used, as well for certain precipitation hardened aluminum alloys. University and independent laboratory metallurgical studies were funded and performed by several C[O.sub.2] blast cleaning early adopter rubber molding companies in the mid-1990s. The specific findings in these reports are proprietary to the companies that paid for the studies. In general, the conclusions drawn were: No metallurgical changes to the martensitic or alloy hardened structure of the tool steel molds were found after many cycles of C[O.sub.2] pellet blasting; and the sub-surface thermal gradient in tool steel was negligible (low thermal conductivity), so the mechanism to promote micro-cracks was considered non existent. In aluminum molds, which have higher thermal conductivity and generally smaller thermal mass. the cool down (thermal gradient) was greater and more pronounced, but no changes in the precipitation hardened structure were discovered after many blasting cycles. Aluminum mold surface erosion was also considered negligible.

Mold surface condition

Abrasive blast media, like plastic or glass beads, typically leave a bare metal appearance after residue removal, even on steel molded rubber products molds. This like new appearance is deceiving because it is achieved at the expense of removing a small amount of metal from the mold surface, and by imparting a much rougher (more micro peaks and valleys) surface finish rote the mold from the chiseling effect of thousands of abrasive impacts. The rough surface creates an anchor pattern that was not present in the original mold surface. This causes fouling residue to adhere and accumulate at an even faster rate than it did on the original mold surface. This mold surface erosion will be discussed later in more detail, but it is evident that what appears to be a clean mold surface is a step toward decreasing the useful life of a very expensive production tool.

Flashless/trimless tooling

Marry rubber product molding companies are currently using, or are considering using, precision designed tool steel flashless/trimless molds. Rubber flash at the tool parting lines, of course, requires additional manufacturing processes to remove it, requires additional quality assurance (inspection) steps to monitor tire level and effectiveness of flash removal, creates scrap robber which must be handled and accounted for, which all adds to the overall production cost. Also, some types of rubber parts have such critical surface finish and dimensional tolerances that only flashless/trimless mold technology can be used to produce them. Flashless/trimless tooling virtually eliminates part flash and it's associated margin reducing production costs, but the capital cost of the tooling is much higher, and the precision mating surfaces of the molds must be maintained without misalignment, knick damage, or erosion for the life of the tooling. Abrasive particle blast cleaning of flashless/trimless tooling typically cannot be tolerated because of the risk of tooling erosion and damage. Non-abrasive C[O.sub.2] particle blasting can be used for hundreds and even thousands of cleaning cycles with no tool wear experienced.

Non-abrasive aspects of C[O.sub.2] particle blasting

C[O.sub.2] particle blasting does not abrade or erode the surface of common mold materials like tool steel and hardened aluminum. Since C[O.sub.2] particle blasting only removes the residue on the mold's surface and not any surface metal, any dark stains from cured molded rubber products compounds will remain on the mold's surface, especially on steel tooling. For lowing C[O.sub.2] particle blasting, a functionally clean, residue tree molded rubber products mold may not at first appear clean by the old standard of a bright, bare metal surface. The proof of the mold's cleanliness will be seen when the first molded rub bet products are cured and inspected for the sharpness of the molded-in details (especially under-cuts and acute angle seal surfaces), and rubber surface gloss level.

Many rubber molds must be seasoned with release agent immediately after abrasive cleaning. This usually involves running three to five cycles of parts with sprays of release agent in the cavities after each cycle until enough of a very thin layer of release coat fills in the valleys of the roughened surface. All of the parts run during the seasoning process are scrapped because of surface gloss defects, etc. Using C[O.sub.2] pellet blasting to clean molds will, over time, shorten or even, as reported in some instances, eliminate the mold seasoning time and resulting scrap. This is because the mold surface will not he severely roughened during each cleaning cycle, and because the energy level of C[O.sub.2] pellet blasting can be adjusted to allow for the retention of the seasoning coating.

As stated earlier, chemical stains in the mold metal typically cannot be removed by non-abrasive C[O.sub.2] pellet blasting. Severe chemical stains which can leach out of the more porous mold metals to cause blooming on the parts may be beyond the ability of C[O.sub.2] pellet blasting to counteract. Severe chemical staining of the molds may, in some cases, be significantly reduced over time by using C[O.sub.2] pellet blast cleaning which does not continually open up the surface pores like abrasive blasting does. As the surface roughness goes down and the pores close up over time, stain leaching and stain related defects should diminish as well.

Mold erosion and damage studies

The most well known and widely used method of rubber and tire mold cleaning is abrasive particle blasting. This method is very cost effective, easy to install and maintain, and relatively easy to use. All forms of abrasive blasting must be done in an enclosed structure to prevent the distribution of fine airborne abrasive dust particles within the factory environment, and to capture and recycle the spent media. The most popular abrasives used to clean robber molds are glass, plastic, metallic and ceramic beads. These media have gained acceptance in the rubber industry because they are regarded as only mildly abrasive. Other particle blast media used in the rubber industry vary in degree of abrasiveness and subsequent mold erosion, and include silica sand, steel shot, walnut shells, bicarbonate of soda and abrasive impregnated sponge. All of these abrasive blast media can typically be captured and recycled for use in more than one cleaning session. However, all of them eventually break down (pulverize) into a fine dust which must be disposed of in compliance with federal regulations. The fact that all of these media types are considered abrasive means that ultimately the molds will be eroded to a point where they must undergo extensive rework or be scrapped. Abrasive particle blast mold cleaning is, at best, a compromise between a cost effective cleaning method and reduced mold life/continuously declining rubber product quality.

Abrasive, or mildly abrasive blasting causes other problems as well. The fine silica imbedded dust residue from sand or glass bead blasting, or the imbedded plastic dust from plastic media blasting (PMB), can alter the surface of the rubber mold enough to prevent the proper chemical bonding of certain release agents. These mold release agents, which depend on a completely metallic surface to bond to for providing many cure cycles worth of release, are actually pulled off of the mold surface and rendered ineffective in fewer cure cycles because of the bond blocking caused by grit residue imbedded in the mold surface. The same type of chemical bond blocking can occur when attempting to apply various mold coatings for long term product release capability. In general, all chemicals applied to a metal mold surface react much faster and more efficiently when all of the metal surface is available and not masked by grit residue.

Other non-abrasive mold cleaning technologies that exist, or are emerging, include laser ablation, chemical flushing and mechanical adhesive bonding of residual rubber. These methods are very much in the developmental stages and do not offer a viable solution to the immediate and near term needs of the molded rubber products industry. Of all the generally accepted off-the-shelf rubber mold cleaning technologies currently available, only solid C[O.sub.2] particle blasting has been acknowledged to be non-abrasive, cost-effective and not produce a secondary waste stream or bond blocking dust residue on the mold surfaces. Tables 1 and 2 present data from two mold erosion studies that were conducted by a major tire manufacturer in 1991, and again hi 1996. Table 1 shows the results of the 1991 tests, where C[O.sub.2] pellet blasting was examined to see what erosion effects were apparent in four types of tire mold materials (steel, forged aluminum and two types of cast aluminum) using three different C[O.sub.2] pellet blast nozzle configurations.

All samples were blasted with C[O.sub.2] pellets at 300 psig blast pressure. Pellet mass flow rate was 250 pounds per hour. Three different single-hose system blast nozzles were used. Negligible effects of blasting were seen on the steel and forged aluminum samples, while both cast aluminum showed minimal to severe erosion. Nozzle design had the most significant effect on erosion rate and cleaning cycle time.

The 1991 and 1996 mold erosion studies were conducted at significantly higher blast pressures (250 to 300 psig) and at higher particle velocities than required to clean typical tire and rubber molds with today's technology. This is because C[O.sub.2] pellet blasting nozzle design technology has advanced significantly from the 1991-1996 era. Nozzle and delivery system improvements now allow molded rubber products manufacturers to completely clean rubber molds at blasting pressures between 70 and 90 psig. Pellet flux density (the distribution of C[O.sub.2] particle impacts per unit area per unit time) at the mold surface is now better than it has ever been because of the aerodynamic advances in the single-hose blast delivery system. Rubber-mold specific blast nozzles now exist to focus the blast energy only where it is needed in the complex geometry of robber mold cavities. The rubber mold fouling residue removal capability of C[O.sub.2] particle blasting is significantly better than it was only three or four years ago, at much lower air pressure and overall kinetic energy delivered to the mold surface. Presently, C[O.sub.2] particle blast induced mold erosion, even for cast aluminum molds, is considered negligible, and a significant number of molded rubber products manufacturers are now specifying and purchasing C[O.sub.2] particle blast mold cleaning systems to replace their existing abrasive blasting systems and methods.

Also, C[O.sub.2] particle blasting can effectively remove most imbedded grit residue left on the mold surface from abrasive cleaning methods, as well as remove grit residue from vents and microvents. This allows mold release agents to work better. and can even lead to the reduction of the amount of mold release required, substantially decreasing the rate of deposit of mold fouling substances.

Another benefit related to the low blast air pressure and volume requirements for today's single hose C[O.sub.2] particle blasting systems is that it reduces the costs associated with the required compressed air system. Earlier, high pressure based single hose C[O.sub.2] systems required expensive, dedicated air compressors. Current single hose direct acceleration systems use shop air at 70 to 80 psig and between 150 and 200 scfm. Most past and current two-hose C[O.sub.2] systems require two or more times the air flow volume, even at low pressure, than a state-of-the-art low pressure single hose system. In a compressed air system, high pressure and high flow volume dramatically increase operating costs in terms of equipment costs and energy (electricity) consumption. Any cost to benefit analysis for a C[O.sub.2] blast rubber mold cleaning system should include the compressed air system fixed and operating costs.

Selecting the correct system for cleaning and maintenance

Direct acceleration system vs. inductive systems Kinetic and thermal energy effects

Solid C[O.sub.2] particle blasting systems are available in two basic configurations. The least complex and least costly to produce is the inductive, two-hose system, sometimes called an inductive or venturi system. These systems are typical of sand, PMB and glass bead blasting systems, and most C[O.sub.2] particle blast systems available today employ this method. The blast media are sucked into a chamber in the handheld applicator on gun by the venturi effect, then propelled out of a short nozzle by a high volume flow of compressed air. Because these systems rely on the creation of a strong suction to bring the blast media from the storage hopper to the nozzle, the length of the interconnecting dual blasting hose is typically limited to fifteen feet or less. In the two hose system, one hose is the media suction hose, and can be constructed from lightweight material. The other hose is the compressed air delivery hose, which is typically heavier, to withstand pressures as high as 200 psig or greater.

In the two hose systems, the media particles are moved from the hopper to the gun chamber by suction, where they drop to a very low velocity before being induced into the outflow of the nozzle by the large flow volume of compressed air. Since the blast media particles have only a short distance in which to gain momentum and accelerate to the nozzle exit (usually only 8 to 12 inches), the final particle average velocity is limited to between 200 and 400 feet per second. So, in general two hose systems, although not as costly, are limited in their ability to deliver contaminant removal energy to the surface of a mold. When the need for more blasting energy is required, these systems must be boosted at the expense of much more air volume required, usually higher blast pressure required as well, with much more nozzle back thrust, and very much more blast noise generated at the nozzle exit plane.

The other type of solid C[O.sub.2] media blasting system is like the pressurized pot abrasive blasting system common it] the sand blasting and PMB blasting industries. These systems use a single delivery hose from the bopper to the nozzle applicator in which both the media particles and the compressed air travel. These systems are more complex and a little more costly than the inductive two hose systems, but the advantages gained greatly outweigh the extra initial expense. In a single hose solid C[O.sub.2] particle blasting system, sometimes referred to as a direct acceleration system, the media are introduced from the bopper into a single, pre-pressurized blast hose through a scaled airlock-feeder. The particles begin their acceleration and velocity increase immediately, and continue to gain momentum as they travel the length of the hose. At the end of the hose, the spray nozzle gun actually consists of a convergent-divergent (isentropic flow) nozzle which exchanges pressure differential across the nozzle for a huge increase in air and particle velocity. C[O.sub.2] particle velocities have been measured and substantiated in excess of 700 feet per second, and up to as high as 950 feet per second at the nozzle exit plane. This is accomplished at less than one third of the flow volume required by the most aggressive two hose systems.

In addition to the lighter weight and less cumbersome hand held applicator and hose of a single hose system, the contaminant removal energy delivered to the surface is considerably higher than that provided by a two hose inductive system. Even with solid C[O.sub.2] particle blasting, a significant component of the contaminant removal energy is the kinetic energy per unit of area delivered to the surface. Since kinetic energy is a function of mass and velocity of the particles in the relation [K.sub.e] = 1/2 [mv.sub.2], it can be seen that a two-fold increase in particle velocity, given equal particle mass and equal nozzle spray area, effectively increases the impact energy delivered to the surface by a factor of four. A three-fold particle velocity increase, from 300 to 900 feet per second, increases the blast impact energy nine times.

Table 3 shows the relative cleaning performance between a single hose system and a two-hose system at typical factory air blasting pressure. The term [C.sub.vbe] is called the blast energy coefficient, and represents comparative capability of a C[O.sub.2] particle blast system to remove a volume of soft pine wood from a test specimen within a controlled interval of time.

As the data in table 3 illustrate, the typical blasting energy of a single hose system compared to a two-hose system at equal blast pressure is about three times greater.

C[O.sub.2] blasting media types

Solid carbon dioxide blasting media are currently available in two forms, discrete rice grain sized pellets which are produced by pressure extruding and cutting strings of dry ice, and sugar granule-sized flakes of dry ice produced by mechanically shaving the face of a large block of dry ice. To understand the differences in mold cleaning performance between the two C[O.sub.2] media forms, some technical background discussion is appropriate.

Traditional abrasive particle blasting, and even the mildly abrasive blasting technologies like PMB and glass bead, rely on the intrinsic surface hardness and geometry of the media, and the work available at the surface resulting from the kinetic energy of the media acting through the surface hardness and geometry.

The hard, sharp abrasive particles actually break into smaller pieces that ricochet into mold surface features for additional residue removal action. The total kinetic energy of abrasive media particles is therefore spread out over more than a single impact per particle. With solid C[O.sub.2] media, however, the particles completely disintegrate and sublimate to C[O.sub.2] vapor upon the initial impact, so all of the solid C[O.sub.2] particles' kinetic energy is spent in one impact per particle. There is no ricochet or secondary impact effect in solid C[O.sub.2] particle blasting. Therefore, C[O.sub.2] particle blast mold cleaning performance is determined by a parameter called flux density. Flux density is defined as the number of particle impacts at the mold surface per unit of area per unit of time. In other words, for two C[O.sub.2] particle blasting systems with nozzles of equal exit plane area, assuming that the particles from each system possess sufficient and equivalent kinetic energy, the system that can deliver more particles to the surface in the same amount of time, or the same amount of particles in less time, will generally remove fouling residue faster and more completely.

C[O.sub.2] particle dynamics

As mentioned earlier, C[O.sub.2] particle blasting harnesses two types of energy to accomplish mold fouling residue removal. C[O.sub.2] particle size directly influences the level of kinetic (velocity or impact) energy and thermal (temperature gradient or thermal stress) energy available at the surface. Large pelletized C[O.sub.2] media are typically 3 mm in diameter and between 5 mm and 8 mm in length. The sugar grain sized flake particles resulting from shaved dry ice block are roughly spherical and 0.5 mm to 1 mm in diameter. In the pelletized C[O.sub.2] system, by the time the pellets are accelerated and travel through the blast hose and the nozzle, they are fractured into roughly uniform sized irregular spheres of dry ice about 2 mm in diameter. Given that the solid C[O.sub.2] is of the same density in both particles, the fractured pellet spheres possess about four times the mass of the individual shaved flakes or granules. Referring back to the kinetic energy equation, each pellet, if traveling the same velocity of each granule, delivers four times the impact energy at the surface. Since the fractured pellet spheres in the single hose blasting system typically travel three times faster than the shaved block granules in the two-hose system, the kinetic energy increases by a factor of 4 x [3.sup.2] = 36. This is the significant underlying factor in the ability of the high velocity, single hose C[O.sub.2] pellet blasting systems to dislodge and remove the robber residue anchored into undercuts and corner profiles, and the vents of rubber molds.

Thermal energy is dependent upon the mass (number and size of particles) of solid C[O.sub.2] delivered to a given area of the surface per unit of time. There is a tremendous latent heat transfer as the solid C[O.sub.2] changes phase to vapor C[O.sub.2] at the mold surface (246 btu per pound of solid C[O.sub.2]). This heat exchange with each impacting C[O.sub.2] particle occurs within a few milliseconds, and the heat is given up mostly from the thin layer of residue, and some from the surface of the mold. It is this instantaneous surface only heat transfer effect which imparts the thermal stress into the residue to fracture it from the mold surface. Having already described the particle velocity and mass delivery characteristics of the single hose, isentropic nozzle systems relative to the two hose inductive nozzle systems, it is evident that the single hose system delivers more thermal mass per unit of area per unit of time. Therefore, it produces the best thermal shock or residue fracturing effect. If the C[O.sub.2] blasting system cannot deliver this effect efficiently and instantaneously, and if the nozzle traverse rate over the surface is reduced to make up for a lower thermal mass delivery rate, the effect will be lost as the mold cross section begins to lose heat from too much C[O.sub.2] striking it too slowly. This is typically why two-hose inductive systems fail to give rise to the thermal fracturing effect in molded rubber products mold cleaning applications. It is a matter of too little kinetic and thermal energy available in a given instant on the mold surface to be fully effective.

Table 4 shows the results of a comparison study in which mold sidewalls on both sides of a two-mold, two-piece tire press were cleaned, one side with a single hose, C[O.sub.2] pellet based system, and one side with a two-hose, shaved C[O.sub.2] block system, using the same operator.

Mold vents

Mold vents that become clogged with robber or other fouling residue prevent the flow of trapped air or outgassing from the compound being cured to exit the mold cavities. Trapped air or other gases in the cavities cause non-fill and flow line defects in the parts, and the defective parts must be scrapped. Keeping the vents continuously clear of robber and fouling residue is a major concern for most robber product manufacturers to reduce or eliminate the scrap rate. The high kinetic energy of the C[O.sub.2] particles in direct acceleration single-hose C[O.sub.2] blasting systems have proven very effective at clearing vents as small in cross-sectional diameter as .030 inch. The rapid kinetic chiseling action on the residual rubber/release agent chemical residue in the vents tends to dislodge and blow out the plug of material, even while the surface of the mold is being cleaned. Even vents that are normal to the surface, like tire mold pin vents (micro vents), are readily unplugged with the C[O.sub.2] pellet blasting process.

Mold cleaning methods for C[O.sub.2] technology

Special nozzle configurations

C[O.sub.2] particle blast mold cleaning is a line of sight process. The C[O.sub.2] high speed particles must impinge on the fouled surface at as close to 90[degrees] as possible. Additionally, the exit or tip of the blasting nozzle must be kept at 1.0 to 1.5 inches from the surface being cleaned to obtain the maximum energy from the C[O.sub.2] particles. These fundamental requirements, plus the fact that most rubber presses have four to eight inch throat openings with 20 inches or more of depth (length of the tooling), dictate that the C[O.sub.2] high speed nozzles and the hand-held applicators must be compact, able to turn the flow of high speed C[O.sub.2] particles at angles of 90[degrees] or greater, produce little or no nozzle backthrust and provide the ability for the mold cleaner to see the details of the molds as he cleans them. At the same time, the nozzle design must accomplish all of this using the lowest air CFM and producing the least noise possible, without sacrificing significant blast energy delivered to the surface of the mold.

The single hose, direct acceleration method of C[O.sub.2] particle blasting provides the correct physical attributes to allow the design of high energy flow turning and retro jet rubber mold cleaning nozzles. State-of-the-art computational fluid dynamics computer modeling and design programs are used to optimize blast energy while minimizing air flow and associated noise. A high energy flow turning nozzle for a five or more inch press opening is illustrated in figure 2.

[FIGURE 2 OMITTED]

Noise associated with manual blasting

Noise created by the C[O.sub.2] particle blasting equipment is another factor to be considered. All compressed gas (air) based particle blasting technologies are inherently noisy. The power level of the noise generated at the blast nozzle exit is largely a function of the compressed air outflow volume and velocity. To a lesser extent, another component of the total noise is also created by the aerodynamic interaction of the individual C[O.sub.2] pellets or particles with the air stream. In molded rubber mold cleaning operations, the noise from the nozzle is efficiently re fleeted back to the operator by the flat surfaces of the mold tooling and/or the inside walls of the press. Noise, specifically the sound pressure level (SPL) in decibels (dbA), is a very real concern and problem to overcome in manual C[O.sub.2] particle blasting.

Very significant advances have been made in the last two years to allow operators to use C[O.sub.2] particle blast to clean rubber molds in the presses, and still meet the OSHA regulations requiring less than 84 dbA SPL exposure for an eight hour period in a day. In the single hose constant acceleration system, the physics of isentropic flow have been enhanced by state-of-the art aerodynamic theory and design practice to produce media delivery systems (hose, applicator and nozzle) which provide maximum acceleration and velocity to the particles with minimum shock or turbulence at the nozzle exit. Thus, cleaning performance is high, and generated noise is very low, typically below 102 dbA at the nozzle exit. Studies have indicated that, by using the new low-noise single hose C[O.sub.2] particle blasting systems, while the operator wears an approved blasting helmet or headphones and state-of-the-art ear plugs (dual hearing protection), the noise (SPL) field to which the operator's cars are exposed is well below the required 84 dbA for eight hours of continuous blasting per day.

Table 5 shows the results of noise level data taken during in-the-press mold cleaning tests at a major tire manufacturer in mid-1996. The data present SPL levels measured at the operator's ear level, outside the blasting helmet resulting from blasting with a single hose system at various pressures. Noise level at the operator's ear were within the OSHA standards for several hours of mold cleaning time.

Redeposition of mold fouling residue

It is certainly true that the use of solid C[O.sub.2] as a surface cleaning media creates no significant secondary waste stream. Over time, however, the fouling residue dust which leaves the mold surfaces will redeposit on other parts of the presses and factory production area. Although redeposited residue build up may take weeks or months to even become noticeable, it is in the best interest of the rubber parts manufacturer to deal with if first hand. To date, the most effective and proven method to curtail residue redeposit is to provide adequately sized (CFM) air extraction hoods or air returns located close to the presses. The normal factory air handling system can capture most of the airborne residue particles mad bring them to a central filtering station.

Emerging mold maintenance technologies

Coated molds

The most promising R&D work occurring to reduce mold fouling and parts sticking in the molds is the development of permanent Teflon-based coatings for aluminum and steel mold surfaces which significantly minimize residue adherence and buildup. Early testing with proprietary coatings applied to production tire molds has shown that the molds can remain unfouled for more cure cycles than identical uncoated molds. When the coated molds eventually accumulate residue build-up, the fouling can typically be removed by reduced energy C[O.sub.2] particle blasting, without damaging the coating layer. Once again, with the advent of coated molds, the single hose direct acceleration C[O.sub.2] particle blast systems will offer the most benefit because these systems can utilize blast nozzles with widths up to six inches wide and at the low pressure and kinetic energy level required for the coated molds. Cleaning coated rubber molds can be fast and easy at very low noise, air pressure and C[O.sub.2] pellet flow rates by employing the very wide nozzles available only with single hose systems.

Laser mold cleaning

Laser rubber mold cleaning technology is now available in its developmental stages. Laser technology has shown adaptability for cleaning the surfaces of passenger and light truck tire molds and other rubber product molds, even while they are in the press, but it is questionable if it is able to consistently clean deep cavities, undercuts, vents and hard to access complex mold tooling. Many trial demonstrations and longer term tests of the laser mold cleaning technology have been accomplished in the tire industry in the last two years. In the tire industry, laser mold cleaning has been determined to be non-viable for several reasons, including personnel safety, high cost, inability to thoroughly clean the molds, inability to clean precision details within the molds, lack of industrial robustness of the equipment, and high electric power consumption. Laser technology may evolve to address these problems, but this may be years away.

Additional uses for C[O.sub.2] particle blasting technology

The very same portable C[O.sub.2] particle blasting systems which are used primarily for molded rubber mold cleaning have many other proven uses in rubber manufacturing facilities. Currently, users of this technology are applying it to the cleaning and maintenance of internal mixers, extruder screws and barrels, sprues, runners, gates, ejection pins and general cleaning of the presses during downtime maintenance. When assessing existing and newly emerging molded rubber mold cleaning technologies, the possibility of applying C[O.sub.2] particle blasting technology to many other areas of the rubber products manufacturing process, and the overall favorable impact to product quality and manufacturing productivity, should not be overlooked.

Summary

From the standpoint of fixed (purchasing the system equipment) and operating (electricity, compressed air, C[O.sub.2] pellet media) costs, cost-to-benefit ratio studies conducted by major molded rubber products manufacturers have proven that C[O.sub.2] particle blasting technology is currently the best choice for rubber mold cleaning.

The single hose direct acceleration solid C[O.sub.2] particle blasting system is preferred over a two hose inductive system as the most capable for maintaining robber molds in an unfouled condition throughout a robber product production run. It is the high level of kinetic energy provided which is capable of removing fouling residue from deep cavities, undercuts, sharp part details, and for removing the quickly built-up residue resulting from robber-to-metal bonded parts and excess release agent. Further, it is the thermal effect of the C[O.sub.2] media for quick removal of the glass-like fouling residue that gives the single hose system its double punch for quick, efficient and complete rubber mold cleaning.

Table 1--1991 tire mold material erosion rate study results--C[O.sub.2]
mold cleaning abrasion trials 9-19-96

Material        Test         Average      Maximum          Rate of
             conditions    change in    change in   change of sur-
                             surface      surface      face rough-
                           roughness    roughness         ness per
                          ([micro]m)   ([micro]m)     clean cycles

Steel            1             -0.82          0.2            -.036
                 2             -0.26          1.8            -.038
                 3             -0.94          0.0            -.014
Forged           1              0.24          1.0            -.005
Aluminum         2              0.88          1.6             .022
                 3              0.28          1.1             .001
Cast             1              2.88          5.2             .132
Aluminum A       2             -0.08          0.6             .014
                 3              1.10          3.4             .027
Cast             1              2.30          4.0             .102
Aluminum B       2              1.90          1.9             .183
                 3              6.85         6.85             .078

Material        Test        Average     Maximum
             conditions   change in   change in
                             coupon      coupon
                             weight      weight
                            (grams)     (grams)

Steel            1             0.01       -0.02
                 2             0.00         0.0
                 3           0.02 *       -0.02
Forged           1             0.00         0.0
Aluminum         2             0.02       -0.02
                 3             0.02       -0.02
Cast             1            -1.57       -5.43
Aluminum A       2            -0.60       -4.02
                 3            -1.02       -4.55
Cast             1            -0.93       -2.01
Aluminum B       2           0.18 *      0.18 *
                 3              0.0         0.0

* Weight gain attributed to oxide formation and/or foreign material

Table 2--1996 tire mold material erosion rate study results--compressed
air usage and cost considerations

Sample    Pre-cleaning   Post-clean-   Weight     Cleaning
   #         weight      ing weight     loss     cycle time
            (grams)        (grams)     (grams)   (minutes)

   1         301.64        301.54       0.10         2
   2         304.43        304.40       0.03         2
   3         295.85        295.78       0.07         2
   4         302.37        302.28       0.09         2
   5         298.71        298.62       0.09         2
   6         298.71        298.64       0.07         2

Sample     Number     Equivalent
   #     of clean-     cleaning
         ing cycles    duration

   1         18        6 months
   2         18        6 months
   3         36         1 year
   4         36         1 year
   5         72        2 years
   6         72        2 years

For purposes of this trial, the following parameters were used:
10-day pull schedule; C[O.sub.2] pellets;
250 psi blasting pressure (compressed air); 360 production days;
3" x 5" engraved 2618-T6 aluminum coupons; 36 cleanings per year;
two minute cleaning cycle.

Note: current segmented mold cleaning cycle times are as follows:

Top sidewall: 5 min.; bottom sidewall: 6 min.; tread: 16.5 min.

Table 3--comparing the blast energy coefficient of
single hose vs. two hose C[O.sub.2] particle blasting systems

Nozzle     Nozzle     Delivery      Pres-    Pellets   Traverse
 model   capability    system     sure psi    lb./hr   in./sec.
523 SF      High     Single hose     80        160       0.75
508 SL     Medium    Single hose     70        200       0.75
EA-145      High      Two-hose       80        200       0.75

 Wood      Blast      Swath in.   Depth In.  Envelope   Power
removed    energy                              in.      index
 in**2      Cvbe
 0.37      0.278        1.2         0.11        20       3.05
 0.186     0.140        0.8         0.19        11       2.65
 0.135     0.101        0.8         0.11        23       1.11


Table 4--comparative results of tire mold cleaning with single hose
C[0.sub.2] pellet based system vs. two-hose, shaved C[O.sub.2] block
system

Single hose C[0.sub.2] pellet blasting system

Mold cavity                    269
Top half sidewall cycle        4 min. 10 sec.
Nozzle                         Isentropic (low press.)
Blasting pressure              60 psi
Pellet feed rate               50%

Mold cavity                    269
Bottom half sidewall cycle     4 min. 25 sec.
Nozzle                         Isentropic (low press.)
Blasting pressure              60 psi
Pellet feed rate               50%

Total cleaning cycle           8 min. 35 sec.
Total pellet usage             27.6 lbs.

Two-hose dry ice block shaving blast system

Mold cavity                    270
Top half sidewall cycle        11 min. 40 sec.
Nozzle                         Inductive (round)
Blasting pressure              70 psi
C[O.sub.2] block feed rate     60%

Mold cavity                    270
Bottom half sidewall cycle     11 min. 0 sec.
Nozzle                         Inductive (round)
Blasting pressure              85 psi
C[O.sub.2] block feed rate     60%

Total cleaning cycle           22 min. 40 sec.
Total C[O.sub.2] block usage   44 lbs.

Table 5--noise (SPL) generated while cleaning molded rubber products
mold sidewalls with a single hose C[O.sub.2] pellet blasting system

Trial #     Mold code     Cycle time    Threshold   Threshold
Segmented                 (min.:sec.)    setting     setting
                                        (80 dbA)    (90 dbA)

1           P225/50R16       13:53        98.08       97.61
2           P225/50R16        8:25        97.73       97.14
3           P225/60R16        7:00        97.49       97.10
4           P225/60R16        6:49        97.67       97.08
Two-piece
1           LT265/75R16      13:03        86.48       74.37
2           P275/60R15        8:27        95.93       96.28
3           P275/60R15        7:50        95.91       96.38

Trial #      Nozzle     Pellet
Segmented   pressure   flow rate
              psi         (%)

1              40         55
2              50         55
3              60         55
4              60         55
Two-piece
1              50         55
2              55         55
3              60         55

This article is based on a paper given at the September 1999 Rubber Division meeting.