Compound variables affecting roll compounds exposed to service fluids.

Compound variables affecting roll compounds exposed to service fluids

A major problem in the graphic arts industry is the service life of the rubber compound used on printing rolls. A rubber-covered roll consists of a metal core with a uniform covering of rubber throughout the length of the roll. With use, changes in the rubber cover cause the roll to take on a concave shape; i.e., the diameter of the roll in the middle of its length is less than at the journal ends. In the printing process, this change is reflected in greater ink usage and poorer finished print quality. To compensate for this developing concavity, the operator will adjust the printing rolls by moving them closer together. This "squeezing" action results in a more uniform ink film on the roll. When this type of adjustment can no longer be made, the service life of the roll is at an end.

Printing rolls are subjected to harsh chemicals during operation and maintenance. Shellac and varnish from printing inks and glycerol from the fountain solution cover the roll during the printing operation. Harsher chemicals such as chlorinated solvents, oxygen-containing cleaning fluids, aliphatic and/or aromatic hydrocarbons contact the roll during maintenance and cleaning. Generally, chlorinated solvents are the most destructive to rubber, but they perform the cleaning task very efficiently. The less destructive materials also accomplish the cleaning function but not as efficiently. For the most part, press-operating personnel prefer materials that rapidly clean the roll. Even though chlorinated materials perform this task the best, the suppliers of graphic arts maintenance materials are moving away from their use for reasons of industrial hygiene and environmental safety.

A second consideration in service life is the proper formulation of the roll-cover compound. The rubber compounder needs to be aware that an overzealous maintenance person may combine too much of a harsh chemical along with abrasive mechanical action thereby significantly decreasing the service life of the roll. Roll cover formulators need to have an intimate knowledge of the cleaning agents and how they can affect polymers. This combined knowledge will enable them to "build" maximum service life into the roll compounds they design.

A third aspect is that design of experiment techniques have become an integral part in the engineering and manufacturing of rubber products. Today, the effective compounder utilizes experimental design techniques to gain maximum information with limited experimentation. Using these statistical techniques, he uncovers useful details that may not be readily apparent in his experimental results.

A search of the literature for systematic studies of how operational and maintenance chemicals affect roll covers produced no significant information. Discussions with roll manufactures, suppliers of printing materials and press servicemen led to the conclusion that the interaction of chemicals and polymers was the major factor in roll service life. Thus, a project was initiated using experimental design techniques in an attempt to isolate the most destructive factors. Determining these factors by experimental design techniques should result in a graphic arts roll with a longer service life.

Experimental In the early stages of this investigation, there were many roll compound factors to consider - too many to use traditional "one-at-a-time" experimentation. As mentioned earlier, solvents, oils, aqueous solutions, etc. from one end of the chemical spectrum to the other come into intimate contact with graphic arts rolls. Roll manufacturers have no way of knowing how their rolls will be used once they leave the factory. A consensus from various roll makers indicated that a number of effects may appear during chemical attack in the form of changes in size, shape, weight, surface appearance, physical properties and quality of performance criteria. The ideal response, after exposure, would be to maintain the original roll dimensions. It was important to know how to take a roll that would withstand as many chemicals as possible. The experimental approach was:

* select a typical graphic arts roll recipe and expose it to various chemicals to establish a point of reference,

* use experimental design techniques to identify ingredients and levels that are least affected by these chemicals,

* based upon the information gained from the designed experiments, predict and evaluate the "ideal" recipe.

Recipe selection The first task was to select a starting point formulation typically used in the graphic arts industry (table 1). A 30% nitrile polymer was selected because it was low enough in acrylonitrile (ACN) to accept the high levels of plasticizer required for achieving a low durometer roll. It is the ACN content that allows the finished product to resist chemical attack.

This particular polymer had an 80 Mooney viscosity which helped to maintain good physical properties in a low hardness compound. Vulcanized vegetable oil was added to facilitate the plasticizer absorption and to aid the finish grinding of the roll to its final dimensions. Dioctyl phthalate (DOP) at high levels was used for its good compatibility with the polymer. Epoxidized soybean oil (SOE) offered low extractability. Diatomaceous earth was added for its ability to absorb plasticizer rather than to add reinforcement. The accelerators were chosen based upon their ability to provide adequate scorch protection and good physical properties.

Mixing/testing All mixing was done in a Model "B" laboratory internal mixer with a 1573 [cm.sup.3] capacity. The mixing procedure is outlined in table 2. Physical properties were determined on slabs prepared and cured according to ASTM standards [ref. 1] for 30 minutes at 154 [degrees] C. Tensile, Mpa was 1.89; elongation, % was 360; Shore A hardness was 30 and rebound, % was 65.

Immersion testing was conducted for four hours at room temperature. Even though the immersion testing time was excessive under real working conditions, it was intended to reproduce the effects of repeated cleaning rather than a single exposure to the chemical. Thus, testing under these conditions provided a good idea of the volume swell and weight change that would be expected. The particular chemicals selected were based upon what press operators might use more frequently than others (table 3).

Experimental design Having established representative volume swell and weight change factors with the formulation, the next task was to choose an experimental design that would help improve chemical resistance.

Fractional factorial 

Initially, seven factors were chosen for study: 1) percent ACN, 2) factice type, 3) filler type, 4) plasticizer ratio, 5) benzothiazole, 6) disulfide level, and 7) sulfur level (table 4). A particular concern was the choice of percent ACN level. It is well known that the higher the ACN level, the lower the swell in organic solvents. Thus, why not choose 38 or 40% ACN polymer? The answer to this is that the compatibility and acceptance of plasticizer in higher nitrile polymers works against the rubber formulator. The lower the percent ACN in a nitrile polymer, the greater the acceptance of the high plasticizer loading required for a 20-30 durometer compound. A two-level, L-8 screening design [ref. 2] was selected to evaluate these seven factors. With the validity of the responses in questions, a replicated study was run resulting in each formula being mixed and tested twice. A random being mixed and tested twice. A random order was chosen for mixing and testing all formulations. Design Cube and ANOVA TM software were used to statistically analyze the results.

Our analysis of these data revealed excessive random variation in immersion data, which could have two sources. The most probable sources of variation were experimental error and inconsistent test techniques. The need to more carefully examine laboratory test procedures became obvious. It was determined that it would be necessary to assess volume swell and weight change immediately upon removal from the solvent to eliminate random error due to time effects. A possible source of excessive experimental error was confounding effects. A confounding effect can result when an interaction is mistaken for a main effect. These effects can often be seen when using an abbreviated series of experiments such as found in Fractional Factorial techniques when more than three factors are assigned in an eight-run series. Although it was not learned in the course of the first experimental design what the main factors were affecting the roll formulation, direction for the next step in the project was obtained. It appeared that high percent ACN level improved chemical resistance.

Full factorial

Using the adjusted formula on table 5, the differences in levels of three ingredients were expanded while four ingredients were eliminated from further experimentation. The remaining factors studied were: percent ACN, factice level and plasticizer ratio. A standard order, eight-run, full factorial design served as the second experiment (table 6). This level four design eliminated the "confounding" factors and allowed a clearer picture of the main effects and interactions. Along with these changes, laboratory procedures for reporting volume and weight change responses were more precisely defined to eliminate error. All of these decisions were based on data generated by the fractional factorial experiment.

Results and discussion The full factorial design produced some very definitive results. Because chemical resistance is critical in the graphic arts industry, analysis of room temperature, 60 minute immersion data was our major response focus. Blankrola II and Wash V120 solvents served as immersion media. Samples were monitored for both volume swell and weight change after immersion. Analysis of variance (ANOVA TM) results are summarized available upon request. It was clear that Factor A (percent ACN) was the greatest contributor to volume swell in these cleaning solvents and accounted for 63 and 73% of the variation observed. Percent ACN accounted for 68% of the weight change variation in both solvents. Plasticizer ratio (Factor C) contributed 15.1% and 22.7% respectively to volume swell change in the two cleaning solvents. Surprisingly, the plasticizer ratio had an effect on weight change in Wash V120 (8%) but had no effect on weight change in Blankrola II. Two interactions were significant contributors to volume swell and weight change in the Wash V120 immersions. A x B interaction - percent ACN and factice - accounted for 8.6% of the variance in volume swell and 7.6% of the variance in weight change. The A x C interaction - percent ACN and plasticizer ratio - accounted for 10.4% of the variance in volume swell and 9.9% of the variance in weight change. Alone, Factice D (Factor B) level had no statistically significant effect on the measured responses. Fortunately, the interactions agreed with the main factor effects.

Through examination of the main factor effects graphs for immersion data, it was clear that the 35% ACN level consistently produced lower volume swell and weight change in Wash V120 and Blankrola II solvents. It is also evident that the low level DOP/SOE ratio (50/20) also reduced these values. The difference between these two effects are shown in table 7. Interaction effects on Wash V120 weight change are shown in figure 1. These interpretations supported the preferred main effects observed so far. Although factice levels generally had no effect when running at the 35% ACN level, the lowest percent weight changes were achieved by using the high factice level and at the already suggested low DOP/SOE ratio (50/20).

The next response considered was volume shrinkage after immersion and a subsequent dry-out period. ANOVA TM results were quite dependent on the immersion solvent. The combination of variables studied accounted for over 90% of the variance observed in both solvents. ACN level is the critical factor for low volume shrinkage in Blankrola II accounting for 67% of observed variation. The plasticizer ratio accounted for almost 20%. Plasticizer ratio proved critical in Wash V120 accounting for 52% of the total variation. Reviewing the main factor effects in figures 2 and 3 suggested that the high ACN and high factice levels produced the lowest shrinkages. Their effects were consistent with the previously discussed volume swell and weight changes. For low volume shrinkage after dry-out, the high level DOP/SOE 20/50 ratio is preferred, but where a reduction or a minimum volume swell and weight change is desired, the 50/20 ratio of DOP/SOE is needed. Thus, a compromise must be made.

Conclusion For optimum chemical resistance, a 35% ACN polymer with 30 parts of factice was chosen because of the contrasting requirements observed in the full factorial experiment. It was decide to run confirmation experiments with both ratios of DOP/SOE plasticizer (50/20 and 20/50) (table 8). After 4 hours (table 9), immersion testing indicated that the 20/50 plasticizer ratio in Formula D gave the better results. Volume swell, volume shrinkage and weight change results were lower in most instances with this ratio. When these tensile slabs were oven aged, however, a greasy bloom appeared on the surface of this compound. This bleeding occurred only on the 20/50 ratio and was disappointing in view of the good immersion results, especially shrinkage. Although bleeding was not anticipated, 70 parts of plasticizer in a low fillerloaded compound based on a 35% ACN polymer is straining the limits of compatibility.

At this point, several options were considered: 1) maintain the 20/50 DOP/SOE ratio and increase the filler level to absorb more plasticizer, 2) reduce the ACN content below 35%, 3) introduce a more soluble epoxidized soybean oil, 4) reduce the SOE by altering 20/50 DOP/SOE ratio to 30/40 or more, or 5) changing plasticizer(s) to one(s) that might give better values at the 50/20 ratio and not alter our previous balance of properties. Rather than use the 20/50 ratio and risk bleeding, the 50/20 plasticizer ratio in Option 5 was selected.

Into the 50/20 DOP/SOE plasticizer ratio, two new phthalate plasticizers were introduced to reduce volume swell and weight change without greasy bloom. Experimentation was conducted with diisodecyl phthalate (DIDP) and diisononyl phthalate (DINP) as potential replacements for DOP. Both are branched alcohols as is DOP, but they are recognized as having fewer potential health problems [ref. 3]. Using either one of these materials as a substitute for DOP should also allow the continued use of a 35% ACN polymer.

After mixing the formulations with the new plasticizers (table 10), immersion testing indicated that DINP had lower volume swell and weight change in comparison with DOP at the 50/20 ratio. Side-by-side comparisons of DOP and DINP showed that DINP was better than DOP for volume swell and weight change and no bleeding at the 50/20 ratio. It should be noted, however, that 20/50 DOP/SOE was still the best combination for low volume swell and weight change in the cleaning solvents evaluated, but this combination bleeds. Based on volume swell, volume shrinkage, weight change and hardness after immersion in the chosen media, the "ideal" recipe in this study was Formula G, table 10.

Summary Through a minimal amount of experimentation, the major factors causing undesirable volume swell, volume shrinkage and weight change in printing rolls have been determined;

For low volume swell

* 35% ACN level polymer

* 50/20 DOP/SOE ratio plasticizer

For low weight change

* 35% ACN level polymer

* 50/20 DOP/SOE ratio plasticizer

* High level factice (30 parts)

For low volume shrinkage

* 35% ACN level polymer

* High level factice (30 parts)

* 20/50 DOP/SOE ratio plasticizer

For a balanced formulation

* 35% ACN level polymer

* High level factice (30 parts)

* 50/20 DOP/SOE ratio plasticizer

* DINP in place of DOP plasticizer

The major lesson drawn from this study was that a roll compound can be continuously improved by using simple experimental design techniques. Using this "statistical" knowledge, a roll compound can be designed that will have a longer service life. [Figures 1 to 3 Omitted] [Tabular Data 1 to 10 Omitted]

References [1]Annu. Book ASTM Stand. 09.01, Standards 395, 412, 1054 and 3182. [2]Genichi Taguchi, "System of Experimental Design," Vols. 1 and 2, Kraus International and American Supplier Institute, New York, 1987. [3]Bernard F. Schneider, ed., Recent Advances in Phthalate Research, Princeton Scientific Publishing Co., Inc., Princeton, NJ 1987.