Using real-time impedance measurement to monitor and control rubber vulcanization.

Knowing the point at which sufficient vulcanization occurs during the production of rubber products would allow the producer to minimize cycle time while reducing product variation. The objective of this article is to familiarize the rubber industry with impedance monitoring as a means of measuring rubber vulcanization. This article will discuss the background and science of impedance curing monitoring, and report the findings of a laboratory study in which this technology was evaluated in a REP injection molding press. The laboratory study was a cooperative effort between Signature Control Systems and REP International.

Background

Rubber cure parameters are typically established through in iterative process based on a combination of rheometry data and operator experience. Rheometry data give an approximate view of the compound behavior, which is often not representative of what occurs in the mold. As is often the case with injection molding, the rheometric measurement is a poor physical model of the actual production process. Therefore, the rubber injection process is typically established through an iterative process based on operator experience. A standard "end of cure" time is chosen based on the known part characteristic.

Problems arise from this method because the actual mold conditions may vary significantly from the conditions that produced the standard cure time.

The material may have significant variability in different batches or even portions of the batch, or the material may age. Other in-mold variables, such as actual temperature and material heat-up rate, may differ from the condition used during the mold set-up.

Any of these conditions requires a security margin that produces the undesirable effect of having a cure time that is not optimum for the material being processed. In some cases, the result may be a product that does not meet quality specifications, resulting in waste of the material. In other cases, the cure may be held excessively long, resulting in longer-than-needed cycle time, and therefore reducing the maximum output capacity of the manufacturing process. Additionally, overcure can also produce poor product quality by degrading the part characteristics.

Impedance cure monitoring offers an extremely advantageous alternative to this recipe-based approach. Impedance monitoring (also referred to as dielectric cure monitoring) is an emerging technology in the rubber industry. The technology has its roots in the production of aerospace composites, and has been used in this and other industries to a limited, but growing extent. The technology has also been used extensively in rheometry applications. Significant scientific study has been devoted to the topic. Excellent references include Kranbuehl (ref. 1), Persson (ref. 2) and Khastgir (ref. 3).

The fundamental principle of impedance cure monitoring is the ability to detect the degree of vulcanization due to changes in the mobility of ions and the rotational mobility of dipoles in the presence of an electric field. Simply stated, as crosslinking occurs during vulcanization, ion motion and dipole mobility are typically restricted. This allows the producer to monitor the rubber cure process on a molecular level.

Of specific interest to the rubber industry is the capacitance (or permittivity) curve. As established by McCrum, Read and Williams (ref. 4), the inverse of the capacitance curve is directly correlated with the elastic torque curve. The ability to monitor capacitance in real-time, therefore, allows the rubber manufacturer to monitor the development of crosslink density in real-time. The technology allows the manufacturer to automate the production process, based on real-time monitoring of the development of crosslink density.

Impedance cure monitoring allows the manufacturer to monitor these properties in-situ, with a rugged, non-invasive sensor. The patented SCS sensor is inserted into the mold in order to monitor the process. The sensors are encased in ceramic, and are capable of operation at temperatures up to 285 [degrees] C. This equipment has been designed for continuous use seven days a week, 24 hours a day.

A rule base can be developed for each product type that describes the optimal cure characteristic. The control system can be programmed to automatically open the press based on certain key characteristics. Additionally, the unit can be programmed to extend press cycle by recognizing an undercure condition.

The control system is able to respond to variability in raw material, environmental characteristics, press temperature, etc. The system fundamentally measures the actual degree of cure in real-time, eliminating recipe-based cure times. The result is shorter cycle times and improved product uniformity.

Impedance monitoring: Conductance and capacitance

The sensing technology creates an electrical circuit with the SCS sensor acting as one plate of a capacitor, while the other side of the molds acts as the other plate of the capacitor. The product, sandwiched between the sensor and mold wall, acts as the dielectric in the capacitor. A low-level AC voltage is applied to the sensor, resulting in a complex current flowing through the material to the grounded platen. This current consists of both an in-phase component (conductance) and an out-of-phase component (capacitance). During the cure, the dielectric properties of the material change and the changes in capacitance and conductance can be monitored. The SCS system uses this data to signal optimum cure.

According to Von Hippel (ref. 5), changes in the capacitance signal are driven primarily by the ability of dipoles to orient themselves to the oscillating electric field. As resin polymerization increases, dipoles become steadily more restricted in their ability to align to this field. The result is a decreasing capacitance signal. Capacitance monitoring provides an excellent measure of both the rate of polymerization and the degree of cure. Most importantly for the rubber industry is the previously discussed relationship between the capacitance curve and the elastic torque curve.

Changes in the conductance signal are driven partially by the same dipolar effect, but are also sensitive to the ability of ions to move within the material. Again, polymerization typically restricts ion motion, but chemical reaction or phase changes that result in increased ion activity can cause an increase in the conductance signal.

The most significant impact on the conductance signal in rubber production is the presence of carbon black. The inclusion of carbon black in a rubber vulcanizate results in a steadily increasing conductance signal during cure, caused by direct conduction paths of linked carbon particles in the matrix (ref. 6). Obviously, conductance monitoring allows for a direct measure of resistivity of the material, which is an important property in some applications. Additionally, it provides for a method to determine the uniformity of carbon black loading from one cure or batch to the next.

Multiple excitation frequencies may be used in the monitoring process, especially in the early phases of analysis. Dielectric materials have peak responses that vary with the frequency of the applied electric field. An optimum excitation frequency can often be found that provides the cleanest signal and the most reliable response. In many cases, however, the response is fairly uniform across a broad band of frequencies.

Laboratory study

SCS and REP collaborated to investigate the use of impedance measurements for the purpose of monitoring the production of test parts in a REP injection molding press. The material used was a mixture of SBR (styrene-butadiene rubber) and NBR (acrylonitrile-butadiene rubber). The study was broken down into two major parts:

1. Part I of the study was devoted to development of the derived torque curves (inverse capacitance signal), while performing numerous parametric variations of key press settings.

2. Part II of the study reviewed the ability of the system to measure the proper end-of-cure, based on the sensor signal.

Part I: Development of parametric crosslink curves (inverse capacitance curves)

Introduction

Several key press settings were identified that were known to have significant impact on the optimum cure time. These variables were then parametrically varied to produce sets of crosslink curves, obtained in real-time in the mold, which show how the variable impacts the speed of the cure. Press temperature, injection speed, injection temperature and material thickness were all varied in this manner, and the resulting cure was monitored.

The monitoring equipment used was an SCS multi-frequency laboratory analysis system. The system is based around a ceramic-encapsulated IMS-400 impedance sensor, which is flushed mounted in the mold. Data acquisition and signal generation are performed by a National Instruments data acquisition card (AT-MIO-16 E-2) cabled to a National Instruments signal conditioning amplifier (SCXI-1140). The software used to analyze and display the data is a proprietary SCS program written with the Bridgeview programming language. The laboratory analysis system analyzes and displays data only. It has no built in feature recognition or closed-loop control functions. These functions are available in Signature Control Systems' SmartTrac system, which is capable of simultaneous closed-loop control for up to six presses.

Prior to beginning the variation experiment, a frequency sweep was conducted to select the sensor frequency which provided the best apparent correlation with the elastic torque. The material provided the optimum response at a frequency setting of 9 kHz. Unless stated otherwise, all derived curves were obtained at a 9 kHz excitation frequency. All crosslink density curves in this report were generated in real-time during cure.

The "reference cure" for these tests was as follows: Thickness = 20 mm; injection speed setting = 15 mm/s; injection temperature settings = 70 [degrees] C and 80 [degrees] C (injection screw temperature and injection piston temperature, respectively); mold temperature = 185 [degrees] C. Unless stated otherwise, all cure parameters can be assumed to be at this setting. The part produced was a disk of rubber with a radius of 200 mm. The sensor itself is approximately 10 mm in width, and was positioned at the 130-140 mm area in the disk. The sensor was placed in this position to facilitate testing. In an actual production application, the sensor would be placed as close to the centerline of the part as possible.

For comparison purposes, MDR 2000E rheometry curves for the NBR material are provided as Attachments 1 and 2. Attachment 1 was taken at 176 [degrees] C, and Attachment 2 was taken at 196 [degrees] C.

Mold temperature variation

Figure 1 shows the typical 9 kHz inverse capacitance plot taken at 185 [degrees] C for this study. The resemblance to the torque curve of Attachment 1 is unmistakable. Note the initial dip in the signal due to melt, and the subsequent characteristic curve shape as crosslinking develops in the material.

[Figure 1 ILLUSTRATION OMITTED]

Figure 2 shows crosslink density curves for the material with parametric temperature variation from 175 [degrees] C through 205 [degrees] C. Note the clear advancement in the speed of cure as temperature is increased. Also, note the indication of reversion that occurs as the temperature is increased to 195 [degrees] C and above. This also correlates directly with the torque curve of Attachment 2, which shows reversion occurring at 195 [degrees] C.

[Figure 2 ILLUSTRATION OMITTED]

Injection speed variation

Figure 3 shows crosslink density curves for the material with parametric variation of the injection speed. The settings were varied between 5, 15, 25 and 35 mm/s to develop this particular plot. Note the impact that higher injection speed has on speeding the completion of cure. Higher injection speed causes significantly more preheating of the material and advances the time to reach the curve peak by as much as a full minute. Figure 4 is a zoom view of the curve peaks so that the difference can be more easily observed.

[Figures 3 and 4 ILLUSTRATION OMITTED]

Injection temperature variation

Figure 5 shows crosslink density curves for the material with parametric variation of the injection temperature settings. The injection screw temperature and injection piston temperature settings were varied through the range of 60, 70 [degrees] C; 80, 90 [degrees] C; 90, 100 [degrees] C; and 100, 110 [degrees] C, respectively.

[Figure 5 ILLUSTRATION OMITTED]

Figure 6 provides a zoom view of two of the plots to observe the curve peaks more clearly. Crosslink density curves show that injection temperature variation had little effect on the speed of the cure, with the exception of the highest temperature setting (100, 110 [degrees] C). The plot shows that crosslink density was still slightly increasing or had just reached its peak for all the other settings at 260 seconds into the cycle. At the highest temperature setting, the crosslink density peaked at approximately 211 seconds, indicating that the highest setting is high enough to cause significant precure in the piston chamber, and therefore shorten the time to end-of-cure. This high injection temperature is an undesirable condition in that it causes a lack of uniformity in the product due to the precure.

[Figure 6 ILLUSTRATION OMITTED]

Material thickness variation

Several different thicknesses of material were produced. Figure 7 shows a plot of the crosslink density curves (9 kHz inverse capacitance) for the initial 160 seconds of cure for three different thicknesses. The materials in this plot were varied between 5 mm, 20 mm and 50 mm in thickness. It can be observed that the steeper curve slope corresponds to the thinnest material, indicating the most rapid state of cure. Likewise, the thickest piece of material has the most gentle slope, which indicates a slower cure progression. The plots here show the expected impact of material thickness on optimum cure time. However, it should be noted that there were difficulties in measuring the optimum cure state of the 50 mm part. These difficulties and the problem of thick parts in general are discussed in detail later in the report.

[Figure 7 ILLUSTRATION OMITTED]

Repeatability

A series of 23 cures was performed at the reference condition. A small amount of expected variation was observed in the impedance curves, due to normal variation in materials and minor variation in processing parameters. However, curve shapes observed were highly repeatable and characteristic of normal elastic torque curve shapes. Figure 8a shows inverse capacitance curves for four consecutive short duration cures (approximately two minutes). Figure 8b shows the same plot for four consecutive long duration cures (approximately four minutes, 30 seconds). It can be observed that the curve shapes are highly repeatable between cures.

[Figure 8 ILLUSTRATION OMITTED]

Part II of this article will appear in the January 2000 issue of Rubber World.

References

(1.) Kranbuehl, eds. Runt and Fitzgerald. Dielectric Spectroscopy of Polymeric Materials, American Chemical Society, 1997.

(2.) Persson, A novel method of measuring cure - dielectric vulcametry, Plastics and Rubber Processing and Applications 7 (1987), 111-125.

(3.) Khastgir, A comparative study of step curing and continuous curing methods, Rubber World, January 1994.

(4.) McCrum, Read and Williams. Anelastic and dielectric effects in polymeric solids, Dover Books, 1967.

(5.) Von Hippel. Dielectric Materials and Applications. Cambridge, Technology Press of MIT, 1954.

(6.) McCrum, Read and Williams, 1967.

Richard Magill, Signature Control Systems and Stephane Demin, REP International can be programmed to extend press cycle by recognizing an undercure condition.