Improved rubber processing for sealing systems using in-mold dielectric sensors.

Knowing the point at which sufficient vulcanization occurs during the production of rubber products would allow the producer to minimize cycle time while simultaneously reducing product variation. This article reports the findings of a technology trial in which intelligent closed-loop control methods were evaluated while injection molding HNBR seals. In particular, this article details the methods by which improved compression set properties were obtained through the use of in-mold dielectric sensors. The laboratory study was a cooperative effort between Signature Control Systems (SCS) and Federal-Mogul.

Background

Many previous studies have been performed detailing the use of dielectric cure monitoring as a method of monitoring the progression of vulcanization in real-time. Previous SCS papers (refs. 1 and 2) and papers by Khastgir (ref. 3), Kranbuehl (ref. 4) and Persson (ref. 5) provide extensive detail regarding various approaches to the topic. Fundamentally, the use of in-mold sensors allows for real-time monitoring of vulcanization, allowing an intelligent control system to determine when the part is properly cured, rather than relying on the more traditional recipe-based rheometer approach.

Problems may arise from the recipe method because the actual mold conditions may vary from batch to batch, or run to run. The material may have significant variability in different batches, or the material may age. Other machine variables, such as injection rate or barrel temperature, may differ from the presumed normal condition used during the mold set-up.

Impedance cure monitoring (also referred to as dielectric cure monitoring) offers an extremely advantageous alternative to this approach. A rule base can be developed for each product type that describes the optimal cure characteristic. A rule base is simply an algorithm that is used to interpret the sensor data. The control system can be programmed to automatically open the press based on certain key characteristics. Additionally, the unit can be programmed to extend the press cycle by recognizing an undercure condition.

Impedance monitoring

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 (figure 1).

[ILLUSTRATION OMITTED]

A low-level AC voltage is applied to the sensor, resulting in a complex current flowing through the material to the grounded mold surface. This current consists of both an in-phase component and an out-of-phase component, from which conductance (loss factor) and capacitance (permittivity) of the material can be derived. During the cure, the dielectric properties of the material change and the changes in capacitance and conductance can be monitored. Furthermore, it is this change in a material's impedance, its relative value, which the SCS system uses for reference to determine and signal optimum cure.

The capacitance curve is a monitor of dipolar presence and mobility within the material (ref. 6). If dipoles are free to align to the applied field, the capacitance signal is high. If dipole mobility is restricted, the capacitance signal decreases. Dipole mobility is typically restricted as vulcanization occurs (due to increasing crosslink density), although the addition of dipolar pendant side chains may also create unrestricted dipoles that can cause the signal to rise (ref. 7).

Since we have observed that it is possible to monitor the effects of crosslink density on various dipoles, this implies that we should therefore be able to monitor the progression of crosslink density (and modulus) in the bulk material. This can be done through the careful selection of a monitoring frequency that produces a response curve that most directly follows rheometry or step cure modulus data.

Figure 2 demonstrates the effects of curing a compound at varying temperatures. Note that the curve shifts in the manner that we expect as temperature is increased.

[GRAPH OMITTED]

Conductance data can provide a reliable measure of the vulcanization as well. The conductance signal also varies in a predictable manner with the development of crosslink density. Conductance signals can either increase or decrease during cure, depending on a number of potentially competing effects. First, the types and amounts of conductive fillers in the material have a significant effect on conductivity, due to the creation of micronetworks of conductive fibers that form during vulcanization (ref. 8). Another important factor is frictional losses associated with dipolar rotation in the applied field (ref. 9).

Either capacitance or conductance (or both) can be used as a control methodology, provided that appropriate correlation is performed between the signal and the desired part characteristic.

Laboratory study

The trial was conducted on a 50 ton injection press with a single cavity gasket sample mold. Part geometry required a mold modification to create a small witness cavity, next to the actual part cavity. This witness cavity contained the sensor and also provided a small button sample (d = 19 mm, t = 3.8 mm) for material property testing.

Compression set resistance was the material property of choice in this case. Compression set is an important property for gasket applications. Also, previous studies had shown that it was a more discerning property to determining state of cure than, for example, microhardness or tensile properties for this particular material.

The compound is based on a HNBR polymer with low unsaturation, carbon black filled and peroxide cured. The coagent was chosen for optimum compression set and compression stress relaxation (CSR) properties. However, these will only be achieved with fully cured material. Any drop in mold temperature or changes in other processing parameters could result in a significant increase in compression set. In order to compensate for these process variations, the cure time would have to be extended significantly, or the parts would have to go through an oven post cure operation.

Previous studies on this compound have also shown that post curing significantly improved compression set in undercured material, while it remained mostly unchanged in fully cured samples. Undercured parts that were post cured showed similar compression set properties as fully cured parts. Compression set before and after an oven post cure was chosen as a way of checking state of cure in this compound. Undercured samples might show a decrease of compression set after post cure of more than 75% (36 actual points). Acceptably cured samples showed a decrease of 25% (4 actual points), or less.

An intelligent algorithm was developed for the compound and was evaluated in this study. This algorithm measures the vulcanization rate, as represented by the slope of the impedance data. It then performs a calculation for the optimal cure time, based on the measured vulcanization rate.

A series of 25 parts was produced at a normal cure temperature of 199 [degrees] C (390 [degrees] F) using the impedance technology to determine end of cure. Another set of 25 parts was molded using the current recommended cure conditions of 180 seconds at 199 [degrees] C (390 [degrees] F). Each part (button from witness cavity) was tested for compression set resistance after 22 hours at 150 [degrees] C per ASTM D 395 method B. The purpose of the comparative study was to examine if it was possible to produce more consistent part properties with the impedance technology that would be possible using a normal fixed cure time. Additionally, it was of interest as to whether these more consistent properties could be realized while simultaneously reducing the average cure time.

Results

Table 1 shows compression set results, comparing the impedance-controlled cure time parts and fixed cure time parts at constant mold temperature. All samples exhibited acceptable compression set resistance, with the intelligent algorithm producing the lowest standard deviation. Figure 3 depicts the frequency distribution of compression set resistance.

[GRAPH OMITTED]

Table 1 - comparison of intelligent algorithm and fixed
cure time

Sample ID   Intelligent        Fixed        Intelligent     Cure time
            comp. set (%)   comp. set (%)   cure time (s)   saving (s)

1               13.9            11.4            158.1          21.9
2               10.8            11.1            157.3          22.7
3               13.9             8.6            157.8          22.2
4               11.1             8.6            156.3          23.7
5               13.9            13.9            155.3          24.7
6               11.1            11.1            157.8          22.2
7               11.1             8.6            159.5          20.5
8               13.9            11.4            156.1          23.9
9               11.1            11.4            161.3          18.7
10              11.1            11.4            166.4          13.6
11              11.1            11.4            164.3          15.7
12              13.9            13.9            172.3           7.7
13              13.5            13.9            159.3          20.7
14              11.1            11.4            164.3          15.7
15              11.1            11.4            161.3          18.7
16              10.8            11.4            161.8          18.2
17              11.4            11.4            159.3          20.7
18              13.9            11.4            165.3          14.7
19              14.3             8.6            157.9          22.1
20              11.4            11.4            166.3          13.7
21              11.1            11.4            162.8          17.2
22              10.8             8.6            161.3          18.7
23              13.9             8.6            160.3          19.7
24              11.1             8.6            159.8          20.2
25              10.8            11.4            161.8          18.2
Avg.            12.1            10.9            161.0          19.0
Min.            10.8             8.6            155.3           7.7
Max.            14.3            13.9            172.3          24.7
Std. dev.        1.4             1.7              3.9           3.9

Part uniformity was improved by 18% (taken as the percent reduction in standard deviation between the two sample sets), while the average cure time was reduced by 10.5%.

This portion of the study showed that it was possible to create more uniform part properties while simultaneously reducing the average cure time, using the impedance control methodology. These results were obtained while producing parts under nominally "normal" cure conditions. The next area of interest was to investigate how the impedance control methodology would respond to variation that is induced into the cure cycle. These types of abnormalities could include compound variation, mold temperature variation or machine setting variation.

For the purpose of this study, temperature variation was chosen as a convenient method to induce variability into the process. A 20 [degrees] C temperature oscillation was induced into the process. A temperature variation of this magnitude would be certain to significantly impact the mechanical properties of parts being produced with a fixed cure time. Ideally then, the impedance technology would be able to account for the temperature variation and still produce consistent part properties.

The algorithm was then evaluated over this temperature range. Four specimens were molded at each temperature. Two of each were oven post cured for four hours at 150 [degrees] C. Again, compression set properties were evaluated after 22 hours at 150 [degrees] C (table 2). Figure 2 shows the sensor response to the induced temperature variation.

Table 2 - evaluation of algorithm - compression set
before and after oven post cure

                       Barrel      Mold
                        temp.      temp.    Cure             Comp.
                       ([deg-     ([deg-    time     OPC      set
Sample ID             rees C])   rees C])   (s)               (%)

1                       116        198      163    [check]    11.1
2                       116        198      165    [check]    11.1
3                       116        198      165      --       11.1
4                       116        198      161      --       11.1
5                       110        198      194    [check]     8.3
6                       110        198      183    [check]     8.1
7                       110        198      183    [check]    13.5
8                       110        198      169      --       13.5
9                       116        193      134    [check]    10.3
10                      116        193      210    [check]    10.8
11                      116        193      213      --       13.9
12                      116        193      207      --       13.9
13                      116        188      227    [check]    10.5
14                      116        188      235    [check]    11.1
15                      116        188      251      --       11.1
16                      116        188      257      --       11.1
17                      116        183      470    [check]     8.3
18                      116        183      403    [check]    14.3
19                      116        184      373      --       11.4
20                      116        183      368      --       11.8
21                      116        179      780    [check]     8.8
22                      116        179      787      --        8.8
Average                  --         --       --      --       11.1
Average (+ OPC)          --         --       --      --       10.2
Average (w/o OPC)        --         --       --      --       11.9
Std. dev.                --         --       --      --        1.9
Std. dev. (+ OPC)        --         --       --      --        1.8
Std. dev. (w/o OPC)      --         --       --      --        1.6

Decreasing the mold temperature from 198 [degrees] C to 179 [degrees] C resulted in a cure time increase from 164 seconds (2.7 minutes) to 784 seconds (13 minutes). Compression set resistance was acceptable over the entire evaluated mold temperature range. In fact, the standard deviation of the non-post-cured parts remained at approximately the same as in the constant temperature study (1.6 versus 1.4).

Also acceptable was the decrease in averaged compression set after post cure. Decreases varied between zero and 23.7% (zero to 3.3 actual points) (figure 4).

Figure 4 - evaluation of algorithm - compression
set before and after oven post cure

Compression set (%)

163-164 s          11.1   11.1
@198 [degrees] C

172-210 s          13.0   10.6
@193 [degrees] C

231-254 s          11.1   10.8
@188 [degrees] C

371-437 s          11.6   11.3
@184 [degrees] C

780-787 s           8.8    8.8
@179 [degrees] C

Figure 5 shows the relationship between mold temperature and cure time as determined by the impedance control system. Compression set properties remained fairly constant.

[GRAPH OMITTED]

Aside from mold temperatures, other process parame-ters influence the rate of cure in a rubber compound. Shown here is how the impedance sensor detected rate of cure differences for varied screw and barrel temperatures at a constant mold temperature. Material injected at a lower starting temperature of 82 [degrees] C (180 [degrees] F) showed a slower rate of cure than material injected at 115 [degrees] C (240 [degrees] F) (figure 6).

[GRAPH OMITTED]

Conclusion

The laboratory study showed that impedance monitoring of the HNBR compound provides an important new method to monitor the vulcanization process in the mold. Specifically, the impedance data clearly and consistently showed the impact of mold temperature on the vulcanization rate. Additionally, the impedance data accurately reflected the impact of other machine variables on the vulcanization process, such as screw and barrel temperature.

In addition to reflecting the cure state of the rubber during vulcanization, the study also showed that it is possible to use this information for real-time control of the production process. Utilization of this control methodology offers the promise of both reduced cycle times and improved mechanical properties, specifically compression set properties for this compound. The study showed that it was possible to reduce cure time by more than 10%, while still maintaining adequate mechanical properties.

This research also showed that the impedance methodology is robust in dealing with process variations that would normally result in a substandard product. Adequate part properties were maintained over a full 20 [degrees] C temperature variation.

Finally, the study was performed in a production ma-chine and mold, producing an actual part. The impedance monitoring system in use was not a laboratory device, but is commercially available and suitable for a variety of industrial rubber production applications.

References

(1.) Magill and Demin, "Using real-time impedance measurement to monitor and control rubber vulcanization during cure," Rubber World 221 (1999) 3, pp. 24-28, 62.

(2.) Magill, "Real-time control of vulcanizate cure times and properties using in-moM sensors," paper presented at the Spring 2000 Rubber Division meeting of the American Chemical Society, April 2000.

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

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

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

(6.) Von Hippel, Dielectric materials and applications, Cambridge, Technology Press of MIT, 1954.

(7.) See reference 3.

(8.) McCrum, Read and Williams, Anelastic and Dielectric Effects in Polymeric Solids, Dover Books, 1967.

(9.) See reference 6.