·Home ·Table of Contents ·Materials Characterization and testing

On-Line Monitoring of Thermosets Moulding Process Applying Ultrasound Wolfgang Stark, Joachim Döring, Jarlath McHugh, Federal Institute for Materials Research and Testing (BAM), Berlin, Germany Christian Kürten Iserlohner Kunststoff-Technologie, ISK, Germany Contact

1 Introduction

Moulding compounds are thermosetting materials widely used in industrial products, e. g. electrical devices, automotive industry. They are manufactured principally by compression and injection moulding. The manufacturing process is more complex than for non-reactive thermoplastics since two steps must be controlled within a few seconds: the mould filling operation with molten compound and curing reaction in the hot mould. Unfortunately flow and curing behaviour is sensitive to the composition of the moulding compound and depends also considerably on manufacturing conditions. Therefore manufacturers have problems to supervise the quality. State of the art is to control the daily output by a global control of selected products at the end of the shift for example by boiling in dye solution to determine the degree of cure. Under and overcuring are characterised by definite discoloration. A component part control in this way is impossible. Therefore manufacturers of thermosets have been searching for non destructive testing methods for online process monitoring. Two methods come into question: measurement of the change of dielectric parameters or of sound propagation properties during manufacturing process. The area of use of dielectric methods for moulding compounds is limited as most varieties are cured via a polycondensation reaction whereby water is set free. Such results cannot be evaluated. Ultrasound parameters do not have these shortcomings as they depend directly from the change in mechanical properties during the manufacturing process.
In this contribution we will report on the findings in application of ultrasound method in the years followed 7th European Conference on Non-Destructive Testing (ECNDT) 1998 [1,2].

2 Experimental

2.1 Moulding compounds
Moulding compounds are mixtures of reactive resins, hardener and fillers. At room temperature they are solid. For manufacturing they are formulated as granulate. They are prepared to have a melting point in the range of 60...70°C. The chemical crosslinking reaction starts in the range of 100...120°C. In the manufacturing process the moulding compound is first formed in the molten state and then cured. The whole process takes only 20...30 s.
Typical thermosets are resins of phenolic - PF, urea - UF, melamine - MF, melamine-phenolic - MP, epoxy - EP, unsaturated polyester - UP (with glass filler - bulk or sheet moulding compounds BMC, SMC). Thermosets are widely used for their low price and sometimes outstanding mechanical and electrical properties, especially at high temperatures.

2.2 Manufacturing process
Moulding compounds are manufactured by compression and injection moulding. High pressures are needed to ensure that the high viscosity material fills the mould and to avoid bubbles caused by the steam pressure of locked in water.
In compression moulding the granulate is directly fed into the tool or it is preformed and preheated to about 90°C to reduce the cycle time and to improve the filling. Typical tool temperatures are in the range of between 140 to 180°C. By closing the mould the piece is formed. The mould is held under pressure until the crosslinking reaction has reached a definite level. This must not be the full curing stage.
The major difference of injection moulding process is that the moulding compound will be pre-melted in an extruder unit. The molten material will be injected into the closed mould where it remains until the reaction has reached the prescribed level.

2.3 Ultrasonic measuring technique
A more detailed description of the measuring technique is given in [3]. Quantities to be measured are sound velocity and damping. In principle propagation of longitudinal and transversal sound waves can be used but as in the experiments completed transversal waves do not deliver signals in the molten state. Therefore measurements with longitudinal wave sensors are concentrated on.

2.4 Sensors


Fig 1: Sensor for application in the mould e

The ultrasound sensors must resist pressures up to 600 bar and temperatures up to 200 °C. Such sensors based on ceramic and working at a frequency of 4 to 5 MHz are now commercially available [4]. Fig. 1 shows a photo.
The typical diameter of the measuring plane is 8 mm, the length 20 mm. The sensors are set plane into the mould surface.
Two sensors are arranged in opposite positions in both halves of the mould. The through transmission technique offers the advantage that the first received signal always contains the information needed. Reflection technique demands thorough calibration of every mould which can not be achieved under industrial conditions.

2.5 Equipment
Ultrasonic signal generation and evaluation are managed by a computer controlled ultrasonic test equipment. The equipment is based on a commercial device. Using a LabView program the equipment works automatically to register sound velocity and damping during the manufacturing process. More details are given in [3]. For sound velocity calculation thickness of the part will be stored or directly read into from a digital gauge.
A schematic view of the whole arrangement is shown in Fig. 2 exemplified for compression moulding.

Fig 2: Sketch of the experimental set-up for online cure monitoring in a compression mould 1 and 2 - sensors, 3 and 4 - half's of the mould, 5 - digital thickness gauge, 6 - mould temperature and pressure controllers, 7 - NDT ultrasonic test equipment, 8 - PC for data handling

2.6 Modulus and damping calculation
From a more physical point of view the ultrasound quantities can be related to the mechanical quantities by the following.
The modulus M is calculated from

M = rv² r-density ,v-sound velocity
The damping is calculated as a relative quantity related to the amplitude of the cured material because it is not easy to obtain real amplitudes at the twofold transition from the sensors to the moulding compound:
Da-relative damping,x-Sample thickness,A₁-actual amplitude,A₀-final amplitude

3 Results and Discussion

To demonstrate the advantages of online process monitoring some examples will be given. They are divided as in industrial practice into problems of moulding compound composition and influence of manufacturing conditions. More details are given in [5]. After having shown in compression moulding tests that all intentional precipitated irregularities are detectable, an example of online quality control of an injection moulded car engine component will be demonstrated.

3.1 Deviations in moulding compound composition


3.11 Influence of hardener concentration
For phenolic resin hexamethylentetramine (HMTA) is used as hardener. At about 110 °C HMTA breaks up and produces formaldehyde. This reacts with phenolic resin and forms a chemically crosslinked network. For example purposes, one case is demonstrated with a total absence of hardener which is obviously not realistic. Fig. 3 shows measuring results for a standard concentration (13%) and for absent HMTA. 20 mg room temperature granulate is filled into the preheated mould (160 °C). The mould is closed and the ultrasound measurement triggered. The resulting information for modulus and relative damping are given in Fig. 3:

Fig 3: Change of modulus and damping during manufacturing process of phenolic moulding compound (PF) The numbers in Fig. describe the following process: first acoustic contact during closing the mould softening of granulate by temperature increase start of curing reaction final value of soft, non-cured material maximum speed of curing reaction slowing down of curing reaction final value of cured material

It is clearly seen that material without HMTA behaves as a thermoplastic which only softens. Crosslinking reaction is characterised by a distinct increase of modulus.
Adequate information although in an other form can be obtained from the damping. The maximum in damping for material without HMTA accompanied by a reduction in modulus is typical for the dynamic glass-rubber transition in polymers. At a definite stage of softening the relaxation time of mechanical elements passes through the measuring frequency which leads to a maximal movement (resonance effect) with a maximum in friction and hence energy loss. The same process should be seen with HMTA. Here, however the first transition glass-rubber (softening) is closed in by the transition rubber-glass from crosslinking reaction which causes also a peak. So both peaks blend to form one broad peak.
In practice during the production of moulding compound inhomogenities or deviations in hardener concentration may occur. Such deviations were simulated by mixtures with different concentrations of HMTA. Fig. 4 demonstrates that not only dramatic failures such as absent hardener as in Fig. 3 can be clearly detected.

Fig 4: Development of sound velocity (long and trans) during compression moulding, parameter: content of hardener HMTA

As expected there is a good correlation between content of hardener and final modulus in the cured state. The possibility to determine a measurable signal from transversal waves when high concentrations of hardener are used was also investigated. In this case also the measurement starts late after the beginning of the reaction and gives little insight into the complete process.

3.12 Influence of moulding compound humidity
An important problem for moulding compound production and manufacturing is also to hold the humidity at a definite level. A pre-set concentration of several per cent is usual to improve the processibility. Particularly by storage of the compound this concentration can vary. In Fig. 5 the influence of artificially changed humidity are demonstrated. The original material has 2%.

Fig 5: Development of sound velocity during compression moulding, parameter: humidity of the moulding compound

It is surprising that small changes in humidity cause dramatic changes both in flow behaviour and in final modulus. The plasticizer effect of water on the flow behaviour may be expected but the great deviations of final modulus in the cured state are surprising and give an explanation of some up till now not well understood problems in production.

3.13 Influence of mould temperature
Another problem in manufacturing is to hold the mould set temperature constant. As well as uncertainties of temperature regulation as changes in cycle time and also failures in an individual heating circuit may cause deviations in mould temperature and its homogeneity. To investigate this influence and to show the potential for cycle time reduction, the influence of mould temperature was investigated. The results are represented in Fig. 6.

Fig 6: Development of sound velocity during compression moulding, parameter: mould temperature

The higher the temperature the softer the material becomes during the heating-up period (minimum in modulus) and the earlier the crosslinking reaction starts and reaches its final value. The different final values, especially the decrease of modulus with temperature is an indication of the temperature function of modulus in the glass state. The great influence of mould temperature on manufacturing process becomes evident. Deviations from the expected curve give the manufacturer a clear indication in which direction temperature shift occurred. The method can be excellently used to find out which pre-heating period is required after a mould change (some hours for heating are needed !) before the production can be started again.

3.2 Automotive belt pulley production by injection moulding
A significant automotive supplier used the online ultrasound monitoring for component part quality assurance [6]. The product - belt pulley - is shown in Fig. 7 and the production line in Fig. 8.

Fig 7: Belt pulley produced with ultrasonic online process monitoring

Fig 8: Injection moulder of Kendrion Backhaus, Kierspe, Germany equipped with ultrasound online monitoring for component part quality management

During production the sound velocity is controlled continuously. The latest 25 curves are permanently displayed on the monitor to look for deviations. The latest curve is marked fat or coloured. The whole set of measuring results is permanently stored for documentation and can be transferred to the central production server. It can be handed to the consumer as quality documentation. By an extra expert software non acceptable group of curves leaving a preselected area (see side lines beneath the curves) can cause programmed actions. The simplest is to stop the production and to release alarm. In a more developed variant the machine parameters can be changed to correct the deviations. In this way for example the mould opening time could be adapted to maintainable changes in moulding compound properties to reach the desired state of curing.
Fig. 9 demonstrates a collection of 25 curves direct from production control.

Fig 9: Example of sound velocity curves from online process monitoring, 25 last cycles are shown, latest is marked, the lines mean control levels for alarm and automatic process stop

4 Summary

We have demonstrated that NDT ultrasonic technique is a powerful tool for online process monitoring. Using several examples we showed that moulding compound variations as well as deviations in machine parameters are reliably detected. Therefore moulding compound manufacturer and processors can employ such equipment for solving their everyday problems. The whole equipment which means sensors and ultrasonic NDT equipment with adapted software are now commercially available [4]. A trial period in production of belt pulleys will end shortly. This resulted in some improvements of the equipment to be realised in the future to enhance the ruggedness for industrial application. The results from this test phase are quite impressive.

Acknowledgements

The authors thank Dr. G. Splitt (Branson-Krautkrämer, Hürth) for the good co-operation in the development of sensors and measuring equipment. We want to thank J. Bartusch BAM for the development of the adapted software. M. Rath (TU-Berlin), S. Robisch (Technical College Iserlohn) and R. Koenig (Kendrion Backhaus, Kierspe) were involved in experimental testing. Special thanks to Kendrion Backhaus for the photos from the belt pulley and machine. The association of European thermoset producers and manufacturers AVK-TV, Frankfurt/M., especially Dr. U. Bültjer (secretary) and M. Stahl (Bakelite AG, Letmathe), we wish also to thank for financial support form AVK-TV and continual encouragement.

6 Literature

1. Stark, W., Bovtun, V., Döring, J., Kürten., Ch. Monitoring of Curing Reaction of Polycondensating Thermosets at Press and Injection Moulding in: Proceedings 7th European Conference on Non-Destructive Testing (ECNDT) Copenhagen 26-29 May, 1998 and http://www.ndt.net/abstract/ecndt98/141.htm
2. Döring, J., Stark, W., Splitt, G. On-line process monitoring of Thermosets by Ultrasonic Methods in: Proceedings 7th European Conference on Non-Destructive Testing (ECNDT) Copenhagen 26-29 May, 1998 and http://www.ndt.net/abstract/ecndt98/159.htm
3. Döring, J., Bartusch, J., McHugh, J., Stark, W. Contribution to Ultrasound Cure Control for Composite Manufacturing, Proceedings, 15^th World Conference for NDT, Roma 15.-21. Oct 2000
4. http://www.isk-iserlohn.de/usplus/index.htm
5. Rath, M., Döring, J., Stark, W., Hinrichsen, G. NDT&E International 33 (2000) 123-130
6. Stark, W., Döring, J., Kürten, Ch., Robisch, S., Koenig, R. Lecture, Annual Polymer Physics Meeting of the German Physical Society, 13.-15.3.2000 Potsdam

© AIPnD , created by NDT.net

|Home|    |Top|