Clear benefits

IN MODERN glass technology, precise control of melting conditions is of fundamental importance. The viscosity of a melt in particular has a decisive influence on its further processing, quality and homogeneity. In the laboratory, viscosities are often determined using rotational viscometers: a spindle is rotated at a constant speed in the sample, the required torque is measured and the viscous resistance, i.e. the viscosity, is determined.

In our customer's project, this concerns the melting of quartz glass. Viscosity is not only a measure of fluidity but also allows conclusions to be drawn about the internal state of the melt, for example, cavities (air bubbles) or incomplete material bonding.

Focusing on precision glass components to improve the manufacturing process for quartz glass in terms of quality and process reliability, the challenge was that the glass is first melted at high temperatures and then has to reach a defined viscosity zone for further processing. Only in this range can it be guaranteed that the melt is homogeneous (no air bubbles), the connection or fusion of different material batches has taken place completely, and the transition to the shaping or quenching/pre-forming process starts reliably.

Reliable monitoring

The objective was to reliably monitor the viscosity of the quartz glass melt in order to detect deviations in the melting process at an early stage and to establish process-reliable feedback for control and quality assurance.

The viscosity of molten glass plays a key role in processing, as it determines the flowability and thus the moulding behaviour, bubble removal, degassing and homogenisation. In the case of quartz glass, the conditions are particularly demanding with: high temperatures and viscosities having a strong influence on the viscosity curve. 

The TorqSense SGR541 was chosen because of its robustness against extreme process conditions (temperatures, thermal shocks, possible vibrations) in the melting range of quartz glass, with separate sensor unit and electronics. Calibration was performed using reference samples or model melts in order to map the torque ↔ viscosity relationship in detail in the specific process setup.

The sensor was integrated into the process – typically between the motor/agitator (or melt mixing system) and the impeller (or melt stirrer) – and connected to the process control system in real time. Since torque transducers can be sensitive to lateral forces, double bearings should be used to avoid transverse forces.

The relevant target values for viscosity were defined together with the customer: e.g., a range from x to y Pa·s at temperature T, at which the melt is considered homogeneous and no bubbles are mobile.

Reference curves were created using laboratory or pilot melts with known compositions and viscosity values. This allowed the sensor output to be mapped to absolute or relative viscosity.

Torque data

The sensor continuously provided torque data from which the viscosity was derived. Trend analysis enabled early detection of deviations – for example, if the viscosity did not fall within the target range, which could indicate trapped bubbles or incomplete mixing of the material batches.

As soon as the measured viscosity deviated from the target curve, measures were taken immediately: adjustment of the temperature, extension of the stirring time, addition of degassing (refining), or even return of the batch for reprocessing.

The introduction of monitoring using SGR541 had the following positive effects: The melt could be continuously monitored for homogeneity and freedom from bubbles. Deviations were detected early and corrected; and the data obtained (torque/viscosity curves) provided valuable insights into the behaviour of the quartz glass melt, enabling future process optimisations.

Tony Ingham is managing director of Sensor Technology

www.sensors.co.uk