Three new tools have been developed that could be used to improve process control in the manufacture of automotive glass.

No one can argue that the glass industry consumes a great deal of energy—least of all the glass manufacturers, who have recently been forced to deal with increasing energy bills. The question is, what can be done to change the situation?

In 1997, a project was initiated under the U.S. Department of Energy’s Office of Industrial Technologies involving Visteon (then Visteon Glass Division of Ford Motor Co.) and Pacific Northwest National Laboratory (PNNL). The aim of the project was to develop tools that could be used to improve process control in the manufacture of automotive glass—which could, in turn, allow significant reductions in the production of waste glass, and therefore energy consumption, in automotive glass plants.

PNNL and Visteon decided to focus on the development of three tools: numerical models simulating the glass forming process, a sensor for measuring residual stress within finished products, and a temperature sensor capable of real time temperature measurement throughout the thickness of a glass sheet. These three technologies have been developed to the point of commercialization and may soon be available for use within automotive glass manufacturing plants.

Modeling Glass Properties

The value of accurate modeling in the automotive glass industry cannot be overstated. Product changes occur every year as new vehicles are introduced and new types of glass are developed. Improved numerical models can allow new products to be developed without costly experimental testing within the production plant, and therefore with less waste glass, which must then be reprocessed. For this reason, improved modeling can reduce energy consumption and its concomitant pollution, while making domestic glass producers more economically competitive.

Numerical simulations of the forming of automotive glass products have been done in the past using commercial finite element codes and proprietary industrial software. However, success has been limited due to a problem unique to glass. The role of radiative heat transfer in glass sheets has been difficult to correctly model since glass is opaque in some regions of the electromagnetic spectrum and transparent in others.

PNNL and Visteon developed a stand-alone numerical code that solves the radiation transport problem and calculates its contribution to the temperature distribution within a glass article as it is being formed into a final product. With this improved temperature information, finite element models can more reliably predict final glass properties (such as geometry and residual stress distribution). The PNNL radiation heat transfer code, called RAD3D, is a fully three-dimensional code that is now near completion and will soon be available as a commercial product.

Measuring Mid-Plane Stress

Another concern with automotive glass manufacturers is the quality of their tempered glass products, which are installed as sidelights and backlights. Compressive stress at the glass surface results in greatly improved strength. The ratio of the tensile stress near the glass center and the stress at the surface determines the size distribution of resulting fragments if the window were to fracture, so this ratio has importance in regard to safety. For this reason, the stress distribution is a measure of the final glass quality.

Stress in glass is typically measured using optical methods. Stress introduces a birefringence (light dividing) property into the glass, and this birefringence (and therefore the stress) can be measured by observing the polarization changes in an initially polarized light beam as it propagates through the glass. Surface stress can be measured with existing optical instrumentation. However, measuring the stress near the mid-plane of a glass sheet has traditionally been very difficult.

Working with Visteon, PNNL developed a method for measuring this stress by creating thermal diffraction gratings within the glass, and subsequently scattering a probe optical beam from these gratings. By this method, the probe beam can be made to interrogate the mid-plane of the glass, and the stress is determined by changes in the optical probe beam polarization, just as in existing methods for surface stress measurement.

A diagram of the method is shown in Figure 1. Two laser beams of very short time duration are retroreflected through the glass, so that optical standing waves are created within the glass. The glass becomes slightly heated as it partially absorbs the energy in the beams, and this creates a pair of diffraction gratings (similar to the standing waves) within the glass. The gratings only exist for a short time before thermal conduction destroys them, so the probe beam, also of short time duration, has to be introduced immediately after the gratings are created.

When the probe beam is introduced into the glass at the appropriate angle (called the Bragg angle), a portion of it will scatter along the glass mid-plane. The scattered beam will undergo polarization changes as it traverses the mid-plane. The second grating then scatters a portion of this beam out of the glass and to a polarization-sensitive detector. By comparing the polarization of the doubly deflected light with the polarization of the directly transmitted probe beam light, the user can determine the stress at the mid-plane.

This method has been demonstrated on tempered glass products and has been found to provide accurate estimates for the mid-plane stress. However, the method is not simple. It requires three pulsed laser beams and a polarization-sensitive detector. Advanced versions of the sensor require movable mirrors to adjust the position of the second thermal grating. However, it is currently the only demonstrated method available for measuring the mid-plane tensile stress in tempered glass. PNNL has advanced the method to the point where a stress measurement can be performed within approximately one minute after initial setup and alignment.

Such a device would enable the automotive glass manufacturing industry to verify product quality in off-line measurements. Currently, only destructive methods are available for determining the stress distribution within finished products, and these methods provide only qualitative results. Further development of the method is needed to advance the sensor to the commercial realm, and PNNL is seeking industrial partners to join in such a venture.

Determining Glass Temperature

The production of tempered glass requires the rapid cooling (quenching) of formed glass from a temperature near the softening point to one where the glass is essentially solidified. The quenching rate determines the stress distribution of the final product. Modeling can predict the desired quenching rate for a particular product, and direct on-line temperature measurement can verify that the desired quenching rate is being obtained within the plant. It is necessary to provide a thermal sensor that can quickly and accurately determine temperature throughout the thickness of the glass sheet.

While no such sensor is currently available from commercial sources, PNNL and Visteon developed an optical sensor that shows promise in solving this problem and thereby contributing to better process control within a plant. The sensor relies on the presence of iron oxide additives used in automotive glass for control of ultraviolet and infrared light transmission. It has been found that when auto glass is irradiated with short pulse ultraviolet light of a certain frequency band, the glass fluoresces. The emitted light is mostly red, and the source of the emission is the iron additives. After initial excitation with the short pulse ultraviolet light, the persistence or duration of the fluorescence emission depends on the glass temperature. Near room temperature, the duration is approximately 100 times longer than the duration near the softening point of glass (near 600?C).

The temperature throughout the glass thickness can be determined by imaging the fluorescence emission from various depths within the glass onto separate detectors and independently observing the persistence times for each detector. A diagram of the experimental arrangement used in the experiments is shown in Figure 2. These experiments have verified the concept.

To reach industrial practicality, there is a need to reduce the measurement time to one second or less and improve the temperature resolution capability. This goal can be achieved by using a higher power illumination laser. PNNL will continue work toward improving this temperature measurement method, and is seeking other interested parties in the commercialization of a final instrument.

Improving the Manufacturing Process

These research activities aim to develop new tools that the automotive glass industry can use to improve process control within their plants. The commercialization of these tools and their eventual general availability will allow more energy efficient production of glass products for the many sectors of the U.S. glass industry.

For More Information

For more information about the technologies discussed in this article, contact Moe Khaleel at PNNL, MS K2-18, PO Box 999, Richland, WA 99352; (509) 375-2438; fax (509) 375-6605; or e-mail moe.khaleel@pnl.gov.