GB flag iconENCN flag iconZH

Assessment of glassware

There are a number of test procedures used to assess the strength and quality of glassware.

Glass is a hard and brittle inorganic material which has been used by humans since at least 3,000 BC. Although glass does occur naturally, we are more familiar with its man-made form which is produced by fusing together a mixture of the main ingredients of silica (from sand), soda and lime by heating. Other ingredients may also be added to impart desirable properties, such as sulphides and metallic oxides to produce different colours and effects. In its molten state, the glass can be blown or moulded into the required shape. In the home, glass is widely used as drinking vessels, ovenware such as casserole dishes, placemats and chopping boards. SATRA offers a suite of tests to ensure that all glassware is supplied safe and fit for purpose.

The annealing process

During manufacture, the molten glass is formed into the desired form, after which the item undergoes a process called ‘annealing’ to relieve any internal stresses that may have been created. Glass items that have not been sufficiently annealed may contain significant levels of stress, making them prone to cracking or shattering. As a brittle material, glass products require careful handling during manufacture and afterwards in transport and use, to prevent such failures from occurring.

These failures can be due to sudden temperature changes, mechanical impacts or surface damage such as scratches. For items that have been very poorly annealed, spontaneous cracking may occur: some test items have been observed to shatter while just standing in our laboratory.

Internal stresses are caused by the rapid cooling of an item, as different thicknesses of a glass product cool at different rates – more quickly on the surface and more slowly within its structure. The last area to cool will often retain the most stress. To relieve this stress, glass items are heated to their ‘annealing point’ or ‘stress relief point’, which is a temperature sufficiently high to relieve the stress (but not enough to deform the product) before they are gradually cooled under controlled conditions.

Generally, the slower the cooling process, the better the quality of the annealing. For annealing point temperatures close to 1,500ºC, this process may take 24 hours.

Assessment of the state of anneal

SATRA uses a polarimeter when assessing the degree of stress within annealed glass products. The method is non-destructive and provides two modes of analysis. In the first mode, a ‘qualitative’ method, polarised white light is shone through the test sample, which is examined through a viewer to detect the presence and location of stresses and any flaws or imperfections which may cause premature failure. Stress in the test sample is visible as patterns of coloured fringes. An example of this mode of analysis is shown in figure 1, where a glass vessel is being viewed through its base. Any flaws, such as bubbles, stones (vitreous lumps from incompletely-dissolved batch materials) or foreign materials can be detected in this mode, and the surrounding area is examined for any flaws or unusual stress patterns. Even small flaws can be the cause of possible failure if they possess different thermal expansion properties and expand at different rates to the surrounding glass. The majority of glass failures propagate from sites of existing damage or flaws.

Figure 1: A polarimeter in use

Finding the Real Temper Number
The behaviour of polarised light passing through glass is dependent upon the stress present. By studying the behaviour with a polarimeter, it is possible to obtain a measure of the stress.

Essentially, a polarimeter is made up of two elements: a polariser and an analyser, between which the item to be examined is placed. The measuring process involves rotating the analyser until specified conditions are achieved. The ‘Apparent Temper Number’ – a value on a scale ranging from 1 to 10 – is then obtained from the angle of rotation of the analyser using a conversion table.

The final step is to calculate the ‘Real Temper Number’ by dividing the item’s Apparent Temper Number by its thickness, and then multiplying by 4.06 if the thickness was measured in millimetres or 0.160 if it was measured in inches. Thus, a thin glass item will have a higher Real Temper Number than a thicker one that has been given the same Apparent Temper Number. It is worth noting that while Apparent Temper Numbers lie on a scale with a maximum of 10, there is no theoretical upper limit on the Real Temper Number.

The second mode is a ‘quantitative’ method, where polarised monochromatic (sodium yellow) light is passed through the test sample which is then examined through the viewer. In this mode, the magnitude of the stress can be measured and expressed in units of ‘Real Temper Number’. Real Temper Number is the measure of the degree of stress within an item and is dependent on the magnitude and distribution of the stress, the thickness of the glass and the composition of the glass. SATRA recommends that all glassware has a maximum Real Temper Number of 2.

Thermal shock resistance

Thermal shock resistance is the ability of an item to withstand a rapid and significant temperature change. This is assessed at SATRA using method B of BS EN 1183:1997, where an item is heated and then plunged into a tank of cold water. If the test sample survives the initial test without failure, then the procedure is repeated with the oven temperature increased to produce a higher temperature differential. Further cycles increase the temperature difference each time until failure occurs – for example, the glass cracking or shattering. Ten test samples are assessed and the results expressed as the highest temperature at which 50 per cent of the items fail. In addition to the temperature of failure, the spread of results is of interest, and a wide spread of failure temperatures would indicate high variability in product batches. SATRA advises that all items of glassware are tested for thermal shock resistance and higher results should be obtained for items that will be significantly heated during normal use. For glass ovenware items, SATRA recommends that these are able to withstand a temperature difference of at least 150ºC and that cups, mugs, teapots and coffee pots can withstand at least 90ºC.

There is normally significant correlation between the observations made in the annealing assessment test and the results of the thermal shock resistance test. For example, a glass vessel that shows strong stress patterns in its thicker base section will often show cracking propagating from these sites in the thermal shock resistance test (figure 2).

Figure 2: Thermal shock resistance testing on a glass vase

Testing services

SATRA can test many types of transparent or translucent glassware, such as wine glasses, oven dishes, vases and candleholders. This is done in accordance with SATRA’s state of anneal test method and polariscopic assessment and BS EN 1183 Method B: ‘Thermal shock endurance’. The examination of the stress in glass when used in conjunction with thermal shock resistance tests can give a clear picture of the product’s projected performance. SATRA recommends the use of these tests not only to examine the performance of products at the early stages of manufacture or selection, but also as part of an ongoing due diligence programme.

How can we help?


Using SATRA’s comprehensive testing facilities, you can be confident that your product is fit for purpose and safe before it reaches the marketplace. Please email with all your glass testing enquiries.