Spontaneous Fragmentation of Tempered Glass
One of the effects of the fracture characteristics of tempered glass is the occurrence of what are known as ‘spontaneous fractures’. These are fractures of the tempered glass due to no immediately obvious reason.
The causes of fracture of tempered glass, listed in order of likelihood, are as follows:
- Impact damage (deliberate or accidental)
- Damage before installation (during handling)
- Poor installation (tight glazing, missing setting blocks)
- Poor design (insufficient clearances, structural movement)
- Inclusions in the glass (e.g. nickel sulphide)
This is the most common cause of breakage of tempered glass, often from a deliberate impact (e.g. throwing stones). It is; however, difficult to identify, because the ‘star’ pattern seen in fractured annealed glass is overwhelmed by the innate fracture pattern of the tempered glass, which also often falls from the frame. The impacting object may be small (a pebble or air gun pellet) and may not end up in the vicinity of the broken glass fragments.
Edge damage often occurs when the glass is being handled. While the majority of such damage either causes instant failure or is insignificant, some edge damage can lead to delayed fracture.
Tempered glass is manufactured in such a way that the outer surface has a built-in compressive stress while the interior of the glass has a compensating built-in tensile stress. Since glass always breaks from tensile stresses at the surface, tempered glass is stronger because the built-in compressive stress has to be overcome. However, if a crack does occur and penetrates to the inner tensile stress zone, the tempered glass will fracture violently and disintegrate into smaller particles.
It is possible for edge damage, e.g. from impact or edge chips, to penetrate close to the inner tensile zone, but without causing complete fracture. In many cases of such damage the tempered glass will remain in a stable state permanently, but, on occasion the damage is such that it destabilises the built-in stresses around the damaged area to an extent which can allow a static fatigue mechanism to operate and which causes complete fracture to occur at some time after the initial damage was caused. The time of fracture could be from a few seconds up to several months after damage. This type of occurrence is relatively rare; most tempered glass fractures at the time it is damaged or not at all.
Most framing systems are designed so that the glass is adequately supported giving no local high spots or stress concentrations on the glass. However, poor installation, or later attachments to the frame (e.g. blinds), may create local high spots or stress concentrations which may, under repeated wind loads, cause weakness to arise in the tempered glass, eventually leading to fracture.
While poor glazing is the cause of some fractures, bad design may make it impossible to install the glass correctly. Bolted fixings without sufficient tolerance allowance, or holes too small to get bushes in, are prime examples. Some framing systems may have inherent faults, such as protruding screw heads or rivets and even structural movement of lightweight structures like space frames may take up the edge clearance allowed round the glass if not properly assessed and allowed for in the design.
Of the various causes of ‘spontaneous fracture’, only that associated with the presence of foreign particles in the glass is more likely to cause fracture in tempered glass than in other forms of glass, because they can disturb the very high built-in stresses in tempered glass.
While there are several (rare) types of foreign particles which may cause ‘spontaneous fracture’, one type in particular, Nickel Sulphide, is directly associated with the tempering process.
In 1962, Ballantyne of the CSIRO (Building Research, Melbourne) published a report indicating that the cause of many spontaneous breakages was nickel sulphide (NiS). NiS is a complex material, which undergoes a phase change (a change in crystalline structure), at 380oC, which is accompanied by a change in volume. The a-NiS, which is stable above 380oC, has a smaller volume than the b-NiS, which is stable below 380oC.
The tempering process in glass requires the glass to be heated to around 620oC followed by rapid cooling. Any NiS in the glass is converted to the a phase at the higher temperature, but the rapid cooling does not allow time for the conversion back to the b phase. The NiS is thus ‘frozen’ into the tempered glass in an unstable form.
Over a period of time the a phase slowly converts back to the b phase, the conversion being accompanied by an increase in volume of 2%-4%. If a particle (inclusion) of NiS is sufficiently large and is in the central (tensile stress) zone of the tempered glass, then the expansion caused by the conversion can exert sufficient excess stress to cause a crack to propagate, leading to disintegration of the pane.
Source of NiS
NiS is a contaminant in the glass. Sulphur compounds are unavoidable, but the nickel can be reduced by careful control. Sources of nickel contamination have been found in the raw materials, the fuels, and the component parts of the melting tank structures and the steel components of equipment in the melting tanks. Reputable glass manufacturers have taken action to reduce nickel contamination over the 30 years since NiS was shown to be a cause of spontaneous fracture. The occurrence of NiS in glass is now considerably lower (by at least an order of magnitude) than it was in the 1960’s.
For example, When Pilkington first assessed the occurrence of NiS in their UK manufactured glass, in 1965, the occurrence of critical NiS particles was around 1 in 500kg of glass produced. By 1988, the occurrence of critical NiS particles in their UK manufactured float glass was down to less than 1 in 13000kg.
Detection of NiS
NiS inclusions are difficult to detect. The size of the inclusions is below the limits for extraneous particles detectable by quality control systems in glass manufacturing plants. They are relatively infrequent and there is no easy method of finding a NiS inclusion other than at the origin of a fracture, which has been caused by one.
Factors Affecting Breakage’s Due to NiS
Three main factors have a considerable effect on breakage due to NiS inclusions.
- The rate of conversion from the a phase to the b phase is temperature dependent. At higher temperatures the chemical reaction of the conversion will proceed at a faster rate. Thus glass which is subjected to higher service temperatures will show higher initial rates of breakage due to NiS This will affect especially solar control glasses and spandrel panels, which tend to show comparatively higher rates of breakage.
- The volume of the NiS inclusion must be sufficiently large. In practise, NiS inclusions, which have caused breakage, have been measured with sizes between 0.04mm and 0.45mm in diameter. The size of inclusion will also affect the time needed to generate failure stress, since, in a larger inclusion, total conversion is not necessary to develop a critical volume increase.
- The initial central stress in the tempered glass also contributes to the probability of breakage. The higher the built-in stress, the smaller is the size of inclusion required to cause fracture. Theoretical studies indicate that inclusions larger than 0.04mm diameter are required to fracture fully tempered glass.
Heat Soak Testing
The effect (a) above can be used to construct a further process, which has been called heat soak testing. This essentially requires the glass to be heated to a high temperature (but less than 380oC) and left for a period of time long enough to fracture a large proportion of the panes which may otherwise fracture in service.
Saudi American Glass strongly recommends that all Tempered Glass is Heat Soak Tested in order to greatly reduce the risk of Spontaneous Breakages resulting from Nickle Sulphide Inclusions ( NiS ).