Unsupported browser

For a better experience please update your browser to its latest version.

Your browser appears to have cookies disabled. For the best experience of this website, please enable cookies in your browser

We'll assume we have your consent to use cookies, for example so you won't need to log in each time you visit our site.
Learn more

Shattering illusions

  • Comment
High-profile accidents involving glass failure have opened the debate on safety. So how safe is the glass in our buildings?

It is bad enough working out what type of glass to specify at the best of times. But recent uncertainty about the properties of various types of glazing has been heightened with a spate of documents and documentaries identifying incidents of glass failure in high-profile buildings.

The key issue is the use of toughened glass in buildings; more specifically, the use of sheets of toughened glass in high-level sloping overhead locations.

The problem has been defined as one of minute particles of nickel sulphide, called 'inclusions'. These are inherent within the central tension zone of tempered glass and can cause 'spontaneous fracture' within individual panes.

Nickel sulphide particles form within the glass and range in size between 0.076-0.38mm and therefore preclude the use of practical inspection methods to spot them in float glass.

Effectively, at this moment in time, nickel sulphide inclusions are undetectable until failure. Upon failure, the expanded nickel sulphide particles tend to cause a crazing pattern with a 'butterfly' or 'figure of eight' pattern at its centre. Current glass technology cannot eliminate the possibility of nickel sulphide inclusions. Nickel may arise in minute particles from stainless steel equipment used in the transportation of materials or production processes; sulphur may intrude into the batch through particulate transfers from diesel vapours or other oil-based lubricants in the vicinity.

Effectively, given the current level of sterilisation of factory conditions - and assuming that improving the sterile environment will prove inordinately costly for the beneficial returns - nickel sulphide will be present in most glazing products.

A number of high-profile cases of late have raised fears of using toughened glass overhead.

However, when the hype has died down it becomes clear that most have nothing to do with nickel sulphide.

Wilkinson Eyre Architects' Stratford Station in east London caused a flurry of anxiety when a 5 x 2m pane of glass 'spontaneously' fell onto the platform. Less well documented was the statement by the client, London Underground, that this incident was a result of accidental damage and that the contractor has replaced the glass with no legal overtones or cost penalties.

Similarly, cracks in the glazing at London's Imax cinema by Avery Associates Architects have subsequently been identified as having been caused by 'mechanical' damage, while the National Glass Centre in Sunderland by Gollifer Associates had glass fractures attributed to locals firing air guns at the main facade. There are very few cases of spontaneous fracture directly attributable to nickel sulphide inclusions.

The one case cited in the recent BBC TV programme Glass Houses as having this problem is Neathouse Place in London's Victoria. Chris Huntley, for the tenant Broken Hill Proprietry, states categorically that the building has had six panes that 'have fallen out as a result of nickel sulphide inclusions'. He says: 'Because of the angle of the glazing this has caused some concern for passers-by underneath the external incline.'

Funnily enough, the building won the Glassex 'Best Building' award in 1997. With no comment available from the architect, it is only appropriate to say that the jury is out on the actual cause of this specific case.

The Health and Safety Executive does not have details of accidents caused by high-level glazing falls, nor does it identify the likely incidence of 'near misses' (see box). In fact, no one seems to know the actual likelihood of a spontaneous breakage occurring.

Glass Houses focused on the Luker family, whose sons had suffered a minor injury from falling glazing, but there seems to be no other coherent evidence of casualties.

Ann Wright, assistant producer, says that 'no one will come clean' on the issue and so assumes that the incidence of the problem is greater than the evidence would suggest.

However, Dr Alan Woodward of Pilkington's Automotive, a significant market leader in toughened glass technology used for windshields, confirmed that he is not exercised by the problem at all.

Even though windshield glass is thinner and the potential risks from glass falling out are significantly reduced (in terms of impact damage rather than from any resultant accident caused), he believes that the issue of spontaneous breakage is 'based more on supposition than fact'. Within the automotive industry, more so than in construction, 'there is no real way of knowing the cause of glass damage'.

In buildings, he recognises that nickel sulphide contamination of glass 'used to be a real problem 20-30 years ago, but with major improvements in glass composition and manufacturing techniques, things have improved dramatically'. But what are the risks?

A technical representative at Pilkingtons, the major market force in glazing to the building industry, told me that, in relation to accidental occurrences after completion, he had 'never come across a single incident' of anyone being killed, injured or hurt by falling glass.

Robert J. Miller, president of Groves Glass Technologies, US, says:

'Nickel sulphide contamination has not been a problem in the US for at least a decade or so and when we did have a problem it occurred mainly in the heat absorbing glasses.

'There were several high-rise buildings that experienced this type of failure in Arizona. I am not aware of any injuries and certainly no deaths occurred because of this particular phenomenon. Arizona code requires tempered laminated for all overhead installations, for example skylights and slopes, and I agree with that. It is more expensive, but worth the peace of mind.'

Alan Keiller, research engineer at Curtain Wall and Cladding Technology (CWCT), explains: 'Heat soaking glass accelerates the change of state of nickel sulphide particles and, in laboratory conditions, reduces the likelihood of fracture to just 1 per cent of ordinary toughened glass'.

He adds: 'Given that there is estimated to be one critical nickel sulphide particle per four tonnes of glass, for 10m-thick glass one pane in 1,600 toughened glass panes (1m 2each) might spontaneously break;

and one pane in 16,000 heat soaked toughened glass panes might break.'

If there is no telling when this 1:1600 risk will occur, is walking through a mall atrium like playing Russian roulette? Not really, when you factor in that even with nickel sulphide inclusions, panes might not break.

Furthermore, buildings may only be inhabited for eight hours or so, leaving two-thirds of the day for accidental breakage to pose no threat to the public. Breakages may be located in non-threatening areas.

Over an entire building the glazing failure might occur in safe locations in the glazed cladding - it need not occur in the overhead glazing, but could be at the walls. After all, 1,600m 2is the equivalent to a 350m 2four-storey, fully glazed box, of which only one-fifth will be roof.

Pilkington's 'Glass and Safety:

Technical Bulletin' recognises that vertical toughened glass, even if subject to spontaneous failure due to nickel sulphide, will remain in situ and will tend not to fall out of its frame.

BS 5516: 1991, 'Code of Practice for vertical and sloping patent glazing', is the only regulation or standard that gives recommendations of the types of glass to be used in overhead conditions. It states simply that 'the risk of spontaneous fracture of toughened glass may be reduced by heat soaking'. There is no specification requirement for vertical glazing.

However, there are dissenting voices. PPG of Pittsburgh says: 'Heat soaking of fully tempered glass is a significant waste of energy, is not completely effective and, therefore, provides little additional information for estimating the probability of breakage caused by nickel sulphide stone inclusions in fully tempered glass which has been heat soaked.

At the end of the day, the higher costs for heat-soaked or laminated glass may be being charged to reduce an over-inflated problem.

Ultimately, it would seem that the Luker family was just unlucky, unfortunate - and extremely rare.


Float glass Molten glass is fed across a bath of molten tin. Because, during production, the surface of the molten glass cools quicker than the centre, the resultant internal stresses may cause the glass to break. 'Annealing' limits these stresses by submitting the glass to controlled cooling and stabilisation. Breakages in float and annealed float glass produce large, jagged shards.

Toughened glass In order to make it resistant to breakages, a glass sheet is heated to a temperature just below its softening point and then immediately cooled by jets of cold air.

This allows the surface to crystallize while the inside solidifies with greater compression than at the surfaces.

Toughened glass is two or three times stronger than untempered glass. If broken, the glass shatters into tiny pieces with blunt edges, although these may combine into larger clumps.

Heat soaked glass Glass is raised to a controlled temperature (approximately 300degreesC) and maintained at this stable temperature for a set time, between two and eight hours. This agitates any nickel sulphide inclusions to prompt the glass to fail within the oven rather than in situ. It is a method of minimising nickel sulfide hazards in fitted glass.

Laminated glass Laminated glass comprises two (or more) sheets of glass with a viscous plastic layer sandwiched between the glass panes. The glass can be ordinary float glass or variations of different glass types. Under simultaneous heating of the already processed layers of glass and special plastic, lamination occurs. When laminated safety glass breaks, the pieces remain attached to the internal plastic layer and the glass remains transparent.

Breakages produce localized damage, although the panes are held together and tend to remain in the frame.


According to statistics available from the Health & Safety Executive and the Department of Trade and Industry, for nearly all age groups the incidence of accidents involving glass has nearly doubled.

However, a cursory glance at the characteristics of the accidents show that they merge the serious and the trivial, the accidental and the litigious (see table).

There are no statistics to record accidents caused by falling glazing, especially in a work environment (there is probably no opportunity for falling glass in the domestic environment).

While there are very serious injuries, especially related to the loading and handling of glass in situ, these government statistics indicating the rise in 'glass accidents'suggest an increased reporting of accidents, rather than an increased incidence of accidents themselves.


When the Crystal Palace was being built 150 years ago, many people expressed concern about the imposing and challenging nature of the all-glass construction. But the cautionary mood was dismissed as doom-mongering at a time of national aspiration.

The scheme went ahead with impressive speed.

The Royal Commission was set up in January 1850 to set the competition for a 700,000 sq ft building.

The competition was won in June and architect Joseph Paxton completed the full working drawings in nine days.The drawings were published in June 1850.

Because of the fear of the (then unknown) effect that large crowds moving about inside might have on the structure, resonance tests were set up with workmen and then with army sappers marching up and down. The structural movement was shown to be negligible and the construction continued apace until completion on 1 May 1851.

The materials included 550 tons of wrought iron, 3,500 tons of cast iron, 202 miles of sash bars, 30 miles of gutters and 900,000 feet of glass.

As John McKean says in his book Crystal Palace (Phaidon Press, 1994), people worried that the building was too unstable, that hail and lightning might destroy it, that the gun salute to Queen Victoria would shatter the glass, and that 'thousands of grand ladies who were to be in their seats. . .

would be cut to mincemeat'. He adds: 'But essentially it is not the satisfaction of having defied the critics which attracts. It is that - with its tall, slender columns of only 8 inches diameter, and its light openness - it holds to the appearance of fragility, the seemingly unsafe, the frisson of danger. Here again is the thrill of the sublime'.

It would probably never have passed a CDM risk assessment analysis.

  • Comment

Have your say

You must sign in to make a comment

Please remember that the submission of any material is governed by our Terms and Conditions and by submitting material you confirm your agreement to these Terms and Conditions.

Links may be included in your comments but HTML is not permitted.