NDT.net • Apr 2005 • Vol. 10 No.4

In-Situ Inspection of Inclusions in Toughened Glass Panels of High-Rise Buildings

Li Xiang*, Fang Zhong Ping, Reading Ivan, Zhao Liping, Chow Siew Loong
Singapore Institute of Manufacturing Technology, 71 Nanyang Drive, Singapore 638075
phone 65-67933253; fax 65-67916377;

*Corresponding Author Contact:
Email: lixiang@SIMTech.a-star.edu.sg

© SPIE - The International Society of Optical Engineers.
This paper was originally published in the SPIE Proceedings vol. 5852
The 3rd Int' Conf' on Experimental Mechanics at 29.11.- 1.12.2004 in Singapore
Back To Session: Smart Structures and Non-Destructive Testing


Transparent toughened glass panels are widely installed in high-rise buildings. There is a growing need for inspection to detect the presence of detrimental inclusions of Nickel Sulfide. These inclusions can cause toughened glass to shatter, possibly causing property damage or injury. Optical equipment has been developed which can detect the inclusions in-situ. Light is coupled into a glass panel and propagates along the glass by total internal reflection. An inclusion in the glass will cause the light to scatter. Once an inclusion is found, it will be observed at higher magnification and the detailed image will be processed. By the analysis of its key features, the inclusion type can be identified. The coupling medium is made of a transparent, soft and deformable material. The equipment can be attached to a glass panel by vacuum suction. The optical system can scan the whole glass panel with a constant force spring as anti-weight structure. The whole system is fast, convenient and highly effective. A patent has been filed for this apparatus.

Keywords: toughened glass panel, nickel sulfide, total internal reflection, inspection, scattering


Toughened glass panels are widely used in our daily life, e.g, the external surface of multistory buildings, the screens of office cubicals/shower rooms, the surfaces of desks etc. A skyscraper has up to 10000 pieces of such panels, where a typical unit is 3mX2m in size and 300kg in weight. All these glass panels are tempered for strength and safety. Unfortunately, many glass panels are subject to spontaneous shattering. There are frequent reports of shattering in public areas, such as mass transit railway stations, high rise residential buildings, exhibition halls, shopping malls etc. According to our surveys, in one office building, this shattering happened at least once, and sometimes up to four times, per month this year. In the units of residential high-rises in one city, at least one shower room glass screen shatters per week. This is a very well known issue which has existed decades of years.

The breakage is due to the presence of a certain inclusion material nickel sulfide[1]. NiS is a residue in the process of glass. It has two forms, the hexagonal structure - a phase, and rhombohedral structure - b phase [2]. When the temperature is above 390 C, the a-phase is stable; when the temperature is lower, e.g room temperature, then the b phase is a stable form. When glass is melted, NiS appears as the a-form. During the tempering process, it dose not transform due to the high cooling speed. The a-NiS is frozen in the glass [3]. At room temperature, the a-phase will transform to the b-phase. The process is slow and may take months to years. The volume of the b-NiS is about 2.2~4 % higher than a-NiS [4]. Due to the expansion, an internal spot-wise stress is generated and flaws appear around the inclusion. Since the thermal dilatation coefficients of glass and NiS are different, the flaws will eventually lead to the breakage of the glass after an incubation period.

Since the breakage of glass is a potential accident source, some countermeasures are proposed. For example, adhering of two glass panels together, building a metallic net inside glass panels etc. Recently a so-called Heat Soak (HS) process has been proposed as an European Standard, which is a destructive test, eliminating glasses bearing critical NiS, by a two hour heat treatment at 290 C[5].

The HS process has not been applied by all the manufacturers. For those glass panels already mounted in high-rise building, there is still stringent demand to find the NiS inclusion effectively and efficiently. Some methods have been proposed, such as a photographic technique [6], Raman characterization [7], laser scanning [8] etc.

A new method is presented in this paper, which can in-situ inspect for the presence of inclusions by light coupling and scattering, measure the critical parameters and identify NiS according to appearance features.


2.1. Critical parameters of NiS inclusions

Only when the parameters of a NiS inclusion meet the critical conditions could it cause the shattering of a glass panel. These conditions will be the judgment criteria of the inclusion inspection system.

The first parameter is the position of the NiS in the thickness direction [3]. Fig 1 shows the stress distribution in a section of a glass panel. The center area (1/4~3/4 in thickness direction) is the tensile zone, the other areas (0~1/4 of thickness from both sides) are the compression zone. When NiS sits in the compression zone, the expansion of NiS will be suppressed by the surrounding stress and will not initiate breakage. When NiS sit in the tensile zone, the volume increment due to phase transformation and thermal dilation will cause deterioration of the stability of material structure, initiate micro flaws and finally lead to breakage.

Fig. 1. Stress distribution in section of glass panel

The second parameter is the size of a NiS inclusion [9]. Based on mechanical model of the stresses around spherical inclusions with thermal expansion, it has been demonstrated that there exists a critical diameter Dc of the inclusion to cause the spontaneous fracture of glass. The critical diameter depends on the residual stress s0 around the inclusion (the level of tempering at the position of the inclusion inside the glass).

( 1 )

(1) there the stress intensity factor K1c =0.76Mpa*m0.5 is a material constant of the glass, and the hydrostatic pressure P0=615MPa. The theoretical minimum diameter to destroy the glass is 50 µm. In practical the smallest critical inclusion found has a diameter of 60 µm [3].

These two critical parameters are the judgment criteria in our system to define the presence of detrimental NiS.

2.2. In-situ inspection of inclusions in a glass panel

The first step to find the critical NiS is to inspect for all the inclusions inside the glass panel. Besides NiS, there are also other inclusions, mainly bubbles.

The main feature of the inspection method is the illumination technique. There are many sources of contamination on the glass surface, and some glass even has a printed pattern in the appearance, these factors cause serious problems for an inspection system, as inclusions submerged below these disturbances are very difficult to find.

If the glass panel is illuminated from all the four sides, the beam will travel inside the glass and when an inclusion is reached it will cause the light scattering, which is very easy to observe. But such side illumination is impractical for in-situ application as there is no space to place the lighting, and due to the light absorbance of the glass, the effective lighting area is about 0.5m in length (depends on the power of light, thickness of glass etc), which means a large glass panel, say 2mX3m, can not be fully illuminated.

Fig. 2. schematic of principle

We use a special medium to couple the light from the front side of the glass panel, as shown in Fig 2. The light beam is directed into the glass through the media within a designed incident angular range. The refracted light in the glass meets the total internal reflection (TIR) condition, and propagates inside the glass panel, so that it can illuminate a large area. Any inclusions and defects in the illuminated area will emit scattered light so that they can be easily detected. The surface contamination will not be illuminated and will not cause visual noise.

Assume the refractive index of medium n1 is greater than that of the glass panel ng, as shown in Fig 3a. θi is the incident angle, θr is the refractive angle. θc is the TIR critical angle from glass to air, θc1 is critical angle from the medium to the glass. θi' is the incident angle when θr= θc. According to Snell's law of refraction, we have,

θc1= sin-1(ng/n1)

If θi'< θi< θc1, the light will be directed into the glass and propagate inside the glass by way of TIR. In case n1g, the incident angle should be θi θi'=sin-1(1/n1).

When more than one medium is employed, e.g. when applying liquid between the fist medium and glass panel as shown in Fig 3b, we need to analyze layer by layer according to their refractive index following the above 2 steps. The following conditions are given as an example, n1> n2, n2< ng, n1< ng, where n1, n2 and ng are the refractive indexes of medium 1, 2 and glass. The incident angle should be in the range of θc1> θi1> θ1', where θ1'=sin-1(1/n1), θc1=sin-1(n2/n1).

a. single coupling medium, n1>ng
b. multi layer of coupling media
Fig 3. design of incident angle in single or multi layer of coupling media

It's necessary to consider the energy lost caused by reflection at the interfaces of the media. According to Fresnel's formulae, the parallel and perpendicular reflectivities are,

( 2 )

( 3)

When θi is close to θr, Rll and R are near to zero, that means when the two refractive indexes are close, the reflectivity is very small. We select silicon rubber as the coupling medium, whose refractive index is 1.407 at 600nm (Fig 4), while that of tempered glass is 1.502. Silicon rubber is soft, deformable and can be easily molded to various shapes.

Fig. 4. refractive index of silicon rubber

2.3 Measurement of inclusion position and size

Once an inclusion is found, we can measure its position and size with a microscope system, as shown in Fig 5. To measure the position, we focus the objective onto the surface of glass and then focus on the inclusion. If the travel of objective is h, then the depth of inclusion is ngh, which corresponds to the optical path in the glass panel. For the lateral measurement, the diameter of the inclusion is same as that of its virtual image whose size can be measured directly by the microscope.

Fig 5. measurement of position and size of inclusions Fig 6. position measurement with camera

Since the field of view (FOV) of a microscope is quite small, using a microscope to measure all the found inclusions will consume a lot of time. We can use a tilted camera system for preliminary measurement of the inclusion position, as shown in Fig 6. This will filter out the inclusions which are not in the tensile zone. Thereafter we only need to measure the size of inclusions in the central zone with a microscope. Fig 6. position measurement with camera

2.4 Identification of NiS

In the laboratory, nickel sulfide inclusions can be identified by their Raman spectrum (Fig 7) and energy dispersive spectroscopy (EDS) of SEM (Fig 8).

Fig. 7. Raman spectrum of NiS. Fig. 8. EDS of NiS

For the in-situ inspection, we employ a much convenient method with image comparison and optical features analysis. NiS has mostly a spherical form, sometimes is slightly elliptic. NiS is dark, non transparent. The surface of NiS is rough due to the crystallization during the cooling process [3]. Fig 9 shows a SEM image of NiS, which is on the surface of a broken glass piece. Fig 10 shows a microscopic image of NiS, which locates in the center zone of glass panel. This signature analysis can filter out most of the non-NiS inclusions.

Fig. 9. SEM image of NiS Fig. 10. Microscopic image of NiS


We developed a scanning inspection system. The coupling media was molded in a rolling shape, which can scan the glass panel in a much faster speed, as shown in Fig 11. The equipment can be located to glass panel by vacuum suction. A constant force spring was used as anti-weight mechanism for the convenience of scanning operation. The whole system is fast, convenient and highly effective. Fig 12 shows the scattering light from an inclusion inside a patterned glass panel.

Fig. 11. Scanning inspection system Fig. 12. The scattering light caused by inclusion in a patterned glass panel

The optical appearance features are used to help identify NiS. The size and location are used as judgment criteria of inclusions which could cause glass shattering. Inclusions other than NiS were also found, as shown in Fig 13. In practice, we find most of the inclusions are air bubbles, which are smooth and transparent, very easy to distinguish.

Fig 13. Various inclusions in glass panels


The in-situ glass inclusion inspection system has been successfully developed. Several buildings have been checked with this system; glass panels with critical NiS were found and replaced. The equipment has been demonstrated as a fast, convenient and efficient tool. It can also be further developed as a quality control tool in glass manufacture, as an alternative of heat soak process. Patent for this technology and apparatus has been filed under the Patent Cooperation Treaty (PCT).


The authors would like to acknowledge the support of our colleagues Ms Xie Hong and Ms Ng Fern Lan for testing and providing the technical data of Raman spectrum, EDS, refractive index etc.


  1. M.V. Swain, A Fracture Mechanics description of the Microcracking about NiS Inclusions in Glass, J.Non-Crystal. Solids 38&39(1980), pp.451-460
  2. D.W.Bishop, Micro Raman Characterization of Nickel Sulfide Inclusions in Toughened Glass. Materials Research Bulletin 35(2000), pp.1123-1128
  3. A. Kasper, Fundamentals of Spontaneous Breakage Mechanism Caused by Nickel Sulfide, Proceedings Book of Glass Processing Days 2003, pp696-699
  4. M.V. Swain, Nickel Sulphide Inclusions in Glass: An Example of Microcracking Induced by a Volumetric Expanding Phase, J. Mat. Sciende 16 (1981), pp.151-158
  5. A. Kasper, Safety of Heat Soaked Thermally Toughened Glass: How Exactly must the Standard Conditions of the Heat Soak Process be Complied with? Proceedings Book of Glass Processing Days 2003, pp670-672
  6. T.J. Ford, Spontaneous Fracture of Glass due to Nickel Sulfide Inclusions: Risk Management and Development of a non Destructive Testing Systems and Technology, CWCT Services, Bath, UK, 1997, pp.99-105
  7. D.W. Bishop, Raman Spectra of Nickel (II) Sulfide, Material Research Bulletin, Vol.33 (1998), pp1303-1306
  8. Edwards, Detection of Inclusion in Glass, International Application Published Under the Patent Cooperation Treaty (PCT), WO/01 18532 A1, 2001
  9. C.C. Hsiao, Spontaneous Fracture of Tempered Glass, Fracture 1977 Vol.3, pp.985-992
© NDT.net |Top|