Glass Fragmentation and Direction-of-Force Analysis

When a force is applied to a glass surface, the glass deforms elastically: the atoms in the silicate network are pushed closer together (compression) or pulled further apart (tension) without permanently rearranging. Because glass is brittle and lacks the plastic deformation mechanisms available to metals, it cannot permanently accommodate large strains. When the tensile stress at any point exceeds the fracture toughness of the glass (typically around 0.7-1.0 MPa m0.5 for soda-lime glass), crack propagation begins.

Cracks propagate through glass by sequential breaking of Si-O bonds at the crack tip. The Griffith criterion (first published for glass in 1921) states that a crack will propagate when the elastic strain energy released by crack extension exceeds the surface energy required to create the two new fracture surfaces. In practice, this means cracks travel fastest through regions of high tensile stress and are arrested or deflected by regions of compressive stress. In thermally tempered glass, the compressive surface layer arrests a surface crack before it can propagate through to the tensile core, which is why tempered glass is so much stronger than annealed glass in surface-loading conditions.

The velocity of a crack propagating through glass is bounded above by the Rayleigh surface wave speed, which for soda-lime glass is approximately 3,200-3,400 m/s. In practice, cracks do not approach that theoretical upper bound; they reach terminal velocities of approximately 500-1,500 m/s, depending on the stress intensity at the tip, and undergo a velocity-limited regime that saturates at roughly half the Rayleigh wave speed before dynamic instabilities set in. As the crack approaches terminal velocity, it becomes unstable: the crack face begins to roughen, bifurcating into multiple paths (the "mist" and "hackle" regions visible in the fracture surface under magnification). This velocity-dependent roughening is the basis for several features the forensic analyst uses.

The stress state in glass ahead of a crack tip is three-dimensional. For a through-crack propagating in the plane of a plate, the crack front experiences Mode I (opening, tensile, perpendicular to the crack plane), Mode II (sliding, in-plane shear), and Mode III (tearing, out-of-plane shear) loading simultaneously. Most glass fractures are dominated by Mode I, but the Mode III component produces the twist hackle and rib features that carry directional information.

When a localised force (a stone, a fist, a bullet, a hammer blow) strikes a glass pane, the initial fracture pattern consists of two crack families. Radial cracks originate at or near the impact point and extend outward toward the edge of the pane in approximately straight radial lines, like spokes on a wheel. Concentric cracks form approximately circular or elliptical arcs centred on the impact point, intersecting the radial cracks. The formation sequence matters: radial cracks form first (while the glass surface at the impact point is under tensile stress as it bends under the load), and concentric cracks form second (as the glass rebounds or as secondary stress waves interact with the already-established radial cracks).

The sequence rule is important because it governs the fracture termination behaviour at crack intersections. A crack that is propagating will stop when it encounters a free surface, including the surface created by a pre-existing crack. Therefore, concentric cracks will terminate when they meet a radial crack (the radial crack was there first and created a free surface), but radial cracks do not terminate at concentric cracks. At any T-junction in a glass fracture pattern, the stem of the T represents the later crack and the crossbar represents the earlier crack.

This intersection rule forms the basis of sequence-of-impact analysis when multiple impacts are present in the same pane. If a second impact creates a new radial crack system, those radial cracks will terminate when they encounter the radial or concentric cracks from the first impact. A systematic survey of T-junction orientations across the fracture pattern allows the analyst to reconstruct which impact occurred first, second, and subsequent. This is used in scenes where the order of firearm discharges or the sequence of blows is disputed.

The number and orientation of radial cracks provides information about the applied force direction and magnitude. For perpendicular impact (force vector exactly normal to the glass surface), the radial cracks are distributed approximately uniformly around the impact point. For oblique impact, the radial crack distribution is asymmetric: more radial cracks form on the side toward which the force vector is inclined, and the impact point (or "origin" of fracture) may be offset from the geometric centre of the radial system. In practice, the impact point is identified as the origin of the radial crack system, where the highest density of radial cracks converges.

The 3R rule is the most widely cited interpretation principle in glass fracture analysis: it is mechanistically founded, experimentally verified, and practically applicable with a hand lens at a crime scene. The statement of the rule is: "Radial cracks form at Right angles to the Reverse side of the glass on which the force was applied."

To apply the rule, the analyst must identify the two faces of a glass fragment and examine the fracture surface (the edge of the fragment, not either face) under a hand lens or stereo microscope. On this fracture surface, looking at the cross-section of a radial crack through the glass thickness, the fracture surface curves from one orientation to another: it is roughly perpendicular to the face that was under tension during crack propagation. The 3R rule states that this perpendicular-to-face configuration corresponds to the reverse side, meaning the tension was on the face away from the impact.

The physical basis is straightforward. When a sharp-tipped impact bends a glass pane, the near (impact) surface is pushed into compression at the point of loading, while the far surface bends into tension. Glass initiates fractures in tension, not compression. The radial cracks therefore originate on the far (rear) surface of the glass, under tension, and propagate toward the near (front) surface. Since the crack starts perpendicular to the principal tension at the far surface, the fracture surface of a radial crack will be roughly perpendicular to the rear face of the glass near the rear surface.

The practical application requires care. The rib features (discussed below as Wallner lines) or contamination can make face identification difficult on small fragments. The relationship is directional: the right-angle condition applies at the face opposite the force, not at the impact face. A correct reading requires the analyst to (1) identify both glass faces, (2) examine the fracture surface, (3) determine which face the fracture surface runs roughly perpendicular to, and (4) conclude that face was on the opposite side from the impact. Standard forensic-science reference texts used by CFSL examiners in India and FBI Laboratory examiners in the US include annotated practice images for this observation.

The 3R rule applies to radial cracks. Concentric cracks have the opposite relationship: the fracture surface of a concentric crack is roughly perpendicular to the impact face (near surface), because concentric cracks propagate in a stress state in which the near face is under tension as the glass springs back from the initial impact deformation.

This complementary relationship (radial cracks right-angle to the far face, concentric cracks right-angle to the near face) allows an analyst to determine both the direction of force AND which face was struck, using only small fragments that preserve the fracture surface. This is significant in cases where the original pane is no longer available and all that remains are fragments recovered from a suspect's clothing or tools.

The fracture surface of a glass crack is not featureless. Under magnification, it shows a sequence of surface textures that record the history of crack propagation: the smooth mirror region near the crack origin, the misty zone as the crack accelerates, and the hackle region where the crack bifurcates at high velocity. Superimposed on these velocity-dependent zones are Wallner lines: curved, concentric rib features that are produced by the interaction of the propagating crack front with elastic stress waves reflected from the edges of the glass plate.

The physical mechanism of Wallner-line formation is as follows. As the crack propagates, it radiates elastic waves into the glass ahead of it. These waves are also reflected from boundaries (edges, pre-existing cracks, inclusions). When a reflected wave reaches the crack front, it momentarily deflects the crack path slightly out of the dominant fracture plane. This deflection is frozen into the fracture surface as a rib or step, producing a visible mark that corresponds to the position of the crack front at the moment the wave arrived. Because the Wallner lines are records of the crack front position at successive instants, their curvature maps the shape of the propagating crack front, and their origin point (where the ribs converge to a focal point) locates the crack initiation site.

The Wallner-line method allows an experienced analyst to identify the crack initiation point on a fracture surface, even when the geometry of the surrounding fracture pattern is incomplete. This is valuable when only isolated fragments are available. The FBI Laboratory glass comparison protocol, published by Koons, Buscaglia, Bottrell, and Miller in the Forensic Science Handbook, 2nd edition (2002), includes Wallner-line examination as part of the standard glass fracture characterisation procedure.

Conchoidal hackle marks are ridges on the fracture surface that appear in the mist and hackle zones. They are oriented perpendicular to the crack front direction, providing a direction-of-propagation vector at any point on the fracture surface. A curved conchoidal hackle line indicates that the crack front was curving at that point; a straight hackle line indicates the crack was travelling in a straight line.

Taken together, Wallner lines and hackle marks give the analyst a directional toolkit that complements the 3R rule: while the 3R rule identifies which face was struck from the orientation of the fracture surface, Wallner lines and hackle marks identify where on the fracture surface the crack started and which direction it was travelling.

When a small, hard projectile strikes a glass plate perpendicularly, the initial contact stress generates a Hertzian contact cone. The Hertz contact theory predicts that the maximum principal tensile stresses in the glass around a concentrated contact load form a cone opening downward from the load point on the impact face. This cone-shaped stress distribution produces a cone crack that opens on the far face of the glass, with the apex of the cone at the contact point on the near face.

In forensic practice, cone fractures appear in two main contexts: tool or blunt-object impacts on glass, and high-velocity (bullet) projectile impacts. The geometry differs between these cases because the loading conditions differ. A slow-moving tool or object creates a symmetric Hertzian cone with the apex at the impact point and a roughly circular "shell chip" exit crater on the far face. The direction of force can be inferred from the orientation of the cone: the impact was from the side where the apex is located.

High-velocity bullet penetration produces a more complex fracture system. The bullet's kinetic energy generates both a cone fracture and a radial-concentric fracture system. The entry hole on the impact face is typically smaller than the exit hole on the far face, and the entry side shows a bevelled cone shape with the bevel opening toward the exit direction. This "entrance bevelling" is directly analogous to the bevelled wound margins seen in bone gunshot wounds in forensic pathology: the mechanism is the same (a high-velocity projectile creating a conical stress field that produces the widest fracture opening on the exit side).

The distinction between entry and exit sides in a bullet-penetrated glass pane is one of the most commonly requested glass reconstruction tasks in shooting-scene investigation. In the US, FBI evidence-response team protocols specify documenting the glass hole geometry before any fragment collection. In the UK, Crime Scene Investigation procedures under the Association of Chief Police Officers (now the National Police Chiefs' Council) guidelines specify the same. In India, CFSL examiners documenting firearm-related glass damage follow the Bureau of Police Research and Development (BPRD) crime-scene examination guidelines, which include scaled photography of glass-penetration holes before recovery.

Many forensic glass scenarios involve multiple impacts to the same pane. Drive-by shootings, repeated hammer blows in a burglary, multiple stone throws, or secondary impacts in a traffic accident can all produce overlapping fracture systems. Determining the sequence of impacts is a direct evidentiary contribution: it can establish which shot was fired first in a shootout, whether a window was broken from inside or outside, or which vehicle struck which in a chain-collision.

The fracture termination rule is the primary analytical tool. Because a propagating crack will terminate when it encounters a free surface (the edge of a previously formed crack), any crack from a later impact will be arrested at the cracks from an earlier impact. By mapping the T-junction terminations across the entire fracture pattern, the analyst builds a relative sequence: impact A was before impact B if B's cracks all terminate at A's cracks, and A's cracks run without interruption to the pane edge or to a still-earlier fracture boundary.

In practice, three or four impact sequences can usually be resolved from a well-preserved pane. More than four impacts tend to produce complex superimposed fracture fields where definitive T-junction identification becomes unreliable. The analyst should document the specific T-junction evidence supporting each sequence determination and acknowledge when the fracture pattern is too complex for confident sequential ordering.

Low-velocity versus high-velocity impact produces different radial-concentric patterns that provide information about the nature of the impacting object. A slow hammer blow (velocity less than approximately 5 m/s) produces a relatively small number of widely spaced radial cracks (typically 6-12) with fewer concentric cracks, and may produce a cone fracture or crush zone at the contact point. A high-velocity projectile (handgun round at 300-400 m/s, rifle round at 700-900 m/s) produces many radial cracks (15-30 or more), creates the Hertzian cone fracture described above, and typically produces a circular entry hole with radial cracks extending from its circumference. The fracture density and the presence or absence of a cone fracture at the impact point can therefore help distinguish bullet penetration from blunt-object impact, though this inference is probabilistic rather than definitive.

Forensic standards for this analysis are provided by the SWGMAT guidelines in the US (the Scientific Working Group for Materials Analysis, which preceded OSAC), the ENFSI Glass EWG ENG3-2013 in Europe, and the FSNI and private laboratory standard operating procedures in the UK. The ANZFSS guidelines in Australia and New Zealand align with the ENFSI approach. In India, CFSL standard operating procedures for physical evidence examination include glass fracture documentation as a defined analysis type.

R v. Hoey was a major criminal prosecution in Northern Ireland relating to the Omagh bombing of August 1998, which killed 29 people and injured over 200. Sean Hoey was charged with multiple counts of murder. Among the forensic evidence presented was glass-fragment evidence. The case concluded in December 2007 with acquittal on all charges, and the judgment by Weir J included extensive critical commentary on the quality of the forensic evidence.

The criticism directed at the glass evidence in the Hoey case was not that glass-fragment analysis is inherently unreliable, but that specific aspects of the evidence as presented were methodologically questionable. Particular concerns raised by the defence, and partially accepted by the judge, included: the potential for secondary transfer of glass fragments (a fragment moves from a primary surface to a secondary surface without direct contact with the source object), the lack of adequate controls for background glass-fragment levels in the environment where evidence was stored or processed, and the inadequate statistical framework presented for the significance of the glass matches.

The secondary transfer concern is significant for all glass-fragment casework. When glass breaks at a scene, fragments scatter widely. Some land on nearby surfaces (primary transfer). If a person or object subsequently contacts those fragment-bearing surfaces, fragments may transfer again (secondary transfer). The glass evidence against a suspect may therefore reflect presence near a glass-breaking event at some remove from the event itself. The ENFSI ENG3-2013 guideline requires the reporting analyst to address secondary transfer explicitly in any opinion about association between a glass fragment population and a source pane.

Background glass-fragment levels matter because glass fragments are ubiquitous in everyday environments: glass fragments from vehicle traffic, packaging, and household breakage accumulate on clothing and surfaces at low but non-zero levels. The significance of finding glass fragments on a suspect's clothing depends critically on whether background levels have been measured and controlled. The Bayesian likelihood-ratio framework for glass evidence, developed by researchers including Curran, Buckleton, Triggs, and Walsh (published extensively in Science and Justice from 1996 onwards, including their 1998 paper on transfer probabilities) addresses this formally: the likelihood ratio compares the probability of the observed glass evidence if the suspect had contact with the source pane against the probability of the observed evidence given only background sources.

Following the Hoey case and related UK reviews of glass-evidence practice (including commentary in the Law Commission's 2011 report on Expert Evidence in Criminal Proceedings in England and Wales), glass-fragment evidence in UK courts is now more consistently accompanied by explicit statements of secondary-transfer probability, background glass-level controls, and Bayesian likelihood-ratio calculations where the statistical framework is appropriate.

Glass fracture analysis begins at the scene. Before any fragment is moved, the fracture pattern of any intact or partially intact glass should be photographed in situ following the four-tier photography protocol (overall, mid-range, close-up, comparison) with scale and orientation markers. For a window pane remaining partially in the frame, this means photographing both faces (inside and outside), the fracture patterns on both faces, and close-up views of representative T-junctions and impact zones with a millimetre scale in frame.

Fragment collection must maintain the distinction between fragment populations from different locations. Fragments from the scene floor, fragments from the immediate vicinity of the broken window, fragments from the exterior ground, and fragments from any suspect vehicle or person are collected as separate, sequentially numbered exhibits. Mixing fragment populations destroys the ability to map fragment transfer and assess primary versus secondary contact.

For in-frame pane remnants, the analyst should mark or photograph the positions of key T-junctions and cone fractures before removal, because removal may cause additional fracturing that obscures the original pattern. The use of adhesive tape or transparent film backing applied to the pane surface before removal stabilises the fracture pattern and allows the pane to be removed in one piece, transported, and re-examined in the laboratory.

Laboratory examination of the fracture pattern uses a stereo microscope (typically at 10x-40x magnification) for Wallner-line and hackle-mark identification and for T-junction documentation. The fracture surfaces are examined in oblique illumination to maximise the visibility of surface texture features. Digital photomicrography with calibrated scale bars provides the permanent record.

Jurisdictional standards for glass-scene documentation include: the SWGMAT and OSAC guidelines in the US (OSAC Task Group on Glass Evidence, active from 2015), the ENFSI Glass EWG and Crime Scene Investigation EWG joint guidance in Europe, the New Zealand Institute of Chemistry and Australian and New Zealand Forensic Science Society (ANZFSS) protocols for the Australasian region, and BPRD and CFSL standard operating procedures in India. All frameworks converge on the requirement for systematic photographic documentation, sequential exhibit collection, and chain-of-custody continuity from scene to laboratory.