Glass Evidence: Fracture Patterns and Refractive Index

The Indian market is dominated by soda-lime glass (about 90% of architectural and bottle glass), with smaller volumes of borosilicate (laboratory, ovenware), tempered (vehicle side and rear windows, public buildings), laminated (windscreens, premium architecture) and wired (older institutional buildings, fire-rated doors). Each type fractures differently, weighs differently in density, and carries a slightly different refractive index range. Misclassifying a fragment at the bench delays the case by routing the sample to the wrong instrument.

• Soda-lime. Standard architectural and container glass. RI around 1.515 to 1.520. Fractures into sharp, irregular shards with clear radial and concentric crack systems. The default case material in Indian SFSLs.
• Borosilicate. Higher boron content, lower thermal expansion. RI around 1.470 to 1.480. Used in lab glassware and high-end cookware; rare in routine casework, common in chemistry-related cases (drug labs, fire scenes).
• Tempered. Soda-lime that has been thermally toughened so it shatters into small cuboidal "dice" when broken. Used in car side and rear windows under IS 2553 Part 2. The dicing pattern destroys radial/concentric information; direction of force has to be inferred from the location and shape of the impact crater.
• Laminated. Two soda-lime layers bonded with a polyvinyl butyral (PVB) interlayer. Windscreens and high-security glazing. Cracks spread but the pane holds together; the interlayer keeps fragments in place and preserves the radial/concentric pattern beautifully.
• Wired. Glass with embedded wire mesh, used in fire-rated doors and older institutional windows. Cracks but doesn't release fragments; useful for direction-of-force work because the mesh holds the crack geometry intact for weeks.

When a force strikes one face of a pane, the opposite face goes into tension first. Glass is strong in compression and weak in tension, so cracks open on the tension face. These radial fractures run outward from the impact point. As the radial cracks open, the impact face goes into bending and the loops between radial cracks fail in their own tension regime, producing the concentric (rib-mark-bearing) fractures. Radial cracks form first; concentric cracks form second.

The fracture surface itself carries a record of this sequence. Fine secondary cracks called Hackle lines fan along the fracture face, perpendicular to the moving crack front. On a radial crack, the Hackle lines fan toward the impact face (because the crack started on the tension/reverse face and travelled toward the compression/impact face). On a concentric crack, the Hackle lines fan the other way. Reading Hackle direction under a stereomicroscope at 10 to 30x is how the lab confirms which crack belongs to which generation when the impact site itself is missing.

High-velocity impact (bullet, projectile above about 100 m/s) produces a small clean crater with extensive radial cracks and limited concentric development. Low-velocity impact (hammer, fist, thrown stone) produces a wider crater with well-developed concentric cracks and lower radial-density. The ratio of radial to concentric crack length is one of the parameters Indian SFSL ballistic examiners record when they're asked to distinguish a thrown brick from a fired projectile in a property-damage case.

The 3R rule states that Radial fractures show Right-angle Rib marks on the Reverse side of the force. The "Rib marks" are the visible curved striations on the fracture surface that, on a radial crack, are perpendicular (right-angled) to the face opposite the impact. The mnemonic reduces a fracture-mechanics relationship to a form that can be applied reliably at the bench or in the field.

The application is procedural. Lay the recovered fragment under a stereomicroscope so the fracture surface is visible. Identify a radial crack (a crack running outward from the impact site; usually longer and straighter than concentric loops). Look at the Rib marks on the fracture surface. The side of the pane on which those Rib marks sit at right angles is the reverse side, which is the side opposite the force. Flip the conclusion: the force came from the other side.

1. Identify the impact origin
   Find the point where radial cracks converge. If the impact crater is missing, work backward from the longest radial cracks; they converge on the origin geometrically.
2. Pick a clear radial crack
   Choose a radial fracture (not a concentric one). Mount the fragment so both glass faces are visible at the fracture surface under 10 to 30x magnification.
3. Read the Rib marks
   Look for curved striations on the fracture surface. On a radial crack, the Rib marks meet one face at right angles and curve smoothly into the other face.
4. Apply the 3R rule
   The face on which the Rib marks form right angles is the Reverse side (opposite the force). The other face is the impact side. Direction of force is from the impact side toward the Reverse side.
5. Cross-check on a concentric crack
   On a concentric crack the geometry inverts: Rib marks form right angles on the impact side. Reading a concentric crack as if it were radial is the single most common mistake.
6. Record under SFSL format
   The standard SFSL report states 'direction of force inferred from radial fracture mechanics, 3R rule applied, confirmed on N independent cracks.' Indian appellate courts have accepted this language when N is two or more.

When a pane has been struck more than once, the sequence of impacts is recoverable. Radial cracks from a later impact stop at radial cracks from an earlier impact, because the earlier crack has already released the local tension field. The terminating crack is the later one; the through-running crack is the earlier one. This is the rule used in a 2024 Delhi SFSL case where a shop-front pane had been struck twice and the defence claimed the second strike had been by a different person. Sequence analysis showed the two strikes were within a few seconds (the radial cracks of the second strike terminated cleanly at the first crack network, with no evidence of intervening relaxation), which contradicted the defence chronology.

A glass fragment's refractive index is measured by immersing it in an oil of variable RI and heating the oil until the fragment's edge becomes invisible (matched RI). Modern Indian SFSLs use GRIM3 instruments (third-generation Glass Refractive Index Measurement), which combine a hot-stage microscope, monochromatic sodium D-line illumination at 589.3 nm, and digital edge-detection software to find the match temperature automatically. Reported precision is ±0.00002 RI units on a single fragment, which is fine enough to separate two windowpanes from different production batches.

The traditional workflow used a Becke line method (visual observation of the bright line at the fragment edge as the oil RI shifted). The modern workflow uses video-thermal-analysis with sub-pixel edge detection. The underlying physics is identical; the precision is about a hundred times better, and the analyst-to-analyst variance has dropped enough that the technique survives cross-examination in a way it didn't in the 1990s.

Typical Indian casework RI values: soda-lime architectural glass 1.515 to 1.520; tempered automotive 1.516 to 1.519; borosilicate 1.470 to 1.480; lead crystal 1.545 to 1.560. The overlap between architectural and tempered automotive is real, which is why RI alone never resolves a hit-and-run case. RI is the screening step; elemental analysis and physical fit do the discrimination.

Density measurement complements RI. The float technique uses a bromoform/bromobenzene mixture in which the analyst adjusts the ratio until the fragment is neutrally buoyant; the density at neutral buoyancy is reported to ±0.0001 g/cm³. The density gradient column variant pours a graduated bromoform/bromobenzene column once, drops the fragment in, and reads the equilibrium depth against calibration beads. Soda-lime glass sits around 2.49 to 2.51 g/cm³; tempered automotive at 2.49 to 2.52; borosilicate at 2.23 to 2.28. The bromoform/bromobenzene system is being phased out at several Indian SFSLs over health concerns; lithium polytungstate solutions are the slow replacement.

Elemental analysis is where the modern discrimination lives. Four instruments do most of the work:

• XRF (X-ray fluorescence). Non-destructive, surface-sensitive. Picks up the major and minor elements (Si, Ca, Na, Mg, Al, K) and several trace elements (Fe, Ti, Ba, Sr, Zr). Good for class-level discrimination; routine in Indian SFSLs since around 2015. Sample sizes from 0.5 mm fragments upward.
• ICP-MS (inductively coupled plasma mass spectrometry). Destructive, solution-based. Detects trace and ultra-trace elements (La, Ce, Hf, U) at parts-per-billion. The best single instrument for separating two pieces of glass from different manufacturing batches. CFSL Hyderabad runs ICP-MS as the discrimination step when RI alone returns a class match.
• LIBS (laser-induced breakdown spectroscopy). Micro-destructive, near-instant. A focused laser pulse vaporises a microgram of glass; the emission spectrum gives elemental composition. Faster than ICP-MS, cheaper to run, lower precision on trace elements. Increasingly deployed at state SFSLs as a screening tool before the sample goes to ICP-MS.
• SEM-EDX. Electron microscope plus energy-dispersive X-ray spectroscopy. Combines morphology (the SEM image) with bulk elemental analysis. Useful when the fragment is too small for XRF and the analyst wants to see surface contamination (paint smear, biological residue) before deciding what else to run.

A representative Indian workflow for a vehicular case: RI by GRIM3, density by gradient column, XRF for class-level major elements, ICP-MS for trace discrimination, and an SEM-EDX check on any surface anomaly. Total turnaround at a well-resourced SFSL is 7 to 14 working days; at a state SFSL without ICP-MS, 21 to 45 working days, often longer when the lab is queuing behind narcotics samples (which take statutory priority under NDPS).

Glass fragments collected at the scene go into folded paper packets (druggist folds), then into a labelled paper envelope, sealed with the IO's signature and the SOCO's signature across the seal. The packet then goes into a rigid container (small cardboard box, not a plastic tube) for transport. Glass-on-glass packaging is forbidden because the recovered fragments will abrade against the container walls and lose the very fracture-edge features the lab needs for physical fit work. Plastic packaging is discouraged for the same reason at a smaller scale, and because static charge can pull micro-fragments out of the packet during handling.

Recovery from clothing follows a tape-lift protocol over a clean paper backing. Recovery from hair uses careful combing onto a paper sheet. Recovery from the body (cuts, embedded fragments) is done at autopsy under the medical examiner's supervision; the embedded fragments are described in the post-mortem report and forwarded to the SFSL with a separate seal.

1. Scene survey
   Photograph the broken pane in situ before any recovery, with scale and orientation. Note which face was inside/outside, which face shows the impact crater, and the position of recovered fragments relative to the source.
2. Source-pane sampling
   Recover at least 10 fragments from the intact source pane, including pieces from the immediate margin of the impact (which carry the richest fracture-edge information). Label as control.
3. Target sampling
   Tape-lift clothing onto clean paper, comb hair onto paper, recover loose fragments from vehicle interior with forceps. Each location gets its own packet and seal.
4. Roadway sampling
   For hit-and-run, photograph the spread of fragments on the road surface before recovery. The spread pattern indicates direction of vehicle travel. Recover from at least three locations across the spread.
5. Transport and chain
   Each packet enters the chain logged under the Section 105 BNSS videography and the seizure memo. See [Chain of Custody](/topics/crime-scene-management/chain-of-custody) for the seal-impression and signature rules that survive cross-examination.
6. FSL intake routing
   Glass goes to the Physics/Trace division at the SFSL, not to Biology or Chemistry. Misrouting at intake adds 5 to 10 days to turnaround. The intake clerk's correct classification depends on the SOCO's label.

A representative SFSL case from a 2024 vehicular fatality in Pune: a victim died on an unlit road; a damaged windscreen-fragment cluster was recovered 18 metres from the body, alongside a paint smear (handled in the companion Paint Evidence workflow). The SFSL matched RI (1.518), density (2.501 g/cm³), and trace ICP-MS profile (Hf, La, Ce within 6%) between scene fragments and a suspect vehicle's broken windscreen recovered three days later. A physical fit was achieved on a single 7 mm fragment whose curved edge slotted onto the source pane's residual margin. The combined class and physical-fit evidence carried the conviction.