Density-Gradient Separation and Heavy-Mineral Analysis

The principle is straightforward. A soil or sediment sample, pre-treated to remove organic matter and sieved to remove the clay and coarse fractions (usually the 63-500 micrometre fraction is optimal), is added to a separation funnel containing a dense liquid. With a density of 2.85-2.90 g/cm3, the liquid is denser than quartz (2.65 g/cm3), feldspar (2.55-2.76 g/cm3), and most carbonates, but less dense than the heavy minerals of interest. The sample is stirred gently and allowed to settle. Light minerals float; heavy minerals sink.

The two fractions are separated by opening the stopcock on the separation funnel, draining the heavy sink fraction onto a filter, washing away the liquid with acetone (for bromoform) or distilled water (for SPT), and mounting the mineral grains on a glass slide for microscopy. The yield depends on the original heavy-mineral content, which varies widely. Some clean quartz sands from highly weathered tropical soils yield less than 0.01% heavy minerals by mass. Lithic-rich sands from young glacial or volcanic deposits can yield 5% or more.

For particles finer than about 63 micrometres, gravity settling in a funnel is too slow and inefficient. Centrifugal separation spins the sample in a dense liquid so that settling is accelerated. The same light-float, heavy-sink logic applies, but the effective g-force separates fine grains within minutes rather than hours.

Column density-gradient separation layers liquids of different densities in a tube, then adds the sample at the top. As grains settle, each stops at the level whose density matches its own. The result is a column with bands of minerals separated by density, which can be aspirated in sections to recover discrete density fractions. This is more informative than a single-threshold separation when the sample contains minerals across a wide density range, and it has been used in forensic work to distinguish mineral suites that share a similar bulk density but differ in detail.

Different parent rocks produce different heavy-mineral assemblages. Granites are rich in zircon, monazite, and biotite. Mafic and ultramafic rocks (basalt, gabbro, peridotite) produce chromite, pyroxene, olivine, and chrome spinel. Metamorphic schists and gneisses are sources of garnet, staurolite, kyanite, and amphiboles. Sedimentary recycling progressively eliminates unstable minerals and concentrates the ultra-stable species (zircon, tourmaline, rutile), so a recycled aeolian sand has a very different suite from a fresh glacial sand derived from the same parent rock.

In forensic casework, the analyst counts several hundred grains per sample and records the relative proportions of each species. Two samples from the same sediment source should produce similar proportions within counting statistics. Two samples from different geological formations should differ in their suites, particularly in the relative proportions of garnet versus amphibole, or the presence or absence of diagnostic species like staurolite or chrome spinel.

Provenance indices reduce a full mineral count to a single number that summarises one aspect of the sediment's history. The ATi and ZTR indices, developed in sedimentology by Andrew Morton and colleagues at the British Geological Survey, have been applied in forensic contexts because they encode weathering and recycling history that can help distinguish between candidate source areas.

A soil with ATi near zero has lost nearly all its apatite: it has either been subjected to prolonged acid weathering in a forest soil, or it has been recycled through a sedimentary rock before being re-deposited. A soil with ATi near 100 has retained abundant apatite and is therefore young, fresh, and probably derived directly from an igneous or metamorphic rock nearby. Similarly, a ZTR approaching 100% means that everything less stable than zircon, tourmaline, and rutile has been dissolved away, which is the signature of a deeply weathered tropical laterite or a many-times-recycled aeolian deposit.

Raymond Murray, an American geologist at the University of Montana, was among the earliest to systematically document the forensic application of soil mineralogy in casework, publishing with colleagues from the 1970s onward. His 2004 book Evidence from the Earth became a standard reference. Murray's cases included distinguishing mining-district soils by their unique accessory mineral assemblages, and linking soil from a suspect's vehicle to a remote rural scene via unusual mineral combinations not found in the intervening highway soils.

Kenneth Pye and colleagues at Royal Holloway University of London and later Kenneth Pye Associates developed a rigorous laboratory protocol combining multiple analytical methods, with heavy-mineral analysis as one layer of a multi-property comparison. Their work on coastal and fluvial sand in UK cases demonstrated that heavy-mineral suite differences between beaches a few kilometres apart could be resolved with sufficient grain counting.

Lloyd Croft, working in New Zealand, applied heavy-mineral analysis to cases involving volcanic soils. Volcanic soils carry distinctive suites of pyroxene, olivine-group minerals, and volcanic glass that are absent in non-volcanic parent materials. A questioned sample carrying volcanic heavy minerals in an area dominated by sedimentary rocks is geologically diagnostic, because no common depositional process would produce that assemblage at the questioned site.

A heavy-mineral count on its own has limited meaning. It becomes forensically significant when it is compared against a reference collection that maps the spatial variation in mineral suites across the relevant area. In practice, this means the examiner needs to know whether the particular combination of garnet, hornblende, and zircon found in a questioned sample is common across the region or restricted to one geological unit.

Published geological surveys (British Geological Survey, United States Geological Survey, Geological Survey of India, and national equivalents worldwide) map bedrock and superficial deposits and often include descriptions of mineralogy in the accompanying memoirs. These provide the baseline. Where the forensic question demands greater geographic precision, the examiner must collect additional reference samples and build a site-specific comparison set. This is time-consuming but necessary when the case hinges on distinguishing sources within a small geographic area.

One limiting factor is that many transported soils have mixed provenance. A river terrace deposit may contain heavy minerals from several source rocks because the river crosses different geological units upstream. A mixed assemblage is harder to link to a specific site than a simple suite. Recognising mixing, and whether the questioned sample has a similarly mixed signature, is part of the interpretive skill.