From the laboratory’s perspective, soil presents a series of problems in terms of the homogenisation and pre-treatment of the sample, all of which can impact significantly on the final analytical result.

This article reviews some of these issues and how laboratories deal with them, and provides an insight into what consultants or contractors should request from their laboratory, bearing in mind the site requirements or risk assessment data needed.

There are guideline documents and standards detailing different sample preparations (see reference list), but they are not definitive, because there are options depending upon the use for which the data is intended – human health risk assessment or waste disposal, for example – and this interpretation of requirements should be made by the consultant or contractor prior to sending samples to a laboratory.

This rarely happens. In addition, many consultants or contractors are not aware of the variations employed by different laboratories in the preparation methods.

As an example1, to demonstrate the possible impact of some of these different preparations, we can compare the arsenic concentration in a soil, utilising three pre-treatment methods:

1. Sieving and analysing the < 2 mm fraction 66.6 mg/kg 2. Sieving and analysing the < 10 mm fraction 30.6 mg/kg 3. Analysing the whole sample 20.0 mg/kg

This soil consisted of 70% of material > 2 mm and 30% of material > 10 mm. *data provided by Professor Clive Thompson, ALcontrol Laboratories

If we use the Environment Agency Soil Guideline Value (SGV) for arsenic for residential land of 32 mg/kg, then protocol 1 would consign this soil for hazardous disposal, and protocol 3 as non-hazardous. Protocol 2 is borderline, and would need further analysis on additional samples.

This can obviously affect whether a site is determined as contaminated under Part IIA. It can cost thousands in possibly unnecessary remediation or waste disposal work, can blight a site entirely, or a site can potentially remain toxic to future users if the contamination is not removed or treated. The risk of future potential litigation should not be underestimated.

Preparation protocols

The main areas of difference in protocols are listed below:

• The sample is dried between ambient temperatures and up to 105°C • The sample is tested as wet, as received soil • The sample is not mixed • The sample is mixed by manual stirring or mechanical methods • The sample is crushed to anything from <200 microns to <10 mm • The sample is analysed without removing anything • The sample has large stones picked out by hand, and the residue analysed • The sample is sieved and the < 10 mm fraction is analysed • The sample is sieved and the < 2 mm fraction is analysed

A more detailed discussion of each of these areas is now considered.

Removal of stones/analysis of fines fraction

The discussion as to whether soils should be analysed in their entirety or whether it is preferable to analyse just the fines fraction has lasted for more than 20 years. The published standards differ on this issue, with the older standards (e.g. BS 1377) relating to agricultural soils, stating that analysis should be performed on the fines, as the main concern is the potential plant uptake.

Other standards appreciate that some contamination may be associated with the larger particles, e.g. oil soaked gravels, tarry lumps, or contaminated concrete.

Obviously, if the contaminants are concentrated in the fine fraction, then just analysing this will provide much higher results than analysing the whole sample.

Alternatively, if the contamination is associated with the larger lumps, then excluding these will decrease the results overall.

The Environment Agency’s Standing Committee of Analysts (SCA) published a guide to soil sample preparation in 2005 in the Blue Book series of Methods for the Examination of Waters and Associated Materials (MEWAM).

This document is entitled ‘The preparation and pre-treatment of soils, contaminated land, and similar materials’. It states:

“The removal of extraneous constituents (namely, large stones, pieces of plastic or other debris), prior to subsampling and drying, may adversely affect the resulting analysis by significantly reducing the contamination under investigation, and hence, the determinands of interest.”

This document, and the Environment Agency’s Monitoring Certification Scheme (MCERTS) standard for soils, states that any material removed from the sample must be recorded in the final report, but it is acknowledged within the industry that this does not happen in many cases.

However, if the main concern on the site is human health risk assessment, then the consultant or contractor will be more concerned with airborne dust, and therefore it would be more appropriate to analyse the fines fraction. Sieving is usually performed on the dried sample, because it is difficult to force wet soils through a sieve, and wet sieving may lose some of the parameters of interest.

The decision to sieve through a 2 mm or 10 mm sieve is currently at the discretion of the laboratory, or can be stipulated by the client, but many soils may not contain a large proportion of < 2 mm material, and this can cause problems in obtaining sufficient sample to test. More commonly, the large stones or lumps are picked out by hand, and the rest of the sample is then mixed and subsampled for analysis. There is little quality control with this method, as it is very subjective and depends on the experience and consistency of the technician, which may vary significantly.

Material which is undergoing assessment for waste disposal, whether by leaching tests or by analysis of the solid material, should always be analysed as the whole sample, with nothing removed.


The consultant or contractor on site should select and mix the soil to provide as representative a sample as possible from the area under investigation.

Some on site staff take composite samples, whereby a scoop of soil is taken from six or seven close locations (e.g. around a trial pit at one depth horizon), and these are then mixed together, and a sample of this mixture sent to the laboratory. This is sometimes referred to as cluster sampling.

Composite sampling from samples taken at widely spaced intervals should be discouraged, as this does not provide meaningful information with respect to the spread of contamination on site.

It is important that sufficient soil should be provided to the laboratory, and this is usually a 500 g plastic tub for inorganic analyses, a 250 g glass jar for organics, and duplicate 60 g glass jars for volatiles.

If leach testing is required, or any physical tests (e.g. particle size distribution), then a minimum of a second tub will be required.

Some samples such as gravels or railway ballast are usually sent in much larger quantities (up to 10 kg), and the laboratory will often invoice an additional charge for the increased time it takes to mix or homogenise a suitably representative subsample.

In the laboratory, the soil sample is tipped into a tray, and mixed, either by coning and quartering (BS 1377 technique), or by other means more suitable to the sample matrix.

Gravel samples or concrete lumps can be broken down using a jaw crusher, which reduces the particle size to < 10 mm. Clay samples can be mixed with a paddle mixer, or kneaded similar to dough. Samples containing chunks of fibrous material may require cutting or shredding to provide more manageable pieces. When mixed, the sample is usually split into quarters, one of which is used for the wet tests, and one of which is sent for drying and crushing. At ALcontrol, the remaining quarters are returned to their original container for cold storage.

The description above is what should happen, but some laboratories do not mix samples, and simply weigh out directly from the jar or tub, rightly or wrongly assuming that the site staff have already sufficiently mixed the sample.

The exception to the above is with samples for analysis of volatile determinands. These are not subjected to any further mixing or homogenisation, because these processes would cause the loss of volatile components. Therefore a subsample for analysis is weighed out directly from the volatile sampling jar, usually first discarding the top centimetre or two, so the sample is taken from the centre of the jar to minimise any loss of volatile components.

It is important for sample preparation staff to be adequately trained in taking representative samples of all the fractions present within the soil when actually weighing out the fraction for analysis.

To dry or not to dry

There are many determinands which would be lost or compromised by drying or crushing, and these include:

• Volatile organics • EPH and PAHs • Cyanide • Hexavalent chromium • Acid soluble sulphide • Phenols

These samples should be analysed on a wet, as received sample, with no drying or crushing. It should be noted, however, that these samples can rarely be completely homogenised, and therefore any replicates will always demonstrate a higher level of variation than tests performed on a dried and crushed sample.

For other analyses, it is possible to dry the sample prior to crushing, and the drying can be conducted at various temperatures. Most laboratories will use ovens set at 30°C, 35°C or 40°C.

For correcting analyses performed on wet samples, then the recorded weight loss will be used as a moisture content. However, if a true moisture content is required, then the sample should be dried at 105°C.

The MCERTS for soils standard does not actually provide a recommended temperature for oven assisted drying, it simply states:

“The temperature at which the drying is to be undertaken shall be appropriate to the parameter being determined to ensure that the parameter does not undergo degradation or is lost from the sample.

If the sample is to be described as air dried, then the temperature shall not be more than 30°C.”

If samples are dried, then the laboratory must supply validation data to confirm that there is no loss on drying of the parameter of interest.

MCERTS requires that all results are reported on a dry weight basis, so even if tests are performed on a wet, as received soil, then a separate aliquot of soil must be dried to determine the water content, and the original result corrected accordingly.

It is also a requirement of MCERTS that the temperature utilised for drying is included in the final report, and that whether the test is performed on a wet or dry sample is also clearly defined.


The dried subsample can be crushed, and there are a number of systems used for this, ranging from rough crushing by hand, using a pestle and mortar, to heavyweight mill crushers, which can deal with much larger samples, and produce a fine homogenised powder down to a particle size of < 200 microns.

The mill crushers consist of a tungsten carbide dish, with a series of concentric pucks inside. The dried sample is tipped into the dish, the lid clamped down, and the machine switched on.

It rapidly vibrates the dish and the pucks crush the sample to about 200 microns within two minutes. The resulting fine, dry powder is much easier to work with in the laboratory, and will produce much better replicates than tests which are performed on the as received material, but unfortunately it cannot be used for all tests.


It is apparent that the diversity of soil preparation and pre-treatment methods is a source of inconsistency if data is compared between laboratories.

It is important for consultants or contractors to understand how their laboratory is processing samples, and to realise that no one protocol is suitable for all samples on all sites.

The Environmental Industries Commission (EIC) is currently setting up a Sample Preparation subgroup, within the Contaminated Land working group, and has issued a survey with respect to how much knowledge consultants or contractors currently have with respect to the practices employed by their laboratories. The data from this survey will be evaluated and made available through the EIC.

It is crucial that any consultant or contractor specifies the sample preparation protocol when submitting samples to a laboratory, to avoid potential expensive repercussions, and this should be discussed at the contract review stage of a tender or quotation.


BS 10175:2011 The Investigation of Potentially Contaminated Sites – Code of Practice (this revision has only just been published) BS 7755-3: 1995, ISO 1464:1994 Soil Quality – chemical methods. Pre-treatment of samples for physico-chemical analyses BS 1377: Parts1 – 9: Testing of soils for engineering purposes BS ISO 14570:2003 Soil Quality – Pre-treatment of samples for the determination of organic contaminants Environment Agency MCERTS Performance Standard for Laboratories undertaking Chemical Testing of Soil: 2003


Hazel Davidson, Technical Marketing Manager, ALcontrol Laboratories

Hazel Davidson has worked for ALcontrol Laboratories for thirty years, initially as an analyst, but then in a series of managerial roles. Special projects included the integration of several laboratory acquisitions, relocation of the laboratories from Chester to Hawarden, a Phare project in Bulgaria and Romania (implementing quality systems), and a UN project involving training for Iraqi environmental scientists in Jordan.

Hazel participates on several industry committees (BSi, MCERTS, SCA and EIC), is a frequent speaker at conferences and runs several seminars each year for ALcontrol clients, as well as providing general technical support, both internally and externally.

About ALcontrol

ALcontrol provides testing and analytical services for soil, water, food, oil, asbestos and air to help clients demonstrate compliance with regulations and achieve their health, safety and environmental goals.

Providing millions of tests per year, with more than 2,000 employees in 30 laboratories across 11 European countries supporting a global customer base, ALcontrol is Europe’s largest independent provider of environmental analytical services.

Speed, accuracy, reliability and efficiency are all key to the successful delivery of analytical services, so ALcontrol provides all of its customers with live access to their laboratory data through the web-based ‘@mis’ scheduling and reporting service.

Further information on ALcontrol’s full range of testing and analytical services is available at

Published: 10th Jun 2011 in AWE International