Methodology and Proximate Analysis

Opening Statement

The preparation for analysis of a small amount of material (the sample), representative of the bulk food, and normally supplied to the analyst, can be divided into the following stages:

1. The first stage for most food matrices is to prepare a weighed and calibrated aliquot – the sample for analysis – in preparation for quantitative extraction of the compounds of interest (analytes).

2. For some food samples, the material has to be rendered accessible to the extracting agent – preparation for analysis.

3. The next stage can then be either (a) removal of the analytes from the sample matrix or (b) removal of interferents from the matrix – the extraction. In each case, the analytes are in a form to be recognised and quantified unambiguously in subsequent examinations.

4. The final stage is to examine the extract, which will normally contain matrix components other than the target analytes, using various of chemical and physical methods to make qualitative or quantitative measurements of the analytes – the analysis.

One contemporary objective in the development of analytical methodology is to automate the whole assay, and there are two ways forward. The classical extraction procedure can be given over to robotic control, or the information about the chemical composition can be "extracted" directly from the sample matrix by remote sensing. Perversely, remote sensing makes extraction redundant. Thus, it is necessary at the outset to recognise that the future analysis may be an extractionless, remote sensing, robotic operation. Although considerable progress has been made already towards these goals, as a "hands on instrumental analytical chemist" the modus vivendi for this monograph was to present classical and modern experiential and methodological data in ways that may be a helpful record and also serve as a transitional reference of methodology to facilitate the advancement of the era of robotic analytical workstations.

Food Sample for Analysis

Foods (and drinks) are nutrient-containing substances that can be metabolised into body tissue and into energy to sustain body tissue. In modern parlance foods are largely solid, and drinks are largely liquid. It is convenient to refer to all nutrient sources as food – the nutrient-carrying matrix – and to consider the removal of compounds from a sample of food as an extraction. However, the English language has many words to express the idea of removing something from the whole. In analytical chemistry, for example, there is no clear distinction between a separation method and an extraction method, and it gets worse because chemists also fractionate, purify, isolate, partition, disperse and distribute components of mixtures. Here, an extraction is thought of as an operation on a sample of food that concentrates the target components, normally by removing them from the bulk of the food sample, often in preparation for further examination such as chromatographic separation. In analytical chemistry, a separation is seldom carried out on the raw material (however, see Chapter 8, Section 2, direct injection), but on an extracted or cleaned up sample for analysis. In addition, there are many procedures associated with extraction that in themselves do not actually remove anything from the sample. These processes are dealt with in Chapter 2 (Sample Preparation for Extraction), and are treated as extraction aids.

The natural origins of human foods are biologically diverse, ranging widely in texture and composition – from nutmeg to oysters. The extremely complex endogenous composition of food is made even more complex in the modern environment where so many extrinsic, additional items – additives such as antioxidants, contaminants from agriculture such as herbicides and industrial adulterants such as hydrocarbons from petroleum – may also be present. This extends the quantitative range of analyses practised by food analysts from the gram amounts encountered in proximate analysis (Section 6) to picogram and even lower amounts of highly toxic contaminants e.g. PCBs. To cover more than 12 orders of magnitude requires an enormously diverse armoury of techniques.

Analysis of Foods

It is usually a concern over the chemical composition or contamination of food and the effect this has on its value to the consumer that generates the need for analysis. The quality of food is based on the natural composition, the balance between the nutrient and the anti-nutrient composition. The health and prosperity of early civilisations depended upon their ability to refine their food supply in the short term by removing toxic materials using extraction methods, or in the long term through crop selection and plant breeding.

History of Food Extraction

Many extraction methods were invented to remove sufficient quantities of toxins (anti-nutrients) from the biological source to make the material acceptable and safe to eat. Notably, nature historically used toxins in sources of human and animal foods to maintain the balance between the survival of the browser and the browsed! These practices were incorporated into the culture of the technology employed in the early analytical laboratories.

The natural processes used to extract moisture in order to increase the "shelf-life" of food and the early uses of extraction methods to concentrate important components, e.g. essential oils, formed the bases for methods of analysis as the science of measurement began to develop. Historically, the extraction of bulk components from food made use of physical processes, such as pressure, to remove the juice or oil from the pulverised pulp. Warm air or sun drying of tomatoes or fish extracted sufficient water to reduce bacterial attack to an acceptable level in preparation for storage. Solvent extracts of essential oils from the pulverised plant, seed, or nut were concentrated by distillation in simple stills. Spicy and resinous plants were solvent refluxed in fractionation columns and valuable components separated and extracted in this way. The modern method of supercritical fluid extraction (SFE) uses ultrapure carbon dioxide as solvent, thus eliminating the fear of toxic residues in the extract. Cold-pressing methods are still used to produce high quality extracts of citrus fruits, and hydrodistillation, the steam distillation of an aqueous solution of the food matrix, was practised, especially on powders, from earliest times. There are many other examples where extraction from the bulk material was used to refine our food supplies.

Analytical data defining food quality and the methods used to obtain them have to be validated; several regulatory bodies oversee this process (FDA, FSA, AOAC, FAO, WHO, etc.). In 1963, the FAO and the WHO set up the Codex Alimentarius Commission to develop food standards, guidelines and codes of practice. Their web-site is Codex@fao.org. Ultimately, there are definitive methods that can be applied to analysis that provide an acceptable degree of confidence and that are universally recognised. For example, The Canadian Health Protection Branch, Health Canada published The Compendium of Methods for Chemical Analysis of Foods, which was extant till 1995 and is currently being updated. The e-mail contact is Xu-Liang Cao@hc-sc.gc.ca. Details of the extraction from the food matrix of the compounds of interest will be given in the method. These data are already correlated and readily available to the practising food analyst and only illustrative examples will be further discussed. It is the historical context, background principles, general practice, and the development of emerging and tentative experimental methods leading to the ultimate automated assay that form the major part of this study.

The ways in which the methodology has been used are illustrated in the examples from the scientific literature chosen to cover a range of commodities and analytes. Modern databases contain huge amounts of information on extraction methods for food analysis and the reader may wish to base further searches for information on keywords found in the appropriate sections of this manual of methods.

Stages in Food Analysis

Stages that may be required in the analysis of foods are:

1. Setting the protocol.

2. Sampling the food.

3. Preparing the sample in readiness for extraction of the chosen analyte or compound of interest (COI), including standardisation.

4. Extraction of the COI.

5. Separation from, or removal of, substances interfering with the detection of the COI in the extract.

6. Detection (recognition or visualisation) of the COI.

7. Identification and/or quantification of the COI.

8. Recording the information.

Items 3–5 are the subjects of this monograph.

Defining the Stages in Food Analysis

Protocol. It is important to have a clearly defined protocol and to adhere to it, so that the analysis can be reported unambiguously, verified by the analyst, and, if necessary, reproduced for verification by other analysts.

Sampling. This is the process of preparing a representative portion of the whole food for analysis. This sample is usually re-sampled by the analyst (the sample for analysis) and may need treatment before the target compound(s) can be extracted. If quantitative results are required, an internal standard (e.g. an isomer with similar chemical properties, but distinguishable from the analyte by GC retention time, or an isotope distinguishable by its mass spectrum) may be added to allow any subsequent losses to be compensated for during the analysis.

Preparation of the Sample for Extraction. The definition of sample preparation is ambiguous in the literature, often covering all processes up to and including the separation stage. The definition of sample preparation for extraction here is "the execution of procedures necessary to prepare the original sample for extraction." Such processes include grinding, digestion, and centrifugation. Occasionally, mainly for liquid foods, no preparation for extraction is required.

Extraction. There can be no hard and fast rule, but having entitled this monograph "extraction methods" the definition of an extraction process used in collecting together relevant subjects was "one where a part of the sample is removed from the whole starting sample." It may be the part containing the compounds of interest or it may be an unwanted part being discarded, leaving a less complex and usually more concentrated remainder for further study.

Direct Analysis without Extraction. The analyte can be in sufficient concentration, and free from interference from the matrix, usually in a liquid food, that no extraction stage is necessary. For example, a sample of the matrix can be injected directly into the separation stage, e.g. HPLC. This possibility becomes more feasible with the use of guard columns and as the resolving power of the separation and detection stages improve. The use of chromatography-MS and electrophoresis-MS, especially MS" and HRMS methods means that many potential interferents can be circumvented without the need to remove them by extraction. Alternatively, a colorimetric reaction can visualise and quantify the analyte in a crude extract, as is the case when the biuret reaction is used to measure protein content.

Separation. The term separation is reserved for chromatographic and electro-phoretic processes where the main objective is not to remove or extract something for further stages of analysis, but to finally resolve components of mixtures for detection and identification.

Exceptions are made in the case of preparative separation, which was the forerunner of two-dimensional separation, where the fractions are collected for manual transfer to a further stage of analysis. This can be seen as a preliminary extraction – and again with multistage chromatography, where each stage serves to fractionate the mixture, presenting one or every fraction extracted as the input to further stages of analysis. Multistage or two-dimensional chromatography is capable of extremely complex, on-line, automated separations and can be seen as a combined extraction and separation system.

Because the separation stage is not in the remit for this monograph, a more pedantic stance has been taken to draw a distinction between the two stages than is necessary when dealing with the analytical process in toto. Nevertheless, on several occasions the separation process has been viewed as a micropreparation process simply to raise the prospects for automated microanalytical methods to be developed.

Detection for Identification and/or Quantification. The signal recorded when a component of the sample is registered (recognised, standardised or calibrated) above a base line, and the signal content is converted into qualitative or quantitative information.

Direct Detection by Remote Sensing of the Food Sample. If the analyte can be recognised (detected) in the food sample prepared for analysis by a sensing probe, then the required analytical data can be extracted directly without the need for physical, chemical, or biochemical intervention. This can be considered to be "virtual extraction or virtual removal of interference." Such methods have the potential to be rapid, economical in time and resources, and ideal for automation.

Recording the Information. For those wishing to refresh on recording and general good analytical chemical practice, a text such as Fundamentals of Analytical Chemistry by D.A. Skoog and D.M. West provides a thorough treatment. Some further discussion of the stages of food analysis with relevance to the extraction techniques that form the major part of the work will follow.

2 Sampling

Introduction

Assuming that the strategic arguments have been addressed and the reason for undertaking the analysis defined, the first analytical procedure is to obtain a representative sample of the bulk material. If the sampling process is inaccurate, all the subsequent, and often expensive extraction, separation and identification stages will be in jeopardy. Sampling is not covered in this monograph, but a few general comments will serve to put into perspective the material upon which the extraction will be made.

There are several different problems that can beset the sampling for analysis procedure. For example, if a new cultivar of broccoli has been created in the greenhouse on a limited experimental scale, and there are only one or two small florets on each plant, it may be permissible only to use a small part of it for, say, glucosinolate analysis. Therefore, high efficiency is required of the extraction and high sensitivity of the analytical methods employed. Even so, the sample represents only a single plant and the results should not be expressed otherwise.

A sampling problem of a different kind is generated when it is necessary to choose a representative amount from a 30-ton truck loaded with carrots from the field; how does one ensure that the relatively small sample of material needed for pesticide analysis is representative of the whole assignment? Also, once the representative sample has been taken, how should it be stored so that no changes in its composition occur until it is analysed? These problems and many more are dealt with in textbooks on sampling and standardisation, for example from the ACOL series, Woodget and Cooper's, Samples and Standards.2

Standardisation and Validation of Methods

If known quantities of standard reference materials (SRMs) – ideally isomers of target compounds – are added and thoroughly mixed into food samples at the outset (standard addition), subsequent methodology can be calibrated against losses occurring during handling, to provide quantitative measurements of composition that can lead to the validation of the analytical procedure.

Recovery, Sensitivity and Limit of Detection

Measures of method performance, such as recovery, the limit of detection (LOD), and quantification (LOQ), are generally based on the use of standard addition and on the assumption that the additional standard material behaves like the natural substance in any physical and chemical treatments employed.

As far as the extraction process is concerned, the total recovery specified for the whole analysis includes the efficiency of the adsorbing medium, etc., but, like all other parts of the assay, any losses that do occur are compensated for by the standard addition process. In practise, losses during extraction should be kept to a minimum, and for high sensitivity to be achieved in trace component analysis it is important to have as near a loss-free system as possible. With modern treated surfaces in separation columns and measuring instruments, and with the use of bonded stationary phases, there is less unwanted (irreversible) adsorption. Once receptor molecules of the target compounds have filled all the active absorption sites, any remaining molecules can proceed to the detector. The limit of detection is expressed as the threshold sensitivity of the detector to the remaining molecules, and is given a signal-to-noise ratio, e.g. 3:1. The LOQ is the lowest concentration of an analyte that can be determined with an acceptable precision and accuracy.

Precision, Accuracy, Reproducibility and Repeatability

Measures of reliability include the extraction stage, and errors of analysis need to be accounted for. Replication of the sampling and standardisation of the procedure is normal when quantitative measurements are being made, and a statistical evaluation of the reliability of this stage will be an integral part of the precision, accuracy, reproducibility, and repeatability of the whole analytical procedure. Most analytical methods provide this information and, therefore, it is assumed here that extraction is one of the processes, but not necessarily the limiting process, represented in the values arrived at for the whole method.