Application Notes
Published: 01 Dec 2024 · Last updated: 03 Aug 2026
The palm oil industry is one of the most productive vegetable oil producers. This is helped, in part, by the oil palm having the highest yield (kg/ha/yr) of all the vegetable oil plants. Nevertheless, there is scope for increasing oil production further by measuring the oil content at various stages in the process from the fields through to the mills.
Benchtop Time Domain Nuclear Magnetic Resonance (TD-NMR) is a well-established technique that has been successfully adopted and used to measure oil content in food and agricultural products over many decades. TD-NMR instrumentation uses permanent magnets and is therefore easy to maintain. Additionally, TD-NMR measurements have many advantages over traditional methods including being non-destructive, rapid, repeatable, and solvent-free. As with near infrared (NIR) spectroscopy, TD-NMR is often calibrated to "match" the oil and water content results generated by solvent extraction and oven drying for which there are ISO methods.1-2 However, unlike NIR, it is also possible to set up a primary calibration using the oil of interest, as the signal is directly and linearly related to hydrogen content. The ability to calibrate against a pure oil standard has been applied to snack foods3 and fish products.4
TD-NMR has numerous advantages over NIR and traditional extraction methods including:
Figure 1 shows a good correlation between the TD-NMR results, using a 100% palm oil calibration, against Soxhlet extraction across a wide range of concentrations. Thus, it is possible to measure the oil content of not only dried palm fruit, or mesocarp, but also of the residues during crushing and extraction using just one calibration; thus TD-NMR can be used as a tool to maximise oil yield by minimising losses during processing.
Fig. 1 — Comparison of NMR and Soxhlet results for palm fruit and the residues during various stages of processing.
For these reasons, benchtop NMR has many potential uses in the palm oil industry which are described in further detail below.
Measurement of oil in palm mesocarp is mainly of interest to plant breeders wishing to measure the oil yield from their crops using bunch analysis. Oil content is normally measured by Soxhlet extraction (after oven drying) which limits the sample throughput. Although the mesocarp must be dry prior to NMR analysis, the sample throughput is much greater than Soxhlet which is restricted by the analysis time as well as by the number of samples that can be analysed at any one time. In addition, NMR has been shown to be just as accurate as the hexane extraction method but does not require the use of solvents or other chemicals, fume cabinets and expensive disposal procedures.5-6
The method is simple; the samples are weighed in a vial then conditioned at 50°C for 20 minutes to ensure that all the oil has melted prior to analysis. The measurement takes just 16 seconds making the time per sample short when measured within a large batch of samples. The instrument may be calibrated against, and as a consequence can obtain results equivalent to, the reference extraction method. Alternatively, a primary calibration can be obtained using crude palm oil for the measurement of total oil content.
It has been shown that extraction methods don't always remove all the oil, whereas TD-NMR is able to measure all the oil in the dry tissue, even that which remains after Soxhlet extraction.
A comparison of the results generated from Soxhlet extraction and an NMR oil calibration for dried palm mesocarp is given in Figure 2.
Fig. 2 — Comparison of TD-NMR vs. Soxhlet extraction results for dried palm mesocarp shows offset and scatter due to oil not extracted by Soxhlet.
Note that the NMR results are typically higher than those from Soxhlet; this is because NMR measures all the oil, whereas Soxhlet does not measure the unextracted oil. The latter is not constant and accounts for scatter in the data. This scatter cannot be corrected using a calibration bias, so the unextracted oil content in the Soxhlet residue was measured using NMR and a correction was applied to the individual Soxhlet results (Figure 3).
Fig. 3 — Results from TD-NMR vs. Soxhlet for dried palm mesocarp corrected for unextractable oil show direct 1:1 correlation and reduced scatter.
The most noticeable feature is that the scatter is reduced, as indicated by the improved correlation (R2). Secondly, there is now a direct correspondence between the NMR and corrected Soxhlet results; that is, the line goes through the origin. This further demonstrates that not only can NMR accurately measure oil content in the residue, for optimization of the crushing/extraction process, but also it can measure "total" oil content from a sample of the oil-producing crop.
It is important to measure the oil content of pressed palm fibre in mills to maximise the efficiency of the extraction process, whether it be pressing or solvent extraction. However, this is a challenging measurement for many techniques as the oil content is relatively low and non-uniform and, as a consequence, the sampling variation can be large. It is important to note that this sampling variation, as well as incomplete extraction, can also lead to errors in the reference method. However, given that NMR is non-destructive, it is possible to measure the same sample by extraction after NMR analysis to avoid discrepancies due to sampling.
Figure 4 shows the correlation between the NMR signal/mass and the reference values determined on the same samples by Soxhlet (after drying). Therefore, as outlined above, any discrepancies between the results and calibration line are likely to be due to errors caused by the reference method itself.
Fig. 4 — NMR calibration for oil in pressed palm fibre; the correlation coefficient and standard deviation are 0.99 and 0.28% respectively.
Palm kernel crushers would also like a rapid measurement of oil content to determine how much can potentially be extracted. Nevertheless, palm kernels are not commonly measured because the hard nut is very difficult to crush into smaller pieces. Crushing is needed to allow the solvent to penetrate the sample for better extraction leading to a more accurate result. As the moisture content is relatively low, it is possible to measure the oil content of palm kernels by NMR without drying. Furthermore, because radio frequency radiation penetrates the whole sample, palm kernels can be measured by NMR without crushing. The only requirement is that they are conditioned at a higher temperature to mobilise the oil prior to analysis.
NMR can also measure oil and moisture content in the palm kernel cake, which is useful for monitoring the efficiency of the extraction process. In addition, NMR can measure the oil content of palm kernel meal, a by-product used for animal feed.
The melting profile of an edible oil is a function of the oil composition and is an important property as it defines the applications for which the oils/fats can be used as ingredients. Fortunately, TD-NMR can also measure the proportion of solid fat in the palm oils/fractions conditioned at different temperatures. Given this is a fundamental measurement, it is not surprising that various standard methods exist which are used by oils/fats companies world-wide to characterise their raw materials.7 The Direct method (AOCS Cd 16b-93, ISO 8292-1, IUPAC 2.150) is the most commonly used due to its simplicity and precision.
Figure 5 shows that the melting profiles of palm olein and palm stearin obtained by TD-NMR, are very reproducible.
Fig. 5 — The melting profiles from two sub-samples of palm olein and palm stearin.
Benchtop NMR can be used at various stages of the palm oil production, from measurement of oil yield of the crops through to processing and characterisation of the raw materials.
© Oxford Instruments 2026