Food Safety and Quality Control

Explore innovative quality control and analytical methods in this Food Safety and Quality Control Resource Guide

Food safety and quality control requires careful monitoring from the point-of-origin through distribution to consumers. Analytik Jena offers solutions to simplify the analysis process, to help meet regulatory directives, and increase analytical performance.

In this eBook, you'll learn about:

Food safety and testing methods
How to accelerate high-throughput analysis of drinking water
Methods to determine metals in dietary supplements with high-resolution continuum source solid atomic absorption spectrometry (HR-CS Solid AAS)
How to determine what trace elements and minerals are in edible oils using high-resolution ICP-OES
Identifying food fraud

Agencies and consumers demand more vigilance, which requires increased analytical performance. Download the free resource guide now.

36345_FoodSafety_AnalytikJena_final

Food Safety and Quality Control

Resource Guide

Improving Food Safety Testing Methods

Speed Up Your Work— High-Throughput Analysis of Drinking Water with ICP-MS

Determination of Metals in Dietary Supplements Using HR-CS Solid AAS

Determination of Trace Elements and Mineral Contents in Edible Oils and Fats by HR ICP-OES

Identifying Food Fraud with Analytical Testing

Food Safety and Quality Control Resource Guide

Introduction

Food Safety and Quality Control Resource Guide

Improving Food Safety Testing Methods

Agencies and consumers demand more vigilance...and that takes new technology by Mike May

“With global food safety testing services projected to reach nearly $20 billion by 2022, the food industry is met with increasing demands from both the consumer and regulatory agencies to maintain a safe food product,” says Lynette Johnston, food safety extension associate at North Carolina State University in Raleigh. “At the same time, research has improved analytical methods, offering faster turnaround times for results with higher sensitivity for low-level contaminants and detection of contaminants in complex matrices.”

Ongoing safety concerns drive this field. “Foodborne diseases in the U.S. cause nearly 48 million illnesses each year, with 128,000 hospitalizations and 3,000 deaths,” says Johnston. Nonetheless, she adds, “Improved surveillance and detection methods have helped improve responses from the food industry and government officials.” Consequently, Johnston says, “information obtained from outbreak investigations

and recalls has allowed identification of previously unrecognized safety concerns in the food supply.”

New government regulations also force vendors to be more careful than ever. “The current onset of the Food Safety Modernization Act, FSMA, for many FDA-regulated food facilities will continue to drive increased testing,” Johnston explains. “For example, FSMA’s Preventive Controls for Human Food rule requires that a facility verify that hazards are being

controlled and corrective actions are taken to prevent contamination.” She adds, “Product testing and environmental monitoring are examples of steps a firm may include for verification.”

New Challenges

Increasing concerns over allergens also fuel the creation of new tests. Mass spectrometry, for example, can be used to prove that something that

Food Safety and Quality Control Resource Guide

is supposed to be free of nuts actually is free of nuts. By looking for a peptide that is unique to an allergen, that peptide can be tracked for its presence in all stages of the manufacturing process.

Specific hazards in foods, such as Shiga toxin– producing E. coli, can also be analyzed with polymerase chain reaction (PCR). Analyses can identify if a bacterium includes both target virulence genes, eae and stx. With specialized PCR tests, food

scientists can quickly know that a bacterium contains both virulence genes, instead of waiting on the seven- to ten-day culturing step that would be needed after conventional PCR.

As this field evolves, some of a scientist’s needs will stay the same. No matter what method is used, the food industry needs with sensitive, accurate, rapid, and cost-effective testing methods.

Food Safety and Quality Control Resource Guide

Speed Up Your Work – High-throughput Analysis of Drinking Water with ICP-MS

Introduction

Drinking water is the world‘s most important resource and most consumed food. Quality control for drinking water is regulated by international and

Challenge

Highly precise high-throughput analysis of 21 elements

(+4 internal standards) in drinking water according to U.S. EPA 200.8.

Solution

The unmatched sensitivity of the PlasmaQuant® MS enables

fastest and most precise analyses resulting in lowest cost per sample.

national regulations and norms. For example, the U.S. Environmental Protection Agency (EPA) has released the Method 200.8 which specifies criteria for the determination of trace elements in waters and wastes by ICP-MS. In Europe, the Council Directive 98/83/EC of 3 November 1998 specifies quality criteria for water intended for human consumption. The analyses are carried out internally by producers and water suppliers or externally by contract labs in order to guarantee safe products free from toxic contaminations. Almost all analyses offered by contract labs follow regulations and standards in order to harmonize the measurement procedures. For those labs requiring high sample throughput each day, not only accuracy, precision and robustness of the method chosen are

important but especially the speed of analysis matters. The more samples that can be analyzed per hour, the lower the cost per sample.

Fast sampling systems were developed for ICP mass spectrometers, dramatically decreasing sample uptake time and thus total analysis time per sample1. The limiting factor is no longer the supply and rinsing time of the sample, but the measuring time that depends on the number of elements to be analyzed.

However, reducing the data acquisition time will directly affect the precision of the results as less averaging is possible due to the time constraint. Therefore, precision is the key parameter for the evaluation of high-throughput methods.

Using Analytik Jena’s patented iCRC technology for interference removal, we demonstrate the measurement of 82 drinking water samples per hour with 21 elements (+4 internal standards). Even with this high sample throughput, highly precise results were achieved (average RSD <2.2 %) according to the U.S. EPA 200.8 regulation. Due to the robustness of the method and the unmatched sensitivity of the PlasmaQuant® MS, an even higher precision with an average RSD of 1.5 % is demonstrated at a very competitive throughput of 60 samples per hour. In combination with the lowest argon consumption on the market, this translates into lowest cost per sample.

Materials and Methods

A PlasmaQuant® MS equipped with Micro Mist (0.4 mL/min) nebulizer, Scott double-pass spray chamber and Fassel torch with 2.4 mm injector was used for the analyses. A fast sample introduction system (oneFAST, ESI) was combined with an ASPQ3300 autosampler to improve the throughput. Featuring the patented ReflexION technology and the self-cleaning pre-quadrupole, the PlasmaQuant® MS reduces contaminations of the ion optics and quadrupole to an absolute minimum avoiding unnecessary cleaning steps and maintenance.

To reduce the costs per sample it is necessary to minimize the time per sample. The sample analysis cycle can be divided into measurement time and non-productive steps such as sample uptake or washout, read delays and autosampler movements (see Figure 1). Those non-productive steps are minimized using the ESI oneFAST system, significantly boosting sample throughput1. The measurement time and thus sample throughput is then defined by the requested precision that needs to be achieved.

Figure 1: Comparison of different steps during sample uptake, measurement and washout with and without ESI oneFAST.

The Aspect MS® software fully controls and monitors all functions and accessories of the instrument. A warm-up routine which can be adapted to the user’s needs automatically performs all required steps reducing instrument downtimes to a minimum. The software allows the automatic optimization of all ion optics, nebulizer and plasma parameters and guides the user through the method development process. It contains built-in quality controls for monitoring the results as required by regulations. Out-of-specification results trigger predefined actions thereby minimizing user actions. The Aspect MS® software provides intuitive data analysis, export and report functions for optimal integration in modern laboratory environments and LIMS systems.

Samples and Reagents

All samples and standards were prepared using high-purity reagents. Samples were diluted by a factor of 2 with deionized water (<0.055 mS, ELGA Lab). Standards and samples contained 1 % nitric acid (Ultrapure, Merck).
Calibration solutions were prepared from a stock solution using the multi-element standard Calibration Mix 2 (Analytik Jena) and single-element standards (ICP grade) for Ag, Sb, Hg, and Mo. The concentrations used for calibration are shown in Table 1.

Concentration of standards [ppb]

Element

Table 1: Concentration of the standards used for calibration.

Be	0.5	5	50	100	500
Al	0.5	5	50	100
V	0.5	5	50	100
Cr	0.5	5	50
Mn	0.5	5	50
Co	0.5	5	50
Ni	0.5	5	50	100
Cu	0.5	5	50	100
Zn	0.5	5	50	100	500
As	0.5	5	50	100	500
Se	0.5	5	50	100	500
Mo	0.5	5	50	100
Ag	0.5	5	50
Cd	0.5	5	50	100
Sb	0.5	5	50	100
Ba	0.5	5	50	100	500
Hg	0.5	5	50
Tl	0.5	5	50	100
Pb	0.5	5	50	100
Th	0.5	5
U	0.5	5

Out of the 21 analytes, 6 representative calibration curves (Be, Al, Mn, Mo, Ba and Tl) are shown in Figure 2. Correlation coefficients >0.99996 were achieved for all elements. The obtained very high correlation coefficients, low RSDs and minor deviations of the individual standards from the regression curve show the excellent quality of the calibration even at high sample throughput.

Figure 2: Representative calibration curves.

Instrument settings and method parameters

A robust method utilizing helium as a collision gas for effective interference removal by kinetic energy discrimination (KED) was developed to achieve high accuracy and robustness at minimum analysis time per sample. Internal standards (6Li, Y, Rh and Ir) were added on-line to the sample via a T-piece at 20 ppb concentration. The method parameters used are listed in Table 2.

Specification

Parameter

Table 2: Method parameters.

Plasma Gas Flow 9 L/min

Auxiliary Gas Flow 1.45 L/min

Nebulizer Gas Flow 1.01 L/min Spray Chamber Temperature 3 °C

RF Power 1450 W

Sampling Depth 5.0 mm

Dwell Time 20 ms (50 ms for Be, As, Se)

Scans per Replicate 7 (peak hopping, 1 pt/peak)

No. of Replicates 6

Pump Rate, Tubings 15 rpm, black/black PVC pump tubing for sample; orange/green PCV tubing for internal standards Stabilization Delay 19 s

Sample Loading Time 7 s

iCRC Gas Flow 120 mL He/min

Detector Attenuation none

Results and Discussion

Table 3 lists the elemental concentrations measured in tap water (Jena, Germany) and in certified reference materials (NIST 1640a and NIST 1643f).

Table 3: Determined concentrations of Jena tap water and certified reference materials.

[ppb]

Be9

Al27

V51

Cr52

Mn55

Co59

Ni60

Cu65

Zn66

As75

Se82

tap water	0.0	2.7	0.5	0.2	1.1	0.0	0.9	63.7	47.7	0.9	1.5
NIST 1640a	2.9	52.8	14.2	38.5	38.7	19.7	23.5	81.2	52.0	7.8	19.3
NIST 1643f	13.1	131.9	35.9	17.5	34.9	24.1	54.6	19.8	70.2	54.0	11.7

[ppb]

Mo98

Ag107

Cd112

Sb121

Ba135

Hg202

Tl203

Pb207

Th232

U238

tap water	0.7	0.2	0.0	0.2	200.0	0.4	0.0	2.4	0.1	5.4
NIST 1640a	45.3	7.4	3.8	5.1	145.7	0.2	1.5	11.7	0.1	24.1
NIST 1643f	118.9	1.0	5.8	56.7	503.2	0.3	6.5	17.1	0.1	0.0

Accuracy

Certified reference materials were measured in order to verify the accuracy of the method. The concentration results were found within 91 and 103 % of the specified value and therewith within the ±10 % range as specified by U.S. EPA 200.8. Additionally, two lab-fortified matrices (LFM, +1 ppb and +10 ppb) were measured to verify the method’s accuracy. The recovery rates of the LFMs were between 91 and 104 % and therefore within the specified range of 70 to 130 %.

It is important to note that good recovery rates for small spikes on high sample concentrations can only be achieved with a precise measurement. For this reason the EPA 200.8 regulation states that the spiked concentration is dependent on the instrument’s sensitivity and that it should be equal to the concentration in the sample. Due to the outstanding sensitivity of the PlasmaQuant® MS and the thereby enhanced precision, it was possible to correctly analyze lab-fortified matrix samples adding only 10 % to the actual sample concentration.

Precision

The relative standard deviation (RSD) was used as a parameter to assess the precision of the measurements. On average, a RSD of 1.5 % was achieved. The results for selenium were less precise due to its inherent low ionization efficiency resulting in only a low count rate. Precision and accuracy of the certified reference materials and lab-fortified matrices are shown in Figure 3. It can be seen that all recovery rates are within 90 to 110 % while the RSD is <2 % for low concentrations

(LFM +1 ppb, NIST 1640a) and ≈1 % for higher concentrations (LFM +10 ppb, NIST 1643f). The required RSD <5 % for a 10/100 ppb tuning solution standard (depending on sensitivity) to verify instrument stability was achieved for all elements.

Figure 3: Recovery rate and precision of certified reference materials and lab-fortified blanks. All measurements are within the range specified. The spike concentration of +1 ppb was <5 % of the sample concentration for Cu and Ba and was excluded.

Speed of analysis

With the PlasmaQuant® MS, 60 samples per hour can be measured with very high precision (mean RSD <1.5 %) allowing to precisely determine LFMs adding only 1 ppb. These results surpass requirements of the U.S. EPA method 200.8 and the European drinking water directive by far, meeting even much stricter requirements and more demanding regulations.

Figure 4: RSD of matrix spikes and certified reference materials measured with two different speeds. The outstanding sensitivity allows to measure 82 water samples per hour with competitive precision.

However, for many applications not only precision and accuracy but also the speed of analysis are important parameters. A higher sample throughput significantly reduces the cost per sample. If the requirements for the precision of the analysis are not the main priority, the spectrometer’s unmatched sensitivity allows to further boost sample throughput to >80 samples per hour still delivering competitive precision (mean RSD 2.2 %) as shown in Figure 4.

Robustness

Figure 5: Recovery rate of the certified reference materials and LFMs as a function of time. Stable and accurate results were measured on all masses for 7 hours proving the robustness of instrument and method.

The recovery rate was evaluated as a function of time in order to assess the stability of the instrument and the robustness of the method. The stability of the recovery rates of the certified reference materials, laboratory-fortified matrices and internal standards were excellent during the analyses over the entire mass range. Therefore, the PlasmaQuant® MS is well suited for routine applications requiring a high sample throughput over a long time (Figure 5).

Conclusion

The costs per sample are directly linked to sample throughput and limited by the precision that needs to be achieved.

The unmatched sensitivity of the PlasmaQuant® MS and the resulting precision advantage allows to achieve the highest sample throughputs in the ICP-MS market. The requirements of the U.S. EPA 200.8 regulation are surpassed by far and the performance is maintained for hours proving the method’s robustness. By consuming only 50 % argon gas compared to a conventional ICP-MS, the mass spectrometer by Analytik Jena has the lowest running costs on the ICP-MS market.

The combination of minimal running costs and highest sample throughput results in the lowest costs per sample, making the PlasmaQuant® MS the ideal solution for customers who need to routinely measure a large number of samples.

References

1 Sample Introduction Accessories for the PlasmaQuant® MS Series (Technote, Analytik Jena)

This document is true and correct at the time of publication; the information within is subject to change. Other documents may supersede this document, including technical modifications and corrections.

Food Safety and Quality Control Resource Guide

Determination of Metals in Dietary Supplements Using HR-CS Solid AAS

Introduction

Challenge

Determination of Pb, Fe, Zn and Se in dietary supplements

Solution

Application of High-Resolution Continuum Source solid AA for determination of Pb, Fe, Zn and Se

The consumption of dietary supplements is widely spread and on the rise. These dietary supplements are generally used without prescriptions, proper counseling or any awareness of their potential health risks. In order to ensure the safety of these products and increase the awareness of the citizen to benefit from these dietary supplements it is of great importance to perform diligent analysis for (toxic) metals present in dietary supplements. While some metals are essential for living organisms because of their responsibility of maintaining the vital body functions, they have fatal effects when they are taken in excess. Research studies have even linked brain disorder symptoms to over dosage. On the other hand, heavy metals such as Pb and Cd are toxic at much lower levels and are known to induce serious diseases.

Atomic absorption spectrometry (AAS) is a robust, reliable and sensitive analytical technique which is well suited for the determination of heavy metals. This analytical technique is therefore widely employed for this purpose.

Instrumentation

The analysis was performed using the High-Resolution Continuum Source Graphite Furnace AAS contrAA 800 along with the solid sampler SSA 600L equipped with a liquid dosing unit.

Table 1: Configuration of the method and the instrument

Parameter

Specification

Temperature 20°C (room temperature)

Spectral observation width 200 pixel

Signal evaluation area

Evaluation pixels 3 (5 for Se)

Baseline fit IBC (iterative background correction)

Auto sampler SSA 600L

Tube type solid tube

Table 2: Method and evaluation parameters

Element / Wavelength

T Pyrolysis [°C]

T Atomization [°C]

Ramp [°C/s]

Time [s]

Modifier

Spectral observation width [nm]

Pb 283.306 nm	1050	2350	1500	4	5 µL Pd/Mg(NO3)2	0.32
Fe 304.266 nm	1150	2450	1300	6	-	0.32
Se 207.479 nm	500*/1100	2400	1500	5	5 µL Pd/Mg(NO3)2	0.26
Zn 307.590 nm	400	1600	1400	5	-	0.33

* oxygen ashing step with air 20 s, subsequently 30 s Ar purge to remove remaining oxygen

Samples and Reagents

Mineral tablets were finely ground and homogenized using a ball mill. Ginseng powder and the content of a capsule were analyzed as received without further sample treatment. A necessary dilution for the determination of Zn was obtained by the addition of spectral carbon powder. For that purpose, 0.04 g of the tablet were homogenized with 1 g of graphite powder using a ball mill resulting in a dilution factor of 25.

The modifier solution was prepared with 0.1% TritonX 100 to ensure wetting of the sample by the modifier.

Results and Discussion

The spectral environment of Fe and Pb shows additional atomic lines of Fe in the respective spectra. However, with a spectral resolution of approximately 2 pm at 200 nm, an interference-free analysis is performed without spectral overlap of other atomic lines. The correction of distinct molecular structures with partial or direct overlap of the analyte line is performed by applying the least squares background correction (LSBC). In this case correction spectra of the molecules that are responsible for the background are recorded, stored in the method and if present subtracted from the sample spectrum (Table 3).

A very insensitive absorption line is required for the analysis of Zn because of high contents in the sample. The available secondary line at 307.590 nm is useful after sample dilution with spectral carbon powder followed by a thorough homogenization.

The solid AA technology allows to directly analyze homogenized solid samples and waives a time-consuming and elaborate sample preparation procedure. Additionally, there is no inherent sample dilution as it is the case for digestion procedures. Analysis time is shortened and sources of contamination i.e. due to chemicals involved are reduced to a minimum.

Figure 1: Calibration curves

Using the solid AA technology higher RSD values are obtained in comparison to the liquid technique. This effect can be mainly ascribed to inhomogeneities in the samples and, in this regard, it is recommended to perform the measurements with an increased number of replicates per sample. Inhomogeneity was observed in particular for the gel-like content of the capsules. Smoke during pyrolysis indicates that there is a high carbon content present in the samples. The implementation of an oxygen ashing step into the temperature-time program facilitates removal of the organic matrix and avoids residues on the sample carrier after the measurement.

Table 3: Least squares background correction (LSBC)

Element

Molecule spectrum

Sample spectrum without LSBC

Sample spectrum with LSBC

SiO2

PO (ca. 4 mg NH4H2PO4)

Table 4: Results from measurement of dietary supplements samples on contrAA 800

Sample	Element / Line	Weight of sample [mg]	Weight of tablet/ capsule [g]	Concentration		RSD (n=4) [%]
				[mg/kg]	content per tablet or capsule

mineral- tablet	Pb 283.306 nm	0.8 – 1.5	1.9689	0.093 ± 0.006	0.183	µg	5.9
	Fe 304.266 nm	0.4 – 0.7	1.9689	2205 ± 18.3	4.34	mg	8.9
	Zn 307.590 nm	0.2 – 0.4	1.9689	2818 ± 47.4	4.55	mg	10.1
ginseng powder	Pb 283.306 nm	2.0 – 2.8	-	< LOQ **	-		-
capsule #1	Se 207.479 nm	1.8 – 4.0	0.3345	76.5 ± 6.16	25.6	µg	18.2
capsule #2	Se 207.479 nm	1.8 – 4.0	0.3275	82.4 ± 6.0	27.0	µg	16.5

**Limit of detection (LOD): 4,84 pg Pb (corresponds to 1,94 µg/kg using a sample weight of 2,5 mg) Limit of quantification (LOQ): 14,5 pg Pb (corresponds to 5,80 µg/kg using a sample weight of 2,5 mg)

Determined by measuring 11 times the blank (empty sample carrier) with modifier against the calibration curve

Conclusion

With its ability of applying correction algorithms for background as well as spectral interferences, contrAA 800 provides not only a robust and sensitive instrument solution, but furthermore allows for simple yet reliable analysis of trace metal contents in any kind of sample. solid AA reduces sample preparation to a minimum avoiding measurement errors and guaranteeing efficient routine operation.

Food Safety and Quality Control Resource Guide

Determination of Trace Elements and Mineral Contents in Edible Oils and Fats by HR ICP-OES

Introduction

Challenge

Reliable and effective assessment of element levels in edible oils

for process monitoring as well as quality and food safety control

Solution

HR ICP-OES for the sensitive and interference-free analysis of

elemental parameters relevant to processes, quality, and safety of edible oils and fats directly from solution in organic solvents

Edible oils and fats, particularly vegetable oils, are essential for a healthy diet because they provide the body with nutrients and are low in cholesterol, rich in unsaturated fatty acids, aid in the absorption of vitamins, and act as a carrier of flavors. Oils and fats are produced from oilseeds as well as from animal sources. Today, the annual production of vegetable oils has gone up well above 200 million tons.[1] Palm, palm kernel, rapeseed, soybean, coconut, peanut, and sunflower

oil are amongst the most widely used oils with the highest annual production

tonnage.

Traditionally the food and cosmetics industries are amongst the main consumers of edible oils and fats. In the case of palm oil, the usage in food products uses a share of 70% of the annual palm oil production, followed by the usage for biofuels and oleochemicals. Palm oil is used in products like margarine, cooking oil, spreads, chocolate, cleaning agents, cosmetics, candles, and many more. Coconut oil has been used for cooking for thousands of years. But it also can be found in

cosmetic products.

After harvesting the oilseeds, extraction methods or milling are applied to separate the vegetable oil from the seed. In the case of palm oil, a milling process separates the crude palm oil (CPO) from the palm kernel, which itself is source to produce palm kernel oil. After liberating the oil, further processing is performed in order to alter characteristics such as color, taste, odor, crystallinity, processability, and shelf life. The magnitude of processing depends on the aimed final use as well as the feedstock quality. Refining processes may contain degumming,

neutralization, washing, bleaching, deodorization, and dry fractionation. These processes employ chemical alteration of oil components by glycerolysis (transesterification), interesterification, or hydrogenation. Throughout the refining process the quality of intermediates as well as final products requires a thorough analytical investigation, not only to determine yield and purity of the products but also to monitor the level of trace elements that are either toxic to the human health or that have adverse effects on the quality or shelf life. Elevated levels of nickel originating from catalysts used during hydrogenation as well as iron and copper which may originate from processing equipment or packaging accelerate oxidation processes in the oil and therefore have tremendous effects on the shelf life of oil containing food products. Also, contents of calcium, lead, magnesium, sodium, and zinc are frequently monitored since they may reduce the process efficiency or cause inferior product quality. From a processing point of view, phosphorous containing compounds such as phosphatides need to be removed prior to the deodorization process step. Phosphorous levels above 1 mg/kg in refined and bleached oils pose the risk of catalyst poisoning as well as odd flavors in the final products and therefore need to be assessed in the according oil intermediates.

Additional to quality and process control, edible oils must comply with food safety regulations concerning toxic trace elements like arsenic, cadmium, lead, mercury, and tin with maximum permitted levels in the low µg/kg range.

Accurate elemental quantification in edible oils and fats requires an analytical methodology that is sensitive and selective. Due to its multi-element determination capability (up to 70 elements), high dynamic linear range and trace element detection capabilities, optical emission spectrometry with inductively coupled plasma (ICP-OES) is widely used for the analysis of oils and fats. The application is described in standard procedures such as ISO 10540-3, ISO 21033, and AOCS Ca 17-01.[2-4] Following these standard procedures, edible oils are diluted in low-viscous solvents (e.g., 1-butanol, kerosene, xylenes) prior to direct aspiration. In comparison to a full mineralization by ashing or digestion, this “dilute and shoot” approach provides the advantages of less sample preparation and handling, use of less equipment, and significantly reduced risk of sample handling related errors. However, the here obtained organic mixtures are challenging sample matrices to be analyzed by ICP techniques. The high load and carbon content of the organic matrix require a robust sample introduction and plasma system, which reliably excites the samples within the ICP and does not suffer from carbon build-up within the torch. Furthermore, carbon-based emission demonstrates an increased risk of spectral interferences and hence inaccurate results. In this regard, high-resolution ICP-OES analyzers offer superior peak separation as well as spectral correction models to resolve even severest interferences. A third challenge of analyzing multiple elements in edible oils via spectrometric techniques is

the wide working range required to measure trace elements for food safety concerns in the same run as medium to high concentration levels of minerals and naturally existing compounds in the oils that may disturb the refining process. DualView ICP-OES systems offer an efficient investigation of traces and major levels from a single measurement without a change of setup or analysis technology.

Within this study, the performance of the PlasmaQuant 9100 Elite high-resolution ICP-OES was investigated for oil samples of different processing stages from crude oils via intermediates to final products. Sample oil specimen including palm, coconut, rapeseed, sunflower, linseed, olive, peanut, sesame, and soybean were investigated with method validation via determination of method detection limits (MDL), spike recovery testing, and long-term stability investigation.

Materials and Methods

Samples and Reagents Sample Preparation

Depending on the fatty acid characteristics (e.g., chain length, degree of saturation), different oils possess different states of crystallinity, ranging from liquid via semi-crystalline to solid types at room temperature. In order to establish a uniform methodology, a solvent study was conducted prior to the analysis. Kerosene, 1-butanol and xylenes were tested for their suitability to prepare stable measurement solutions of the here investigated oils and fats with minimum dilution factors.

Standard procedures usually prefer 1-butanol over kerosene due to its better moisture tolerance and higher achievable pump rates. Since xylenes has comparable physical parameters to 1-butanol, it was included in the study as well. The tests have shown that xylene is the solvent of choice for this sample type. A fivefold dilution of liquified solid samples results in mixtures which are stable for several days without any signs of crystallization. Additionally, pump rates are the same as for 1-butanol dilutions. Samples which are liquid at room temperature were diluted by a factor of two.

Prior to dilution, solid and semi-crystalline samples were liquefied by heating at a temperature of 60 °C. Stock standards and diluted samples were homogenized in an ultrasonic bath for 15 minutes. Yttrium oil-based standard (CONOSTAN, 1000 ppm) was diluted in xylenes to give a concentration of 2 mg/kg (dilution factor (DF): 5) and 4 mg/kg (DF: 2), respectively. These solutions were used on the one hand as solvent for all dilutions and on the other hand to introduce Y as internal standard.

Calibration

The here presented methodology was used to analyze a large variety of edible oil and fat samples by using an external calibration in xylenes as described in ISO 10540-3, ISO 21033, and AOCS Ca 17-01. Calibration standards were prepared from organometallic single (As: CONOSTAN, 100 ppm; Hg: CONOSTAN, 100 ppm) and multi-element standards (S21+K, CONOSTAN, 885 ppm) by diluting with xylenes using concentrations as described in Tables 1 and 2. Blank oil was used as calibration blank and was added prior to the dilution to keep the oil ratio/fraction and therefore viscosity stable within the standard solution.

Table 1: Concentration of calibration standards for the analysis of palm and coconut oils in fivefold dilution

Element

Unit

Std. 1

Std. 2

Std. 3

Std. 4

Std. 5

Std. 6

Ag, Al, Ba, Cd, Cr, Cu, Fe, Mn, Mo, Na, Ni, Pb, Sn, Ti, V, Zn	mg/kg	0.107	0.256	0.524	1.050	-	-
As	mg/kg	0.099	0.259	0.505	-	-	-
Ca, K, Mg, P	mg/kg	-	-	0.524	1.050	5.545	10.12
Hg	mg/kg	0.098	0.256	0.508	-	-	-
Si	mg/kg	0.107	0.256	0.524	1.050	5.545	-

Table 2: Concentration of calibration standards for the analysis of vegetable oils in twofold dilution

Element

Unit

Std. 1

Std. 2

Std. 3

Std. 4

Std. 5

Std. 6

Std. 7

Std. 8

Ag, Al, Ba, Cd, Cr, Cu, Fe, Mg, Mo, Ni, Pb, Si, Sn, Ti, V, Zn	mg/kg	0.125	0.272	0.547	1.061	-	-	-	-
As	mg/kg	0.139	0.281	0.468	-	-	-	-	-
Ca, Mn, P	mg/kg	-	-	-	1.061	4.423	20.99	47.28	95.79
Hg	mg/kg	0.114	0.239	0.490	-	-	-	-	-
K	mg/kg	-	-	-	1.061	4.423	20.99	-	-
Na	mg/kg	0.125	0.272	0.547	1.061	4.423	-	-	-

Instrumentation Instrument Settings

For the measurements a PlasmaQuant 9100 Elite ICP-OES was used in combination with a Teledyne Cetac Oils 7400 stirring autosampler. The instrument was equipped with the organic kit, comprising of a glass concentric nebulizer, a double-pass cyclonic spray chamber, a 1.0 mm id injector tube, solvent resistant tubing, and a 0.4 mL/min nebulizer.

Table 3 describes a standard set of instrumental settings which is suitable to reliably measure all elements included into this methodology. Additional to this standard setup, an optional setting, including the use of oxygen addition to the plasma, is suggested in order to achieve lowest possible method detection limits and highest possible precision for potassium and sodium in oil samples. The effects of this optimized setup are discussed below.

Table 3: Instrument settings

Parameter

Standard settings

Optional settings for improved detectability and precision on K and Na

Plasma power	1450 W		1300 W
Plasma gas flow		15 L/min
Auxillary gas flow	1.75 L/min		0.25 L/min
Nebulizer gas flow	0.35 L/min		0.30 L/min
Oxygen gas flow	0.0 L/min		0.05 L/min
Nebulizer		Concentric, 1.0 mL/min, borosilicate
Spray chamber Double pass cyclonic spray chamber, 50 mL, borosilicate
Outer tube/Inner tube		Quartz/Quartz
Injector		Quartz, ID: 1mm
Pump tubing		Viton (black, black)
Sample pump rate		0.8 mL/min
Delay time		90 s
Torch positionA	-3 mm		0 mm

A Spacing between injector and coil further supresses carbon deposits (injector tip)

Method and evaluation parameters

Table 4: Method parameters

Element	Line [nm]	Plasma view	Integration mode	Read time [s]	Evaluation
					No. of pixel	Baseline fit, Pixel No.	Polyn. degree	Correction

Ag	328.068	axial	peak	3	3	ABC2	auto	Y3
Al	396.152	axial	peak	3	3	ABC	auto	Y
As	193.698	axial	peak	10	3	ABC	auto	CSI4, Y
Ba	455.403	axial	peak	3	3	static	auto	Y
Ca	315.887	radial	peak	3	3	ABC	auto	Y
Cd	214.441	axial	peak	3	3	ABC	auto	Y
Cr	267.716	axial	peak	3	3	ABC	auto	Y
Cu	324.754	axial	peak	3	3	ABC	auto	Y
Hg	184.886	axial	peak	10	3	ABC	auto	CSI, Y
Fe	259.940	axial	peak	3	3	ABC	auto	Y
K1	766.491	radial	peak	3	3	ABC	auto	Y
Mg	280.271	radial	peak	3	3	ABC	auto	Y
Mn	259.372	axial	peak	3	3	ABC	auto	CSI, Y
Mo	202.030	axial	peak	3	3	ABC	auto	Y
Na1	589.592	axial/radial5	peak	3	3	ABC	auto	Y

					Evaluation
Element	Line	Plasma view	Integration	Read time
	[nm]		mode	[s]	No. of	Baseline fit,	Polyn.	Correction
					pixel	Pixel No.	degree
Ni	221.648	axial	peak	3	3	ABC	auto	Y
P	213.618	axial/radial5	peak	10	3	ABC	auto	Y
Pb	220.353	axial	peak	10	3	ABC	auto	Y
Si	251.611	axial	peak	3	3	ABC	auto	Y
Sn	189.611	axial	peak	3	3	static	auto	Y
Ti	334.941	axial	peak	3	3	ABC	auto	Y
V	309.311	axial	peak	3	3	ABC	auto	Y
Zn	202.548	axial	peak	3	3	ABC	auto	Y

1 Optionally to be measured with oxygen addition to the plasma

2 Automated Baseline Correction

3 Internal standard correction using yttrium

4 Mathematical correction of spectral interferences originating from xylenes

5 Due to large variations in P and Na contents, the according emission lines were measured in axial as well as radial plasma observation

Results and Discussion

Palm oil is semi-crystalline, coconut oil is solid at room temperature. Therefore, a fivefold dilution in xylenes is required to obtain stable measurement solutions. Palm oil investigations included the analyses of crude palm oil as feedstock material as well as two different processing intermediates, red palm oil, and white palm oil. Achieved method detection limits (MDLs) well below 15 µg/kg ensure compliance with food safety regulations for toxic trace elements and allow for an efficient monitoring of elements that adversely affect the refining process as well as the product quality. The results of palm oil samples as well as crude coconut oil (see Table 5) show concentrations of elements concerning food safety which are well below the regulated limits. Quality control analysis along the palm oil refining process shows that the levels of quality indicators such as calcium, copper, iron, potassium, nickel, and sodium are reduced to sub- to low mg/kg range. Also, the removal of phosphorus containing compounds by processing steps can be effectively monitored with a level of 3.9 mg/kg in crude palm oil and 2.2 mg/kg in refined white palm oil.

Method validation for solid and semi-crystalline samples was performed by spiking white palm oil samples with 0.3 mg/kg of the target analytes. Spike recoveries in the range of 89% to 114% prove the accuracy of the employed method. Long-term

stability was investigated on a red palm oil sample. Here a 1.0 mg/kg spike showed recoveries between 92% and 108% for an 8-hour measurement with a measurement precision of well below 2% RSD for all investigated elements (see Figure 1).

Table 5: Quantitative results for investigated palm oil (PO) and coconut oil (CO) samples

Element

Line [nm]

MDL1

[µg/kg]

Crude PO

Red PO

White PO

Crude CO

White PO Spike recovery

Mass fraction [mg/kg]

Spike amount [mg/kg]

Recovery [%]

Ag	328.068	1.56	0.05	0.04	<MDL	0.06	0.31	98
Al	396.152	13.0	0.72	0.55	<MDL	0.20	0.31	100
As	193.698	14.8	<MDL	<MDL	<MDL	<MDL	0.30	111
Ba	455.403	0.62	0.09	0.10	<LOD	0.03	0.31	100
Ca	317.933	5.61	25.9	20.0	0.24	4.89	0.31	101
Cd	214.441	1.03	0.06	0.06	<MDL	<MDL	0.31	98
Cr	267.716	0.86	0.04	<MDL	<MDL	<MDL	0.31	102

Element

Line [nm]

MDL1

[µg/kg]

Crude PO

Red PO

White PO

Crude CO

White PO Spike recovery

Mass fraction [mg/kg]

Spike amount [mg/kg]

Recovery [%]

Cu	324.754	1.81	0.06	0.09	0.02	0.02	0.31	100
Fe	259.940	2.09	6.30	3.97	0.12	1.45	0.31	101
Hg	184.886	4.42	<MDL	<MDL	<MDL	<MDL	0.26	104
K2	766.491	13.1	7.93	2.51	0.17	27.7	0.31	89
Mg	280.271	0.95	6.55	2.50	0.06	12.6	0.31	101
Mn	259.372	0.42	0.83	0.35	0.03	0.19	0.31	100
Mo	202.030	3.58	<MDL	<MDL	<MDL	<MDL	0.31	99
Na2	589.592	14.6	3.77	1.36	0.24	3.90	0.31	91
Ni	221.648	2.94	<MDL	<MDL	<MDL	<MDL	0.31	101
P	213.618	10.3	31.7	3.85	2.20	45.7	0.31	114
Pb	220.353	8.57	<MDL	<MDL	<MDL	<MDL	0.31	99
Si	251.611	6.38	1.30	1.55	0.06	0.43	0.31	98
Sn	189.611	15.1	<MDL	<MDL	<MDL	<MDL	0.31	101
Ti	334.941	1.17	0.06	0.06	<MDL	<MDL	0.31	99
V	309.311	1.04	<MDL	<MDL	<MDL	<MDL	0.31	100
Zn	202.548	1.69	0.57	0.33	0.11	0.17	0.31	99

1 Method-specific detection limits obtained from calibration method (DF: 5)

2 Measured with oxygen addition to the plasma

Figure 1: Percentage recoveries of an 8-hour measurement of different elements spiked (1.0 mg/kg) to diluted red palm oil. RSD values were below 1.8% for all elements

Exemplary for edible oils that are liquid at room temperature, rapeseed oil (RO) samples of different processing stages were investigated for their element concentrations. As a twofold dilution in xylenes provides a good stability of the measurement solutions, achievable method detection limits are expected to be below the ones for solid sample types. Investigating MDLs in rapeseed oil provides diverse results. Improved detectability was achieved for the majority of elements whereas some elements such as arsenic or phosphorus did not show significant improvements which may be due to the increased matrix

contents originating from the lower sample dilution. Overall it can be stated that the detectability improves or shows an equal

level compared to a fivefold dilution. Monitoring elements of relevance to food safety as well as to process and product quality concerns show the same behavior as the results of palm oil. As displayed in Table 6, critical toxic elements are well below the regulated limits whereas an increased processing state of the rapeseed oil, from crude RO to refined RO, shows decreasing levels of calcium, copper, iron, potassium, nickel, and sodium. Spike recovery testing at a spike level of 0.26 mg/kg provided good recoveries in the range from 89% to 117%. Long-term stability testing showed recoveries between 92% to 108% for a 16-hour measurement with a measurement precision of well below 2% RSD for all investigated elements (see Figure 2).

Table 6: Quantitative results for investigated rapeseed oil (RO) samples

Element

Line [nm]

MDL1

[µg/kg]

Crude RO

Bleached RO

Half-refined RO

Refined RO

Refined RO Spike recovery

Mass fraction [mg/kg]

Spike amount [mg/kg]

Recovery [%]

Ag	328.068	0.83	<MDL	<MDL	<MDL	<MDL	0.27	87
Al	396.152	22.6	0.16	0.05	0.05	<MDL	0.27	91
As	193.698	15.7	<LOD	<LOD	<LOD	<MDL	0.27	111
Ba	455.403	0.28	0.03	<MDL	<MDL	<MDL	0.27	94
Ca	317.933	1.58	58.6	0.57	0.290	0.20	0.96	114
Cd	214.441	0.34	<MDL	<MDL	<MDL	<MDL	0.27	93
Cr	267.716	0.46	<MDL	<MDL	<MDL	<MDL	0.27	93
Cu	324.754	0.67	0.01	0.003	<MDL	<MDL	0.27	87
Fe	259.940	1.31	0.57	0.03	0.01	<LOQ	0.27	93
Hg	194.159	6.09	<MDL	<MDL	<MDL	<MDL	0.27	102
K2	766.491	26.4	28.2	0.18	<MDL	<MDL	0.96	118
Mg	280.271	0.66	11.9	0.13	0.10	0.09	0.96	93
Mn	259.372	0.10	0.17	<MQL	<MDL	<MDL	0.27	95
Mo	202.030	1.49	<MDL	<MDL	<MDL	<MDL	0.27	92
Na2	589.592	7.55	0.14	<MDL	<MDL	<MDL	0.27	117
Ni	221.648	0.93	<MDL	<MDL	<MDL	<MDL	0.27	94
P	213.618	11.3	164	2.92	0.66	0.48	0.27	97
Pb	220.353	8.08	<MDL	<MDL	<MDL	<MDL	0.27	92
Si	251.611	2.05	0.18	<MDL	<MDL	<MDL	0.27	85
Sn	189.611	3.71	<MDL	<MDL	<MDL	<MDL	0.27	100
Ti	334.941	1.13	<MDL	<MDL	<MDL	<MDL	0.27	93
V	309.311	0.43	<MDL	<MDL	<MDL	<MDL	0.27	93
Zn	202.548	0.49	0.20	<MDL	<MDL	<MDL	0.27	93

1 Method-specific detection limits obtained from calibration method (DF: 2)

2 Measured with oxygen addition to the plasma

Figure 2: Percentage recoveries of a 16-hour measurement of different elements spiked (1.0 mg/kg) to diluted commercial rapeseed oil. RSD values were below 1.5% for all elements

The here developed and validated methodology can be easily extended to other edible oils such as linseed, olive, peanut, sesame, and sunflower oil, which are all liquid at room temperature. Hence, a twofold dilution of the samples can be employed for the measurement against the calibration performed for rapeseed oil. The results for seven commercially available oils are shown in Table 7.

Table 7: Quantitative results for investigated commercial vegetable oils

Element

Line [nm]

MDL1

[µg/kg]

Linseed oil

Olive oil #1

Olive oil #2

Peanut oil

Sesame oil

Soybean oil

Sunflower oil

Mass fraction [mg/kg]

Ag	328.068	0.83	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL
Al	396.152	22.6	0.13	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL
As	193.698	15.7	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL
Ba	455.403	0.28	0.07	0.01	0.01	0.13	0.01	0.07	0.004
Ca	317.933	1.58	57.4	0.06	0.14	10.4	0.20	11.6	1.93
Cd	214.441	0.34	<MDL	<MQL	<MDL	0.002	<MDL	<MDL	<MDL
Cr	267.716	0.46	<MDL	<MDL	<MDL	0.002	0.002	0.01	<MDL
Cu	324.754	0.67	<MDL	<MQL	<MQL	0.01	<MDL	0.002	<MDL
Fe	259.940	1.31	0.34	<MLD	0.04	0.20	<MQL	0.49	0.02
Hg	194.159	6.09	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL
K2	766.491	26.4	19.2	<MDL	<MDL	29.9	<MQL	8.58	<MDL
Mg	280.271	0.66	34.7	0.06	0.06	11.3	0.12	8.20	0.56
Mn	259.372	0.10	0.34	<MDL	<MDL	0.14	<MDL	0.08	0.02
Mo	202.030	1.49	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL
Na2	589.592	7.55	1.12	0.11	0.11	0.48	0.33	0.71	0.37
Ni	221.648	0.93	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL
P	213.618	11.3	120	0.12	0.28	62.0	0.36	33.3	2.64
Pb	220.353	8.08	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL
Si	251.611	2.05	<MDL	<MDL	<MDL	0.05	0.03	0.16	<MDL

Element

Line [nm]

MDL1

[µg/kg]

Linseed oil

Olive oil #1

Olive oil #2

Peanut oil

Sesame oil

Soybean oil

Sunflower oil

Mass fraction [mg/kg]

Sn	189.611	3.71	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL
Ti	334.941	1.13	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL
V	309.311	0.43	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL	<MDL
Zn	202.548	0.49	1.85	<MDL	<MDL	0.21	<MDL	0.45	0.04

1 Method-specific detection limits obtained from calibration method (DF: 2)

2 Measured with oxygen addition to the plasma

The use of most sensitive emission lines is a prerequisite to achieve the best analytical performance in terms of achievable method detection limits as well as high accuracy and precision for trace element detection. In complex sample types such as organic materials, spectral interferences from either the matrix itself or from main constituents may restrict the use of the most suitable lines. In this regard, a high-resolution spectrometer as used in the PlasmaQuant 9100 Elite improves the separation of analyte signal and interferent to an extend that an interference-free quantification of the most sensitive line is possible for almost all elements. In this regard, Figure 3 shows the comparison of spectral data acquired with an average resolution spectrometer (Figure 3, left) and one acquired with a high-resolution spectrometer (Figure 3, right). The high-

P177.436 average resolution (6 pm @ 200 nm)

P177.436 high resolution (2 pm @ 200 nm)

resolution reveals a second signal as a shoulder to the phosphorous line that is unnoticed in the average resolution spectrum. This interference typically stays unnoticed on common ICP-OES instruments, which causes false positive results, especially when looking for trace element contents.

Figure 3: Comparison of P177.436 nm spectra acquired with average spectral resolution (left) and high spectral resolution (right (red: sample, black: spike, blue: Cal. o, green: baseline correction.

Within this method, only arsenic, manganese, and mercury showed insufficiently resolved signals on their respective primary emission lines. Here an easy to adopt spectral correction algorithm such as the CSI software tool enables the removal of severe interferences to make the desired emission lines accessible for routine measurements. Figure 4 shows the as-recorded spectrum of Mn259.372 nm (Figure 4, left) with the manganese emission line situated in a very crowded spectral environment from which an accurate baseline-fitting and peak evaluation is hardly possible. Applying a spectral correction via the CSI software algorithm results in a simple to evaluate spectrum delivering highly accurate results for a previously interfered line (Figure 4, right).

Mn259.372 (uncorrected)

Mn259.372 (CSI corrected)

Figure 4 : As-acquired spectrum of Mn259.372 nm (left) and spectrum after application of spectral corrections via CSI algorithm (right) (red: sample, black: spike, blue: Cal. 0, green: baseline correction.

In the analysis of organic sample types, emission lines in the long wavelength range suffer from elevated background levels and line-rich spectra due to carbon-based emission of the oil and solvent matrix. This mainly concerns the detectability of sodium and potassium. The resolution of the spectrometer allows for an identification of the Na589.592 nm line within a very crowded spectrum (Figure 5, left). By this it is possible to achieve detection limits in the range of 50 µg/kg, which is sufficiently low for most standard applications. However, since sodium and its removal after the neutralization process step plays a very important role in the overall process efficiency of edible oil refining, stringent monitoring of lowest levels may be beneficial to maximize the process yield.

The PlasmaQuant 9100 Elite enables the suppression of carbon-based signals in the spectrum by removing carbon in the sample feed area including the plasma. To do so, a small flow of oxygen can be dosed to convert carbon into carbon dioxide, which can easily by extracted by the ventilation of the system. The effects on the spectral complexity can be observed in

the spectrum displayed in Figure 5 (right) where the background level has dropped by a factor of ten, whilst the signal to background ratio is kept the same. This allows for a tenfold increase in sodium detectability to a detection limit below

Na589.592 (regular plasma conditions)

Na589.592 (oxygen mode)

5 µg/kg. On top of this, the baseline is much smoother, and a more reliable baseline fitting can be applied with the effects of a significantly improved precision in the trace detection range.

Figure 5: As-acquired spectrum of Na589.592 nm (left) and spectrum after application of oxygen (right) (red: sample, blue: Cal. 0, green: baseline

correction.

Conclusion

The here presented methodology describes the analysis of elemental parameters relevant to process monitoring as well as quality and food safety control by a high-resolution ICP-OES, the PlasmaQuant 9100 Elite. To achieve the highest possible detectability, the approach of direct dilution of the samples in xylenes was employed since it keeps the overall dilution factors at a minimum. The challenges for the analysis of the fully organic measurement solutions are perfectly addressed by the features of the PlasmaQuant 9100 Elite, the vertical plasma orientation provided by the V Shuttle torch, the high plasma robustness by the high-frequency generator, the wide working range by the DualView Plus plasma observation modes,

and the high resolution spectrometer. The results clearly demonstrate the enormous application advantages originating from the instrument features. The high-frequency generator in combination with the unique V-Shuttle torch allows for the measurement of almost any sample type including undiluted solvents and high matrix samples. Especially, the option to increase the distance of injector to the plasma offers huge advantages in daily routine of organic applications and reduces

time for maintenance due to practically nonexistent carbon deposits. Furthermore, the user benefits from the possibility of operating the instrument in oxygen mode providing reduction of spectral interferences for certain elements and improving limits of detection. The high spectral resolution allows for using the most sensitive emission lines without compromises

in detectability or precision in the target concentration levels. In combination with a high sensitivity and a robust plasma, exceptional limits of quantification (sub- to low µg/kg range) can be achieved with high confidence in the obtained results. Additionally, software tools such as the automatic background correction (ABC) and in particular the correction for spectral interferences (CSI) significantly reduce the time required for data evaluation and often further improve the instrument’s sensitivity. In summary, the PlasmaQuant 9100 Elite is well suited for the process, quality, and food safety control of edible oils and fats.

References

1 Oilseeds: World Markets and Trade, United States Department of Agriculture, September 2020.

2 ISO 10540-3, First Edition 2002/12/1, Animal and vegetable fats and oils — determination of phosphorus content; Part 3: Method using ICP-OES.

3 ISO 21033, First Edition 2016/05/1, Animal and vegetable fats and oils — determination of trace elements by inductively coupled plasma optical emission spectroscopy (ICP-OES).

4 American Oil Chemist Society, AOCS Recommended Practice Ca 17-01, approved 2001.

Food Safety and Quality Control Resource Guide

Identifying Food Fraud with Analytical Testing

Olive oil is an ingredient at high risk for food fraud, necessitating analytical testing to protect consumers by Michelle Dotzert

Olive oil consumption is associated with a myriad of benefits to human health, attributed to the monounsaturated fatty acids, α- and γ-tocopherols, phytosterols, phenolic compounds, and flavonoids it contains. In ancient Greece, Hippocrates called it

“the great healer,” and Homer referred to it as “liquid gold.” It is an essential component of the widely popular Mediterranean diet, and is marketed to consumers as a healthy option, often with images

of olive groves in idyllic Mediterranean locales. Unfortunately, olive oil is an ingredient most at risk for food fraud, and the daydream associated with this marketing quickly dissolves when the bottle is shown to contain a mix of olive oil and any of several less expensive oils such as soybean, corn, or rapeseed.

Since a story about the adulteration of olive oils was published in The New Yorker in 2007, it has continued to make headlines. However, the American olive oil market remains one of the largest outside Europe,

and a variety of quality tests, methods, and limits have been included in United States Department of Agriculture (USDA) regulations to protect consumers. Other techniques have been developed to support rapid, simplified testing.

Not all olive oil is created equal

Olive oil is very similar to a freshly squeezed juice. Processing begins when olives are harvested by hand or by machines that shake olives from the tree into drop nets below. The olives are then washed and processed within 24 hours to minimize oxidation.

Stone mills—and more recently, motorized mills—are used to crush the olives into a paste, which is then mixed so that various enzymes may begin to produce aromas. The paste is then pressed with a hydraulic press and the remaining liquid is centrifuged to separate the oil from water.

According to the USDA, for a product to be classified as extra virgin olive oil (EVOO), it must be obtained solely by mechanical or physical means, under thermal conditions that do not lead to any alterations in the oil, and no treatments other than washing, decantation, centrifugation, and filtration may be employed. Alternatively, products classified as olive oil must be obtained solely from the fruit of the olive tree (Olea europaea L.).

Olive pomace is a solid residue that results from pressing or centrifugation. Oil extracted from pomace, called crude olive pomace oil, is refined with solvents and other physical treatments prior to human consumption. Olive pomace oil consists of a blend of refined pomace oil and virgin olive oils, and while safe for consumption, lacks the polyphenols found in EVOO.

Why and how oil adulteration occurs

Xylella fastidiosa is a bacterial plant pathogen, commonly referred to as olive leaf scorch, or olive tree leprosy. It colonizes the xylem network within the plant, preventing the flow of water from the roots to the leaves, killing the plant. It was first detected

in olive trees in the Lecce province in Apulia, Italy in 2013, and is estimated to have destroyed up to one million trees. The resulting poor harvest may have contributed to increased sales of low quality or adulterated oil.

Criminal organizations have also been suspected of adulterating olive oil for profit. In May 2019, Europol’s Intellectual Property Crime Coordinated Coalition

and the Italian NAS Carabinieri, and the Tribunal of Darmstadt in Germany arrested 20 individuals and seized 150,000 liters of fake olive oil—sunflower oil with added chlorophyll, beta-carotene, and soybean oil.

What this means for consumers

Olive oil adulteration has implications for the consumer and the company selling the product. Experts advise that the consumers will receive a lower quality product than they paid for, lacking the flavor and health benefits for EVOO. This can ultimately

damage a company’s reputation, and pose legal and financial issues.

Contamination with other compounds can also pose health risks to the consumer. Routine testing is performed for other contaminants such as pesticides and heavy metals, as well as contaminants that

are introduced during processing, such as 3-MCPD (3-monochloropropane-1,2-diol esters) and glycidyl esters. These are heat-induced contaminants that form during the deodorization step of refining, and

can be used to determine if refined and pomace olive oil was added to virgin olive oil.

There are regulations in place in the US and the EU to ensure olive oil quality. For a description of the quality criteria for different olive oil grades, consult USDA regulation §52.1539. In the EU, Commission Regulation (EU) No 61/2011 defines how EVOO must be produced and what the EU quality criteria are. A major component of quality assurance is testing.

Olive oil testing methods

Experts describe two different methods of analysis. The differing oil compositions, such as the fatty acid profile, allows the analysis of the detailed

composition of an oil to determine if it is pure EVOO. If adulteration has occurred, some components will be present in higher or lower concentrations than

in pure EVOO. This detailed composition can be determined with a combination of liquid (LC) and gas chromatography (GC), which are also suitable for identifying organic contaminants. Heavy metals may be tested with inductively coupled plasma mass

spectrometry. Alternatively, infrared (IR) spectroscopy may be used to fingerprint the oil, looking for patterns rather than individual component levels.

GC and LC methods offer some advantages, and are outlined in the USDA regulations. Chromatographic analysis is ideal for separating the individual components of complex mixtures, offering a high degree of resolution and precision, and requires

a small sample volume. However, although these

methods are standardized, they are laborious, with sample preparation taking several hours. This is a barrier for a facility, such as an oil refinery, as they won’t be able to ensure the authenticity of every truckload. In contrast, more rapid techniques, such as IR spectroscopy, enable analysis within minutes and do not require highly trained technical staff.

UV-visible spectroscopy (UV-Vis) can also be used to differentiate between oils in a sample. Lower quality oils may contain unsaturated hydrocarbons that absorb UV light in the 200-300 nm spectral range.

Measuring with UV-Vis can help identify a high absorption within this wavelength range to identify a lower quality oil.

Industry leaders and regulators rely on a range of instrumentation to ensure olive oil—and countless other food products—meet quality standards to protect consumers. Of these, rapid testing techniques like IR and UV-Vis spectroscopy are ideal for identifying fraudulent products at numerous points in the supply chain.

Contaminants

Regulated Drinking Water

Regulated Drinking Water Contaminants Infographic

Regulated by the Environmental Protection Agency (EPA), testing is required for over 90 contaminants in drinking water. Further, the EPA sets water- testing schedules and testing methods that must be followed to ensure that contaminant levels remain below that which protects human health and that can be achieved using the best available technologies.

Download the full infographic compliments of Lab Manager

1

Microorganisms

Contaminants

sources

Microorganisms in drinking water can result in severe acute symptoms and gastrointestinal illness.

Cryptosporidium Giardia lamblia Legionella

Total coliforms( Including E. coli) Viruses (enteric) Heterotrophic plate count Turbidity

Human and animal fecal waste Naturally present

Soil runoff

2

Disinfectants

Long term exposure to disinfectants in water can result in eye and nose irritation, stomach

Contaminants

sources

discomfort, and anemia.

Chloramines Chlorine Chlorine dioxide

Water additive to control microbial growth.

Resource Guide

How to Improve Safety for Food, Water, and Edible Oils

ICP-MS, AAS, and ICP-OES solutions to improve food safety testing methods, speed up your work, and identify food fraud

Food Safety and Quality Control

Improving Food Safety Testing Methods

Identifying Food Fraud with Analytical Testing

1

2

Be	0.5	5	50	100	500
Al	0.5	5	50	100
V	0.5	5	50	100
Cr	0.5	5	50
Mn	0.5	5	50
Co	0.5	5	50
Ni	0.5	5	50	100
Cu	0.5	5	50	100
Zn	0.5	5	50	100	500
As	0.5	5	50	100	500
Se	0.5	5	50	100	500
Mo	0.5	5	50	100
Ag	0.5	5	50
Cd	0.5	5	50	100
Sb	0.5	5	50	100
Ba	0.5	5	50	100	500
Hg	0.5	5	50
Tl	0.5	5	50	100
Pb	0.5	5	50	100
Th	0.5	5
U	0.5	5

Be	0.5	5	50	100	500
Al	0.5	5	50	100
V	0.5	5	50	100
Cr	0.5	5	50
Mn	0.5	5	50
Co	0.5	5	50
Ni	0.5	5	50	100
Cu	0.5	5	50	100
Zn	0.5	5	50	100	500
As	0.5	5	50	100	500
Se	0.5	5	50	100	500
Mo	0.5	5	50	100
Ag	0.5	5	50
Cd	0.5	5	50	100
Sb	0.5	5	50	100
Ba	0.5	5	50	100	500
Hg	0.5	5	50
Tl	0.5	5	50	100
Pb	0.5	5	50	100
Th	0.5	5
U	0.5	5

Be	0.5	5	50	100	500
Al	0.5	5	50	100
V	0.5	5	50	100
Cr	0.5	5	50
Mn	0.5	5	50
Co	0.5	5	50
Ni	0.5	5	50	100
Cu	0.5	5	50	100
Zn	0.5	5	50	100	500
As	0.5	5	50	100	500
Se	0.5	5	50	100	500
Mo	0.5	5	50	100
Ag	0.5	5	50
Cd	0.5	5	50	100
Sb	0.5	5	50	100
Ba	0.5	5	50	100	500
Hg	0.5	5	50
Tl	0.5	5	50	100
Pb	0.5	5	50	100
Th	0.5	5
U	0.5	5