pat

test arrondi

North and South America Sales Headquarters+1 727-733-2447

European Sales Headquarters+31 26-3190500

France Sales Office+33 442-386 588

Germany Sales Office+49 711-341696-0

UK Sales Office+44 1865-811118

Asia Sales Information (All Countries) +86 21-6295-6600

Japan and South Korea+82 10-8514-3797

India, Middle East, Africa Sales Information +31 26-3190500

Australia/NZ Sales Information+61 8-8443-8668 (AU) +64 9-273-2080 (NZ)

Contact Our Sales Specialistsinfo@oceanoptics.com

Table of Contents

The Spectroscopy of Food............................................. 3

Absorbance Spectroscopy............................................ 4

Reflectance Spectroscopy.............................................6

NIR Spectroscopy...............................................................7

Fluorescence Spectroscopy...........................................8

SERS..................................................................................9

Chemometrics.................................................................10

Download our Authentication & Anti-Counterfeit brochure at oceanoptics.com/authentication

2

FLUORESCENCE

ABSORBANCE

NIR SPECTROSCOPY

REFLECTANCE

SURFACE-ENHANCED RAMAN SPECTROSCOPY

Food Security and the Harmful Risk of Fraud

A century ago, most food was sourced locally, and the world’s resources supported a mere ~25% of the current population. Since then, advanced processing of food and globalization of trade has reduced prices, increased selection, and made many seasonal items available year-round. With this convenience has come an increase in food fraud – substitution or dilution with lower quality ingredients to reduce cost, and/or adulter-ation or treatment with banned ingredients.

Food fraud is a global problem, affecting some 10% of our food supply; thousands of tons of substandard food are seized by Interpol each year. Though food fraud is often aimed at deceiving affluent food lovers, severe incidents can result in poisoning, illness or death. Improved methods are needed to validate the safety and authenticity of foods to protect both consumers and the brands they consume. Fortunately, spectroscopy is beginning to offer solutions that are robust and portable enough for deployment into the food chain, at customs, and even for consumers.

Processed foods like powders and liquids are most frequently adulterated; additives like colorants and flavors can be used to mimic the authentic product, and dilution or replacement can be difficult to detect. Premium liquors and spirits are a popular target (see application note p. 4), while lower-grade wines may be passed off as expensive vintages. Illegally produced “honey” represents 7% of all food fraud cases, from falsified origin to presence of banned antibiotics and pesticides. Even meats are at risk for adulteration, with recent scandals involving addi-tion of horse meat and organs to ground beef. Absorbance,

reflectance and Raman spectroscopy systems from Ocean Op-tics have proven useful for authentication and safety testing of these foods, as have fluorescence techniques sensitive to pro-teins, DNA, polyphenols and vitamin content.

Produce is less vulnerable, but can be difficult to test due to the need for nondestructive measurement. Reflectance and through-sample transmission techniques in the visible and NIR work well for fruit and dairy product testing. SERS, meanwhile, is gaining ground for detection of banned pesticides and fungicides on produce, as well as antibiotics and fungicides used in illegal fish farming. Ocean Optics’ highly selective SERS substrates offer the sensitivity and low cost per use to make field testing feasible. We’ve even devel-oped a complete fluorescence-based optical subassembly for an OEM instrument to test against regulated mycotoxin limits in food and feed products (www.ToxiMet.com).

The Spectroscopy of Food

Absorbance SpectroscopyAbsorbance measurements probe the chemical composition of a food directly via the wavelengths absorbed when light is trans-mitted through it. Absorbance is most often used for liquids, as Beer’s Law simplifies quantitative analysis. The optical density and water content of the sample must be considered in order to determine the right pathlength for measurement. Samples with high absorption or scatter like colloids (milk) or suspensions (mixed vegetable juice) may require chemometric analysis. Temperature is also important for aqueous samples, and for oils that might be solid at ambient conditions. It is also possible to look at the light transmitted through whole foods, which has proven a viable method for inspecting fruit for ripeness, internal rot, pests and defects.

√ Vegetable oils: identity, adulteration, acid value, peroxide value

√ Saffron: ISO 3632 quality method to measure crocin, picrocrocin and safranal

√ Orange juice: soluble solids content, pH, adulteration with grapefruit juice

√ Mushrooms: detection of bruises, evaluation of freshness

√ Wine: quality, phenols, tannins, methanol content

√ Milk: fat, protein and casein content

Application Note: Authenticating SpiritsChallenge: Top-shelf liquor commands premium prices, tempting resellers and serving establishments to pass off substandard or diluted product in high-end brand bottles

Solution: UV-Vis absorbance signatures allow the brand, integrity, and distillery of origin of a spirit to be verified by comparison against a known library in a turnkey system

4

High-end spirits and liquors are often subject to tampering and counterfeiting, with impact to brand valuation and revenue. We developed a UV-Vis absorbance system capable of measuring the “brand profile,” of spirits like rum, whisky and vodka in the field, acquiring a spectral signature that uniquely identifies not only the brand, but the grade or distillery of origin. Through extensive measurements in our applications lab, we created brand profiles for each spirit of interest, and developed a partial least squares analysis method. Data can be acquired by a non-

expert in the field and automatically compared to a known library. Authenticity is validated with a high confidence factor in under 5 seconds through a traffic light-style soft-ware interface to provide clear, yes/no validation. The Spirit Sampler has also been adopted by several distilleries for routine batch characterization and measurement of the EBC color index. When configured with an autosampler and barcode scanner, it can be used to analyze a large number of samples for more efficient and accurate process control.

Application Note: Detecting Olive Oil FraudChallenge: Global demand for olive oil now exceeds supply; sales of adulterated and diluted olive oil are on the rise, yet no convenient method of validating quality exists

Solution: Absorbance spectroscopy offers a quick and easy method to detect dilution with lower grade olive oil, or adulteration with other oils, flavors and colorants

Dilution of extra-virgin olive oil with lower grades is the most common type of olive oil fraud. EVOO’s characteristic dark green color persists even in 50:50 mixes, making it difficult to authenticate by eye. Absorbance spectra, however, show a clear progression correlating to the fraction of EVOO in the

Chlorophyll is often added to lesser-grade edible oils and sold as authentic olive oil. We simulated this adulteration with rapeseed oil, comparing it to a low-grade olive oil. While the two were difficult to distinguish by eye, the absorbance spectra are clearly different. Chlorophyll increases the absorbance of the rapeseed oil from 415-510 nm enough to alter its color, but cannot mim-ic the true spectrum of olive oil. Software analysis based on simple peak height ratio comparisons would allow adulterated rapeseed oil to be easily distinguished from pure olive oil.

From validating product grade and identity to detecting the presence of artificial colorants and flavorings, absorbance spec-troscopy is a powerful method for rapid quality testing of liq-uid food products. Compact sensors like the Spark-VIS offer a high-performance, low-cost option for deployment of food validation measurement systems throughout the supply chain.

The cost of raw olive oil has reached its highest levels in a decade, doubling in some areas in 2014 due to back-to-back years of crops devastated by drought, pests and bacterial disease in Spain and Italy. The gap between supply and demand motivates producers to favor quantity over quality, impacting the price and integrity of extra-virgin olive oil (EVOO) in particular. By some estimates, 70% of EVOO on American store shelves has been adulterated in some way to reduce cost. It may be cut with lower grade olive oil, or with substitutes like sunflower, rapeseed, soybean and canola oil. Chlorophyll is added to imitate the dark green color of EVOO, while beta-carotene helps mimic its taste.

Absorbance spectroscopy using visible light is able to quantify even small differences in color and composition in oils and oil blends indiscernible to the human eye and palate, sounding the alarm on substandard olive oil. We assembled a modular absorbance system based on our compact Spark-VIS spectral sensor to measure various oils and mixtures in a cuvette from 380-700 nm.

mixtures tested, from pure EVOO to 100% low-grade olive oil (see graph at bottom left). The absorbance peak at 670 nm is a useful feature for this analysis, as it is quite pronounced in EVOO. Features from 380-500 nm also vary with EVOO fraction, offering sufficient spectral differences between EVOO and low-grade olive oil to enable effective chemometric analysis for authentication purposes.

Reflectance SpectroscopyReflection yields similar information to absorbance when applied to solid foods, with the depth of penetration being dependent on factors like the wavelengths used, sample composition and structure, and illumination source intensity. It has the benefit of being non- contact and noninvasive, and can be configured for a wide range of samples. Bulk samples like meats and fruit are usually measured in a stand-off configuration, but a window is typically used for measurement of ground samples and powders. Rotating the sample is used to increase the measurement area and reduce error due to sample nonuniformity for coarse or large-particle samples like cereals, soybean or cornmeals, crushed processed foods and powdered foods or drinks. The greatest amount of information can often be extracted about food quality or integrity when visible and NIR reflectance measurements are combined to cover 400-2500 nm, which can be achieved using a tungsten halogen lamp with Flame-S-VIS-NIR and NIRQuest512-2.5 spectrometers for detection.

√ Fruit: variety, ripeness, sugar content, soluble solid content (SSC), internal pests

√ Whole grains: protein, moisture, oil content

√ Meat: distinguishing pasture-fed from concentrate-fed livestock

√ Crabmeat: substitution with surimi-based imitations and lower-grade crabmeat

√ Purees: adulteration of strawberry or raspberry with lower-cost apple

6

Application Note: Spotting Adulterated SpicesChallenge: High-value ground powders like spices are frequently adulterated, as it is often difficult for consumers and importers to evaluate spice quality based on sensory input alone

Solution: Reflectance spectra can reveal features signaling common spice adulterants, providing rapid quality assessment without the need for chemical analysis

Spices are being used more than ever as consumers seek to reduce salt intake and expand their palates. Sold most often as ground powders, spices like turmeric, paprika and ginger can easily be cut with fillers, from less expensive spices to flours, corn starch and even sawdust. In extreme cases like the large-scale 2005 recall of chili powder due to adulteration with the red dye Sudan 1, toxic or potentially carcinogenic substances are added to hide aging of a product or the presence of fillers.

Modular reflectance spectroscopy offers a rapid, cost-effective alternative to chemical analysis, and allows the ground powder to be measured directly. To demonstrate this, we looked at ginger, a spice sometimes cut with allspice to reduce cost. The two spices were mixed in various ratios. We pointed a reflection probe at 45° to the underside of a glass dish filled with spice, measuring the spectrum using a tungsten halogen lamp and Flame-S-VIS-NIR spectrometer. Though both spices are much more reflective at longer wavelengths, allspice possesses a characteristic dip in its spectrum at ~670 nm, visible even at levels as low as 10%

of the ginger-allspice mixture. This spectroscopic analysis demon- strates the speed and simplicity with which the purity and authenticity of products can be validated within the food supply chain.

NIR SpectroscopyNear infrared spectroscopy is particularly well-suited to nondestructive analysis of bulk, high-moisture samples like fruit, fish, meat and grains. While more challenging to interpret and analyze than visible or infrared, NIR spectroscopy probes the vibrational overtone absorption of chemical bonds, predominantly CH, OH and NH, and is therefore sensitive to most chemical constituents in a given food. The resulting spectra are often broad, overlapping and complex, necessitating chemometric analysis to unlock its secrets. An-other benefit is that light at NIR wavelengths penetrates fairly deeply with less scattering, allowing internal composition to be probed via nondestructive reflectance and through-sample transmission techniques. When processed using a well-developed chemometric model, a simple NIR reflectance spectrum of a food can be used to predict complex characteristics, such as an apple’s ripeness, sweetness or storage duration.

√ Fruit quality: screening for core rot, internal pests and ripeness

√ “Gluten-free”: sorting unprocessed grains with NIR and machine vision

√ Adulterated ground beef: detecting mutton, pork, organs and fillers

√ Fraudulent labeling of fish: identification of fish species without DNA testing

√ Chicken quality: detecting thawed versus fresh; artificially boosted water content

Application Note: Measuring Protein in SoybeansChallenge: Rapid measurements of protein content are needed for high-throughput sorting of soybeans at harvest, and for cost-efficient process feedback during soybean meal production

Solution: NIR reflectance measurements offer the speed and accuracy to measure protein at both points in the supply chain, enabled by chemometric analysis and automation

Soybeans are valued for their oil and as a source of protein-rich meal for animal feeds. The protein content of soybeans is heavily dependent on the nitrogen content of the soil in which they are grown, and also on farming methods. During the soybean harvest, a high volume of soybeans must be sorted by protein level in a short period of time. While traditional analytical methods cannot keep up, NIR reflectance spectroscopy has emerged to provide high-throughput, automated analysis in the form of easy to use instruments offering clean operation in a small footprint.

Soybeans fed continuously through the InLAB NIR 512 system from TecnoCientifica pass by a window, through which reflection measurements are made by a NIRQuest512 spectrometer (900-1700 nm). The spectra undergo PLS multi-parameter chemometric analysis to yield the protein level, with excellent agreement between predicted and measured values.

Also, the same system can be used for online process monitoring of soy meal and oil production. Effective monitoring ensures extraction of the maximum oil content while still meeting require-

ments for water, protein and residual oil content in the soybean meal. While up to tens of thousands of dollars can be lost per hour in delayed feedback from conventional or even “rapid” onsite lab testing, online NIR spectroscopy allows instant, continual process optimization and quality validation.

Image courtesy of TecnoCientifica, www.tecnocientifica.com

Fluorescence SpectroscopyFluorescence spectroscopy utilizes the native fluorophores in food products to detect the presence or concentration of those compo-nents. Fluorescence is already used as a detection method for HPLC and CE in chemical analysis, but offers a wealth of information when used as a standalone analytical technique. Fluorescence spectra are composed of broad, overlapping emission bands contain-ing information about the sample components, which must be extracted through mathematical analysis. Fluorescence analysis may look for peaks to change in intensity, or to shift in wavelength or bandwidth. Different excitation wavelengths may be used to build up a fluorescence excitation-emission matrix (EEM) that can be reduced to a 3D component space for correlation to concentration via chemometrics. It can be performed in right angle or front face fluorescence configurations, adapting to provide rapid, sensitive, nondestructive analysis of a wide variety of food types.

√ Food and animal feed: mycotoxins in nuts, grains, dried fruits and spices

√ Fruit: fecal matter contamination of peel

√ Cheese: variety, manufacturing method, geographic origin and degree of aging

√ Meat: quantification of fat, bone, cartilage and connective tissue

√ Eggs: freshness

Application Note: Assessing Honey PurityChallenge: With attractive nutritional and medicinal benefits, and its appeal as a traditional product, pure honey is a blossoming target for adulteration with sugar syrups and more

Solution: Characterization of fluorescence spectra for pure, adulterated and treated honey enables rapid, accurate discrimination of honey samples

8

As a premium, natural product, honey is a frequent target of food fraud, by some accounts ranking just after olive oil and milk as most commonly counterfeited. Honey is adulterated with sugar syrups, starch and water to fool the consumer with a product that is less than pure. Detecting adulterated honey is a challenge due to natural variations in its composition arising from nectar differences.

Fluorescence spectroscopy is an effective method for distin-guishing pure from adulterated honey, as various honey constituents and adulterants have identifiable fluorescence response. Researchers have confirmed this using an Ocean Optics spectrometer for fluorescence detection and various laser wavelengths for excitation to make laser induced fluo-rescence (LIF) measurements of honey. We performed similar measurements on honey samples using a QE Pro-FL spectrom-eter and 365 nm LED (LLS-365) as the fluorescence excitation source. Differences in fluorescence intensity and subtle differ-ences in spectral shape were observed for all samples.

Variations in the shape of these spectra are attributed to the flavonoid composition of the nectar used to make the honey. Measurements could be expanded to use a range of laser and LED wavelengths to optimize the excitation wavelength for the fluorescence detection of honey adulterants, with setups configured for the lab or line.

SERS Surface-enhanced Raman spectroscopy can be used for trace detection of harmful compounds in foods, from pathogens to antifungal dyes, antibiotics and pesticides. Trace contaminants and residues used in the growth, processing and storage of food products have been shown to be toxic to humans, with effects ranging from digestive problems to death. SERS uses metal colloids or nanostruc-tures to enhance the Raman effect by a factor of up to 100 trillion, extending the detection ability of Raman spectroscopy to ppm or ppb concentrations. SERS substrates may be gold or silver-based, and can be made even more specific through functionalizing ligands specific to a particular pathogen. Ocean Optics SERS substrates offer better performance than expensive patterned substrates at a fraction of the price, and are compatible with our IDRaman mini handheld detection system for rapid, on-site testing of food products with little to no sample preparation.

√ Pesticides: organophosphorus and sulfur-containing pesticides, tricyclazole, malathion

√ Pathogens: Salmonella enterica and Staphylococcus aureus on fresh spinach

√ Restricted antibiotics: enrofloxacin, ciprofloxacin and chloramphenicol

√ Banned antifungal agents used in aquaculture: crystal violet and malachite green

√ Clenbuterol: an illegal growth-promoting drug used in food animals

√ Sudan 1: a carcinogenic, mutagenic dye used to boost color of chili powder

Application Note: Finding Melamine in MilkChallenge: In the 2008 Chinese tainted milk scandal, melamine was added to watered-down infant formula, fooling tests designed to ensure protein content met acceptable levels

Solution: Surface-enhanced Raman spectroscopy (SERS) offers a rapid, field-portable alternative to chromatographic and other screening techniques for melamine and other contaminants

9

An estimated 300,000 babies became ill in China after being fed milk powder adulterated with melamine, a flame- retardant plastic most often used in industrial products. Six infants died, and more than 50,000 were hospitalized with symptoms of kidney stones, kidney damage and malnutrition. Products from more than 20 companies were involved, demon-strating the prevalence of economically motivated food fraud in industry. Melamine contamination has also been discovered in eggs, possibly the result of use in animal feed, spurring industry to investigate methods for its detection.

Highly sensitive and molecule-specific, SERS is capable of detecting even trace levels of melamine in milk products and other foods, with minimal sample preparation. We tested infant formula doped with melamine at several concentrations, applying it directly to our low-cost inkjet-printed SERS substrates. Unlike other testing methods, no separation of components was required prior to analysis. Samples were tested using a 785 nm QE Pro modular Raman system, with pure formula used as the

dark reference. Strong SERS signals allowed melamine to be seen and quantified via a calibration curve down to 10 ppm levels, showing the power of this technique to perform sensitive, quantitative detection of melamine in milk products.

ChemometricsChemometrics is a powerful tool for analyzing the optical spectra of complex chemical systems like foods, extracting information about chemical composition or identity that would otherwise require laboratory analysis and/or expert evaluation. It can be used to address a wide range of issues related to food quality and safety and be applied at all points in the supply chain, from producer to consumer.

Application Note: Predicting Apple SweetnessChallenge: Determining apple sweetness currently requires testing with a refractometer to measure the sugar content in degrees Brix (°Bx), but destroys the fruit in the process

Solution: NIR reflectance measurements analyzed with chemometrics can be used to build an accurate model to predict the sugar content of apples rapidly and nondestructively

10

Light at NIR wavelengths penetrates deep into the tissue of fruits, probing the glucose, fructose and sucrose absorption bands that span 1400-1600 nm. We collected diffuse reflec-tion measurements for 76 Ginger Gold apples using a tung-sten halogen light source and Flame-NIR spectrometer (950-1650 nm). Both ripe and unripe apples were used, and their Brix values determined in a separate lab analysis. The apple spectra were scaled to compensate for slight differences in the measurement geometry and apple shape, followed by mean-

centering (subtracting the average from each spectrum to isolate spectral variations). PLS analysis was performed against the sweetness in °Bx, using two-thirds of the apples as the training set and the balance as a test set for cross-validation. Best model performance was achieved by including 5 compo-nents, yielding an error of 0.3°Bx. When compared to the 4°Bx variation seen between ripe and unripe apples of the same variety, this demonstrates excellent resolution for evaluating apple sweetness accurately and nondestructively.

Application Note: Identifying Apple VarietyChallenge: Some species, varieties or types of foods are sold at a premium price, but resellers and consumers have no way to validate their identity without expert evaluation

Solution: Chemometric analysis offers a rapid, noninvasive method to accurately classify many foods based on their reflectance spectraw

Chemometrics is particularly useful for analysis of reflectance spectra of foods, allowing nondestructive testing of bulk, high-moisture samples like fruit, fish, meat and grains. While absorbance measurements of liquids typically use a known pathlength to correlate absorbance to concentration, reflec-tance measurements probe an unknown depth and include scattering from within the sample. Chemometrics uses sophis-ticated mathematical approaches to determine the concentra-tion of a component or identity of a sample, even when spec-tra involve hundreds of individual components and/or complex light-matter interactions.

Chemometrics requires its own special type of calibration using a large number of samples for which the analyte or identity is known. Chemometric software tools utilize machine learning algorithms to develop a model, and thus the calibration process is usually referred to as “training,” followed by validation of the model using an additional “test set” of samples. Once created, the chemometric model can be used to analyze a new spectrum to predict the quantity or quality for which it was developed, such as fat content or type of meat. While establishing a model requires significant initial investment, it is quickly returned in the savings to be gained from rapid, nondestructive predictions in the field, eliminating the need for lab analysis.

A number of mathematical methods are available for chemo-metrics through commercial software packages. Quantification is well suited to Partial Least Squares Regression (PLS), which determines sets of spectra (the “components”) that can most effectively explain the variations in the concentration of the analyte. Support Vector Machine (SVM) is popular for classi-fication, while discrimination often uses Principal Component Analysis (PCA) to compare a sample against a standard.

Online batch quality monitoring enabled by chemometrics

To determine whether we could identify an apple’s variety, we collected 576 reflectance spectra of Ginger Gold (GG), Golden Delicious (GD), Granny Smith (GS), Fuji (FF), Honeycrisp (HC) and Macintosh (MC) apples. A tungsten halogen light source and visible Flame spectrometer were used to probe skin pigments like carotenoids (420-500 nm), anthocyanins (540-550 nm) and chlo-rophyll (600-700 nm). Spectra were pre-processed, scaling with SNV and mean-centering to isolate spectral differences. To visu-alize, we performed PCA, identifying two principal components that contributed most to spectral variations between the apples.

Different apples “scored” different principal component values. When plotted in two dimensions, the apple varieties fell into three groups: green, yellow-green and red apples, illustrating visually how chemometric analysis extracts meaningful infor-mation from complex spectra taken on a large test group of samples. To take classification one step further, we performed SVM analysis on the same set of data, developing a model for

specific apple variety that yielded a cross-validation classifica-tion error of only 0.3%. In fact, the model was able to classify all 144 apples in the test set accurately, despite many being similar in color.

Modular Spectroscopy of FoodsSpectroscopy can be a powerful tool for authentication and quality control of foods, providing rapid, nondestructive and quantitative results. Our flexible, modular products are easily mixed and matched, allowing our Application Scientists to configure a system with the best spectrometer, light source and accessories for your wavelength range, measurement technique and application.