Introduction
Fruit juice is made from a wide variety of fruits, including oranges, apples, grapes, cranberries, grapefruits, tomatoes, bayberries, and pineapples. The major components of fruit juice are water, sugars, and organic acids with lesser amounts of amino acids, vitamins, and phenolic compounds. In the United States, the term “fruit juice” can only be legally used to describe a product which is 100% fruit juice. A blend of fruit juice with other ingredients is referred to as a “juice cocktail” or “juice drink”. Some pure fruit juices are blended together as well, making blend monitoring an important component to observe and measure for proper taste and flavor. Sugar and acidity measurements are the most important constituents and are strictly regulated in marketed juices. Soluble Solids Content (SSC, expressed as °Brix), glucose, fructose, and sucrose are the most common sugar measurements while acidity measurements include Titratable Acidity (TA) and pH. As is the case with many liquid foods, improper or overly lengthy storage leads to oxidation, which can reduce nutritional value and even present potential health hazards. Adulteration is an emerging problem in the food and beverage industry and fruit juice is no exception. New methods for fraud and adulteration in fruit juice are continuously being developed and likewise, new methods for detection must also progress to keep up. The most common forms of adulteration in fruit juice include water dilution, artificial sweetener addition, and the addition of lower quality products and fruit juice. In the case of water, a test for SSC and pH can determine if a fruit juice has been diluted with pure water but adding water spiked with sugar and citric acid can make the adulterant undetectable by simple testing methods. There is a need to measure and determine these quality parameters at all stages of the fruit juice manufacturing process from the initial analysis of sugar and acidity parameters all the way to determining if the final product of fruit juice is what is actually being labeled and marketed. Currently, methods for testing these parameters such as HPLC are expensive, laborious, and time-consuming, especially when implemented in a process setting. There is a need for fast, cost-effective, and real-time monitoring of parameters at all stages of the fruit juice manufacturing process. One such method that has been examined is NIR spectroscopy.
Analytes
- Glucose
- Fructose
- Sucrose
- Soluble Solids Content (SSC, expressed as °Brix)
- Titratable Acidity (TA)
- SSC/TA Ratio
- Classification of Tomato Juice (Fresh and Aged For One Month)
- pH
- Identification of Blueberry Beverage
- Saccharin Adulteration
- Classifying Natural and Synthetic Lime Juice
Summary of Published Papers, Articles, and Reference Materials
Measurement of chemical parameters in fruit juice has been studied using NIR spectroscopy and the results of most studies have demonstrated the potential of using NIR spectroscopy as a replacement for expensive and time-consuming wet chemistry methods. A study is presented that used stock solutions of glucose, fructose, and sucrose and NIR spectra to create calibration models for measuring these parameters in various kinds of fruit juice. Results were excellent and show the feasibility of this method as a quality control parameter or to detect adulteration or contamination in fruit juice. Likewise, one study examined measuring the same three sugar parameters in bayberry juice by correlating NIR spectra to HPLC reference values. The results were good for sucrose and fructose but correlation was lower for glucose. This is most likely due to a small range of glucose in the reference values and results should improve with a wider range of samples in the calibration set. SSC, TA, and SSC/TA ratio are considered three primary components affecting the taste of fruit juice and measuring these parameters using NIR spectroscopy of apple juice was examined. Results were good considering the different types of samples were both clear and cloudy. Oxidation is known to have an effect on the nutritional components of fruit juice and NIR spectroscopy was examined as a method for determining changes in quality in tomato juice that were stored for a month. Clear spectral differences were observed in the two groups of samples and classification analysis discriminated with 100% accuracy. SSC and pH tests confirmed that chemical changes did take place in the samples over a period of a month, confirming the validity of the analysis. Orange juice is one of the most popular fruit juices and one study measured SSC and pH quality parameters using NIR spectroscopy. Correlation was excellent for both parameters. Blueberry beverage is another popular fruit juice and NIR spectra were successfully used to classify four different types of it: a single blueberry juice (by definition with no additives present) and three other blueberry beverages (allowed to have additives). Adulteration is a major issue in the fruit juice market and new types of frauds are created regularly. One study examined detecting and quantifying saccharin adulterant in different commercial fruit juices. Both detection and quantification were proven feasible using NIR spectra and regression models. Another form of adulteration is presenting a synthetic product as natural and samples of lime juice were procured for this purpose. Both pure and synthetic samples were scanned using a NIR spectrometer and various classification algorithms and data mining techniques were applied. The best method accurately classified the samples at a 97% rate. Overall, the studies discussed here prove that NIR spectroscopy and regression models can be used as a quality control tool and method for adulterant detection in fruit juices.
Scientific References and Statistics
Rapid Analysis of Sugars in Fruit Juices by FT-NIR Spectroscopy – Rodriguez-Saona, Fry, McLaughlin, Calvey, Carbohydrate Research 336 (2001) 63-74
NIR spectroscopy was examined as a method for measuring glucose, fructose, and sucrose in fruit juice, which are all important sugar parameters for determining quality as well as detecting adulteration or contamination. Analytical grade samples of glucose, fructose, and sucrose were used to prepare stock solutions at concentration levels from 0 to 8 g/100 mL. FT-NIR spectra of the samples were collected from 10000 cm-1 to 4000 cm-1 at 2 cm-1 intervals operating at 8 cm-1 resolution. Sixty-four scans were collected per reading and averaged into one spectrum. Three separate methods were used for the spectra collection: transmittance using a 0.5 mm cell, transflection using a reflectance accessory, and reflectance using a fiberglass paper filter. Various pretreatments were performed on the spectral data and Partial Least Squares (PLS) regression models for glucose, fructose, and sucrose were created for all three sets of spectra.
Transmittance | |||
Glucose | R2= | 0.9990 | RMSEP= 0.086 g/100 mL |
Fructose | R2= | 0.9998 | RMSEP= 0.038 g/100 mL |
Sucrose | R2= | 0.9994 | RMSEP= 0.069 g/100 mL |
Transflectance | |||
Glucose | R2= | 0.9960 | RMSEP= 0.171g/100 mL |
Fructose | R2= | 0.9968 | RMSEP= 0.157 g/100 mL |
Sucrose | R2= | 0.9980 | RMSEP= 0.134 g/100 mL |
Reflectance | |||
Glucose | R2= | 0.9345 | RMSEP= 0.704 g/100 mL |
Fructose | R2= | 0.9568 | RMSEP= 0.585 g/100 mL |
Sucrose | R2= | 0.9648 | RMSEP= 0.501 g/100 mL |
Calibration results were best using transmittance spectra and were far better for transmittance and transflectance than for reflectance. In order to validate the models, commercial samples of both apple and orange juice were obtained. These samples were scanned using the same collection parameters as the stock solutions in transmittance and transflectance mode. They were not scanned in reflectance mode due to the lower correlation in those models. Variation was incorporated by scanning samples a week after the original spectra collection and also by spiking samples with additional glucose, fructose, and sucrose. The PLS models were used with the sample juice spectra to predict concentration of the three sugars and these values were compared with reference testing of the samples by High-Performance Anion-Exchange chromatography. Prediction results were best for all three parameters using transmittance spectra but excellent for both. The results in this study prove the feasibility of measuring sugars in fruit juices using NIR spectroscopy and calibration models created from stock solutions of glucose, fructose, and sucrose. Such a method provides many advantages to fruit juice manufacturers. Testing is rapid, non-destructive, accurate, and has the ability to measure all three sugar parameters at once instead of individual expensive chromatography tests for each parameter.
https://www.sciencedirect.com/science/article/abs/pii/S0008621501002440
Quantification of Glucose, Sucrose, and Fructose in Bayberry Juice – Xie, Ye, Liu, Ying, Food Chemistry 114 (2009) 1135-1140
Bayberry has been cultivated in Southeast China for more than two thousand years and the annual output is around three hundred thousand tons. It is known for health benefits and especially in treating gastrointestinal issues. Bayberry is processed into many different forms, including sweets, jam, juice, wine, or canned in syrup. It is abundant in sugar components like glucose, fructose, and sucrose. NIR spectroscopy was examined as a method for measuring quantitative determination of these three sugar components in bayberry juice. A total of one hundred twenty samples cultivated from the same harvest were used for the study. In order to create a range of values for the parameters of interest, samples of various bayberry species were obtained from different geographical regions. The samples were centrifuged to remove solid particles and scanned in a 1 mm quartz cuvette using an FT-NIR spectrometer in transmission mode. Resolution was 1 cm-1 and sixty-four scans were averaged per spectrum. Reference tests were conducted using HPLC for the three sugar parameters and Partial Least Squares (PLS) calibration models were created correlating the NIR spectral data to glucose, sucrose, and fructose.
Glucose | R2= | 0.855 | RMSEP= 0.0625 g/100 mL |
Sucrose | R2= | 0.993 | RMSEP= 0.0866 g/100 mL |
Fructose | R2= | 0.967 | RMSEP= 0.114 g/100 mL |
Correlation coefficients were very high for sucrose, high for fructose, and reasonable for glucose. The likely reason for the lower glucose correlation is a much smaller range of values for the samples and the fact that the RMSEP is lowest for glucose while also having the lowest correlation coefficient makes a strong case for this reasoning. There are studies for other types of juices for these three parameters that have shown nearly equal correlation, indicating that glucose correlation will improve with a sample set encompassing a larger range of values. The results here demonstrate the potential to use NIR spectroscopy and calibration models to measure glucose, sucrose, and fructose in bayberry juice, offering an alternative to current expensive and time-consuming reference tests as well as the ability to measure all three components simultaneously.
https://www.sciencedirect.com/science/article/pii/S03088146080129
Evaluation of Quality Parameters of Apple Juices Using Near-Infrared Spectroscopy and Chemometrics – Wlodarska, Khmelinskii, Sikorska, Hindawi Journal of Spectroscopy, Volume 2018 Article ID 5191283
NIR spectroscopy was examined as a method for measuring two important taste related parameters in apple juice: Soluble Solids Content (SSC, expressed as °Brix), Titratable Acidity (TA), and the ratio of these two parameters (SSC/TA). SSC indicates sweetness of fresh and processed fruit products, TA is related to organic acid contents that contribute to sour taste, stabilize color, and extend shelf life, and the SSC/TA ratio is related to the overall taste and is used as an index of sensory acceptability of fruit taste. Commercial samples of apple juice were procured for the study. They included clear and cloudy juices reconstituted from the concentrate, direct juices that were pasteurized, and freshly squeezed juices. Thirty juices from fifteen different producers were included and all of these were used in duplicate from two different production batches. FT-NIR spectra were collected from 12500 cm-1 to 4000 cm-1 at 8 cm-1resolution and averaging sixty-four scans per spectrum. Samples were centrifuged before collection and six replicated spectra were collected for each of the juices. Reference tests were conducted for SSC using a refractometer and TA using a pH meter, measuring all samples in triplicate and averaging the three values. Various pretreatments were used to process the NIR spectra and a selective wavelength algorithm was chosen to correlate the spectral data to the parameters of interest.
SSC | R2= | 0.881 | RMSEP= | 0.277 °Brix |
TA | R2= | 0.761 | RMSEP= | 0.239 g/L |
SSC/TA | R2= | 0.843 | RMSEP= | 5.04% |
The interval Partial Least Squares (iPLS) algorithm was chosen to create the calibration models. It works in the same manner as a normal PLS model correlating the NIR spectra to each parameter of interest but uses an iterative approach and selects specific wavelength ranges for model optimization. The results obtained from the models are consistent with literature data for measuring these parameters in apples and other types of fruit juices. Lower predictive ability is expected for acidity measurements compared to sugars due to a smaller concentration and lower spectral sensitivity of acids. It is also likely that using different types of juices increased prediction error and lowered the correlation. A larger universal sample set or data sets specific to individual types of juice should improve results. The results of this study do demonstrate the feasibility of using NIR spectra and calibration models for measuring these important quality parameters in apple juice.
https://www.hindawi.com/journals/jspec/2018/5191283/58
Use of Near-Infrared Spectroscopy and Least-Squares Support Vector Machine to Determine Quality Change of Tomato Juice – Xie, Ying, Journal of Zhejiang University Science B, 2009, 10(6):465-471
Tomatoes are grown worldwide and are the second most consumed vegetable in the world. Tomato juice is rich in organic acids, sugars, vitamins, and natural pigments. Vitamin C can be easily oxidized with exposure to air and pigments can decompose, reducing the nutritional components in juice. One hundred fully ripened tomatoes from a single harvest were procured for the study. Each sample was squeezed and centrifuged and two separate datasets were used for each sample. The first dataset was used immediately after centrifuging. NIR spectra were collected and reference values for Soluble Solids Content (expressed as °Brix) and pH were obtained. The process was repeated for the second dataset, but the samples were first stored for a month in airtight bottles and refrigerated. For both datasets, FT-NIR spectra were collected from 800 nm to 2400 nm using 4 cm-1resolution and averaging thirty-two scans per spectrum. External conditions were reproduced as closely as possible for data collection on both datasets. Difference in the reference values for SSC and pH between the datasets were evaluated and different classification algorithms were used to classify the NIR spectra based on immediate collection and the data collected a month later.
Classification Results: | |
Least-Squares Support Vector Machine (LS-SVM) Method | 100% Accuracy |
The reference tests on SSC and pH for the two datasets showed a marked change in both parameters, indicating that the juice samples were changing over time. Visual inspection of the spectra showed clear differences in the wavelength range from 2190 nm to 2270 nm. While it is difficult to quantify the exact chemical changes that create the differences, one possibility is O-H and C=O stretching of the carboxylic group of citric acid or other acids affected by oxidation. Four separate classification algorithms were tested and the LS-SVM method showed the best results with a 100% accuracy. These results were proven by a validation set that chose the correct classification every time. The precision and accuracy shown here indicate that NIR spectroscopy can be used as a tool to control the quality change of tomato juice during storage.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2689559/
Measurement of Soluble Solids Contents and pH in Orange Juice Using Chemometrics and Vis-NIRS – Cen, He, Huang, Journal of Agricultural and Food Chemistry, 2006, 54, 7437-7443
Orange juice is considered an important beverage for consumers and Soluble Solids Content (SSC, expressed as °Brix) and pH are two of the most important chemical parameters. SSC is comprised of mainly sugars, such as fructose, sucrose, and glucose. Pure fruit juice typically contains between 9% and 15% SSC. Acids are important sources of nutrition and freshness and these include citric, tartaric, and malic acids. pH also plays an important role in variation of color, microbial control, taste, and authentication of food. Measuring these parameters is crucial to determining orange juice quality and NIR spectroscopy was examined as a method for replacing traditional reference tests. Eight commercially available brands of orange juice were procured for the study. Each brand included different samples produced at different dates and some samples were diluted to change the values for SSC and pH. In all, one hundred four total samples were used in the study. Spectra were collected from 325 nm to 1075 nm at 3.5 nm resolution. The juice was poured into glass sample containers. Three spectra were collected for each sample and the container was rotated 120° twice after the first spectrum was collected. Reference values were obtained for SSC using a refractometer and for pH using a pH meter. Various data pretreatments were applied to the spectra before modeling and Partial Least Squares (PLS) calibration models were created correlating the spectral data to SSC and pH.
SSC | R2= 0.98 | RMSEP= 0.73 °Brix | |
pH | R2= 0.96 | RMSEP= 0.06 |
Due to noise in the spectra, only the wavelength range from 400 nm to 1000 nm was used for the models. The PLS calibration models proved the feasibility of the measurement. Correlation coefficients were high and prediction results showed error well within the acceptable range for the models to be used as a quality control tool. Visible spectral differences were apparent as well in wavelength absorbing regions related to changes in SSC and pH. This study provides a basis for using Vis-NIRS spectroscopy as a fast, non-destructive method for measuring these parameters simultaneously as opposed to separate reference tests.
https://pubs.acs.org/doi/abs/10.1021/jf061689f
Identification of Blueberry Beverage Using VIS/NIR Spectroscopy – Li, Wu, Ma, et al., MATEC Web of Conferences 139, 00050 (2017)
Blueberries are known as a strong antioxidant and contain many substances that are known for their potential health benefits, such as hypertension treatment, preventing and curing inflammation, and even inhibition of cancer cell growth. Blueberry juice is popular for its flavor and these potential health benefits. Identifying different varieties of blueberry juice is a complex and diverse process and there is a need for a rapid method to identify blueberry juice and beverage. VIS/NIR spectroscopy was examined as a method for this purpose. Four different kinds of blueberry beverage were procured for the study. One was identified as blueberry juice which by definition means there are no additives present. The other three are defined as blueberry beverage which means the presence of additives is acceptable. Samples were scanned from 350 nm to 1850 nm. The sampling interval varied based on the wavelength range: 1.4 nm from 350 nm to 1000 nm and 2 nm from 1000 nm to 1830 nm. Spectral resolution was 3 nm at 700 nm and 10 nm at 1400 nm. The samples were scanned at 2 mm pathlength and thirty scans were collected and averaged per spectrum. This process was repeated three times for each sample and the three total spectra were averaged into one spectrum per sample. One hundred of the samples were used to build a classification model for classifying the four types of samples and forty were used as a validation set to prove the feasibility of the model.
Classification Accuracy (10 of each sample) 100% | 100% |
Principle Component Analysis (PCA) was first performed to determine if clear grouping between the 4 types of samples could be determined. Analysis showed the following wavelength areas were relevant for the grouping: 420 nm to 430 nm, 490 nm to 500 nm, 570 nm to 580 nm, and 1350 nm to 1365 nm. Based on input data from PCA, a Multilayer Perceptron (MLP) neural network was created to analyze the forty sample validation set. All forty samples were classified correctly based on their variety. The results here show that blueberry beverage can be classified from spectral data and a classification calibration model.
https://www.researchgate.net/publication/321536402_Identification_of_Blueberry_Beverage_Using_VisNIR_Spectroscopy
Applications of FT-NIRS Combined With PLS Multivariate Methods For the Detection & Quantification of Saccharin Adulteration in Commercial Fruit Juices – Mabood, Hussain, Jabeen, Food Additives and Contaminants: Part A, 2018, Vol. 35, No. 6, 1052-1060
Detecting adulteration in foods that are high in carbohydrates can be difficult because there are a variety of commercial sweeteners that match the concentration profile of major carbohydrates. One potential method for detecting commercial sweetener adulterants in fruit juice is FT-NIR spectroscopy and this study examined detecting and quantifying saccharin adulteration for this purpose. Six different commercial fruit juices were obtained for the study. Each sample was spiked with saccharin ranging from 0.10% to 2.00% w/v at varying intervals. The pure samples with 0% saccharin were used as well and in total, one hundred ninety-eight samples were used. Eighteen were pure juice samples and the remainder were spiked with saccharin. FT-NIR spectra were collected for all samples from 10000 cm-1to 4000 cm-1using a sealed cell at 0.20 mm pathlength and 2 cm-1resolution. Various pretreatments were performed on the spectral data for model optimization. Three separate modeling algorithms were used: Principle Component Analysis (PCA) and Partial Least Squares Discriminant Analysis (PLS-DA) for classifying adulterated and non-adulterated samples and Partial Least Squares (PLS) for quantifying amount of saccharin present based on the FT-NIR spectra and reference values for saccharin.
PLS-DA Classification: | |||
R2= | 0.979 | RMSEP= | 0.067 |
PLS | |||
R2= | 0.970 | RMSEP= | 0.88% of 0% to 2% Saccharin range (w/v) |
Visual examination of the spectra showed a clear difference between the samples with saccharin present and 0% saccharin. PCA analysis first showed a clear grouping between these two sets and in order to confirm the validity of determining the presence of saccharin from the spectra, a PLS-DA model was created. This algorithm assigns two arbitrary values (0 and 1 in this case) to separate data sets and predicts a value for classification purposes. The high correlation and low RMSEP prove that this calibration could be used to determine the presence of saccharin in fruit juice. The PLS model performed a quantitative measurement for the amount of saccharin present in the samples based on the FT-NIR spectra and reference values. Correlation was high and a validation data set proved the feasibility of the calibration. The results were especially good considering the different types of fruit juice samples used in the study and provide a basis for using a universal model with many types of fruit juice to identify and quantify the presence of saccharin adulterant using NIR spectroscopy.
https://tandfonline.com/doi/abs/10.1080/19440049.2018.1457802?
Combined Data Mining/NIR Spectroscopy for Purity Assessment of Lime Juice – Shafiee, Minaei, Infrared Physics and Technology 91 (2018) 193-199
Adulteration of fruit juice is a common practice and is always evolving to avoid detection. The major components of fruit juice are water, sugars, and organic acids with lesser amounts of amino acids, vitamins, and phenolic compounds. The demand for lime juice for both direct consumption and as a cooking supplement has led to various types of adulteration, even including completely synthetic products which can potentially result in health hazards for the consumer. While acidity and sugar measurements can detect dilution with water, which is the simplest form of adulteration, it is becoming more common to use water containing sugar and citric acid to avoid detection by simple tests. There is a need for a simple, fast, and cost-effective method to determine whether a lime juice sample is natural or synthetic and NIR spectroscopy was examined for this purpose. Thirty-four pure lime juice samples were procured from different orchards by picking lime fruit and using a juicer machine and filtration. Thirty-eight samples of synthetic lime juice were purchased from a market and subjected to quality assessment in the laboratory to confirm their synthetic nature. Samples were scanned from 350 nm to 2500 nm at 1 nm intervals and averaging 25 scans per spectrum. Spectral resolution at 1000 nm was approximately 3 nm. This process was repeated six times for each sample and the six spectra were then averaged into one spectrum per sample. Various pretreatments were performed on the spectral data for model optimization. Different classification algorithms and data mining techniques were applied to the spectra to determine the feasibility of classifying natural and synthetic lime juice.
Principle Component Analysis (PCA) Classification Accuracy | 66% |
Support Vector Machine (SVM) Classification Accuracy | 97% |
Genetic Algorithm (GA) Classification Accuracy | 93% |
PCA is the standard initial chemometric algorithm for classification analysis and to determine any outliers in a group of samples. Cross validation tests using the generated PCA model showed a poor grouping between the two sets of samples. SVM is a more advanced data mining algorithm that analyzes specific features in the data for deviation reduction and uses selective areas for input to perform a classification analysis on the data. GA is a variable selection method that is based on principles of natural and genetic selection of the data. In this case, the SVM algorithm tested a 97% accuracy rate when analyzing the spectra of both data sets for classification purposes. The study verified the capability of using NIR spectroscopy in combination with data mining and powerful classification methods to accurately discriminate between natural and synthetic lime juice.
https://www.sciencedirect.com/science/article/abs/pii/S135044951730662X
Commercial Reference
Contact one of Galaxy Scientific’s Applications Specialists to discuss this information in further detail.