Introduction
Milk is the material for all dairy based products and in addition to liquid milk and cream, it is the raw material for butter, cheese, casein, ice cream, lactose, and yogurt. It is estimated that about 37% of all milk produced in the United States is consumed as fluid milk and cream, while the remainder is further processed into other products. The average composition of cow’s milk is 87.2% water, 3.7% milk fat, 3.5% protein, 4.9% lactose, and 0.7% ash. While technological advances have increased the yield from milk producing animals as well as helped maintain quality and efficiency during transport, there is still a need to develop rapid, non-invasive, cost-effective, and environmentally sound methods for quality testing of milk during manufacturing. In addition to manufacturing optimization, frequent feedback of milk parameters can provide valuable information that can benefit herd management. Fat, protein, lactose, urea, and somatic cell content (SCC – an indication of the likelihood that the milk does not contain harmful bacteria) values are typically collected once or twice a month and used for estimation of breeding values. More frequent collection could provide significant benefits for dieting and herd management. One potential benefit is getting information about the physical structure and crude fiber in the ration having a direct effect on fat in the milk. Another is protein and urea contents helping develop conclusions about the balance of energy feed supply and protein concentration in the diet. Other parameters of interest in milk include total solids, non-fatty solids contents, freezing point, and acidity measurements such as titratable acidity (TA) and pH. Adulteration is a big issue in the food and beverage industry and milk products are no exception. There are strict regulations on the labelling and grading of milk based on numerous factors such as fat content, conditions of manufacturing, and other processes like pasteurization, homogenization, and vitamin fortification. Misrepresenting milk as labelled is one form of adulteration. Another form of adulteration is adding a portion of a cheaper product to a more expensive and higher quality product. Such adulteration can not only reduce the nutritional value of a product, it can also cause health risks. This can occur in all dairy products, including milk and milk powder. Current methods for testing these parameters 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 milk manufacturing and transport processes. One such method that has been examined is NIR spectroscopy.
Analytes
- Fat
- Protein
- Lactose
- Urea
- Somatic Cell Count (SCC)
- Total Solids
- Non-Fatty Solids Contents
- Freezing Point
- Titratable Acidity (TA)
- pH
- Camel Milk Adulteration
- Milk Powder Adulteration
Summary of Published Papers, Articles, and Reference Materials
Measurement of chemical parameters in major constituents of milk for quality control purposes has been studied using NIR spectroscopy. The results of most studies have been promising. One study performed in-line measurements during the milking process to determine the feasibility of using NIR spectroscopy for various parameters in milk: fat, protein, lactose, urea, and somatic cell count (SCC). SCC results were considered good enough for screening purposes while all other parameters showed results that can be used for real-time measurement during the milking process. Fat content is considered the most important parameter in milk and currently, two separate reference methods are used for determining fat in milk: Rose-Gottlieb and Gerber, both of which use toxic chemicals, require highly trained technicians, and are inadequate for real-time measurement. NIR spectroscopy was used to create calibration models for fat with both these methods performed to determine reference values. Results were excellent using both methods. Goat milk samples were used with an FT-NIR instrument to measure different parameters of interest: fat, protein, lactose, total solids, non-fatty solids contents, freezing point, titratable acidity (TA), and pH. Results were validated and proved good enough for real-time measurement for all parameters except freezing point, TA, and pH. However, the range of values for these three parameters was very small but the results were still good enough for screening purposes. Adulteration in both milk and milk powder is considered a big problem in the dairy industry and two separate studies used NIR spectroscopy for adulterant determination. The first used camel milk samples adulterated with cow milk, which is a much cheaper product. Results showed that the presence of cow milk adulterant in camel milk could be determined as well as quantified using NIR spectra and calibration models. The second study used different samples of pure milk powder and various adulterants to determine the feasibility of finding the presence of an adulterant in milk powder. Three different spectrometers were used for this study: Two FT-NIR benchtop instruments and a handheld NIR instrument. Results were excellent for the two FT-NIR instruments but much worse for the handheld NIR instrument, most likely due to a shorter wavelength range and lower spectral resolution for the handheld instrument.
Scientific References and Statistics
Accuracy of In-Line Milk Composition Analysis with Diffuse Reflectance Near-Infrared Spectroscopy – Melfsen, Hartung, Haeusserman, Journal of Dairy Science 95: 6465-6476
Daily monitoring of composition changes in milk can assist in monitoring cow health and be used for detection of nutritional imbalances. The composition of milk can be affected by a large number of factors, such as breed, nutrition, seasonality, lactation, and health. NIR spectroscopy was examined for in-line measurement of the fat, protein, lactose, urea, and somatic cell count (SCC) in milk during the milking process. Eighty-four composite milkings were used for the study and a total of seven hundred eighty-five partial milkings from the composites were scanned with an NIR spectrometer using diffuse reflectance. NIR spectra were collected every 500 ms from 851 nm to 1649 nm during the milking process. A measuring cell was used and the layer thickness of the milk was 30 mm. After 2 kg of milk was collected, the spectra collected during each 2 kg were averaged into one single spectrum. A bypassed portion of each milk sample was sent to the laboratory and reference values were obtained for the parameters of interest. Partial Least Squares (PLS) regression models were created correlating the NIR spectra to fat, protein, lactose, urea, and SCC.
| Fat | R² = 0.998 | RMSEP= 0.09% |
| Protein | R² = 0.99 | RMSEP= 0.04% |
| Lactose | R² = 0.96 | RMSEP= 0.05% |
| Urea | R² = 0.89 | RMSEP= 15.24 mg/L |
| SCC (Log) | R² = 0.90 | RMSEP= 0.15 |
The results obtained in this study confirmed the feasibility of the calibration models used to measure the parameters of interest. Fat, protein, and lactose all showed high correlation and a low SEP. The results for these three calibration models are good enough for real-time on-line monitoring of milk using the procedure documented in this study. Results for urea were not as good but still good enough to be considered suitable for screening purposes. It must be noted that the concentration of urea measured here is extremely small. The likelihood of a direct measurement of a parameter on the order of ppm using NIR spectroscopy is very low. The calibration model is most likely correlating to another parameter which may or may not be affected by a change in the urea concentration. While an indirect correlation is acceptable when using NIR spectroscopy, such calibrations must be examined and validated carefully to prove the model is valid. More work will be required to use this model in a real-time setting. In the case of SCC, previous studies have indicated that SCC is a difficult parameter to measure using NIR spectroscopy because the parameter is not defined from explicit chemical bonds. Most likely, other constituents are changing and the SCC calibration is at least partially using these changes for model correlation. An examination of predicted values for SCC using the NIR calibration model and reference values indicates that a successful classification could be performed for various ranges of SCC values from NIR spectra. Overall, the results in this study prove the feasibility of monitoring fat, protein, and lactose in milk in an in-line setting using NIR spectroscopy and calibration models. The potential for measuring urea and SCC was demonstrated as well, but further work will be necessary to implement these calibrations in an in-line environment.
https://www.sciencedirect.com/science/article/pii/S0022030212006509
Accuracy of the FT-NIR Method in Evaluating the Fat Content of Milk Using Calibration Models Developed for the Reference Methods According to Rose-Gottlieb and Gerber – Mlcek, Dvorak, Sustova, Szwedziak, Journal of AOAC International Volume 99, No. 5, 2016
FT-NIR spectroscopy was examined as a method for comparative purposes using two separate reference methods for the fat content in cow milk to build the calibration models: Rose-Gottlieb and Gerber. The Rose-Gottlieb method treats samples with ammonia to dissolve protein and ethyl alcohol to help precipitate proteins. Fat is then extracted with diethyl ether and petroleum ether. After the ethers are evaporated, the residues are weighed. The Gerber method adds sulfuric acid to separate proteins. The separation is facilitated by using amyl alcohol and centrifugation. Fat content is read using a specialized butyrometer. Thirty individual cow milk samples were procured for the study. Samples were scanned in reflectance mode using a transflectance cuvette with a 0.2 mm pathlength. Spectral range was from 10000 cm-1 to 4000 cm-1 with a spectral resolution of 8 cm-1 and 100 scans averaged per spectrum. Both reference methods were performed in separate independent laboratories to determine fat content for each sample. Two Partial Least Squares (PLS) regression models for fat were created using both sets of reference data and the NIR spectra.
Rose-Gottlieb:
| Fat | R² = 0.993 | RMSEP= 0.133% |
Gerber:
| Fat | R² = 0.996 | RMSEP= 0.095% |
The results show that both calibration models work very well and the values obtained from both models differ in fat content values by parts per hundred. There were lower deviations observed in the Rose-Gottlieb models but there was no statistically significant difference between the two data sets. Fat is one of the most valuable components of milk and current analytical methods depend on acids and solvents as well as chemical equipment and trained technicians. NIR spectroscopy offers a fast, non-destructive technique for the measurement of fat in milk samples without the use of toxic chemicals.
https://www.ncbi.nlm.nih.gov/pubmed/27324807
Analysis of Goat Milk by Near-Infrared Spectroscopy – Drackova, Hadra, Janstova, et al., Acta Vet. Brno 2008, 77: 415-422
FT-NIR spectroscopy was examined as a method for determining various parameters of interest in goat milk. The parameters measured in this study are protein, fat, lactose, total solids, non-fatty solids contents, freezing point, titratable acidity (TA), and pH. Sixty samples of goat milk each taken from separate bulk tanks were procured for the study. Samples were scanned from 10000 cm-1 to 4000 cm-1 in reflectance mode using a transflectance cell with 0.1 mm pathlength. One hundred scans were collected per reading and averaged into a single spectrum. Traditional reference values for the parameters of interest according to standards were obtained for each sample. The reference values and NIR spectra were used to create Partial Least Squares (PLS) calibration models correlating the NIR spectra to each parameter.
| Protein | R² = 0.92 | RMSEP= 0.094% |
| Fat | R² = 0.951 | RMSEP= 0.124% |
| Lactose | R² = 0.997 | RMSEP= 0.011% |
| Total Solids | R² = 0.940 | RMSEP= 0.260% |
| Non-Fatty Solids Contents | R² = 0.873 | RMSEP= 0.159% |
| Freezing Point | R² = 0.935 | RMSEP= 0.003°C |
| Titratable Acidity | R² = 0.952 | RMSEP= 0.295 SH (Soxhlet Henkel) |
| pH | R² = 0.835 | RMSEP= 0.057 |
All calibration models showed good results and were tested using cross-validation to determine prediction error for each parameter. The models for protein, fat, lactose, total solids, and non-fatty solid contents showed good enough results for real-time measurement of these parameters using the calibration models. In the case of freezing point, titratable acidity (TA), and pH, the models were less robust and prediction results can only be considered good enough to use the models for screening purposes. However, this most likely occurred because of the very small range of values: less than 0.03°C for freezing point, less than 4 SH for TA, and less than 0.5 for pH. More samples and a larger range of values should improve the results. Overall, the results here proved the feasibility of measuring these parameters using NIR spectra and calibration models.
https://pdfs.semanticscholar.org/548d/0627b4b9de94abceac0e9e80cb8045b436f3.pdf
FT-NIRS Coupled With Chemometric Methods As A Rapid Alternative Tool for the Detection & Quantification of Cow Milk Adulteration in Camel Milk Samples – Mabood, Jabeen, Hussain, et al., Vibrational Spectroscopy 92 (2017) 245-250
Camel milk is considered to have high nutritional value in comparison to milk produced by cows. It is a rich source of Vitamin A and C, has a high content of potent immunoglobins, and does not contain lactoglobulin and A1 casein, making it suitable for consumption by people with bovine dairy allergies. Thus, it sells for a higher market price and is subject to adulteration using cheaper forms of milk. FT-NIR spectroscopy was examined as a method for determining cow milk adulteration in samples of camel milk. Three samples of camel milk were procured for the study and prepared in triplicate form. Each separate camel milk sample was adulterated with different percent levels of cow milk adulterant: 2%, 5%, 10%, 15%, and 20%. Including the three pure samples, a total of fifty-four samples were created. 70% of the samples were used as a calibration set to create the model and the remaining 30% was used as a validation set. All samples were scanned from 700 nm to 2500 nm at 2 cm-1 spectral resolution using a 0.2 mm pathlength sealed cell. Two calibration models were created: Partial Least Squares Discriminant Analysis (PLS-DA) to determine the presence of the cow milk adulterant and Partial Least Squares (PLS) to quantify the amount of adulterant present.
PLS-DA
| Presence of Adulterant | R² = 0.973 | RMSEP= 0.0801 |
PLS
| Amount of Adulterant | R² = 0.926 | RMSEP= 1.32% |
Both calibration models showed good results and proved the feasibility of the measurement. In the case of PLS-DA, an arbitrary value of 0 was assigned to the pure camel milk samples and 1 to samples spiked with 10% cow milk adulterant. The model predicts a number and a threshold of 0.5 was chosen to determine the presence of adulterant. A predicted value less than 0.5 indicates no adulterant and a predicted value greater than 0.5 indicates an adulterant is present in the sample. The high correlation and low RMSEP show that this model can be used to determine the presence of cow milk adulterant in camel milk. In the case of PLS, the results show that the model can predict the amount of adulterant present within an accuracy of 1.32%. It must be noted that while the models showed good results, using them in a practical setting for different kinds of milk and adulterants would require a much larger sample set. Natural products often show variability in NIR spectra due to many factors, such as region of origin, different types of food fed to animals, different soil for plant growing, and so forth. Incorporating different samples encompassing any potential variability is important when building calibration models. Predictions performed on the validation set proved that the models could work in a real-time setting, using the PLS-DA model as a detection tool and the PLS model as a quantification tool, providing information that would be very difficult to find using conventional methods for adulteration detection.
https://www.sciencedirect.com/science/article/abs/pii/S0924203117300668
Non-Targeted NIR Spectroscopy and SIMCA Classification for Commercial Milk Powder Authentication: A Study Using Eleven Potential Adulterants – Karunathilaka, Yakes, He, et al., Heliyon 4 (2018) e00806.
NIR spectroscopy was evaluated as a method for rapid screening of commercial milk powder products as authentic or being mixed with known and unknown adulterants. Milk powder is a known target of adulteration, ranking only second to olive oil according to the USP database on food fraud and economic adulteration. Milk powder adulteration can cause adverse health effects, as highlighted by two incidents of milk and wheat gluten adulteration with melamine. One incident was linked to renal failure in cats and dogs due to pet food adulteration in the United States. The other was linked to infant formula in China, causing thousands of cases of renal complications in children as well as at least six confirmed deaths. A number of commercially available milk powder samples were procured for the study, representing manufacturing in sixteen different states and twenty-four different companies and brands. Eleven different milk powder adulterants were selected for the study based on their history or potential use as adulterants. These can be divided into 4 categories: 1. Low Molecular Weight, Nitrogen-Rich Compounds, 2. Plant Proteins, 3. Inorganic Salts, 4. Non-Fat Solids. Some of the adulterants were blended as well to increase the scope of the potential adulterants that could be used. Different levels of all adulterants were mixed with the pure milk powder samples before NIR spectra were collected. Three different spectrometers were used for the study: Two benchtop FT-NIR instruments and a handheld NIR device. Principle Component Analysis (PCA) was performed on all three sets of data and SIMCA classification models were created for determining the presence of an adulterant in the milk powder. After model creation, separate test sets of both pure and adulterated samples were scanned for model validation.
FT-NIR Spectrometer #1
Instrument Parameters:
| 12500 cm-1 to 4000 cm-1 | 64 scans per average | 16 cm-1 spectral resolution |
100% Correct Adulterant Classification and Concentration (%w):
| Melamine | 0.6% to 2.0% |
| Dicyandiamide | 2% |
| Aminotriazole | 0.4% to 2.0% |
| Biuret | 0.2% to 2.0% |
| Soy Protein Isolate | 5% to 20% |
| Pea Protein Isolate | 5% to 20% |
| Calcium Carbonate | 2% |
| Maltodextrin | 2% to 20% |
| Sucrose | 7% to 50% |
FT-NIR Spectrometer #2
Instrument Parameters:
| 10000 cm-1 to 4000 cm-1 | 32 scans per average | 16 cm-1 spectral resolution |
100% Correct Adulterant Classification and Concentration (%w):
| Melamine | 0.4% to 2.0% |
| Aminotriazole | 0.4% to 2.0% |
| Biuret | 0.2% to 2.0% |
| Cyanuric Acid | 2% |
| Soy Protein Isolate | 5% to 20% |
| Pea Protein Isolate | 2% to 20% |
| Maltodextrin | 5% to 20% |
| Sucrose | 10% to 50% |
Handheld Spectrometer
Instrument Parameters:
| 6266 cm-1 to 4167 cm-1 | 10 scans per average | 11 nm optical resolution |
100% Correct Adulterant Classification and Concentration (%w):
| Biuret | 0.4% to 2.0% |
Results for all three spectrometers are shown above. Both FT-NIR benchtop spectrometers showed 100% specificity and accuracy for determining the presence of an adulterant in milk powder if the adulterant was at a sufficiently high percentage, which varied based on the type of adulterant present. In the case of the handheld NIR device, results were much worse. This is most likely due to a narrower wavelength range and lower resolution than the benchtop FT-NIR instruments. The results here prove the feasibility of using FT-NIR spectrometers as a tool for determining the presence of adulterant in milk powder and show that FT-NIR spectrometers are much better suited for such analysis than handheld NIR spectrometers.