FT-NIR analysis for measuring meat products sector - Food and Beverage - Food
According to the Organization for Economic Co-operation and Development, worldwide consumption of beef, poultry, and pork will exceed a combined 308 million tons in 2018. Additionally, Newsweek reports, "Over the last 50 years, the average amount of meat eaten per person globally has almost doubled, from around 50 pounds in 1961 to 94 pounds in 2014."
Overview Applications and Industries Served
Quality assurance of meat production is critical, not only for price and brand protection but also for safeguarding human health. Moreover, consumer awareness of quality assurance amongst brands is a significant factor in consumers' decisions to purchase meat products. The implications of poor product quality spreading in today's social media environment, as well as the financial consequences of a product recall can be devastating to a company's bottom line. Since traditional testing methods are costly and time-consuming, new and objective quality control methods are needed. Developing rapid, non-invasive, cost-effective, and environmentally sound methods tor quality testing has become a priority tor meat companies. One such method with the potential for both laboratory and real-time testing of meat products is FT-NIR spectroscopy.
Near-infrared spectroscopy is not a primary technique so it must first be calibrated using a known set of data that includes the near-infrared spectrum as well as reference values for the analytes of interest typically determined by traditional lab methods. To produce a calibration for moisture, for example, several dozen samples varying across the expected range of moisture would first be measured with a near-infrared spectrometer and then have their reference values produced with a traditional laboratory method such as loss on drying or Karl Fischer titration. Once a sufficient set of data has been collected, a near-infrared calibration can be produced, and the laboratory method is no longer needed. In the same manner, other analytes can be calibrated such as protein content, fat content, fatty acid profile, amino acid profile, and ash. Once the calibration has been established, all the parameters can be measured simultaneously, without consumable costs, and in only a few seconds with FT-NIR spectroscopy.
This application note will show a simple example of testing meat for basic macronutrient content using FT-NIR spectroscopy. A calibration is shown for fresh poultry meat that analyzes protein, fat, and moisture content. The calibration model is then tested for accuracy by comparing the output of the near-infrared analysis to a known reference value for each sample.
Four hundred and fifty-one samples of fresh chicken which were each ground and placed in a 100 mm diameter sample cup with a low-OH quartz window. The cup was then placed on the integrating sphere of an FT-NIR spectrometer to measure the diffuse reflectance spectrum of the sample. A sample spinner accessory was used to rotate the sample while the spectrometer was scanning. This allows a larger volume of the sample to be averaged which reduces the inherent variability in inhomogeneous samples like meat. To expand on why a large sampling area is important, consider what would happen if only a 2 mm square area were measured by the spectrometer. The small area on the sample could consist of entirely fat or entirely protein which would then produce a skewed quantitative result that is not representative of the entirety of the meat. Galaxy Scientific's QuasIR™ 3000 uses a large 20 mm diameter sampling spot and then further increases the measured area by rotating the sample cup while the measurement is taking place.
Two important parameters of a near-infrared spectrometer are resolution and number of scans to average. The resolution setting determines the smallest peak in the data that will be visible, similar to how the pixel resolution of a digital camera determines the smallest feature that can be observed in a picture. The number of scans averaged effects the signal-to-noise ratio of the resulting spectrum. If too few scans are averaged, then the noise level will be high and vital features will not be distinguishable. If too many scans are collected, then the collection time will be longer than necessary and will waste the time of the operator. FT-NIR spectrometers, in comparison other technologies, achieve both high resolution and high signal-to-noise ratios in very little time. This experiment used a resolution of 8 cm'1 and averaged 10 scans. With these parameters, the scan was completed in less than 6 seconds. Examples of near-infrared spectra from two different ground chicken samples collected with this method are shown in Figure 1.
