Monitoring Sugar Content in Cereal with NIR Analysis

Introduction

The high nutrient density of breakfast cereals (especially whole grain or high in fiber) makes them an important source of key nutrients. Individuals consuming breakfast have shown better overall nutrition profiles, improvements in cognitive functioning, and might be less likely to be overweight. Breakfast is widely recommended as part of a healthy diet, with breakfast cereals being a popular breakfast choice especially for children, due to its variety and ease. However, most breakfast cereals also contribute to sugar intake. Higher sugar intake by children has been shown through meta-analysis to be associated with a higher risk of obesity (Te Morenga et al., 2012).

Overview

Nutrition labels for some commercial products may not reflect the most accurate sugar content information (Ventura et al., 2011; Walker et al., 2014). Walker and Goran (2015) reported that nutrient label data of breakfast cereal either underestimated or overestimated actual sugar content by 20%. Given the abundance of data linking added sugar consumption to disease risk it is imperative to determine actual sugar content to allow consumers to make informed diet decisions for processed food products that contain added sugars (Walker and Goran, 2015).

Sucrose is the main coating material used in breakfast cereals providing sweetness (Solis- Morales et al., 2009) and minimal stickiness and hygroscopicity as compared to high fructose corn syrups. High performance liquid chromatography (HPLC), enzymatic analysis, electrochemical or spectro-photo metric methods have been used in quantitative and qualitative analysis of sucrose (Low- ell & Kuo, 1989; Kumar et al., 2010). Among these methods, HPLC and enzymatic analyses are accurate, sensitive and specific, but costs are prohibitive for large number of samples (Texeira et al., 2012; Gomez et al., 2007).

How NIR Works

NeoSpectra spectral sensors are a great potential alternative for the analysis of sugar content in cereal, with a lot of benefits that enable wide adoption by different types of users. Being miniaturized, low cost, and able to be produced in large quantities, it can be used by producers, quality checkers, health watchers, dietitians, and even consumers to verify the correctness of the nutrient label data. This application note shows that the NeoSpectra spectral sensors can estimate sucrose levels in breakfast cereals providing equivalent results to that of HPLC methods.

Testing Methods

Cereals coated with sucrose (n=60) were manufactured by a leading Ohio snack manufacturer and before NIR measurements were blended to obtain a homogeneous particle size. Cereal samples were transferred to glass petri dishes and placed on a rotating stage for spectral averaging of heterogeneous samples to ensure reproducible measurements (Figure 1). It should be noted that different optical setups/optical heads can be configured such that the samples may not need to be blended and rotated.

The NIR spectra was collected for 20s and quantitative algorithms were generated using Partial Least Squares regression. The sucrose content was analyzed using a high performance liquid chromatography (HPLC) (Shimadzu Scientific Instruments, Inc. Columbia, MD) equipped with a refractive index detector. The sugars were separated on a stainless steel, 7.8 mm ID x 300 mm Aminex® HPX-87C carbohydrate column under isocratic conditions at 80°C using HPLC grade water with a flow rate of 0.6 mL/ min for 30 minutes.

The sucrose levels in the cereal samples ranged from 0 to 30 g/100g covering a wide range of concentrations resembling commercial breakfast cereals with reported averages of 22.5±12.6 g/100g (Pombo-Rodrigues et al., 2017).

Figure 2 shows the NIR spectrum collected using the Neospectra Module. Characteristics bands were identified at 4390 cm^-1 associated with the aliphatic C-H bands, 4420 cm^-1 due to C-H absorption in the carbohydrate bands, 4880 cm^-1 due to the amide absorption in protein, 5180 cm^-1 related to O-H absorption in the water bands, and 5807 cm^-1 due to the aliphatic C-H bands. There was also a unique band at ~4800 cm-1 indicative of O-H groups in crystalline sucrose Kays et al. (1998).

The PLSR regression graph (Figure 3) is an excellent correlation between the estimated sucrose concentrations by NIR and the HPLC reference analysis. The optimum number of factors giving the minimum standard error of cross validation (SECV) was two and gave SECV for sucrose of 1.2%, correlation coefficient (r) of 0.99.

The performance of the regression algorithms for sucrose content using the NeoSpectra Module were superior to those previously reported for bench-top FT-NIR (SEP=1.5%; Wang and Rodriguez-Saona, 2012), and another commercial handheld analyzer (1.5%, Lin et al, 2014). Furthermore, the NeoSpectra Module showed much better performance to bench-top dispersive units that have reported SEP of 2.8% on 84 branded cereals (Baker and Norris, 1985).

Conclusions

The NeoSpectra spectral sensors provide the best performance for determination of sucrose levels in breakfast cereals. With no sample preparation and fast results (20s) to allow the industry to monitor sugar content in commercial products to meet nutrition labeling requirements and allowing consumers to make informed diet decisions as growing concerns to risk of obesity from excessive sugar consumption.