Miller Magazine Issue 108 / December 2018

56 MILLER / DECEMBER 2018 COVER STORY ters, i.e., protein, moisture, oil and carbohydrates. The accuracy of NIR to measure these components in whole wheat grains is often poor. Grinding the grains into a powder and collecting the Near Infrared Reflectance spe- ctra from 1000 to 2500nm, will improve the accuracy. Particle size of the powder is commonly correlated with measurements such as hardness, sedimentation and Ze- leny and thereby explains why NIR Reflectance analysers are better for performing these measurement over NIR Transmission analysers. Falling Number is a measurement of the extent of ger- mination in a seed. Low Falling Number, i.e., less than 250 seconds, indicates a high degree of germination whe- re the enzyme Amlyase is released and starts the break- down of the starch in the endosperm. NIR spectroscopy measures total starch and thereby should provide a means of measuring Falling Number. However Falling Number is a viscosity analysis technique that relates the reduction in viscosity due to the breakdown in starch. The starch con- tent as measured by NIR does not correlate well with the viscosity measurement and at best can be used as a rough estimate of the true Falling Number value. Wheat and durum can exhibit white or yellow spots in the seeds due to high starch levels. In durum there is a measurement called Vitreousness that measures the transparency of the seeds. Starch spots inside the seeds decreases the transparency and are considered a nega- tive quality parameter. High vitreousness measurement is associated with hardness and exhibits higher yield in semolina than low vitreousness durum. NIR Transmission analysers are capable for measuring Vitreousness. Alt- hough Vitreousness is related to the protein and starch content of the seeds, the actual measurement is related to the transparency of NIR light through the seeds. There are more NIR measurements of quality parame- ters in flour and semolina than for whole grains. These include; Protein, Moisture, Starch, Ash, Water Absorpti- on, Starch Damage, Dough Stren- gth, and Dough Stability. Protein, Moisture and Starch measurements using NIR are straight forward. Ash which is a measure of the minerals in the flour, have no direct spect- ral bands in the NIR region. There are indirect measurements of Ash using the Visible spectral region, i.e., 520nm, which relates to the brand content of the flour. Several NIR analysers provide a wider spe- ctral coverage in order to collect the 520nm band. The brand content can also measured in the NIR region using the 700-720nm region which captures the tail of the red absorp- tion bands from the Visible spectral region, i.e., 350-680nm. Ash con- tent is a good indicator of the degree of milling of the grain and separation of the non endosperm components of the grains. It is arguable whether Ash content of the flour directly affects the baking qualities of the flour, where as the inclusion of non endosperm components in the flour do affect baking quality. The reference method for Ash is weight of the residual after placing a sample in an Ashing Oven for many hours. Although a slow met- hod as compared to NIR, the Ashing Oven method is hi- ghly sensitive and data is expressed as 0.001%. The NIR method for Ash is generally measured to only 0.01%. Starch Damage, Water Absorption, Dough Strength and Dough Stability are parameters that are measured using a Farinograph, Extensiograph and Amylograph. These analysers are all based on measuring the rheometry or viscosity of a flour water mixture. As the mixture for- ms a dough, the strength of the dough is measured. Ty- pically these procedures take 30 minutes to perform and can only be done one at a time. Flour millers use the data as guides to the baker as to the water take up and mixing performance of the flour when made into a dough. These are all physical measurements that at first thought should not be measured by NIR analysers. However calibrations for all four parameters have been developed across many brands of NIR analysers. The reason is that these parame- ters are Influenced by the protein, starch and water con- tent of the flour. NIR chemometric calibration software uses multi variable linear regression processes to develop mathematical models to describe chemical components in food and agricultural products. As such, NIR calibrati- ons for these four parameters can be rationalised due to the multi component interactions that exist within flour from protein, moisture and starch. Figures 3 and 4 show NIR calibration plots Water Absorption and Starch Dama- ge using a NIR Transmission analyser. Semolina is used to make pasta. The Protein, Moisture and Ash are critical parameters that are measured by pas- ta manufacturers. Figure 5 shows a typical calibration plot Figure 3. NIR Calibration Plot for Water Absorption in Flour