Mid-infrared (MIR) spectroscopy is a technique whereby the composition of a substance, whether it is a solid, a liquid, or a gas, is analysed using light in the mid-infrared spectrum (roughly 600 cm⁻¹ and 1500 cm⁻¹).
As part of the VistaMilk project, this technique is used to determine the composition of the milk produced by a dairy cow. Milk composition data can further be used to gather information about a cow’s welfare, and the effect that the cow has on the environment.
When a dairy cow is milked, a sample of that milk is analysed using a MIR spectrometer. The spectrometer generates light in the mid-infrared light region which passes through the milk sample, and measurements at 1,060 different frequencies in the MIR spectrum are generated. Commonly, each measurement records the amount of light at a given frequency which is absorbed by the sample (absorbance).
When a scientist reviews a set of measurements taken by the spectrometer, it is often visualised on a graph like this one.
By studying this graph, a scientist can determine the components of the milk. For example, if peaks are higher than the average around the 1750 cm⁻¹ frequency or around the 2900 cm⁻¹ frequency, we can determine that the sample contains high levels of fat.
The MIR spectrum is often used to determine the levels of key components in a milk sample, such as fat, protein and lactose. This information can have a variety of uses. For example, it may enable a milk processor to decide the optimal use for the milk from which the sample was derived. Employing MIR spectroscopy may thereby allow the processor to avoid more expensive and labour-intensive methods of analysing the composition of the milk.
We can take the method a step further. Scientists have developed a number of prediction equations to infer different types of information from MIR measurements taken from a milk sample.
Applying a prediction equation to our MIR measurements can reveal additional information, like the body condition score of the cow that produced the milk, or even the methane emissions that can be expected from the cow.