Are you considering what to look for in your next NIR instrument? Senior Manager Lars Nørgaard and Program Manager Magnus Tunklev reflect upon the importance of Resolution versus Signal-to-Noise ratio.
By Magnus Tunklev and Lars Nørgaard,
A frequently asked question in the NIR community is the following: “Is resolution or signal-to-noise the most important parameter for an NIR instrument?”
Recommendations depend on the application
It is difficult to come up with a clear recommendation since it depends on the application for which the NIR solution must work. As the NIR spectrum of a particular molecule is a series of modes arising from the different ways the molecule can twist, stretch and bend, the molecule is not identified by a single mode alone but by the total sum of modes spread out over the spectra. This means that unless you really want to know if an exact bending is present or not, high resolution will simply add noise without adding any useful information of its own. This is the case for the majority of applications in food-agri, where the sample matrix consists primarily of protein, carbohydrates, fat and water, giving rise to broad bands in the NIR spectrum.
Multivariate calibration modeling
Another important aspect is that the use of multivariate calibration modeling delivers a mathematical resolution, making it better to focus on a high signal-to-noise ratio for food-agri NIR applications and then performing post-measurement resolution using an efficient calibration model. The opposite is not possible: if you have analysed your sample with a high resolution spectrometer at the cost of a low signal-to-noise ratio, it is not mathematically possible to increase the signal-to-noise ratio afterwards (given a fixed measuring time).
Resolution is nice if you would like to inspect your spectra – but very often this is not relevant in a practical daily context with a multitude of routine analyses!
If you are interested in further reading, we can recommend two very informative papers by Karl H. Norris (NIR news 9, issue 4, p. 3-5, 1998) and mythbusters Kim Esbensen, Paul Geladi and Anders Larsen (NIR news 23, issue 7, p.17-19, 2012).