The limit of detection (LOD) in spectroscopy is defined by the signal-to-noise ratio (SNR) of the spectral data.

In the context of spectroscopy, the LOD represents the smallest quantity of a substance that can be reliably distinguished from a blank (the absence of that substance) within a specified confidence limit. This parameter is critical in spectroscopic analyses, as it establishes the minimum concentration or absolute amount of an analyte that can be confidently detected.

To determine the LOD, it is essential to first measure the SNR of your spectral data. Here, the signal refers to the response generated by the analyte (the substance you aim to detect), while the noise represents the random fluctuations in the signal. The SNR can be calculated using the formula:

Generally, a higher SNR correlates with a lower LOD, indicating that the analyte can be detected at lower concentrations. A widely accepted guideline is that the LOD is reached when the SNR is $3:1$. This implies that the signal must be three times greater than the noise, which is typically considered the minimum threshold for reliable detection.

The specific methodology for calculating the LOD can vary based on the type of spectroscopy employed and the characteristics of both the analyte and the matrix (the medium containing the analyte). For instance, in atomic absorption spectroscopy, the LOD is often determined by measuring the absorbance of a series of standard solutions. By plotting the absorbance against concentration to create a calibration curve, the LOD can be identified as the concentration that corresponds to an absorbance value three times the standard deviation of the blank measurements.

In addition to the SNR and calibration curve methods, various statistical approaches exist for determining the LOD. One such method is the IUPAC approach, which involves calculating the standard deviation of the response alongside the slope of the calibration curve.

In summary, establishing the LOD in spectroscopy involves assessing the SNR and may include constructing a calibration curve or applying statistical techniques. The precise method utilized can depend on the specific type of spectroscopy and the nature of the analyte and its matrix.