Photometric accuracy

The closeness of measured transmittance or absorbance to a reference value.

Three closely matching measurements of the same sample indicate good short-term repeatability. If all readings are too high or too low, repeating the measurement will not reveal that systematic error.

Accuracy requires a reference value

Photometric accuracy evaluates how closely an instrument's transmittance or absorbance agrees with a reference value. The reference generally comes from certified materials or a traceable comparative measurement, with its own uncertainty and specified conditions of use.

NIST emphasizes that photometric reference materials must be used only at their specified wavelengths, ranges and conditions. Outside the certified conditions, a material does not automatically retain the original reference value simply because its name is the same.

How error enters experimental results

Systematic photometric bias shifts absorbance and propagates into absorbance-based concentration calculations. For short-path measurements, photometric-scale error must also be separated from effective-path-length error: both may appear as a proportional concentration bias, but originate in different parts of the system.

Wavelength offsets, finite bandwidth, stray light and detector nonlinearity all affect measured reference values. Verification should therefore be close to the wavelengths and absorbance levels actually used, rather than treating one condition as representative of the entire spectrum and signal range.

One reference point cannot establish all aspects of performance

Passing a single reference point at moderate transmittance supports a conclusion only near that point and under the corresponding conditions. It does not automatically establish high-absorbance linearity, wavelength accuracy, path-length accuracy or baseline stability. ASTM instrument-performance practices likewise emphasize matching performance tests to the needs of the specific method.

When reporting accuracy, retain the reference-material identity, certificate status, wavelength, bandwidth, temperature, path length, Blank and result uncertainty. Without these conditions, a claim of good accuracy cannot be independently checked.

Accuracy addresses distance from a reference value. Repeatability, linearity and stability answer different questions and cannot substitute for one another.