Читать книгу Patty's Industrial Hygiene, Physical and Biological Agents - Группа авторов - Страница 123

Significant errors in measurement of effective irradiance or dose can result when the spectral response of a broadband detector is not well matched to the spectral weighting function for the biological effect of interest. Under these circumstances, a broadband detector that is calibrated using a monochromatic calibration source will give inaccurate readings when measuring broadband radiation. The deviation of the broadband detector reading from the true effective irradiance depends on the spectral distribution of the incident radiation. Correction factors must, therefore, be developed that are specific to the exposure conditions to be assessed. If the relative spectral response function D(λ) of the broadband detector and a spectral distribution function I_λ of the optical radiation source (representing the normalized spectral intensity, spectral irradiance or spectral radiance) are known, then a correction factor f may be calculated as

(22)

where s(λ) is the spectral weighting function for the biological effect and the summation is over the wavelength range to which the detector responds. It should be recognized that the detector response might not cut off sharply at the limits of the spectral weighting function for the biological effect.

Some practical difficulties with calculating correction factors should be noted. The source spectral distribution function must be obtained from the source manufacturer, and the detector response function must be obtained from the detector manufacturer. The functions s(λ), D(λ), and I_λ may each range over several orders of magnitude. Spectral features such as the tails of a curve, which might appear small when a spectral distribution or response function is plotted on a linear scale, may become important when multiplied by a high value in another function. For accurate calculations, spectral data should be obtained from the manufacturer in numerical form if possible. If numerical data are not available, semilogarithmic plots of spectral distribution or response function can be used to evaluate small peaks, tails, continuum levels, and other features. It should also be recognized that, due to variations in the manufacturing process, the spectral distribution function for the particular source that is actually present in the workplace may differ from representative data provided by the manufacturer. Shifts in wavelengths of emission peaks between the actual and the reported spectral distribution functions may affect the correction factor if a shift coincides with a spectral region where s(λ) or D(λ) is changing rapidly with wavelength.

An alternative approach for determining correction factors for a broadband detector is to compare effective irradiance measurements from that detector to effective irradiance measurements obtained simultaneously using a more accurate method. This may be the only practical approach when the radiation source is not a manufactured lamp for which a spectral distribution function can be obtained. The correction factor will be affected by conditions that modify the spectral distribution of the source. For example, when polysulphone films used to measure erythemal effective dose from ambient solar UV radiation were calibrated against electronic radiometers and dosimeters that had spectral responses well matched to the erythemal reference action spectrum, the calibration curves showed variability related to atmospheric conditions that altered the ground‐level solar UV spectrum (39). Erythemal dose measurements of solar UV determined from polysulphone films are vulnerable to spectral variability in the incident radiation because the action spectrum for polysulphone dosimeters does not reproduce the shape of the erythemal reference action spectrum at wavelengths longer than 300 nm, where most solar UV occurs. Spectral mismatch effects in measurement of solar UV were also seen, to a lesser extent, in an assessment (40) of electronic dosimeters that used an aluminum gallium nitride photodiode detector with a response function that approximates the CIE erythemal reference action spectrum (32); the dosimeters, which had been calibrated in New Zealand in summer under high solar UV conditions, showed deviations of up to 30%, compared to the erythemally weighted irradiance determined from reference spectroradiometer measurements, when tested in northern Germany in autumn (40).

The ACGIH and ICNIRP guidelines for protection of the lens of the eye from UV‐A radiation is based on irradiance or radiant exposure that is not spectrally weighted. Ideally, measurements of UV‐A for purposes of assessing this hazard should be taken with a detector having a flat spectral response from 315 to 400 nm; such a detector might not, however, be available in practice. If the source spectral distribution is known, a correction factor can be calculated for a UV‐A detector using Eq. (22) with s(λ) set equal to unity for the entire range 315–400 nm and set equal to zero outside that range. An evaluation of two commonly used broadband UV‐A detectors found that correction factors of 1.27–1.41 were applicable when measuring a UV‐A phototherapy source (41).