-Pitts and Tredici Rabbit Cornea, -Pitts Human Cornea, 1973 -Pitts, Cullen et al Rabbit Cornea, 1977 Zuclich and Taboada Primate Cornea, 1978
Figure 4. Reported photokeratitis thresholds and the ICNIRP UV exposure limit. Outlying threshold points plotted at the center wavelength of the monochromator are especially apparent in the 300-320 nm wavelength region. The Kurtin and Zuclich 1978 , and Zuclich and Taboda 1978  data are included for purposes of comparison.
slope very similar to that obtained for DNA by Peak, Peak, et al. .
Threshold data in the 300-320 nm region can be particularly misleading if the data were taken using a monochromator with too large a bandwidth. The Pitts and Tredici 1971 paper reports that data were obtained with "a nominal 9.92-nm bandpass which did not exceed 10 nm for all wavelengths." For the Pitts 1973 data, the report specifies that "full band pass did not exceed 10 nm for all wavebands." The Pitts et al  data were obtained with two different monochromators: a single grating monochromator "whose entrance and exit slits were set to provide a 9.96 nanometer full-bandwidth wavelength" (bandwidths were reported as 10.0 nm); and a double grating monochromator "set to pass a full-bandwidth of 6.6 nm; however, all wavebands are reported as 5.0 nm." The Cullen and Perera data were obtained with a 9 nm bandwidth (FWHM). The different published reports are not always clear on the meaning of bandwidths given, i.e., if they are the FWHM, as recommended.
Such large bandwidths do not lead to great plotting errors in the derived action spectrum for wavelengths around 270-280 nm or at wavelengths greater than 330 nm. However, the plotting error is quite noticeable in the 300-320 nm range since the exposure limits in these wavelengths are rapidly increasing—the limits increase by an order of magnitude between 303-310 nm, and another order of magnitude between 310-320 nm. For the threshold points in question, most of the effective dose comes not from the wavelength at the center of the bandwidth, but from the shorter wavelengths. These wavelengths frequently are within the exposure limits.
The slit function was centered at the wavelength of the point being investigated and normalized so that the area under the function corresponded to the value of the exposure dose when the function was weighted by the exposure dose (figure 3). We assumed that the actual arc-spectrum did not significantly vary across the limited bandwidth, which is valid for a xenon arc.
This function was weighted by the exposure dose, and then spectrally weighted by the UV hazard action spectrum S(X)  to determine what wavelengths actually were contributing to the actual photobiological effect (Fig. 5).
Assuming the published bandwidths are the FWHM values, a shift in wavelength occurs for all eight data points examined (Figs 6-10).
For seven of the eight examined data points, the greatest dose comes from a wavelength within the ACGIH/ICNIRP exposure limits. The exception appears to be the "320-nm" threshold for human conjunctivitis from Cullen and Perera's 1994 study. This surprisingly low reported threshold could be due to the contribution of a thermal effect because of the long exposure duration required. Another explanation of this apparent inconsistency might have resulted from short-wavelength stray light in the small, single grating monochromator used in their experiment  Figure 11 shows a possible slit function for this data point accounting for stray light, and the resulting effectiveness function. The stray light was estimated for a single grating monochromator with high spectral scatter . Although the stray light does not shift the peak wavelength from 314 nm, which was determined with the "perfect" slit function, the stray light in this estimation does account for 8% of the effective dose.
E(A) Pitts Human Cornea, 1973
S'(A) Normalized Slit Function
E(A) Pitts Human Cornea, 1973
Was this article helpful?