fluorescence, spectral energy distribution, instrument, glass cylinder, illumination system, measuring instrument, measuring instruments, UV region, UV filter, source, UV excitation, Dipl.-Ing, Authors, HTL F., Griesser Farbmessung FC, illumination, measurements, routine measurements, random tests, xenon lamp, Statistical evaluation, excitation spectra, CIBA-GEIGY, tristimulus values, White Scale, visible region, standard deviation, whiteness, measurement, UV absorption
Reprint and translation from: DIE FARBE 24 (1975), No. 1/6 Felix Gдrtner* und Rolf Griesser*, BASLE: A device for measuring fluorescent white samples with constant UV excitation DK 535.65.084.852 535.665.7 535.668.6 667.2 In all measurements of fluorescent whites the spectral energy distribution of the illumination is of paramount importance for the excitation of fluorescent whitening agents in the UV range. With commercially available instruments, however, the energy distribution cannot be controlled. A device fitted in a spectrophotometer to allow empirical adaptation of the UV component of the xenon lamp markedly improved long-term reproducibility and gave better matching to standard source D65. The instrument has been in daily use for three years. 1. Problem When fluorescent whites are measured using spectrophotometers, the spectral energy distribution of the illumination at the point where the test sample is positioned must be identical with that of the standard source for which the tristimulus values are calculated. In European measuring instrument
s the preferred method
of replicating standard source D65 recommended by the CIE is to fit a high-pressure xenon lamp. Its spectrum is often modified by special sphere coatings and filters to achieve the closest possible match to the spectrum of this standard source (fig. 1) [1; p. 50, fig. 4]. Most users of these instruments cannot check whether or not the sample irradiation in their instrument matches the standard source quoted by the manufacturer. Against this, however, it is fairly easy to ascertain that the quoted energy distribution is not constant1. Various factors can alter energy distribution, e.g. aging of lamps, differences between lamps, soiling of optical system and inner wall of sphere, voltage and current fluctuations, Ambient temperature
, etc. Our primary aim was to ascertain and, where possible, offset the medium and long-term changes brought about by soiling and aging of lamp and sphere coating. The need to measure fluorescent white samples with constant UV excitation arises from the colorist's demand for measurements to be comparable, regardless of when taken. The growing trend towards colorimetric rather than Visual assessment
of whiteness lends force to the colorist's demand.
* CIBA-GEIGY AG, Basle
1 Described in section 2
Fig. 1: Relative spectral energy distribution of standard source X65 and filtered xenon lamp Constant UV excitation is indispensable for rating fluorescent white samples on a white scale using a formula. Relative assessments valid only for the measuring instrument used and for the condition of its illumination system at the time are inadequate in such cases because the white scale rating can be visually checked at any time, albeit only within a given tolerance range defined by the person of the assessor, the prevailing conditions of illumination, and the method of matching . To obtain levels of whiteness that match the white scale when changes occur in the spectral energy distribution of the illumination system and consequently in the tristimulus values, the parameters of the whiteness formula have to be continually matched to the changes of the measuring instrument. The aim we set ourselves, however, was to adapt the illumination system to the method of calculation laid down.
2. test method
Table 1: Instrument specifications
XBO 150 W/1
Monochromator 33 interference filters
Measuring range 380 - 700 nm
Sphere coating aluminium
Sphere diameter 330 mm
Instrument B spectrophotometer XBO 250 W glass prisms 380 - 700 nm based on BaS04 100 mm
Instrument C Colorimeter XBO 150 W/1 3 AGB colorimetric filters visible region of spectrum based on BaS04 150 mm
(Instrument A = Pretema Spectromat FS-3A; Instrument B = Zeiss DMC 25; Instrument C = Zeiss Elrepho)
3 Fluorescent standards are useful for qualitatively studying changes in the spectral energy distribution of illumination systems
, including the ultraviolet region of the spectrum. The steps of the CIBA-GEIGY Plastic White scale containing FWAs provided us with washable fluorescence standards with high light fastness . Initially, over a period of about 12 months
, we investigated the instruments listed in table 1, all of which have a polychromatic illumination system.
Fig. 2: Long-term test with three instruments for whiteness measurement
With each of the three instruments we first measured all Twelve Step
s of the same CIBA-GEIGY Plastic White Scale. We measured the dull side of each plate at 5 different points and averaged the spectral radiance factors or "measured tristimulus values". From the calculated tristimulus values, using whiteness formula 
W = 2Y - 1280.3x - 3555.2y + 1505.9
we worked out the whiteness. Subsequently, as a control, we also measured in the manner described the step containing the highest concentration of FWA each time the instrument was used. The results are shown in fig. 2. The whiteness formula used gives values in CIBA-GEIGY Whiteness Scale units. The visual distinguishing threshold, i.e. the smallest visually perceptible difference between whites, is 5 whiteness units. The Statistical evaluation
of the measurements is shown in table 2. The number of random tests with each instrument depended on how frequently it was used during the test period.
Table 2. Statistical evaluation of degrees of whiteness measured in the tests Instrument A Instrument B Instrument C
No. of random tests n 68
tolerance (95 %)
± 5.4standard deviation
4 The recorded changes are differences of Fluorescence intensity
. They are caused almost exclusively by the tristimulus value Z. The differences in whiteness are a multiple of the distinguishing threshold mentioned. Fig. 3 shows the spectral radiance factors of two measurements whose color stimulus specifications give a visible difference in whiteness of 6.9 units. The performance of each of the three instruments tends to fall off markedly over the test period (fig. 2). Abrupt, pronounced improvements occur after lamp changes, cleaning of optical system and sphere (aluminium coating) or re-coating of sphere (barium sulphate). Fig. 3: Spectral radiance factors of two measurements of the same fluorescence standard with UV excitation of different intensities. They give a whiteness difference that lies slightly above the distinguishing threshold. 3. UV monitoring and stabilizing The true tristimulus values of an unused specimen of the fluorescence standard were determined for D65/2° by the EITLE-GANZ method . Measurements of this standard with all three instruments gave too low a Z value. The UV component hence had to be increased. Fig. 4: Spectral transmission curves for lamp cylinders (left) and filter cylinders (right)
Fig. 5: Instrument A with inbuilt UV filter cylinder
1 Xenon lamp XBO 150 W/1 2 sphere coating 3 measuring aperture 4 reference support 5 beam changer 6 quartz cylinder
7 UV filter cylinder 8 mount 9 raising and lowering rod 10 felt gaskets 11 setting knob to calibrate UV 12 cover
This was possible with instrument A, which was used for all routine measurements and in which the xenon lamp - encased in a glass cylinder - is fitted inside the sphere. This glass cylinder was replaced by a quartz cylinder. Transmission in the near UV region was far better with this cylinder (fig. 4, left). Measurement of the standard using this cylinder gave a Z value that was too high.
We then looked for a way of controlling the UV excitation so that the color stimulus specification of the fluorescence standard would be essentially identical with the one determined by the EITLE-GANZ method. The approach adopted was to partly cover the UV-permeable quartz cylinder with another cylinder having the largest possible UV absorption without materially affecting the visible region of the spectrum.
As the ordinary filter glasses are available only as plane glass, we fell back on glasses that are cylindrical to start with. The choice was limited. Fig. 4 (right) shows the spectral transmission of a few of the glasses tested. The best glass for our purpose was found to be Jena KW. We used a cylinder made of this glass.
We designed a mechanical arrangement enabling the filter cylinder to be displaced vertically from outside when the instrument is ready for operation (fig. 5). It can be set in any position from almost complete coverage to full exposure of the light source
6 At the design stage and during subsequent use the main problems were: a) space for mounting further components was very limited b) achieving a durable bond between glass and metal was difficult. Problem b) was only solved satisfactorily after a great deal of experimenting. The difficulty here is the difference between the coefficients of expansion of the two materials, and the effect of temperature and UV radiation
on the test adhesives2. 2authors' note while proofreading: A system has now been found which requires no bonded joints. The glass cylinder is connected to the raising and lowering rod by metal clips. 4. Results For routine use the instrument is calibrated as follows. First the visible region of the spectrum is calibrated in the usual way against barium sulphate. The fluorescence standard is then applied to the test aperture and the UV blocking cylinder positioned so that a predetermined radiance factor is attained at the wavelength of the fluorescence peak. With this instrument setting, which is valid for the particular fluorescence standard used, the test measurements are carried out, after the calibration has been re-checked in the visible region. Fig. 6 shows the results of test measurements with the converted instrument over a period of 7 months. The statistical evaluation of these measurements is shown in table 3.
Fig. 6: Long-term testing with the converted instrument A
Table 3: Statistical evaluation of the whiteness values measured in the tests with the converted instrument
No. of random tests n mean x tolerance (95 %) standard deviation s lowest value highest value difference
Instrument A with UV calibration 180 237.4 ± 1.6 0.79 236.1 238.6 2.5
The instrument has been in daily use for three years. Long-term reproducibility is completely satisfactory, even with FWAs whose excitation spectra differ from that of the fluorescence standard. This is shown by test measurements carried out at long intervals with several sets of the CIBA-GEIGY cotton White Scale.
7 The xenon lamp is replaced immediately the UV excitation has become too weak, even when the blocking cylinder is fully withdrawn. This used to be the case after a lamp had burned for times between 720 and 1085 hours. Two lamps even had to be changed after 27 and 141 hours. After exposure several times daily to the intense radiation of a xenon lamp, a fluorescence standard has to be replaced by a new one after a few months. There must be a sufficient reserve of fluorescence standards with known tristimulus values to check it and ensure continuity of measurement.
5. Summary In tests with three measuring instruments
fitted with xenon lamps, it was shown that changes occur in the spectral energy distribution of the light sources, and that these changes are an obstacle to measuring fluorescent whiteness with adequate medium and long term reproducibility. An accessory for a spectrophotometer was designed and tested in continuous service for a number of years. With this accessory FWAs with similar excitation spectra to the fluorescence standard used for calibration can be measured in light approximating closely to that emitted by standard source D65. The light used can be kept constant within the limits of human ability
to distinguish differences between whites. The method can be applied to other spectrophotometers and colorimeters provided that the spectral energy distribution of the source has a sufficient excess in the UV region. In the case of instruments with a lamp outside the sphere, the mechanical problem should be easier to overcome.
Literature  Wyszecki, G., Development of new CIE standard sources for colorimetry. Farbe 19 (1970), p. 43-76  Eckhardt, G., Visuelle WeiЯbewertung. CIBA-GEIGY-Rdsch. 1973; Nr. 1, p.10-13  Griesser, R., Instrumentelle WeiЯbewertung. CIBA-GEIGY-Rdsch. 1973; Nr. 1, p. 1425  Ganz, E., Whiteness measurement. J. Color and Appearance 1 (1972); Nr. 5, p. 33-41  Eitle, D., u. E. Ganz, Eine Methode zur Bestimmung von Normfarbwerten fьr fluoreszierende Proben. Textilveredlung 3 (1968), p. 389-392; Nr. 8
Authors' addresses: El.-Ing. HTL F. Gдrtner MeЯtechnik FC 2.64 Dipl.-Ing. R. Griesser Farbmessung FC 6.22 CIBA-GEIGY AG CH-4002 Basel
Manuscript received: 1 December 1974
F Gartner, R Griesser