Detector-based calibration of a solar UV spectroradiometer

Petri Kärhä, Heidi Fagerlund, Antti Lassila, Farshid Manoochehri and Erkki Ikonen

Metrology Research Institute, Helsinki University of Technology, Otakaari 5 A, FIN-02150 Espoo

Kari Jokela, Kirsti Leszczynski and Reijo Visuri

Non-Ionizing Radiation Laboratory, Finnish Centre for Radiation and Nuclear Safety, P.O.B. 14, FIN-00881 Helsinki

Introduction

The accuracy of the Finnish solar UV monitoring network is based on intercomparison of the network instruments with an accurately calibrated reference spectroradiometer. The uncertainty of the lamp-based calibration of the reference spectroradiometer is 3,6 % (k=2) and the overall uncertainty of solar UV irradiance measurements is 8% (at 310 nm) /1/. Hence, the improvement of the calibration methods and development of new standards with lower uncertainties are of primary importance.

Cryogenic absolute radiometers /2, 3/ are currently the most accurate devices to measure optical power. Uncertainties as low as 0,04 % may be achieved in detector calibrations /4/.

In this report we present a new detector-based method for UV irradiance calibration and some critical aspects encountered during the development work. Transferring the accuracy of the cryogenic absolute radiometer to the desired UV-wavelengths requires specific transfer standard radiometers. The characterization of these radiometers requires careful measurements of e.g. aperture area, filter transmittance and spectral responsivity of the detector. In addition to these, accurate temperature control is needed.

Standard detectors of optical power

In the Helsinki University of Technology the cryogenic absolute radiometer is operated twice a year to calibrate secondary transfer standard detectors. The methods used in calibrations have been thoroughly presented in /4/. Usually a trap detector /5/ is used as a transfer standard, but also other type of detectors e.g. pyroelectric radiometers may be calibrated. The calibration is performed using a power- and frequency-stabilized 543,5 nm He-Ne laser as the light source.

The trap detector consists of three photodiodes arranged in such a way that the light undergoes multiple reflections from the photodiode surfaces as depicted in Fig. 1. This arrangement reduces the reflection of incident light to a negligible level which is a desired feature in a radiometric system. In addition to the low reflectance, trap detectors have predictable spectral responsivity in the visible, high sensitivity, low noise characteristics and good spatial responsivity. They are also linear, stable in time and relatively cheap.

The responsivity of the trap detector is not predictable in the ultraviolet range. Therefore a pyroelectric detector with a flat spectral responsivity is also needed in UV calibrations. A pyroelectric detector by itself is not sufficient for UV calibrations for several reasons. They tend to be unstable, insensitive and have a poor spatial responsivity.

Fig. 1 Arrangement of the photodiodes in a trap detector. A three dimensional configuration reduces sensitivity to polarization of the incident light.

A setup to compare detectors in the UV and visible spectral ranges has been constructed using a 1 kW xenon lamp and an irradiance monochromator. The relative spectral responsivity of the trap detector is compared with the flat spectral responsivity of the pyroelectric detector. Combined with the absolute responsivity calibration against the cryogenic absolute radiometer this gives the absolute responsivity of the trap detector at UV wavelengths.

Scheme for detector-based UV calibration

A setup for calibrating the standard lamps used in the calibration of the solar UV spectroradiometer is presented in Fig. 2. A special wavelength selective filter radiometer is formed using a trap detector, a precision aperture and a UV bandpass filter. All the components are packed closely together to form a compact radiometer. A temperature controller keeps the temperature of the bandpass filter stable during the measurements in the vicinity of a high-power lamp.

Fig. 2 Setup for the detector-based calibration of the UV standard lamp.

The photocurrent of the filter radiometer is calculated as

i=A*int(E(lambda)*T(lambda)*R(lambda))dlambda, (1)

where E(lambda) (W cm-2 nm-1) is the spectral irradiance of the lamp, A (cm2) is the area of the aperture, T(lambda) is the filter transmittance and R(lambda) (A W-1) is the responsivity of the trap detector. Parameters A, T(lambda) and R(lambda) can be determined separately. The shape of the lamp irradiance E(lambda) can be measured using the spectroradiometer within the narrow transmission band of T(lambda). Substituting the measured photocurrent into Eq. (1) gives the irradiance of the lamp and thus the calibration of the spectroradiometer within the passband of the filter.

Results and discussion

So far only preliminary results have been achieved using the detector based calibration system.

The spectral transmittance of an interference filter with a center wavelength of 312 nm and a bandwidth of 10 nm (full width at half maximum) was determined using a reference spectrometer with a deuterium lamp as the light source. The spectral responsivity of a trap detector was calibrated in the spectral region 280-340 nm. The areas of three precision apertures with nominal diameters 3,0, 3,5 and 4,0 mm were measured optically using a laser interferometer.

The characterized filter radiometer was compared with a 1 kW halogen standard lamp which is traceable to the spectral irradiance scale of NIST. The measured photocurrent was 8 % higher than the value based on the specified lamp irradiance. This may be considered as a fair first result taking into account the complexity of UV measurements. There are several possible reasons for the difference.

Of main concern are the properties of the interference filter. Great care should be taken in making the transmittance measurements. The filter has steep slopes which are very difficult to measure. Accurate measurements require a good wavelength scale. Also the blocking of the filter outside its passband is critical.

The filter transmittance was measured in the region from 400 to 800 nm to check for any leakage, none was found. However, the resolution of the measurements was 5*10-6, and a leakage of this order of magnitude in longer wavelengths can cause considerable errors. It was calculated that e.g. a leakage with a maximum transmission of 5*10-6 at 700 nm wavelength would cause a 5 % error, assuming a bandwidth of 100 nm for the leakage. This is due to the spectral characteristics of the halogen lamp and the trap detector.

Some error may also be caused by the interreflections between the filter and the aperture and between the filter and the trap detector.

The effect of the aperture size on the responsivity of the radiometer was also studied. It was noted that increasing the aperture diameter from 3,0 mm causes slight decrease in the responsivity (approximately -0,8 % with the 3,5 mm and -1,5 % with the 4,0 mm diameter aperture). This is due to the limited angular response of the trap detector.

Acknowledgements

This work has been supported financially by the Technology Development Centre, Academy of Finland and Centre for Metrology and Accreditation.

References

/1/ K. Leszczynski, K. Jokela, R. Visuri and L. Ylianttila, "Accuracy problems in solar UV radiation measurements" (this conference).

/2/ J. E. Martin, N. P. Fox and P. J. Key, "A cryogenic radiometer for absolute radiometric measurements", Metrologia 21, 147 (1985).

/3/ T. Varpula, H. Seppä and J.-M. Saari, "Optical power calibrator based on a stabilized green He-Ne laser and a cryogenic absolute radiometer", IEEE Trans. Instrum. Meas. 38, 558 (1989).

/4/ P. Kärhä, A. Lassila, H. Ludvigsen, F. Manoochehri, H. Fagerlund and E. Ikonen, "Optical power and transmittance measurements and their use in detector-based realization of the luminous intensity scale", Optical Engineering (in press).

/5/ N. P. Fox, "Trap detectors and their properties", Metrologia 28, 197 (1991)