Which light sources in a city are mainly responsible for the light pollution in its suburban area? This is a fundamental question that must be answered before any steps in lowering the light pollution can be entered. A night-sky spectrometer can help in answering this question. The spectrum of the light dome over a city can be decomposed into typical groups of light sources (high pressure sodium lamps, LED streetlamps, fluorescent lamps, ...). For more information see our Spectrum Analyser software and some published case studies.
Our light pollution spectrometer is based on Shelyak Aply-600 spectrometer and cooled Atik 414EX mono CCD camera. Maximum resolution of the Alpy-600 spectrometer is about 1 nm (25 μm slit) and its first order spectrum (370 - 750 nm) fits perfectly onto 9 mm wide sensor of the CCD camera. The resolution of 1391 pixels (6.45 μm pixel size) is optimal for the resolution of the spectrometer. We use 50 μm slit (ca 2 nm resolution) - that is enough for our application and more light is gathered.
The optical design of the spectrometer is as follows:
Alpy-600 uses grisma as the dispersion element. Grisma is a grating placed on a prisma. Such combination results in on-axis direction of diffracted rays that allows for very simple and compact design. However, such design introduces also a few distorsions in recorded spectra, mainly:
Both distorsions, that must be corrected by software during processing the spectrum, are shown in next picture:
Our spectrometer contains also the Alpy calibration module that allows for fast and comfortable wavelength calibration of recorded spectra and for flat-field correction. The calibration module contains a discharge tube producing known line spectrum and a halogen lamp producing continuous spectrum.
The whole system is completed by simple objective lens with focal length of 100 mm. The diafragma reduces the aperture to f/5 (input aperture of the Alpy-600 spectrometer is f/4, optimal performance is claimed to be for f/5 input beam).
To complete tj spectrometer, we have developed an 3D-printed objective with lens and diafragma (left), connection between the calibration module and the spectrometer (in the middle) and interface to iPANO mount and the holders:
Optimal exposures of our system (binning 1x1) are:
We use Demetra software for spectra processing. Thanks to the calibration module the whole procedure is practically automatic. The only complication is that we can't record the spectrum of a known star for intensity calibration in "Instrument response" phase (we have no astro-telescope in front of the spectrometer). An overcome is to record during daytime calibration session the spectrum of any known continuos light source (FITS file). The known spectrum of the light source (DAT text file) has to be placed into ISIS-database as a synthetic star. Afterwards the IR-curve (FITS file) can be constructed in Demetra in a standard way and used later automatically.
Original uncorrected and uncalibrated spectrum of a white paper illuminated by direct sunlight:
Recorded flat-field (1s), recorded calibration spectrum (30s) and geometrically corrected calibration spectrum:
Calibration spectrum profile after automatic detection of spectral lines by Demetra software:
Fully calibrated and corrected spectrum profile of a white paper illuminated by direct sunlight:
Thanks to to the objective lens, our spectrometer has a narrow field of view (2.7 deg x 1.1 deg) without significant loss of sensitivity. Measured FOV in two directions is as follows:
The narrow FOV is very useful at recording the spectra of the light dome at elevations ca 10 - 20 deg. Narrow FOV eliminates direct light from light sources in the city. Of course, whole objective can be easily removed (unscrewed), if needed. Without objective, the FOV of the spectrometer is about 20 deg x 16 deg.
Spectrum of the light dome over Vienna was measured from Devinska Nova Ves (ca 50 km ost from Vienna) on 14./15.9.2020. The atmosphere was quite turbid (but no clouds), so the light from the light dome was significantly absorbed/scattered on the path from Vienna to the observing site. The spectrometer was oriented toward the centre of Vienna, the elevation of the spectrometer was 20 deg. The CCD sensor was cooled down to 0 deg Celsius. The single exposure time was 5 min, totally 9 spectra were recorded.
Single recorded spectrum is in next figure. Note many hot pixels but very low thermal noise, typical for Atik 414EX cameras.
Calibrated and corrected spectrum profile (the whole job was done automatically by Demetra software):
The result is a very nice spectrum profile without any noticeable noise. The blue part of the spectrum is significantly suppressed due to high atmospheric turbidity.
Some pictures of the spectrometer on iPANO robotic mount and in the transport case:
František Kundracik
Department of Experimental Physics
Faculty of Mathematics, Physics and Informatics
Comenius University
Bratislava, Slovakia