Beating the LED streetlights: IDAS light-pollution suppression LPS-D2 filter

An IDAS LPS-D2 filter suitable for two-inch push-fit telescope camera accessories with an M48 × 0.75 connection thread. The filter has male and female threads on either side, hence it is stackable with other filters and M48 adaptors. The LPS-D2 is also available in 52mm and Canon APS-C clip-filter formats. All images: Ade Ashford.

At a glance

Type: light-pollution suppression filter for low-/high-pressure sodium vapour and LED lights
Coating technology: Ion-Gun Assisted Deposition (IGAD)
Suitability: DSLR and astro cameras
Connection thread: M48 × 0.75 (male and female on either side, hence stackable)
Substrate thickness: 2.5mm
Diameter of filter glass: 49mm
Price: £175 (M48 and 52mm); £185 (Canon APS-C clip filter)
Manufacturer: ICAS Enterprises, Japan

Light pollution is a regrettable fact of life for most of us. By night, the sky over our cities and towns – even villages – is increasingly awash with the glare of unnecessary or misdirected artificial light. This is not only a tremendous waste of energy, but it upsets nocturnal ecosystems and harms human health, disturbed sleep patterns and the disruption of natural circadian rhythms.

For almost three decades, Tokyo-based ICAS Enterprises’ IDAS Division has been responsible for manufacturing some of the world’s most respected interference filters for suppressing light pollution for astronomers. Their LPS-D1 filter made its debut in 1991 at a time when the main sources of artificial illumination in our towns and cities were low- and high-pressure sodium vapour and mercury vapour lamps. Fortunately for astronomers, both sodium (Na) and mercury (Hg) vapour lamps share a common characteristic: they typically emit light in specific and largely narrow wavelength bands of the spectrum – so-called emission lines – that can be removed by an interference filter.

The LPS-D1 was designed for one-shot CCD/CMOS colour cameras and DSLRs to eliminate the glow from low-pressure sodium and high-pressure mercury street lights, while substantially reducing the peak intensities of high-pressure sodium light emissions. Both the IDAS D1 and P2 filters pass the desirable spectral lines of hydrogen-beta, oxygen-III, hydrogen-alpha light from nebulae, plus diatomic carbon (C2, the so-called Swan bands) from comets. The LPS-P2 is virtually identical to the D1 except for a slightly greater red sensitivity encompassing sulphur-II emissions.

The challenge posed by white light LEDs

As many of us up and down the United Kingdom and around the world are now acutely aware, the nature of street lighting is rapidly changing. The mellow yellow glow of low-pressure sodium light is being replaced with the energy-efficient yet brilliant white glare of light-emitting diodes (LEDs). I never thought that I would lament the passing of sodium street lights, but at least their light was relatively easy to mitigate. White LEDs, on the other hand, emit what is largely a continuous spectrum across a swathe of wavelengths (or colours, if you prefer) which is far harder to filter out.

If you consult the accompanying graph that shows the spectral profile of a typical white LED in blue, you will see immediately that it emits its greatest intensity of light – almost 98 per cent transmittance – in a well-defined peak at a wavelength close to 463 nanometres (nm), which is 4.63 × 10–7 metres, or 0.000463mm. Thus, the peak emission of a typical white LED is actually in the violet end of the blue region of the visible spectrum, at wavelengths that research has shown disrupts human circadian rhythms by keeping our brains in an ‘awake’ state.

After the initial peak intensity, the white LED’s transmittance rapidly drops to around nine per cent at a wavelength of about 486nm in the blue–green part of the spectrum. Thereafter, the transmittance rises steeply to a secondary, broader peak intensity of 53 per cent at about 560nm in the yellow part of the visible spectrum before gradually tailing off to zero in the far-infrared. If we were to use a conventional IDAS LPS-D1 or P2 filter on a white LED, then its peak intensity and much of its broader secondary intensity would not be filtered out. Clearly, we need another type of interference filter.

The real-world performance of an IDAS LPS-D2 filter plotted in yellow as wavelength versus percentage transmittance, using actual measured production data. For comparison, the twin-peaked emission spectrum of a typical white light LED is also shown in blue. The LPS-D2 filter has three transmission peaks in the violet, blue–green and red portions of the visible spectrum that include all of the desirable ‘nebula light’ emission spectra (shown as vertical dashed green lines) of hydrogen-beta (Hβ), oxygen-III (O-III), Swan bands (C2), hydrogen-alpha (Hα) and sulphur II (S-II). The vertical dashed red lines are emission spectra of artificial mercury (Hg) light and natural airglow that partially intrude into the LPS-D2 transmission band.

The IDAS LPS-D2 versus white LEDs

I was able to obtain data for the LPS-D2 filter based on a laboratory analysis rather than just rely on the design specification. The accompanying graph is a plot of the filter’s transmission versus wavelength in yellow, superimposed with that of a typical white LED in cyan. Where the white LED’s light intrudes into the D2 filter’s transmission curve is shown in green. Furthermore, the graphic shows the emission spectra of desirable nebula light (vertical dashed green lines), plus residual sources of light pollution that we wish to remove or mitigate (vertical red dashed lines). At the top and bottom of the graphic we see a continuous spectrum showing the approximate colour that corresponds to a specific wavelength; V = violet, B = blue, G = green, and so on.


The IDAS LPS-D2 is clearly very effective at removing the initial and most intense transmission spike from a typical white LED centred around 463 nanometres. However, when we come to capturing the desirable emission spectra of nebulae and comets in the blue– green part of the spectrum – hydrogen-beta, oxygen-III and diatomic carbon – the intrusion of the LED’s light rises from a transmission of nine per cent at the hydrogen-beta line to around 30 per cent at the Swan bands of diatomic carbon. Note that some high-pressure mercury light pollution at 436nm and 546nm will also be passed by the LPS-D2 filter. Similarly, the second transmission peak of the LPS-D2 encompasses some of the white LED’s secondary peak light at around 52 per cent transmittance, so your white balance will have some strong green dominance. Fortunately, low-pressure sodium light pollution is fully suppressed and by the hydrogen-alpha and sulphur-II emission lines the white LED’s transmission is down to just 18 and 13 per cent, respectively.

Ade Ashford has travelled the globe writing about astronomy and telescopes, serving on the staff of astronomy magazines on both sides of the Atlantic.