Caution urged when considering LED light sources for light-sensitive materials

Below is a letter sent to the Green Task Force of the American
Institute for Conservation of Artistic and Historic Works, at their
request, clarifying the potential problems of using LED's to
illuminate art cultural heritage materials.  Members of the task
force suggested we also post this email to the PACIN list serve and
it has also been circulated at the RCAAM distlist.

First let me say that LED's present a low-energy-of-operation
alternative for the illumination of non-light-sensitive museum
materials where the color discrimination of the object is not
critical to its appreciation or understanding and in offices and
non-collection areas.

I believe the existing reported data demonstrates that LEDs are a
potentially damaging light source for light-sensitive museum
materials and are deficient where accurate color discrimination
within the human visible spectrum is required.  The deficiencies of
LED light sources, both for color rendering and for their narrow
emission spectra, is not new information. The CIE data first
explaining these deficiencies in detail was published in 2004.  I
believe the manufacturers are acutely aware of these deficiencies.
Jim Druzik of the Getty Conservation Institute recently returned
from an NIST meeting where the efforts of the manufacturers to
overcome them was frequently discussed. Unless conservators ask very
pointed and direct questions, my sense is that manufacturers have
not initiated discussions that reveal the risks of LED light sources
to light sensitive materials.

It is important to remember that 50 lux, 65 lux or any light level
from an LED source is *not* the same as 50 lux, 65 lux, etc. from an
incandescent source. The spectral power distributions (SPD's) are
different. Guidelines for lighting light-sensitive museum materials
are based upon the spectral power distribution of most
halogen-incandescent sources (e.g. MR-16, Par 38 lamps) and their
fading behaviors with materials of known sensitivities, the ISO Blue
Wool Standards. Since light sensitive materials generally fade first
as a result of exposure to the visible light spectrum, and since
most museums are aware of the need to filter for UV and IR
radiation, the fading and color-shifting behaviors of light
sensitive materials may be significantly accelerated as a result of
absorption of the elevated narrow energy bands within the visible
spectrum delivered by LED sources.

I have studied published SPD's of numerous un-filtered MR-16 halogen
incandescent lamps, 2-phosphor white LED's, a red, green, blue,
amber white appearing LED and a red, green, blue "white" appearing
LED, each taken off a standard spectralon tile at 660 lux.  The
highest power output of a broad, continuous incandescent lamp is
roughly 12 microwatts nanometers while the power of the narrow wave
peaks produced by the LED's are 20% to 400% higher than the halogen
MR-16 at the same light level.

Since it is well established that light-sensitive
materials--typically illustrated by Japanese wood block prints, dyed
textiles, watercolors, pastels, color photographic images,
biological specimens etc.--are damaged (ie undergo color shift such
as loss of coloration--fading, darkening, yellowing or color-shift)
by radiant energy *within* the human visible spectrum, any
light-sensitive material whose damage spectra (within the absorption
spectra as opposed to the reflection spectra) contains any of these
isolated LED output peaks would undergo damage at 20% to 400% faster
rates than if lit at the same light levels with an unfiltered MR 16.

To this information, Jim Druzik, Senior Scientist with the Getty
Conservation Institute and project leader for the joint O'Keeffe
Museum - GCI - University of Texas El Paso light damage prevention
project adds:

   "It has been demonstrated that some LED mixtures will accelerate
    damage more rapidly on natural yellow dyes. The SPD is useful
    for calculating gamut area mapping and CRI, so between those two
    metrics one can determine color rendering. But there are no
    damage-activation spectra of most dyes and pigments. Because so
    much is piled on to narrow bands in LEDs, the possibility of
    hole burning is very possible and trying to be "green" with so
    little knowledge is Russian Roulette and the LED green committee
    is playing with fire. At this time I don't think LEDs are ready
    for prime time both from the point of view of color rendering
    and conservation. Clearly two LEDs with a bridging phosphor is
    better than all LEDs but some manufacturers have driver
    problems, green is mismatched for optimal color rendering, and
    only one manufacturer hits the Energy Star specification.  I
    know all this because I just came back from NIST which is a
    hotbed of research on LED lighting and that's most of what all
    the current research is about.  The one Energy Star qualifier
    does achieve an amazing 65 lumens/Watt but I didn't see it
    turned on so I don't know how good its color is. As you know, it
    could be any CCT yet if displaced off the spectrum locus for
    blackbody illuminants, it'll still could be a disaster.  I have
    it on my "To Do List" to buy a few of the Energy Star qualifiers
    and measure their spectrum.  I just got a piece of NIST software
    which will display two sets of Munsell swatches according to
    their appearance side-by-side with a reference illuminant. That
    might be the way to go first."

Jim Druzik and I will be reviewing the latest Federally confirmed
data at the DOE Energy Efficiency and Renewable Energy testing
laboratory and the Committee of Illuminating Engineering Society of
North America, author of the RP-30 recommended practice for Lighting
Museum and Art Galleries, over the next two weeks. We may find that
there are sources that do less damage to light sensitive materials
than the 2007 lamp data I refer to here. If so, I will update the
AIC-Green Task force, the PACIN listserv and the RCAAM listserv and
post a link to any DOE-EERE-IES joint statement to museums.

The color rendering indexes of these light sources, when tested in
independent studies, has been near 80% at best across the entire
human visible spectrum.  Experimenters have documented considerable
human impairment with color discrimination with materials viewed
under both 2-phosphor white LEDs and RGBA and RGB 4- and 3- source
"white" LED sources. If exposing light-sensitive materials to damage
within exhibitions is to severely limited, then certainly the few
times they are exhibited ought to be done so that people can
accurately discriminate between the colors we are so carefully
trying to preserve.  To expose them to damage *and* prevent an
accurate sense of their visual characteristics seems to be the
biggest crime of all.

We hope this clarifies the need for using extreme caution when
considering LEDs for the illumination of light sensitive materials.
To reiterate our earlier point, for non-light-sensitive materials
where the color discrimination of the object is not critical to its
appreciation and understanding and in offices and non-collection
areas LED's present a low-energy-of-operation alternative. For light
sensitive museum materials, there is no existing data that
demonstrates that they are either safe or effective at the 50 to 70
lux levels commonly used to extend the life of these materials
during exhibition.

Dale Kronkright
Head of Conservation
Georgia O'Keeffe Museum
217 Johnson Street
Santa Fe, New Mexico 87501, USA
505-946-1041
Fax: 505-946-1093


                                  ***
                  Conservation DistList Instance 23:38
                   Distributed: Sunday, April 4, 2010
                       Message Id: cdl-23-38-002
                                  ***