Agent of Deterioration: Light, Ultraviolet and Infrared

Table of Contents

Light Does Not "Contain" Ultraviolet and Infrared

In the museum business, one often hears the expression "the light contains ultraviolet and infrared." This is incorrect and will lead to unnecessary confusion in practical discussions of museum lighting. Light, by definition, is the band of radiation to which our eye is sensitive. Ultraviolet radiation (UV) and infrared radiation (IR) are not visible. They are the bands of radiation on either side of the visible band (ultra means beyond, infra means below). Informally, the term radiation is dropped. We usually speak of ultraviolet and infrared, or simply of UV and IR. Ultraviolet and infrared are not necessary for seeing (except in rare cases of UV fluorescent colours); therefore, they are not part of the dilemma between seeing and damaging, they are simply damaging.

It is correct, however, to state that some light sources emit ultraviolet and infrared, or that museum lighting may cause UV and IR deterioration

The Radiation Spectrum

Figure 1 plots the adjacent bands of UV, light, and IR on the conventional scale of wavelength (in nanometers – nm). The reciprocal scale for photon energy is also shown (in electron Volts – eV) to show how photon energy climbs rapidly in the direction of the UV band.

Radiation bands emitted by various light sources are shown by light grey bands. Bands of radiation blocked by some filters are shown as dark grey bands. Convention assigns the boundary between UV and light at 400 nm, but slight perception begins at 380 nm. This boundary of 380 nm is often used by the window industry in rating the UV characteristics of glazing.

The different types of damage typical of UV, light, and IR result from their different photon energies. The photochemistry that underlies much of the disintegration of materials and production of yellow by-products typical of UV exposure requires energies greater than about 3 eV, whereas the photochemistry typical of colourant fading, as well as the operation of our retina, occurs in a range between about 2 eV and 3 eV. We are fated, in fact, to see in the same band as that which causes sensitive colourants to fade, given the related photochemical phenomena. Infrared photons are not energetic enough to initiate any of the forms of photochemistry driven by UV or light, so their effect is simply a heating of the surfaces that absorb them.

Measuring Light and its Exposure

The technical term for the amount of light falling on a surface is "illuminance," but informal phrases such as "light intensity" or "lux level" are used in the museum literature. The unit is lux (both singular and plural). Old light meters may still use the imperial unit "foot candles." Their readings can be converted to lux by multiplying them by 10 (10.76 precisely). Many companies make light meters, also called lux meters. Some of these meters are especially designed for museums so that they include UV and even RH and temperature measurement.

Figure 2 plots various situations and their lux levels across the vast range of the human eye, from moonlight to sunlight.

Our eye changes from night vision (scotopic) to colour vision (photopic), with a mixed range (mesopic) between. Rate of light damage is proportional to intensity; therefore, it increases 10 million times from moonlight to sunlight, and 1,000 times from good museum levels to sunlight.

The total exposure or dose of light on a surface is the product of light intensity (lux) and time (hours). In museums, the practical unit is millions of lux hours, abbreviated Mlx h, and pronounced "mega lux hours."

Measuring UV and its Exposure

Rather than measure the intensity of UV directly, the convention in museums has been to measure it relative to the intensity of the light, in units of microwatts (of UV) per lumen (of light), abbreviated µW/lm. This ratio is much more useful than the direct measure of UV when characterizing light sources in a museum and characterizing the benefits of any UV filters on these sources. Various companies make UV meters for museums

Although some authors have suggested doing so, it has not been conventional to measure UV exposure in museums. One can express it if needed as a combination of the light exposure in Mlx h and the UV (ratio) in µW/lm, as will be done later in Table 5, Sensitivity to UV.

Measuring IR

There are no museum conventions or common instruments for measuring IR because it is not nearly as important as UV or light to collection damage. To make a simple instrument for measuring the heating potential of IR from a light source, paint the bulb of an ordinary outdoor glass thermometer with a matte black paint. Place the bulb in the light beam near the object and wait until the temperature stops rising (several minutes). To see if the temperature rise is a problem, refer to the section on Incorrect temperature. As a common-sense alternative estimate, place your hand in the light beam (at the point it might strike artifacts) and use a piece of cardboard to alternately illuminate and shade your palm. If you feel a noticeable warming due to the light, then those artifacts identified as sensitive to "temperature too high" in the section on Incorrect temperature will be at risk.

The Benchmark is 50 lux

When guidelines for museum lighting were first explored 60 years ago, colour science had established that 50 lux was enough to ensure that the human eye was operating well within the range of full colour vision (consult Figure 2); therefore, conservation adopted it as the benchmark level for museums. Since then, however, the public has voiced complaints about low light levels in museums. Although our responsibility for the future viewer will always force us to use low light levels for some objects, it is useful to understand the validity of the statement "I cannot see the objects."

A more precise description of our ability to see at 50 lux emerged in the 1980s, centred not simply on whether we could discriminate differences between patches of colour, but whether we could see the tiny details of an object. It emerged that a young person (age 25) viewing a moderately light-coloured object, with a moderate degree of detail, in a moderately complex pattern, in a reasonable period of time, will see all the details almost as well at 50 lux as they will in full sunshine. Unfortunately, they will not see those details as well as they can in sunshine if the object is dark, if the details are very fine, or if the pattern one is looking for within the details is subtle, and the viewing time is limited. Even more unfortunately, someone older (age 65) will need several times as much light to see as well as the youth, even with all necessary optical corrections such as glasses. Recent research has shown that even our ability to discriminate large patches of colour falters as we age.

Adjustments for Everyone to See Better

It is obvious to us all that we see tiny details much better in brighter light, especially if the object is dark, or the details very "soft" (i.e. low contrast), or when one is searching for subtle patterns in these details such as in an etching on handmade paper versus a good facsimile on machine-made paper. Our ability to see objects as real, genuine, and authentic, resides in our ability to see such details. One cannot imagine an institution more devoted to people "seeing the real thing" than a museum; hence, the complaints when they cannot. The question becomes: how much visibility of the real thing should a museum provide, given the steep cost to the lifetime of the objects? And how much more light does this increased visibility require?

If we use the 50 lux benchmark, Table 1 summarizes some simple (and conservative) rules for adjusting visibility for different objects. For a technical summary of the research underlying these adjustments for visibility and the original sources, consult Michalski 1997.

Table 1 does not imply that a museum must make these adjustments, it simply describes the adjustments necessary to maintain good object visibility across various situations. Whether or not one adopts any of these adjustments for visibility depends on the balance with the preservation issues raised in the later sections on deterioration by light and UV. This balance forms the subject of the final section on "Control Strategies."

Adjustments for Older Viewers to See Equally Well

Our visual system is not so much a still camera as a video camera connected to a complex and dynamic processor. As we age, not only do the lenses in our eyes yellow and fluoresce, but more stray light is created from internal scattering, cones and rods decrease in number, and the neural processing deteriorates. This is above and beyond the issues of normal aging that can be corrected with glasses and age-related pathologies that cannot. The factor of times 3 given in Table 1, to give us equal visual access at age 65 as we have at age 25, is smaller than actually necessary, but it does provide most of the benefit.

Lighting Design Mistakes that Reduce Visibility

A "Palette" of Light Sources for Museums

One currently has a daunting range of options for museum lighting. Table 2 summarizes the advantages, disadvantages, costs, and other parameters of currently available light sources.

Colour Rendering Index

Colour Rendering Index (CRI) measures light quality in terms of the viewer's ability to see colours correctly. The scale has a maximum of 100 and no units. CRI is derived by a colourimetric calculation performed on up to 14 different colour samples illuminated by the light source in question, compared to the calculation using daylight or an incandescent lamp as reference. While recognized as imperfect in its correlation with our visual system, CRI is still the best indicator currently available.

There is no international museum standard on what is or is not an "acceptable" CRI, but the Canadian Conservation Institute (CCI) recommends a minimum of 85. Many museums specify greater than 90. That being said, the difference between a compact fluorescent lamp scoring 82, for example, and the guideline 85, is not noticeable by most people in most situations. If such a lamp has major design, cost, and energy advantages, it makes sense to use it. Light sources easily seen as poor, such as the lowest-cost commercial fluorescent lamps, can score below 60.

Note that daylight of CRI 100, after reflecting against a coloured wall or floor, may measure a far worse CRI than light that comes directly from a lamp of CRI 85. If one chooses to illuminate using "bounced" daylight (or any other light source), then the reflector must not be coloured.

Correlated Colour Temperature

Correlated Colour Temperature (CCT) measures the quality of light that passes from "cool" to "warm." This is not a scale of good to bad, unless one is arguing a personal preference for some types of object. The units are degrees Kelvin, abbreviated simply to K. The common terms for this parameter are, unfortunately, contradictory and confusing. A "cool" light source has a high colour temperature and a "warm" light source has a low colour temperature. This comes from our use of the phrase "warm light" to refer to the golden light of sunrise and flames, and "cool light" to refer to the blue skylight that illuminates shaded areas.

With low light levels, as in museums, viewers tend to prefer warmer light similar to that of incandescent lamps, e.g. the 2800 K of standard incandescent lamps, or the slightly higher 3000 K of quartz halogen incandescent lamps. As illumination increases to several thousand lux, preference is for cooler light, 5000 K or higher. The most common energy-efficient lamps (fluorescent lamps) are available in a wide range of colour temperatures. Successful use of compact fluorescent lamps in small museums requires careful attention to colour temperature. Lamps producing warm light, usually marked with 2800 K, or simply "28 K" are generally preferred at lower light levels, as noted earlier. However, lamps producing cooler light (3500 K up to 5000 K) can increase the colour contrast of objects, which may also be desirable. In conclusion, one should always test before making a final choice on colour temperature.

Natural Versus Artificial Light

Proponents of daylight in museums often use the terminological trick of "natural" light to mean daylight, and "artificial" light to mean electric sources, but all light sources are natural, whether glowing stars, glowing filaments, or glowing phosphors. The correct question is whether the CRI is good enough, and, as noted above, both daylight and electric lights can be either good or bad in CRI. The psychological appeal of windows and skylights comes about from the connection to the outside and from the high intensity of the light (when the sun shines). Careful treatment of existing windows by solar screens, blinds, partially closed curtains, and outdoor shutters closed during peak daylight, can reduce the fading and glare risks, while leaving intact the highly desirable visual connection to the outdoors.

Practical generalizations about deterioration by light, UV, and IR

Given the three distinct bands of radiation — light, ultraviolet, and infrared — one can make useful generalizations about the types of deterioration they cause in museums:

• Light fades (or "bleaches" colours). Those colours that fade can disappear within as little as a few hours of direct sunshine, or just a few years at low museum lighting (e.g. some felt tip pen inks, some colour photographs). Those which do not fade may last centuries in direct sunshine (e.g. ceramics, Minoan frescoes). All coloured objects fall somewhere between these two extremes.
• UV causes yellowing, chalking, weakening, and/or disintegration of materials. Chalking of paint media is often mistaken for pigment fading.
• IR heats the surface of objects, and thus becomes a form of incorrect temperature (too high), with all the damage possibilities outlined in the section on Incorrect temperature. IR will not be considered in any detail in this section.

There is some overlap in the forms of deterioration caused by light and UV. Light (especially violet) can cause some of the disintegration and yellowing listed under UV, but only in a few materials, and only very slowly in comparison to UV. In turn, UV does contribute to the fading of colours, but its contribution becomes dominant only for colours that are durable to light.

None of these overlaps reduces the practical reliability of the above generalizations. To reduce the fading of collections due to display lighting, especially the most rapid fading, there is only one option: reduce light exposure. Many museums, private donors, and their framers have assumed that the primary cause of fading is UV, and that a good UV filter would prevent their collections from fading. Some advertisements for UV filters imply the same. For colours that are sensitive to light — the crux of the museum lighting dilemma — UV usually contributes less than half of the fading and often only one tenth; therefore, it does not allow one to think any differently about reducing light exposure. (The exposure scales in the centre of Table 3 quantify this phenomenon.)

Why bother, then, with UV control? Because for many artifacts, such as paintings with permanent pigments or monochromatic prints and drawings, the yellowing and disintegration of the media and support by UV is the major form of deterioration suffered during uncontrolled museum lighting.

Rates of Light Deterioration

Light damages the colours of some objects — most such colours fade (most of the colours in Figure 3a and 3b) and a few darken (the vermilion in Figure 3a). Table 3 summarizes the available data on the rate of this damage. Coloured materials are divided into four broad categories of sensitivity to light: none; low; medium; and high. For each category, the table provides estimates of the time it takes at various lux levels for the fading to start (first be noticeable) and for it to end (almost no original colour left). One can see that although the range, or uncertainty, within a category is very wide, the differences from one category to the next are much wider. Much of the variation in people's perception of whether light fading is a risk or not arises because this range of sensitivity is so dramatic — some colours in old objects that look fragile can indeed last many centuries, while some colours disappear within our own lifetime, or even in just a few years.

(3a) Oil paints, on the left vermilion darkening and; on the right, carmine lake glaze on white, fading. (3b) Drawing inks on paper, all fading. The letters on the samples indicate the following exposures: 0 - unexposed; A - 0.17 Mlx h; B – 1.7 Mlx h; C – 6.2 Mlx h; D – 17 Mlx h; E – 67 Mlx h. Equivalent exposures range from A: 1 day of sunlight or 1 year at 50 lux to D: 8 months sunlight or 400 years at 50 lux. All areas are protected by a UV filter except areas marked with an asterisk (*). Note that the differences between the presence or absence of a UV filter (B vs. B*, C vs. C*, D vs. D*), while sometimes noticeable, are much less significant than differences between different exposures (A vs. B vs. C vs. D).

Table 3: Sensitivity Levels and Fading

The broad sensitivity categories of Table 3 (high, medium, and low) were adopted in a recent international guideline for museum lighting (CIE 2004). They are defined using the industrial lightfastness standards known as the ISO Blue Wools. These are a set of textiles, originally numbered #1 to #8, each about 2 to 3 times as sensitive as the next. High sensitivity was defined as materials rated #1, #2, or #3; medium as #4, #5, or #6; and low as #7, #8, or higher (more were added to the original eight Blue Wools as needed by industry). The Blue Wool numbers are the main route into the literature on colourant sensitivity, as reviewed in Michalski (1987 and 1997) and as summarized in a more detailed version of Table 3 contained in the CIE guideline (2004).

The conversion of a Blue Wool rating into an estimate of the light exposure that will cause just noticeable fading is provided in Table 4, derived from a review of the literature partially described in Michalski (1987). The estimates in Table 4 are the basis of the time to fade estimates in Table 3.

The Blue Wools as an estimate of the range of sensitivities in collections

Museums inevitably ask, what is the range of colourant sensitivities in my collection? The original eight Blue Wools, developed in the 1920s, represented the sensitivity range that the dye and colourant industry knew was reflected in all the coloured goods of the time, whether using natural dyes, synthetic dyes (started in the 19th century), or even pigments. Thus the range of these eight Blue Wools is an excellent estimate of the range of light-fading sensitivities one might expect to find in a mixed museum collection. Of course, some coloured objects are not sensitive at all. Some coloured objects are even more sensitive than #1 because they were not intended to last even as long as a poor-quality textile, e.g. some felt tip pens.

Note in Table 5 the strong role of reduced light intensity, not just the degree of UV filtration, in extending the lifetime of materials on display. When UV is measured as a ratio — microwatts (of UV energy) per lumen (of light) — then the total UV exposure, and the damage that depends on it, is proportional to both the intensity of the lighting and the UV measurement.

Rates of IR Deterioration

Infrared causes heating (considered in detail in the section Incorrect temperature). IR heating usually becomes a problem only with two sources of light: incandescent lamps at high intensity, over 5000 lux, and direct sunlight. In Table 5, nominally about UV damage, the effects of elevated object temperatures by direct sunlight are noted in the rows for average daylight through window glass. Sunlight, or intense incandescent lighting, can warm surfaces 40°C above ambient, or more. This elevates the rate of thermal decay by a factor of 20 or more.

When Light, UV, IR, and Other Agents of Deterioration Mix in the Same Object

Different deterioration phenomena often occur simultaneously: the yellowing or weakening caused by UV can be mixed with similar effects caused by thermal ageing. This thermal ageing is, in turn, accelerated by the high temperatures possible with IR (as noted in Table 5). On top of that, some of the new yellowing may be faded by light (blue light in particular). All of us have seen old framed prints with various patterns of yellowing. These provide an interesting amalgam of agent effects. To begin with, any coloured paints and inks may be faded by light. The paper may be yellowed by UV that the glass does not block, but under the matt, it will be protected. In extreme exposures, the fibres of the paper weaken (often not noticed until such prints are handled or washed during conservation treatment and the image area begins to disintegrate). If the matt is of poor quality, it will emit vapours that cause a narrow band of yellow/brown near the edge of the matt, which is greatly accelerated by IR heating. If a good UV filter is present, the paper in the image area will become whiter, not yellow, but the region under the matt will become uniformly yellow from thermal yellowing, which is accelerated by IR heating. The permutations may be complex, but the conclusions are simple: for organic materials, keeping the light intensity below many thousands of lux will reduce all forms of light, UV, and IR damage. In addition, using low UV light sources will push UV damage of high-sensitivity materials well below the similar forms of damage caused by room temperature itself.

Stages of Control

Control Strategies for Different Degrees of Preservation

Vignette 1. Use of a Window Well in a Historic House for Display

Here, two strategies have been taken: (1) The window has been screened by a translucent fluted plastic board, usually used for graphic signs. This reduces the light intensity by about one half (it also adds to the insulating quality of the window). This is a simple method using a few pushpins, easily removed if the room use changes. Although the glare from the translucent panel is less than ideal, the background fabric immediately adjacent to the objects is dark and matte. (2) Most importantly, zero sensitivity objects (metal stamps) have been selected for this display, or low-sensitivity artifacts (white paper, black ink, unstained wood). The brass stamps are very detailed, dark, low-contrast objects; therefore, seeing them well takes advantage of the strong window light.

Vignette 2: A Local Art Gallery with Basic Track Lighting

Unlike the historic house with its original window wells described in the earlier vignette, a purpose-built display area, such as this small gallery at the Peel Regional Art Gallery, Brampton, Ontario, has full control of the lighting, both ambient and on the artworks. It uses a basic track pattern, one strip along each of the long walls, about 1.5 m from the wall, so that the light beam hits the painting centre at approximately 30° (to the vertical). The end walls are lit from the last portion of track. Note that glare control by the full lamp housings is very good when facing away from the viewer, but less successful when trying to light the left end wall. The spot lamps place the emphasis on the paintings and reduce competition from the walls, but finding spot lamps that yield moderate intensities at close quarters is difficult. Without knowing the artists' palettes, the gallery must assume some high-sensitivity colourants are in use. Because many of the artists selected for this gallery are local and alive, the museum could ask for information on the palettes, obtain their sensitivities, or even advise current artists about low-sensitivity palettes.

Key Readings

1. Commission internationale de l'éclairage (CIE). Control of Damage to Museum Objects by Optical Radiation. CIE Technical Report, 157. Vienna: Commission internationale de l'éclairage, 2004.

2. Ashley-Smith, J., A. Derbyshire, and B. Pretzel. "The Continuing Development of a Practical Lighting Policy for Works of Art on Paper and Other Object Types at the Victoria and Albert Museum." ICOM Committee for Conservation. In, Triennial meeting (13th), Rio de Janeiro, 22-27 September 2002: Preprints. London: James & James, 2002, p. 3-8.

3. Michalski, S. "The Lighting Decision." In, Fabric of an Exhibition, Preprints of Textile Symposium 97. Ottawa: Canadian Conservation Institute, 1997, pp. 97-104.