Ultra-Violet Filters for Artificial Light Sources David Saunders
Download 167.66 Kb.
|
- Navigate this page:
- The ultra-violet content of artificial lamps
- Tungsten incandescent lamps
- Fluorescent lamps
- Tungsten halogen lamps
- Metal halide lamps
- Ultra-violet filters
- Filters for low operating temperatures
- High temperature filters
- Assessment of heat resistant filters
- Practical assessment
- Acknowledgements
- National Gallery Publications Limited Published by order of the Trustees
National Gallery Technical Bulletin Volume 13, 1989
Ultra-Violet Filters for Artificial Light SourcesDavid SaundersIntroductionThe detrimental effects of radiation in the visible and ultra-violet region of the spectrum on works of art are well established. Deterioration may vary from slight loss of colour or flexibility to complete disintegration (note 1). A dual approach has been adopted by most museums and galleries to minimize damage. Firstly, the overall illumination or annual exposure is controlled to a level appropriate for the most sensitive materials present in the object (note 2). Secondly, the proportion of ultra-violet radiation in the light reaching the object is minimized. Because the human eye is not sensitive to the ultra-violet region of the spectrum, this radiation may be eliminated without disturbing colour appearance. Historically, daylight has been a preferred source for lighting works of art. Much research has, therefore, been directed toward the development of daylight control systems and glazing materials which absorb ultra-violet radiation. It is now common practice either to install laminated glazing with an ultra-violet absorbing inter-layer or to apply a film or varnish with ultra-violet blocking properties to existing glazing. As a result the daylight entering many museums and galleries has a lower ultra-violet content than those artificial lamps which had previously been regarded as low ultra-violet emitting sources. The ultra-violet content of artificial lampsThe sections which follow describe those artificial light sources currently available for museum and gallery lighting. Colour temperature, colour-rendering and ultra-violet content data for the lamp types described are collated in Table 1. Tungsten incandescent lampsThe tungsten incandescent lamp has been used widely for lighting in museums and galleries for a number of reasons. The light emitted is of good colour quality and has an ultra-violet content in the region of 65–75µ W lumen-1, depending on the operating temperature of the filament (note 3). The current recommended maximum ultra-violet content of other light sources (less than 75µ W lumen-1) was based upon the emission of the tungsten incandescent lamp which was assumed to have an acceptable ultra-violet content (note 4). Fluorescent lampsThe efficacy of fluorescent lamps compared to that of tungsten lamps had led to their use in a great number of museums and galleries. The ultra-violet content of the light emitted varies according to lamp type, but is generally in the region 40–150µ W lumen-1 (note 5). The theoretical and practical aspects of selecting fluorescent lamps with good colour-rendering properties and an acceptable ultra-violet content have been described in earlier volumes of this ‘Bulletin’ (note 5,6). Tungsten halogen lampsThe tungsten halogen lamp operates upon a similar principle to the tungsten incandescent lamp. A filament of tungsten is heated by an electrical current to a temperature at which it emits light as well as heat. In the tungsten lamp the temperature to which the filament may be heated is limited. If the temperature is increased, evaporation of the filament causes deposition of tungsten metal on the inside of the glass envelope and, eventually, failure of the filament. The introduction of a small quantity of halogen vapour into the envelope establishes a regenerative cycle which prevents metal deposition and allows operation at a higher colour temperature, although filament life is still related to temperature (note 7). The ultra-violet output of the lamp depends upon the colour temperature at which the lamp is operating. The 'bluer' the light produced, the greater the ultra-violet content. At 2800K, which is a typical temperature for a tungsten lamp, the ultra-violet content of the light is 67µ W lumen-1. Raising the operating temperature to 3400 K, characteristic of short-life tungsten halogen lamps, increases the ultra-violet content to 165µ W lumen-1 (see Table 1). Metal halide lampsMetal halide lamps are mercury discharge lamps doped with selected metal halides. Excitation of these metals gives rise to additional bands in the visible region creating a 'white light' (note 8). Most types of metal halide lamp have rather poor colour-rendering properties. As a result such lamps are inappropriate for lighting exhibits when colour is important. Those sources with better colour-rendering indices tend to produce very 'cold' light which is not usually desirable for gallery lighting, although it may be suitable for conservation studios or perhaps as a method of supplementing the natural light passing through a laylight. In situations where colour appearance is not of prime importance, metal halide lamps are available which emit warmer light with moderately good colour-rendering properties. The ultra-violet content of the light emitted varies with lamp type, but can be up to ‘c’.700µ W lumen -1 (see Table 1). Other lampsOther types of artificial lamp are not widely used for display purposes in museums and galleries. The very poor colour-rendering properties of most sodium and mercury discharge lamps have prevented their use in situations where colour is important. The colour quality of the light from high-pressure sodium (SON) lamps is improved slightly in the 'deluxe' version by increasing the sodium vapour pressure (note 9). Although the colour quality of the light produced by 'deluxe' mercury fluorescent (MBF) and mercury fluorescent-tungsten 'hybrid' (MBTF) lamps is improving as new phosphor coatings are introduced (note 10), for the present such sources should be avoided for display. The properties of these lamps are summarized in Table 1. Table 1 Data for selected light sources.
Notes: (1) Correlated colour temperature (CCT) indicates the warmth or coolness of the light produced. Lamps may also be classified as follows. Warm: 3300 K > CCT Intermediate: 3300 K < CCT < 5300 K Cold: 5300K < CCT (2) Ra is the general colour-rendering index; Rw is the lowest (worst) of the eight individual indices. (3) The colour-rendering group is based upon the general colour-rendering index Ra. Group 1a: 100 > Ra ≥ 90 Group 1b: 90 > Ra ≥ 80 Group 2: 80 > Ra ≥ 60 Group 3: 60 > Ra ≥ 40 Group 4: 40 > Ra ≥ 20 (4) Ultra-violet radiation in the region 300–400nm. (5) More detailed data for a wider selection of fluorescent lamps may be found in references (note 5) and (note 6). Ultra-violet filtersThe spectral absorption properties of certain materials have been exploited in the manufacture of film and sheet filters for excluding ultra-violet radiation. Research in this area has aimed primarily at reducing the ultra-violet content of daylight passing into museums and galleries. In order to assess the effectiveness of ultra-violet absorbing materials, a specification for an acceptable filter has been established (note 11). Because some of the materials which comply with the limits imposed by this specification absorb light in the blue region of the visible spectrum as well as ultra-violet radiation, filters containing these substances appear yellow. It has been noted that absorbing the blue, and most damaging, portion of the visible spectrum may reduce certain light-induced damage (note 12). Yellow coloured filters will be of use to protect sensitive objects under circumstances where colour appearance is not important. In many situations, however, colour is of great importance and it is necessary to evaluate the effect of a given filter on transmitted light. Filters which have a slightly yellow appearance will tend to decrease the colour temperature of the transmitted light. The calculated colour temperature shift can be used as an indication of the suitability of a filter. In addition, the filter may absorb a greater proportion of the light in particular regions of the visible spectrum. The colour quality of the light transmitted may be reduced as a result. To assess this effect, the colour-rendering index of the light produced by a specific combination of source and filter may be calculated. Any significant reduction in the colour-rendering index of a light source when the filter is incorporated indicates that the filter may not be suitable. It is useful to make a distinction between those ultraviolet filters which are stable to heat and those which should only be used with lamps which generate little heat or are placed at a distance from a high temperature source. Generally, the latter are most suitable for use with fluorescent lamps or in laylight systems where lamps are located some way above the glazing. Filters for low operating temperaturesBecause fluorescent lamps operate on a discharge principle, little heat is generated. As a result, the ultra-violet content of the light emitted may be reduced simply and efficiently by using an absorbing sleeve or sheet. The sleeve is a thin polymeric film, commonly a polyester or acetate, in which an ultra-violet absorbing compound has been dispersed. The film is either pressed so that it forms a cylinder around a standard tube, or else is contained within a rigid jacket into which the tube is inserted. These films have a similar formulation to the adhesive films applied to exterior glazing to reduce the ultra-violet content of transmitted daylight. Although the best films reduce the ultra-violet content of light from most fluorescent lamps to between 0 and 5µ W lumen-1, there can be considerable variation between batches. The properties of some selected filters are summarized in Table 2. Transmittance spectra for these filters are shown in Fig. 1. A second filter type is a sheet of polymeric material, usually of 2–4 mm thickness, which serves to reduce the ultra-violet content of the light transmitted. The same sheet is frequently moulded to serve as a diffuser. The degree to which ultra-violet radiation is absorbed will depend on the thickness and type of polymer used. Certain types of both acrylic and polycarbonate sheet possess satisfactory ultra-violet absorption properties. These materials may also be used in laylight glazing, providing that no high temperature sources are located in the immediate vicinity. Table 2 summarizes the properties of some filters of this type; transmittance spectra for these filters are shown in Fig.2. High temperature filtersBecause the tungsten incandescent source was used to set the recommended maximum ultra-violet content, it has not been the practice to fit these lamps with a filter. The more damaging UV-A and UV-B radiation, with a wavelength of less than c.320nm is generally absorbed by the outer glass envelope. It was the increasing use of tungsten halogen lamps, particularly those operating at low voltage (typically 12 volt), that prompted us to reinvestigate the variety of materials available for ultra-violet filtration. Because of the high operating temperature, the filament of a tungsten halogen lamp is encased in a quartz rather than glass envelope. The quartz transmits some of the short wave ultra-violet radiation (UV-B) that would be absorbed by the glass envelope of a conventional tungsten source. For this reason it has been recommended that tungsten halogen lamps should be used in conjunction with glass filters (note 13). It will be seen from Table 3, however, that whilst 'plain' glass may absorb all radiation below c.320nm, the ultra-violet content of the light from a tungsten halogen lamp is not greatly reduced. Polymeric films and sheet are usually inappropriate for use with tungsten halogen sources because of the high operating temperature. It would be possible to locate the filter some distance from the envelope, but this would obviate one of the most attractive feartures of this type of lamp – its extreme compactness. Another alternative, which is not always desirable, is to glaze paintings or encase objects with an ultra-violet absorbing polymeric material (note 14). These difficulties prompted a new investigation of a number of heat-stable filters. We have evaluated their efficiency in absorbing ultraviolet radiation and their effect upon the colour temperature and colour-rendering properties of the light transmitted (note 15). The results of this study will also be relevant when considering metal halide and other high temperature lamps, detailed in Tablei, which have a significant emission in the ultra-violet region. Assessment of heat resistant filtersTheoretical calculationsA number of heat-stable materials were examined in a search for potential ultra-violet absorbers. Both borosilicate glass and 'plain' window glass were tested, as well as a number of glasses to which an ultra-violet absorber had been added whilst in the molten state. Two filters supplied by lamp manufacturers and designed to be used in conjunction with tungsten halogen lamps were also tested. Finally, two thin-film interference filters were assessed. The effect of the filters on the colour temperature, colour-rendering and ultra-violet content of the transmitted light were calculated (note 15), and are given in Table 3, whilst their transmittance curves are shown in Fig.3. The filters can be divided into three groups. The first includes the borosilicate glass, plain glass, Thorn ultraviolet filter and Pilkington MR2 glass. These filters have a minimal effect upon the colour temperature and colour-rendering properties of the transmitted light. Unfortunately, the ultra-violet content of the light transmitted by these filters is still above the current recommended level of 75µ W lumen -1. The second group, comprising the Erco ultra-violet filter, and the Oralan and Uvilex 1 glasses manufactured by Schott, reduce the ultra-violet content of the transmitted light to below 75µ W lumen-1. Because these filters also have a considerable absorption in the blue region of the visible spectrum, they decrease the colour temperature of the transmitted light and appear yellow. This is most clearly seen in Plate 10 (p.39), which shows a selection of the filters from Table 3 photographed against a white background. Finally, the two dichroic filters, manufactured by Bausch & Lomb and Balzers, transmit virtually no ultraviolet light and do not significantly change the colour properties of the light. These filters comprise a glass substrate on to which an inorganic coating has been deposited under vacuum. This thin but resilient layer is generally a mixture of metal oxides. Interference effects at the coating interface cause light of particular wavelengths to be reflected whilst light at other wavelengths is transmitted. In the case of the filters tested here, radiation in the ultra-violet portion of the spectrum is reflected whilst visible radiation is transmitted. Practical assessmentIn order to assess the validity of the theoretical calculations, four of the filter materials were compared in a practical study. Blue wool standards Nos.1, 2 and 3 were used as test samples. Three pieces of each blue wool standard were exposed to the light from a 12 volt 100 watt tungsten halogen lamp for 1000 hours. Three samples of each blue wool standard were placed beneath each of the four filter materials to be tested. The four filters selected were plain window glass, Thorn ultra-violet filter, Erco ultra-violet filter and Bausch & Lomb thin-film filter. Finally, three pieces of each blue wool standard were maintained in darkness but otherwise under identical conditions to the exposed samples. This procedure was adopted in order that both test and control samples might be subject to the same degree of heating from the source. The positions of the samples were interchanged at regular intervals to ensure even illumination of all test pieces. Table 2 Data for selected low temperature filters (1).
Notes: (1) Reference source: Thorn Kolor-rite; colour temperature 3930 K; general colour-rendering index (Ra) 89; ultra-violet content 85µ W/lumen. The first line of data gives the properties of the source alone, subsequent lines indicate the properties of the reference source combined with selected filters. (2) The colour temperature shift indicates the effect of the filter upon the warmth or coolness of the light transmitted. A negative colour temperature shift indicates that the colour of the transmitted light is warmer, or yellower, than the incident light. A positive shift indicates that the transmitted light is cooler, or bluer. (3) Ra is the general colour-rendering index of the light produced by the combination of filter and reference source. The greater the colour-rendering index the better the colour quality. (4) Ultra-violet radiation in the region 300–400nm. (5) All sheets tested were 3 mm thick. The colour of each sample was measured prior to exposure and after 1000 hours exposure at approximately 25,000 lux (note 16). The average CIE (Commission Internationale de l'Éclairage) coordinates L*, a* and b* were derived for every measurement and used to calculate the colour difference, ΔE, for each blue wool sample (note 17). Lastly, the colour differences were corrected to compensate for the differing transmittances in the visible region. The values of ΔE obtained in this way appear in Table 4. The colour differences noted for the dark control samples are at the level of experimental reproducibility and probably do not reflect any temperature-induced colour change. As expected, all three blue wool standards are affected by exposure to light, standard No.1 being the most fugitive. Greatest fading occurs when no filter is present. The reduction in fading due to each of the filters follows the same pattern for all three blue wool standards. The greatest reduction in fading is provided by the Bausch & Lomb filter followed by the Erco filter, the plain window glass and finally, the Thorn filter. The degree of protection afforded by each of the filters may be seen to relate to the ultra-violet content of the light transmitted, as indicated in Table 3. Table 3 Data for selected heat resistant filters (1).
Notes: (1) Reference source: 12V, 100W tungsten halogen lamp; colour temperature 3360K; general colour-rendering index (Ra) 99; ultra-violet content 165µ W/lumen. The first line of data gives the properties of the source alone, subsequent lines indicate the properties of the reference source combined with selected filters. (2) The colour temperature shift indicates the effect of the filter upon the warmth or coolness of the light transmitted. A negative colour temperature shift indicates that the colour of the transmitted light is warmer, or yellower, than the incident light. A positive shift indicates that the transmitted light is cooler, or bluer. (3) Ra is the general colour-rendering index of the light produced by the combination of filter and reference source. The greater the colour-rendering index the better the colour quality. (4) Ultra-violet radiation in the region 300–400 nm. (5) All the samples were between 3 and 4 mm thick, with the exception of the Balzers filter which was 2 mm thick. (6) Filters supplied by lighting manufacturers for use with tungsten halogen lamps or fittings.
Notes: (1) Reference source: 12V, 100W tungsten halogen lamp; illuminance 25,000 lux. Samples were exposed for 1000 hours. Sample positions were interchanged periodically to ensure equal illumination. (2) The colour was measured before and after exposure using a Minolta CR–200 chroma meter. Entries in the Table derive from the average of three measurements on each of three samples exposed under identical conditions. The colour difference, ΔE, has been calculated using the CMC (l:c) colour difference formula; see reference (note 16). ΔE has been corrected to compensate for the differences in visible transmittance of the various filters. (3) Filters supplied by lighting manufacturers for use with tungsten halogen lamps or fittings. It is equally clear, however, that none of the filters eliminate fading, which is caused by both visible and ultra-violet light. This emphasizes the importance of combining ultra-violet filtration with a regime that limits the total light exposure. ConclusionsMany of the artificial light sources suitable for museums and galleries emit a significant proportion of ultra-violet radiation. Although light-induced damage cannot be prevented by eliminating ultra-violet radiation alone, a number of studies have shown that radiation in the ultraviolet region of the spectrum is considerably more damaging than visible light (note 18). Since this ultra-violet light is not necessary for accurate perception of form or colour, it should be eliminated by effective filters whenever possible. The problem of selecting efficient heat resistant filters for tungsten halogen lamps has led to an acceptance of higher than recommended ultra-violet levels (note 19). The results above indicate that with appropriate filters capable of withstanding high temperatures, it is possible to reduce the ultra-violet content of the light from a tungsten halogen lamp to a negligible level. Indeed, the Bausch & Lomb filters tested are about to become generally available, whilst other manufacturers are beginning to introduce low voltage tungsten halogen lamps fitted with a vacuum coated front glass to reduce ultra-violet content. If film or sheet ultra-violet absorbers are used to reduce the ultra-violet content of daylight and fluorescent light, and heat resistant filters are used with tungsten halogen and other high temperature discharge lamps, then a paradoxical situation is reached in which the source with the highest ultra-violet emission will be the incandescent lamp; previously considered to have a sufficiently low ultra-violet content to act as a benchmark for the filtration of other sources. There is no reason why incandescent lamps should not be fitted with the same filters as tungsten halogen lamps. It is unlikely that this will be done whilst the accepted maximum ultra-violet light content remains fixed at 75µ W lumen -1. Since the best of the filters described in the preceding sections are capable of reducing the ultraviolet level of all sources to well below 10µ W lumen -1, it is proposed that museums and galleries should in future aim for an ultra-violet content of less than 10µ W lumen-1, particularly for the more light sensitive materials, which would be exhibited at light levels of less than 50 lux under the current recommendations. AcknowledgementsI should like to thank all the manufacturers who have supplied technical data for inclusion in this article, particularly H. Rudolph Inhelder of Wotan Lamps Ltd and Worth D. Holder of Bausch & Lomb Vacuum Coatings Division. Notes and references1. For a review of the types of damage caused by both visible and ultra-violet radiation, see Michalski, S., 'Damage to museum objects by visible radiation (light) and ultraviolet (UV)', in ‘Lighting: Preprints of the Bristol Conference on Lighting in Museums, Galleries and Historic Houses’, The Museums Association (1987), pp.3–16. 2. Thomson, G., ‘The Museum Environment’, 2nd ed., Butterworths (London 1986), pp.22–34. 3. The ultra-violet content of a light source is usually given in microwatts per lumen (µW lumen-1). This indicates the proportion of radiation in the ultra-violet region. To control the amount of ultra-violet radiation incident on an object it is necessary, therefore, both to minimize the ultra-violet content of the light source and to limit the overall illuminance. 4. Thomson, G., ‘The Museum Environment, op. cit.’, p.20. 5. Thomson, G., 'Colour Under Some New Fluorescent Lamps', ‘National Gallery Technical Bulletin, 9’ (1985), pp.5–11. 6. Saunders, D.R., 'Fluorescent Lamps: A Practical Assessment', ‘National Gallery Technical Bulletin’, 11 (1987), pp.86–91. 7. Wolfe, K.R. and Letchford, J. A., 'Tungsten Halogen Lamps' in ‘Lamps and Lighting’, M. A. Cayless and A.M. Marsden (eds.), Edward Arnold (London 1983), pp. 169–82. 8. Hall, R. and Odell, E.C., 'Metal Halide Lamps' in ‘Lamps and Lighting, op. cit.’, pp.236–48. 9. Molesdale, P.D., 'High Pressure Sodium Lamps' in ‘Lamps and Lighting, op. cit.’, p.210–23. 10. Martin, W., 'Mercury Lamps' in ‘Lamps and Lighting, op. cit.’, pp.224–35. 11. The specification for an acceptable ultra-violet filter takes as a reference value the transmittance of the filter at 550nm. An acceptable filter has a transmittance of less than 1% of the reference at both 320 nm and 380 nm and a transmittance of less than 50% of the reference at 400nm. See Thomson, G., ‘The Museum Environment, op. cit.’, p.17. 12. Cox Crews, P., 'A Comparison of Clear Versus Yellow Ultraviolet Filters in Reducing Fading of Selected Dyes', ‘Studies in Conservation’, 33 (1988), pp.87–93. 13. Thomson, G., ‘The Museum Environment, op.cit.’, p.172. 14. The damage factor (in the visible and ultra-violet regions) has been calculated for a number of artificial sources and the effect of protecting the object with a series of filter materials tabulated. See Hilbert, G.S., ‘Sammlungsgut in Sicherheit; Teil 2. Lichtschutz, Klimatisierung', Gebr. Mann Verlag (Berlin 1987), p.76. 15. The computer program developed for this purpose calculates the spectral power distribution of the transmitted light by multiplying the spectral power distribution of a source with the transmittance spectrum of any of the filters for which data is available. The colour-rendering properties, colour temperature and ultraviolet content of the transmitted light are then calculated. 16. Colour measurements were made using a Minolta CR – 200 chroma meter. Values of L*, a* and b* were calculated for CIE illuminant D65. 17. Colour difference ΔE was calculated using the Society of Dyers and Colourists Colour Measurement Committee recommended formula; CMC(l:c) formula. See Clarke, F.J.J., McDonald, R. and Rigg, B., 'Modification to the JCP79 Colour-Difference Formula', ‘Journal of the Society of Dyers and Colourists’, 100 (1984), pp.128–32. 18. For a review of the evidence, see Michalskí, S., ‘op.cit.’, pp.13–14, or Cuttle, C, 'Lighting Works of Art for Exhibition and Conservation', ‘Lighting Research and Technology’, 20 (1988), pp.43–53. 19. Staniforth, S., 'Problems with Ultraviolet Filters', in ‘Lighting: Preprints of the Bristol Conference on Lighting in Museums, Galleries and Historic Houses, op cit.’, pp.25–9. National Gallery Publications Limited Published by order of the TrusteesScientific Adviser, John Mills Chief Restorer, Martin Wyld Editor: Ashok Roy © The National Gallery 1989 All rights reserved. No part of this publication may be transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without the prior permission in writing from the publisher. British Library Cataloguing in Publication Data National Gallery National Gallery technical bulletin. — Volume 13, 1989. 1. Painting — Conservation and restoration — Periodicals I. Title
751.6'05 ND1640 ISSN 0140–7430 ISBN 0 947645 70 5 Designed by James Shurmer Phototypeset by Keyspools Limited Golborne, Lancs Printed and bound in Great Britain by William Clowes Limited, Beccles and London Directory: upload -> rtf rtf -> National Gallery Technical Bulletin Volume 18, 1997 rtf -> National Gallery Technical Bulletin Volume 29, 2008 rtf -> Titian’s 'Bacchus and Ariadne' Arthur Lucas and Joyce Plesters rtf -> Hogarth's 'Marriage à la Mode' David Bomford and Ashok Roy rtf -> National Gallery Technical Bulletin Volume 25, 2004 rtf -> Rubens’s ‘Peace and War’ Ashok Roy rtf -> Monitoring Shock and Vibration during the Transportation of Paintings David Saunders rtf -> Packing: a case Study Sarah Staniforth rtf -> Six Paintings by Corot: Methods, Materials and Sources Sarah Herring rtf -> Colour Under Some New Fluorescent Lamps Garry Thomson Download 167.66 Kb. Share with your friends:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||