Preventive Conservation Research for Plastics on Open Display

Decisions regarding the long-term display of artefacts within historic collections rely on an intimate knowledge of materials behaviour. Deterioration rates may be greatly reduced by careful control of the environment. However, this is not always possible where objects are presented as part of an historic interior, outside of the museum environment. The physical response of Delrin panels has been determined and its implication for allowable relative humidity fluctuations assessed. Degradation and stability vary through the polymer classes, with some synthetic and semi-synthetic types posing particular problems. This illustrates the need for positive identification and characterisation of these polymers. However, the nature of such materials can restrict sampling, particularly where repeated, periodic condition monitoring is desired. It is for this reason that the use of non-invasive, vibrational spectroscopy techniques is finding increasing application within the heritage sector. The ability to quickly identify modern materials enables museum professionals to make informed long-term action plans and more immediate remedial conservation plans. Monitoring of rubber objects has been undertaken and the results have informed the preventive conservation of this material as part of historic interiors. KeyWorDS monitoring, environment, dimensional change, analysis, NIR, FTIR, Raman Thickett and Richardson pp.89_9689 89 28/10/08 14:37:09 90 D AV I D T H I C K E T T & E M M A R I C H A R D S O N collected using diamond attenuated total reflectance (ATR) (Durascope on a Perkin Elmer 2000). These were compared to spectra from a sample of Delrin obtained commercially and matched to the Hummel plastics library. They were found to be an excellent visual match to the Delrin samples and this material was selected as the highest match using three algorithms to compare to the Hummel library. Raman spectra were also collected using a Spectrolab R2001 spectrometer with 785 nm laser at 60 mW, again giving an excellent visual match to the Delrin samples. Delrin was developed between 1952 and 1956 by DuPont and entered volume production in 1956. While not contemporary with presentation of the tunnels, the exchanges were sourced from Osborne House and Hinchley Wood and are of the same physical type originally used in the telephone exchange. The Osborne House exchanges date from 1962 to 1965, explaining the use of a post-1956 plastic in their construction. Examination of the exchanges in 2003 indicated that several of the side panels had undergone plastic deformation and some of the front panels had cracked. Little has been reported in the conservation literature about the physical deformation of plastics and this was investigated further. The reaction time of the plastic sheeting was of particular interest as the ventilation system causes magnified daily fluctuations in RH in the telephone exchange. If the reaction time is significantly greater than 24 hours then these fluctuations would have little effect. One 2 cm piece of Delrin had become detached and its physical response to RH was measured. A frame was made to clamp the piece to a GTX1000 linear variable differential transformer (LVDT). The LVDT is capable of converting very small displacements into a voltage signal. The LVDT output was recorded with a SR008 voltage logger, recording any change in the piece’s length. The RH in the chamber was controlled with glycerol solutions (Miner and Dalton 1953) between 30 and 98%. The RH was continuously monitored using a Rotronic hygroclip probe connected to a Meaco radiotelemetry system, calibrated to NAMAS accredited saturated salts covering this RH range. The reaction time, fitted as an exponential, was found to be approximately 72 hours. In order to confirm the assumption that the large daily variations were of little significance, the measuring rig was placed in the telephone exchange for a 12-month period. The length was found to follow the long-term variation in RH and not to follow the daily variations, confirming the previous reasoning (see Fig. 2). Autocorrelation analysis yielded a maximum correlation between the length and RH changes of 70 hours (Chatfield 1984). Figure 1 Temperature and relative humidity in the telephone exchange at the Dover Secret Wartime Tunnels. Figure 2 Length response to RH of Delrin in the telephone exchange. Thickett and Richardson pp.89_9690 90 28/10/08 14:37:19 91 P R E V E N T I V E C O N S E R VAT I O N R E S E A R C H F O R P L A S T I C S O N O P E N D I S P L AY royal observer corps bunker, york The bunker was decommissioned in 1991 and taken into care by English Heritage. It was opened to the public in 2005. The contents had been largely dispersed, and the bunker is presented with contemporary material from store or purchased on the open market. The composition of the great majority of the items was unknown and the bunker now contains well over 300 plastic objects, some of which have over 30 different components. In order to manage the conservation, both preventive and interventive, identification of the plastics was urgently required. Three techniques were used, all based on vibrational spectroscopy. In an initial study, a series of 50 small samples was taken from objects of particular significance or representative of groups of example telephones and chairs. This provided an overview of the types of materials to be found within the collection. The samples were analysed using FTIR on a Perkin Elmer 2000 using a Durascope with diamond ATR. This technique is well established for application within the heritage field and is not described in detail. Its primary drawbacks relate to its timeconsuming nature and the need for sampling (Keneghan 1995; Garside and Wyeth 2003; Buzio et al. 2004; Shashoua and Johansen 2005). Although non-destructive, the technique is usually invasive and therefore object sampling may be limited to certain discreet regions, thus reducing the representative nature of the analysis. The more recent ATR method of collecting spectra has several benefits for this type of work: it requires minimal sample preparation; the sample size can be relatively small; controlling the force applied to the sample gives good reproducibility of spectra, improving both the search efficiency and quantification; it can cope well with higher absorbing and scattering samples, such as carbon black fillers. Two portable near infrared (NIR) reflectance spectrometers were subsequently employed providing non-invasive characterisation across the collection. NIR spectra are produced by the absorption of radiation at higher wavenumbers than those found in more conventional mid-IR spectroscopy (12800–4000 cm–1 compared to 4000–400 cm–1 respectively). The higher energy vibrations excited are high frequency overtone and combinations of fundamental modes seen in the mid-IR, and are generally dominated by vibrations of molecular bonds with hydrogen, which are notably anharmonic. This makes NIR spectroscopy particularly applicable to the study of organic polymeric material. The complexity of the spectra can preclude simple band assignments. However, the spectra are still characteristic, and comparison with a suitable spectral reference set usually allows for ready identification of unknowns. A Perkin Elmer Fourier transform Spectrum One NTS with an Axiom fibre optic probe attachment was applied in conjunction with a Manfrotto articulated arm for accurate positioning of the analysis area. Spectra were recorded in absorbance with a wavenumber resolution of 8 cm–1, a scan speed of 1 cm–1.s–1 over a spectral scan range of 12000–4000 cm–1 and scan accumulation • • • • of 32. This set-up can achieve sampling depths in the region of a few millimetres, making this a non-contact method of analysis with the possibility of depth sensitivity. One aim was to exploit this characteristic for the non-invasive identification of foam padding under a number of seat covers held within the collection. Polyurethane-based foams have been used extensively in furniture manufacture. Differences in stability have been found with certain types, undergoing oxidation or hydrolysis reactions over time (Garside and Lovett 2006; Kessler and Van Oosten 2005). Unfortunately, on this occasion, depth sampling was not possible due to the tight, heavy weave of the upper fabric. It was, however, possible to identify the upper textile material as a polyester/elastomer blend (Fig. 3), which is of importance for collections care due to the fact that rubbers, both synthetic and natural, are known to suffer significant degradation with age. Further problematic materials identified within the bunker include cellulose acetate laminates to documents and polyvinyl chloride table tops, fire exit signs and cable covers. Materials identification was made possible by spectral matching using Grams7 Spectral ID software to a custom library of over 300 reference samples of synthetic polymers. In the majority of cases matching was performed using the first derivative least squares algorithm, where unknown spectra are matched to the closest match held within the database. In some cases the complicated nature of the spectra and spectral interferences can produce poor correlation and false positive results. It is not possible to apply the analysis as a black box method therefore all assignments were accompanied by visual comparisons and the application of first or second order derivatives. The second NIR spectrometer utilised for comparative purposes was an ASD LabSpec 2500 dispersive NIR instrument with confocal fibre optic probes. Spectra were recorded in reflectance with a wavelength resolution of 3.3 cm–1 over a spectral range of 28570–4000 cm–1 and averaged over 20 scans. Data acquisition was in the region of 2 seconds. The smaller dispersive system allowed ready manoeuvring within the tight confines of the bunker. Coupled with the confocal optics, analysis of inac