Comparative Study of Fourier Transform Infrared Spectroscopy in Transmission,Attenuated Total Reflection,and Total Reflection Modes for the Analysis of Plastics in the Cultural Heritage Field

Abstract

This study was completed within the framework of two research projects dealing with the conservation of contemporary artworks. The first is the Seventh Framework Project (FP7) of the European Union, Preservation of Plastic ARTefacts in Museum Collections (POPART), spanning years 2008–2012, and the second is the Italian project funded by the Tuscan Region, Preventive Conservation of Contemporary Art (Conservazione Preventiva dell'Arte Contemporanea (COPAC)), spanning 2011–2013. Both of these programs pointed out the great importance of having noninvasive and portable analytical techniques that can be used to investigate and characterize modern and contemporary artworks, especially those consisting of synthetic polymers. Indeed, despite the extensive presence of plastics in museum collections, there is still a lack of analytical tools for identifying, characterizing, and setting up adequate conservation strategies for these materials. In this work, the potentials of in situ and noninvasive Fourier transform infrared (FT-IR) spectroscopy, implemented by means of portable devices that operate in reflection mode, are investigated with a view to applying the results in large-scale surveys of plastic objects in museums. To this end, an essential prerequisite are the reliability of spectral data acquired in situ and the availability of spectral databases acquired from reference materials. A collection of polymeric samples, which are available commercially as ResinKit, was analyzed to create a reference spectral archive. All the spectra were recorded using three FT-IR configurations: transmission (trans), attenuated total reflection (ATR), and total reflection (TR). A comparative evaluation of the data acquired using the three instrumental configurations is presented, together with an evaluation of the similarity percentages and a discussion of the critical cases.

Keywords

Polymers Noninvasive Fourier transform infrared spectroscopy FT-IR Noncontact FT-IR Total reflection Contemporary art

INTRODUCTION

The word plastic literally means “moldable.” This term is commonly used to indicate synthetic and semi-synthetic organic polymers that have a high molecular weight and are made up of specific units (monomers) linked to form long chains. The chemical–physical properties of plastics may be designed by varying the production processes and the formulations. Indeed, by starting from the same basic resin, the final formulation may be varied by adding several types of additives (colorants, fillers, plasticizers, etc.), which then make it possible to control precisely the characteristics of the finished material. This exceptional versatility of plastics is indeed the reason for their pervasive and capillary diffusion into every field and aspect of our lives. Nowadays, as culture and art reflect and dip into the historical context and into daily life, plastics have also become part of our cultural heritage and are widespread in art collections, museums, and art galleries.

The first documented use of plastics as artistic materials dates back to the pioneer works of Naum Gabo in the early 1900s.¹ Since then, recourse to synthetic polymers in the art field has constantly increased, reaching a high point in the 1960s. At present, resorting to synthetic polymers is ubiquitous among artists, and plastics are definitely considered to be also a class of artistic materials. In addition to artistic production, the appearance of plastics in museums and galleries is also related to design, ethnographic collections, and technical collections, as well as film archives. Consequently, a notable percentage of artworks and valuable objects consisting of synthetic and semi-synthetic polymers are kept in museums, and these are classified as items to be preserved for future generations. Despite this fact, the problem of the conservation of plastics in art collections had been disregarded until recently because, for a long time and due to widespread misconception, plastics have been considered imperishable materials. The need to adopt systematic strategies for the conservation of plastic objects, in a way analogous to other, more traditional materials in collections, has been completely recognized only in recent decades. Such an increased awareness has given rise to a new research area aimed at filling in the lack of knowledge, practices, and methodologies for preserving plastics in collections.^2–4 Several projects devoted to the safeguarding of contemporary art have recently been launched, such as the Preservation of Plastic ARTefacts in Museum Collections (POPART), Conservazione Preventiva dell'Arte Contemporanea (COPAC), and International Network for the Conservation of Contemporary Art (INCCA),^5–7 and studies on the care of plastics in collections have been published.^8–12 Within these research projects, particular attention has been paid to the development of new tools and techniques specifically tailored to the in situ and noninvasive identification of polymers that constitute museum objects. In fact, a knowledge of the materials constituting an artwork is the essential first step in understanding the causes of its degradation and, if necessary, in setting up a correct approach to its preservation.

Because the integrity of the object is a priority in the art conservation field, noninvasive or micro-invasive analytical techniques are preferred to sampling methods. In addition, the use of portable instrumentation is desirable because this makes it possible to investigate the artworks in situ. However, most of the best-established analytical techniques for the identification of plastics are borrowed from other application sectors (industrial, chemical, manufacturer, etc.), in which the requirements of portability and noninvasiveness are not usually met. State-of-the-art analytical methods used for the identification of plastics in museum collections were drawn up by the POPART Project, in which a comparative evaluation of well-established methods and novel techniques was carried out.¹³ Several analytical methods were examined in this study, and those based on vibrational spectroscopies—in particular, on Fourier transform infrared (FT-IR) techniques—have been found to be among the most effective and suitable for identifying synthetic resin blends in artworks.

In fact, FT-IR analysis in the mid-infrared (mid-IR) spectral region gives researchers considerable ability to distinguish and identify polymers, and meets requirements such as micro-destructiveness (and, in certain circumstances, nondestructiveness), cost effectiveness, and the availability of a wide range of commercial instrumentation. Actually, FT-IR spectroscopy is well established in several applicative sectors (industry, quality controls, etc.)¹⁴ as well as in the cultural heritage field.^15,16 However, literature reporting applications of FT-IR to the identification and characterization of plastics in art collections is relatively scarce,^17–20 and certain fundamental issues still need to be investigated in greater depth.

First, there is a lack of suitable spectral databases of reference materials set up specifically to interpret data acquired from plastic artworks. Indeed, although several FT-IR spectral libraries of synthetic polymers are available, only a few of these include a significant number of polymers of interest to researchers of art collections. In recent years, this issue has been partially tackled, and spectral archives of plastics of specific interest to the art conservation field are being assembled, such as the SamCo archive, developed under POPART, and the Infrared and Raman Users Group (IRUG) database.^21–23 Nevertheless, extensive further work is still needed to build rich and statistically meaningful databases for practical applications.

A second issue concerns the feasibility of noninvasive FT-IR measurements on museum objects and the quality of FT-IR spectra acquired in situ. Indeed, in recent decades, a new generation of high-resolution portable FT-IR devices has been marketed at affordable prices.²⁴ In addition to their portability, these devices offer the possibility of operating in total reflection (TR) mode by recording spectra in a contactless mode, without any need for sampling. It should be pointed out that, until recently, the only way to perform noninvasive FT-IR analysis in the mid-IR region was by means of optical fibers coupled with FT-IR benches.^25–28 However, the use of optical fibers in the mid-IR region has several drawbacks, such as the elevated cost of the equipment and the need for liquid nitrogen to cool down the mercury–cadmium–telluride (MCT) detectors. In addition, chalcogenide glass fibers, which are commonly used to produce mid-IR fiber bundles, strongly attenuate the IR radiation and have their own absorption bands. These absorption bands may mask spectral features of interest for the identification of materials. Thus, the recent availability of portable FT-IR instruments equipped with reflection mode accessories has opened up new applicative perspectives, especially in the field of art conservation, where in situ and noninvasive analyses are in great demand.

Despite these advances, TR FT-IR spectroscopy is still far from being fully exploited as a routine technique for large-scale surveys and in situ investigations of art collections. This is mainly due to the lack of databases of TR FT-IR spectra acquired from reference materials, which are essential for a correct interpretation of spectra acquired on real objects. Commercial FT-IR spectral libraries are usually built using other FT-IR configurations, such as the transmission (Trans) and attenuated total reflection (ATR) modes. In the TR FT-IR configuration, both diffuse and specular components of the reflected radiation are acquired and contribute to the registered spectrum. For this reason, TR FT-IR spectra are intrinsically more complex and are not always directly comparable to the data included in commercial spectral archives, unless suitable data pretreatments are made.

The research presented here aims to investigate the effectiveness of TR FT-IR spectroscopy implemented by means of portable spectrophotometers as a noninvasive analytical technique for routine and large-scale investigations on plastic objects that can be found in contemporary art museums. We carried out a pilot study using a commercial set of 50 reference samples of thermoplastics resins, with the intention of ascertaining the comparability of FT-IR data acquired in the TR mode to those acquired in both the Trans and ATR modes. A comparison of the FT-IR spectra acquired from these plastic samples using the same FT-IR bench in three different configurations (Trans, ATR, and TR) is reported and discussed, with particular focus on critical cases. A collection of TR FT-IR spectra based on this comparative evaluation has been chosen to provide a first nucleus for a TR FT-IR spectral archive of some of the most common synthetic resins, to be used as a guide in interpreting data acquired in the field.

EXPERIMENTAL

Fourier Transform Infrared Experimental Configurations for the Analysis of Plastics. The choice of an experimental FT-IR configuration (Trans, ATR, or TR) depends on the features of the objects and materials being examined. In principle all three of these methodologies are usable with regard to plastics.

The most well-established method for the analysis of solid plastic samples is the conventional Trans mode, which can be used whenever sampling is possible. The Trans spectra are usually acquired from pellets prepared by mixing a small amount (about 0.5–1 mg) of the ground sample with potassium bromide (KBr; about 100 mg). Because plastics are highly variable in softness, flexibility, and consistency, in some cases they can barely be transformed into a thin powder, thus making it impossible to obtain a homogeneous pellet and, consequently, a well-resolved Trans spectrum. When it is difficult or impossible to obtain homogeneous sample–KBr dispersions, less straightforward preparation processes have to be adopted to perform measurements in Trans mode. If the samples are soluble in a volatile solvent, they can be dissolved and analyzed by depositing a few drops of the solution onto an IR transparent material. If the polymer is not easily soluble, Trans spectra may be recorded by resorting to specific instrumental accessories, such as diamond cells. Thus, provided that suitable technical expedients are adopted for the preparation of the samples, high-quality Trans FT-IR spectra may, in principle, be acquired from any type of specimen.

Another well-established method for performing FT-IR analysis is ATR mode. The use of this configuration is widespread for the characterization of polymers and overcomes any drawbacks related to the preparation of KBr pellet samples. However, ATR spectra may be less easy to interpret than spectra recorded in a conventional Trans mode.

In the ATR mode, the penetration depth of the evanescent wave within the sample depends on the optical thickness of the sample; longer wavelengths may undergo relatively stronger absorptions. This fact may cause spectral band distortions due to broadening or peak shifts at the longer wavelengths.²⁹ These effects have to be taken into account when ATR spectra are compared with those acquired in Trans mode. Suitable algorithms are available to correct for such systematic spectral differences. It should be remembered that in ATR mode, high-quality spectra may be obtained only by applying high pressure to ensure an intimate contact between the sample and the ATR crystal. Therefore, even if the ATR mode is in principle noninvasive, this method cannot be used to analyze fragile or precious objects without risking damage to them. On very soft materials, such as some common plastics (e.g., expanded polystyrene), the ATR mode cannot be used profitably. The use of an ATR configuration to distinguish between similar classes of plastics was reported by Enlow et al.,³⁰ and an application to semi-synthetic resins of interest in museum applications was reported by Paris et al.¹⁹

Whenever sampling has to be avoided, as is often the case in museum contexts, FT-IR spectra can be acquired in TR mode. As mentioned, one of the main drawbacks of TR FT-IR measurements is the complexity of the spectra acquired. Indeed, they are often affected by distortions, both in the band shape and in the absorption frequency, which depend mainly on the variable balance between the diffuse and specular reflection components. These spectral anomalies are due to several factors, such as the concentration of the sample, the surface texture, and the refractive index of the materials.³¹ For example, when a strong specular component is present, as is the case when materials have high refractivity, anomalous spectral bands (usually known as derivative-like bands) are observed. These anomalies in the FT-IR reflection spectra are usually treated by applying specific data-processing algorithms, such as the Kramers–Kroenig (KK) transform.³² The KK algorithm makes it possible to reconstruct a pseudo-absorption spectrum, thus making the TR spectra comparable with those acquired in the Trans or ATR modes. However, the use of correction algorithms is not always sufficient to make the TR spectra superimposable on (or comparable with) those registered in Trans and ATR modes from the same specimen. For example, when undiluted materials with a strong absorption coefficient are analyzed in TR mode, intense distortions may appear in the FT-IR spectra. The most evident case is the Reststrahlen band, where a reflection maximum occurs in place of the absorption band. In this case, spectral distortions cannot be corrected by recourse to mathematical algorithms.²⁷

Materials. A collection of 50 thermoplastic resin specimens, representative of some of the most widespread plastics, was characterized using FT-IR spectroscopy in the three instrumental configurations. This sample collection is marketed under the name ResinKit (ResinKit Company, Woonsocket, RI).³³ Each ResinKit specimen, molded as a 10 × 4 cm chip with a variable thickness ranging from 1 to 3 mm, is cataloged by means of a numerical index on its surface and is provided with technical information on its properties. Each chip has two lateral combs whose teeth are removable for different kinds of destructive analyses. The materials included in the ResinKit set analyzed are listed in Table I.

TABLE I.

List of ResinKit specimens and corresponding case typologies.

Number	Name^a	Case
4	Styrene acrylonitrile (SAN)	1
5	ABS, transparent	1
6	ABS, medium impact	1
7	ABS, high impact	1
9	Acrylic	1
10	Modified acrylic	1
15	Nylon (polyamide), type 66	1
16	Nylon (polyamide), type 6	1
18	Thermoplastic polyester, PETG	1
19	Polyphenylene oxide (PPO)	1
20	Polycarbonate (PC)	1
21	Polysulfone (PSU)	1
25	Polyethylene, high density (HDPE)	1
29	Polyvinyl chloride, flexible	1
30	Polyvinyl chloride, rigid	1
42	Styrenic terpolymer	1
50	ABS, nylon alloy	1
2	Polystyrene, medium impact	2
3	Polystyrene, high impact (HIPS)	2
8	Styrene, butadiene block copolymer (SBR)	2
11	Cellulose acetate (CA)	2
12	Cellulose acetate butyrate (CAB)	2
13	Cellulose acetate propionate (CAP)	2
14	Nylon, transparent	2
17	Thermoplastic polyester, PBT	2
23	Ionomer	2
24	Polyethylene, low density (LDPE)	2
26	Polypropylene, copolymer	2
31	Acetal resin, homopolymer	2
34	Ethylene vinyl acetate (EVA)	2
39	Polyester elastomer (PE)	2
49	Polyethylene, medium density (MDPE)	2
1	Polystyrene, general purpose (GPPS)	3
22	Polybutylene (PB)	3
27	Polypropylene, homopolymer	3
32	Acetal resin, copolymer	3
38	Polypropylene, flame retardant	3
40	ABS, flame retardant	3
43	Polymethylpentene (PMP)	3
44	Polypropylene, talc reinforced	3
45	Polypropylene, calcium carbonate reinforced	3
33	Polyphenylene sulfide (PPS)	4
35	Synthetic elastomer	4
37	Urethane elastomer, thermoplastic (TPU)	4
41	Polyallomer	4
48	Thermoplastic rubber (TPV)	4
28	Polypropylene, barium sulfate reinforced	5
36	Polypropylene, glass filled	5
46	Polypropylene, mica reinforced	5
47	Nylon, type 66, 33% glass	5

PBT, polybutylene terephthalate; TPU, thermoplastic polyurethane; TPV, thermoplastic vulcanisate.

Methods. The FT-IR spectra of all the specimens were recorded using a spectrophotometer (Alpha Bruker Optics) equipped with accessories for the three configurations: Trans, ATR, and TR modes.

Transmission Mode . All 50 samples were prepared using the same procedure. A small sample was scraped from a lateral tooth of the chip and was thinly ground and mixed with KBr (Sigma-Aldrich, 99% FT-IR grade) to prepare a pellet. In some cases, depending on the nature of the polymer (e.g., hard or gummy), it was difficult to obtain a homogeneous dispersion of the material in the KBr matrix. Although this aspect could affect the final spectral quality, we decided to adopt the same preparatory procedure for all the materials analyzed to facilitate their comparison with the FT-IR data acquired using the other configurations.

The spectral range investigated was 4000–375 cm⁻¹, with 4 cm⁻¹ resolution, and 64 scans. The Trans spectra are reported in absorbance scale to facilitate their comparison with the ATR and TR data.

Attenuated Total Reflection Mode . The ATR spectra were noninvasively recorded on the surface of the thickest part of each specimen, using a diamond crystal. The spectral range investigated was 7000–375 cm⁻¹, with 4 cm⁻¹ resolution, and 64 scans. All the spectra were processed using the ATR correction tool for ATR diamond crystal in the Opus 7.0.122 software (Bruker Optics). Other post-spectroscopic manipulations (such as normalization) were avoided.

Total Reflection Mode . The TR spectra were noninvasively registered on the central area of the specimen where the thickness was 3 mm. The spectral range investigated was 7000–375 cm⁻¹, with 4 cm⁻¹ resolution. Spectral acquisition was usually made using 64 scans; in some cases, it was necessary to acquire 128 scans to optimize the signal-to-noise ratio.

All the TR spectra were processed using the KK algorithm. This procedure was usually applied in the 4000–375 cm⁻¹ range, but in a few cases the KK algorithm was applied only to selected regions of the spectra.

RESULTS AND DISCUSSION

Because acquisition in Trans mode is traditionally the most widespread FT-IR technique for identifying the materials under investigation, and most FT-IR spectral libraries are based on Trans spectra, the 50 samples were first characterized in Trans mode to build a reference spectral archive. Subsequently, these spectra were compared to the ones acquired in both ATR and TR modes, and the differences and similarities were analyzed. Based on the results, five case typologies were identified.

Case 1. Spectra Acquired in Transmission, Attenuated Total Reflection, and Total Reflection Modes Are All Comparable. This is the most desirable condition because the TR spectra were superimposable on both the ATR and Trans spectra, and this made it possible to identify the polymeric material using TR. In the collection of plastic samples analyzed, 17 specimens fell into this category (34% of the cases). For example, in the spectra acquired from ResinKit no. 25 (polyethylene, high density) shown in Fig. 1, it can be observed that all the spectra have a similar shape and are in good agreement. The slight difference in intensity, which was mainly observed for the absorption bands at 1468 and 721 cm⁻¹, did not affect the spectral interpretation.

Fig. 1.

The FT-IR spectra of specimen no. 25 (polyethylene, high density). (a) Transmission mode. (b) Attenuation transmission reflection mode. (c) Total reflection mode.

Case 2. Spectra Acquired in Total Reflection and Attenuated Total Reflection Modes Are Comparable to Each Other and Differ from the Ones Acquired in Transmission Mode. The spectra acquired in ATR and TR modes evidenced a higher signal-to-noise ratio than those acquired in Trans. Therefore, the Trans spectra were less resolved, although all the most significant bands of the polymers under examination were still evident. Fifteen specimens were placed in this class (30% of the cases). A typical example is sample no. 14 (nylon, transparent); its three types of FT-IR spectra are shown in Fig. 2. The most intense peaks of the amide I and amide II bands at 1640 cm⁻¹ and 1540 cm⁻¹, respectively, are intense and sharp in the ATR and TR spectra, but in the Trans spectrum these bands appear to be much broader. Furthermore, the band at about 3300 cm⁻¹ that corresponds to the N–H stretching is clearly evident in the ATR and TR spectra, but is hidden by the broad band of the water in the Trans spectrum. In some cases, due to the complexity of reducing the plastic material to a thin powder, the difficulty in obtaining a homogeneous dispersion of the analyzed polymer in the KBr pellet was the main cause of the differences encountered in the Trans spectra compared to the ATR and TR spectra.

Fig. 2.

The FT-IR spectra of specimen no. 14 (nylon, transparent). (a) Transmission mode. (b) Attenuation transmission reflection mode. (c) Total reflection mode.

This is a typical example of polymers for which the researcher would have to resort to using different types of sample preparations to increase the quality of the spectral data, such as diamond cells, ATR, and TR measurement modes.

Case 3. Spectra Acquired in Attenuated Total Reflection and Transmission Modes Are Comparable to Each Other, but Differ from the Ones Acquired in Total Reflection Mode. This is one of the most critical categories to be considered. In the present study, this condition occurred in 9 out of the 50 samples investigated (18%). An example is specimen no. 44 (polypropylene, talc reinforced); its ATR and TR spectra are shown in Fig. 3 (no differences were found between its ATR and Trans spectra). We can observe that the absorption bands falling within the 1500–500 cm⁻¹ range are acceptably comparable; that is, both spectra show similar absorption bands at the same wavelengths, although the general shapes of the spectra are slightly different (Figs. 3a and 3b). In contrast, the curves are appreciably different in the 3000–2700 cm⁻¹ and 500–375 cm⁻¹ ranges. In these cases, the application of the KK spectral-transform algorithm made it possible to obtain a good comparability of the TR spectra with the Trans and ATR spectra.

Fig. 3.

The FT-IR spectra of specimen no. 44 (polypropylene, talc reinforced). (a) Untransformed TR mode. (b) Attenuation transmission reflection mode. (c) Kramers–Kroenig-transformed TR mode.

Therefore, no direct comparison was possible between the TR and the Trans or ATR spectra for this category of samples for the entire frequency range. A suitable spectral matching could be obtained among the data acquired in the different configurations only after the application of the KK algorithm in selected spectral regions.

Case 4. Transmission, Attenuated Total Reflection, and Total Reflection Mode Spectra Are All Different and Are Not Comparable to Each Other. This is the least common category, and this condition occurred in five specimens (10%). An example is ResinKit no. 37 (urethane thermoplastic elastomer), whose spectra are shown in Fig. 4. In this case, the Trans spectrum (Fig. 4a) appears to be quite unusual due to the presence of broadened and poorly resolved bands with anomalous intensities. This anomalous spectral behavior may be related to the extreme difficulty we had in obtaining a homogeneous dispersion of the material in the KBr matrix because of the soft and gummy consistency of the polymer sample. In contrast, the typical spectral features of polyester-based thermoplastic polyurethane are present in the ATR spectrum (Fig. 4b).³⁴ Further, the TR spectrum (Fig. 4c) presents absorption bands that could be used to correctly identify the polymer. Note, however, that in this category the application of the KK algorithm did not make the TR spectra more comparable to the Trans and ATR spectra.

Fig. 4.

The FT-IR spectra of specimen no. 37 (urethane thermoplastic elastomer). (a) Transmission mode. (b) Attenuation transmission reflection mode. (c) Total reflection mode.

As shown in Table I, the five specimens grouped in case 4 have diverse compositions and appearances.

Case 5. Spectra Acquired in Attenuated Total Reflection and Total Reflection Modes Are Comparable but Fail to Reveal the Presence of Additives Only Detectable in Transmission Mode. In principle, this condition could be considered a subgroup of case 3. Nevertheless, it deserves particular attention because it points out a substantially different circumstance—here the Trans spectrum is more informative than the ATR and TR spectra. In the present study, this condition occurred in four cases (8%), all of them polymers containing inorganic additives. An example is sample no. 28 (polypropylene, barium sulfate reinforced), whose spectra are shown in Fig. 5. In the Trans spectrum, the absorption bands attributed to barium sulfate (BaSO₄) in the 1200–1050 cm⁻¹ range (sulfate group [SO₄^2–] asymmetric stretching) and the bands at 640 and 615 cm⁻¹ (SO₄^2– bending vibrations) are present. In contrast, these spectral features are absent or show very weak intensities in both the ATR and TR spectra, thus preventing a complete identification of the included additives in the polymer formulation. This result might be explained by the fact that, depending on the portion of the specimen sampled, the Trans mode technique can detect the bulk composition of the investigated samples more easily (if the specimen has been adequately ground), whereas the other two techniques have a penetration depth limited to the first microns of the sample. Because the inorganic fillers and additives used for reinforcement in the plastics may have a minimal concentration on the surface, they are not easily revealed using the ATR and TR modes. Therefore, for the specimens in case 5, the Trans FT-IR technique turned out to be decisive for identifying the fillers and additives (e.g., calcium carbonate, BaSO₄, mica, and glass) included in the formulation of the polymer.

Fig. 5.

The FT-IR spectra of specimen no. 28 (polypropylene, barium sulfate reinforced). (a) Transmission mode. (b) Attenuation transmission reflection mode. (c) Total reflection mode.

Another example is shown in Fig. 6; here the spectra acquired from specimen no. 36 (polypropylene, glass filled) are compared. Once again, the broad absorption band at 1200–900 cm⁻¹, due to the presence of glass, is detectable only using the Trans spectrum.

Fig. 6.

The FT-IR spectra of specimen no. 36 (polypropylene, glass filled). (a) Transmission mode. (b) Attenuation transmission reflection mode. (c) Total reflection mode.

As a general comment with regard to the comparability of the three techniques, we can state that the spectra recorded using both the ATR and TR techniques were often in poor agreement with those obtained using Trans mode in terms of the intensity and shape of the absorption bands; however, this usually did not affect the correct identification of the materials. For a relatively small number of cases (14 specimens, falling into cases 2 and 3) of the 50 analyzed, TR could not be considered interchangeable with and as reliable as Trans and ATR. This indicates that portable FT-IR devices may be used to implement the noninvasive identification of constituent polymers in plastic artworks, provided that a certain margin of inexact attributions is taken into account when TR spectra are compared to spectral libraries built using Trans and ATR. Indeed, the reliability and good quality of the TR spectra acquired without sampling have been confirmed in most cases.

A side outcome of this study performed on the ResinKit set was the demonstration that, in a few cases, FT-IR analysis succeeded in highlighting some discrepancies between the declared composition and the actual formulation of the specimens analyzed. For example, the FT-IR spectrum of sample no. 18, classified in the ResinKit set as thermoplastic polyester, glycol-modified polyethylene terephthalate (PETG), had the spectral features of an alloy of polyethylene terephthalate (PET) and polycarbonate (Fig. 7). This material is available commercially and is used to manufacture such items as greenhouse roofs, automobile instrument panels, wheel covers, snowmobiles, and cellular phones.

Fig. 7.

The FT-IR spectra of three plastics. (a) Polyethylene terephthalate. (b) ResinKit no. 18, thermoplastic polyester, PETG. (c) Polycarbonate.

Another general consideration concerns the possibility of distinguishing polymers that belong to the same polymeric class but have different densities. Based on the results obtained, we found that, irrespective of the instrument configuration used, FT-IR techniques were not suitable for this finer discrimination. This was the case for the following polymeric classes: polyethylene (no. 24, low density; no. 25, high density), polystyrene (no. 1, general purpose; no. 2, medium impact; no. 3, high impact), and acrylonitrile butadiene styrene (ABS) (no. 5, transparent; no. 6, medium impact; no. 7, high impact).

CONCLUSION

A set of 50 thermoplastic resins was characterized using three different FT-IR instrument configurations: Trans, ATR, and TR modes. Our ultimate aim was to investigate the quality and reliability of FT-IR spectra acquired noninvasively using TR mode and portable FT-IR spectrometers. The TR spectra were then compared with those acquired using the Trans and ATR modes, and the results were classified into categories based on their agreement level.

We found that the spectra recorded using both the ATR and TR techniques were often in poor agreement with those obtained using the Trans mode, in both the intensity and the shape of the absorption bands. Conversely, we found in most cases that the spectra acquired using the ATR and TR techniques were definitely comparable, particularly after processing the TR spectra by means of the KK algorithm. However, the results we obtained point out that the application of the KK algorithm should be considered case by case, on the basis of a careful evaluation of the frequency range in which the KK correction is to be made. Moreover, in a few cases the TR spectra were not comparable with those collected using the ATR and Trans modes.

The results obtained reveal that, in the majority of cases, the TR mode may be used advantageously in situ to obtain a noninvasive identification of constituent polymers in plastic objects, especially for large-scale and preliminary surveys of plastic artworks. However, a certain margin (18% according to the results presented here) of inexact attributions has to be taken into account when comparing TR spectra with spectral libraries built using different techniques (Trans and ATR).

We conclude that, considering that the TR FT-IR technique is a noninvasive and portable methodology for in situ measurements, our results are promising for application in the cultural heritage field and in the analysis of plastic artworks.

Footnotes

ACKNOWLEDGMENTS

This work was carried out within the Preventive Conservation of Contemporary Art project (COPAC, 2011–2013) financed by the Regione Toscana (Tuscan Region, Italy) within the framework of the PAR-FAS Regional Project (2007–2013). The research presented is based partially on materials and results obtained within the framework of the EC POPART Project (FP7/2007-2013, Grant Agreement No. 21221), financed by the European Community.

References

Derrick

Stulik

Ordonez

. “Deterioration of Cellulose Nitrate Sculptures Made by Gabo and Pevsner”. In: Grattan

, editor. Saving the Twentieth Century: The Conservation of Modern Materials. Ottawa: Canadian Conservation Institute, 1993. Pp. 169–182.

Shashoua

. Conservation of Plastics: Materials Science, Degradation and Preservation. Oxford: Butterworth-Heinemann (Elsevier), 2008.

Waenting

. Plastics in Art: A Study from the Conservation Point of View. Petersberg, Germany: Michael Imhof Verlag, 2008.

Albus

Bonten

Keßler

Rossi

Wessel

. Plastic Art—A Precarious Success Story. Cologne, Germany: AXA Art Versicherung, 2007.

POPART: Preservation of Plastic ARTefacts in Museum Collections. http://popart.mnhn.fr/ [accessed Jul 5 2013].

La Conservazione Preventiva dell'Arte Contemporanea: Il Progetto COPAC. http://copac.sns.it/ [accessed Jul 5 2013].

INCCA: International Network for the Conservation of Contemporary Art. http://www.incca.nl/ [accessed Jul 5 2013].

Pastorelli

Trafela

Taday

P.F.

Portieri

Lowe

Fukunaga

Strlič

. “Characterisation of Historic Plastics Using Terahertz Time-Domain Spectroscopy and Pulsed Imaging”. Anal. Bioanal. Chem. 2012. 403(5): 1405–1414.

Cucci

Bigazzi

Picollo

. “Fibre Optic Reflectance Spectroscopy as a Non-Invasive Tool for Investigating Plastics Degradation in Contemporary Art Collections: A Methodological Study on an Expanded Polystyrene Artwork”. J. Cult. Her. 2013. 14(4): 290–296.

10.

Lattuati-Derieux

Egassea

Thao-Heu

Balcar

Barabant

Lavédrine

. “What Do Plastics Emit? HS-SPME-GC/MS Analyses of New Standard Plastics and Plastic Objects in Museum Collections”. J. Cult. Her. 2013. 14(3): 238–247.

11.

Cairns

Chalmers

J.M.

Everall

N.J.

Ellison

. “Specular Reflectance: A Convenient Tool for Polymer Characterisation by FTIR-Microscopy?”. Micron. 1996. 27(5): 315–328.

12.

Littlejohn

Pethrick

R.A.

Quye

Ballany

J.M.

. “Investigation of the Degradation of Cellulose Acetate Museum Artefacts”. Polym. Degrad. Stabil. 2013. 98(1): 416–424.

13.

Van Oosten

. “Identification and Characterization of Plastic Artefacts”. In: Lavédrine

Fournier

Martin

, editors. Preservation of Plastic Artefacts in Museum Collections. Paris: Comité des Travaux Historiques et Scientifiques (CTHS), 2012. Pp. 29–109.

14.

Naranjo

A.C.

del Pilar Noriega

Osswald

T.A.

Alejandro

R.-A.

Sierra

J.D.

. Plastics Testing and Characterization: Industrial Applications. Munich: Hanser Publishers, 2008.

15.

Casadio

Toniolo

. “The Analysis of Polychrome Works of Art: 40 Years of Infrared Spectroscopic Investigations”. J. Cult. Her. 2001. 2(1): 71–78.

16.

Derrick

M.R.

Stulik

Landry

J.M.

. Infrared Spectroscopy in Conservation Science. Los Angeles: The Getty Conservation Institute, 1999.

17.

Shearer

C.L.

Doyal

. “Use of FT-IR in the Conservation of Twentieth-Century Objects”. In: Vandiver

P.B.

Druzik

Wheeler

G.S.

, editors. Materials Issues in Art and Archeology 2: Symposium G. San Francisco, CA; April 17-21, 1990. Materials Research Society Symposium Proceedings. Pittsburgh, PA: Materials Research Society 1990. Vol. 185, pp. 813–823. doi:10.1557/PROC-185-813.

18.

Keneghan

. “Assessing Plastic Collections in Museums by FTIR Spectroscopy”. In: Pretzel

and IRUG, editors. Postprints IRUG2 at the V & A: Proceedings of the Second Infrared and Raman Users Group Conference (IRUG2). London, UK: Victoria & Albert Museum, 1998. Pp. 21–24.

19.

Paris

Lecomte

Coupry

. “ATR-FTIR Spectroscopy as a Way to Identify Natural Protein-Based Materials, Tortoiseshell and Horn, from Their Protein-Based Imitation, Galalith”. Spectrochim. Acta A. 2005. 62(1): 532–538.

20.

Chércoles Asensio

San Andrés Moya

de la Roja

J.M.

Gómez

. “Analytical Characterization of Polymers used in Conservation and Restoration by ATR-FTIR Spectroscopy”. Anal. Bioanal. Chem. 2009. 395(7): 2081–2096.

21.

Laganà

Keneghan

. “Which Plastics Are in My Collections? The Need for a Plastic Reference Sample Collection (SamCo)”. In: Lavédrine

Fournier

Martin

, editors. Preservation of Plastic Artefacts in Museum Collections. Paris: Comité des Travaux Historiques et Scientifiques (CTHS), 2012. Pp. 37–42.

22.

POPART: Preservation of Plastic ARTefacts in Museum Collections. “POPART Conference 7-9 March 2012”. http://popart.mnhn.fr/spip.php?rubrique79 [accessed Jul 5 2013].

23.

IRUG: Infrared & Raman Users Group. http://www.irug.org/ [accessed Jul 5 2013].

24.

Mukhopadhyay

. “Portable FTIR Spectrometers Get Moving”. Anal. Chem. 2004. 76(19): 369A–372A.

25.

Miliani

Rosi

Brunetti

B.G.

Sgamellotti

. “In Situ Noninvasive Study of Artworks: The MOLAB Multitechnique Approach”. Acc. Chem. Res. 2010. 43(6): 728–738.

26.

Fabbri

Picollo

Porcinai

Bacci

. “Mid-Infrared Fiber Optics Reflectance Spectroscopy: A Non-Invasive Technique for Remote Analysis of Painted Layers. Part II: Statistical Analysis of Spectra”. Appl. Spectrosc. 2001. 55(4): 428–433.

27.

Fabbri

Picollo

Porcinai

Bacci

. “Mid-Infrared Fiber Optics Reflectance Spectroscopy: A Non-Invasive Technique for Remote Analysis of Painted Layers. Part I: Technical Setup”. Appl. Spectrosc. 2001. 55(4): 420–427.

28.

Ploeger

Chiantore

Scalarone

Poli

. “Mid-Infrared Fiber-Optic Reflection Spectroscopy (FORS) Analysis of Artists' Alkyd Paints on Different Supports”. Appl. Spectrosc. 2011. 65(4): 429–435.

29.

Boulet-Audet

Buffeteau

Boudreault

Daugey

Pézolet

. “Quantitative Determination of Band Distortions in Diamond Attenuated Total Reflectance Infrared Spectra”. J. Phys. Chem. B. 2010. 114(24): 8255–8261.

30.

Enlow

Kennedy

J.L.

Nieuwland

A.A.

Hendrix

J.E.

Morgan

S.L.

. “Discrimination of Nylon Polymers Using Attenuated Total Reflection Mid-Infrared Spectra and Multivariate Statistical Techniques”. Appl. Spectrosc. 2005. 59(8): 986–992.

31.

Kortüm

. Reflectance Spectroscopy: Principles, Methods, Applications. New York: Springer Verlag, 1969.

32.

Lucarini

Saarinen

J.J.

Peiponen

K.-E.

Vartiainen

E.M.

. Kramers-Kronig Relations in Optical Materials Research. Springer Series in Optical Sciences 110. Berlin: Springer, 2005.

33.

The Plastics Group of America. “ResinKit”. 2012. http://www.plasticsgroup.com/resin_kit [accessed July 5 2013].

34.

Wang

Luo

. “A Polymer Network Based on Thermoplastic Polyurethane and Ethylene–Propylene–Diene Elastomer Via Melt Blending: Morphology, Mechanical Properties, and Rheology”. Eur. Polym. J. 2004. 40(10): 2391–2399.