Characterization of Historical Parchment Using Laser-Induced Breakdown Spectroscopy

Parchments are one of the oldest and most durable information media. Some historical parchments belong to the cultural heritage of mankind. Therefore, it is important to develop a rapid and reliable method for parchment identification and characterization of preservation conditions. The advantages of laser-induced breakdown spectroscopy (LIBS) are its portability, low operational cost, short analysis time, and virtual nondestructiveness; also, it does not require sample preparation. Therefore, LIBS is suitable for routine classification in which a large number of samples have to be analyzed.

Parchment is produced from animal skin and consists of a multilayered semisolid matrix. Parchments from various sources were analyzed using LIBS to obtain both surface distribution and depth profile of elements of interest, namely, Na, Ca, Cu, Fe, K, Mg, and Mn. Additionally, internal distribution of fluorophores in parchment layers was investigated by means of laser breakdown ablation (LBA) followed by synchronous fluorescence (SF).

Materials and methods

Parchment samples

A large set of 54 parchment samples (37 new and 17 historical), produced from various animal skins, was assembled.

Clean parchment matrices preparation

Small pieces of parchment were placed in 10 mL of double-distilled water and shaken at room temperature for 24 hr. Two additional cycles were applied using fresh portions of water. The cleaned matrix samples were then air-dried at room temperature for 24 hr.

LIBS and LBA

A fundamental harmonic (1064 nm) Powerlite Nd:YAG laser (Continuum®, Santa Clara, CA) operating at 1 Hz at 7-nsec pulse duration was focused on the sample. The laser energy was adjusted to 20 mJ/pulse. The plasma emission was delivered by a quartz fiber to the SpectraPro-500i spectrometer (Acton Research Corp., Acton, MS, grating 1200 g/mm) coupled with a Princeton Instruments intensified charge-coupled device (ICCD) (Roper Scientific, Trenton, NJ). Typical ICCD delay and integration times were 1.5 μsec and 5.0 μsec.

Depth profiling of the elemental distribution was obtained by applying a series of laser shots to the same location (pulse energy 30 mJ, focused beam intensity 3.5 × 109 W/cm2). The number of shots (100–270) was dependent on sample thickness and conditions.

ICP measurements

Inductively coupled plasma - optical emission spectroscopy (ICPOES) using an Optima 3000V (PerkinElmer, Norwalk, CT) served as reference technique for LIBS measurements. Approximately 30 mg (dry weight) of parchment was ashed at 600 °C and was dissolved in 10% nitric acid. Multielemental standards were used for calibration at the appropriate dilution.

The parchment’s layered structure examination was obtained by means of synchronous fluorescence spectra. First, LBA was carried out by focusing the laser beam on the parchment surface through a cylindrical lens (f = 70 mm). The pulse energy was 310 mJ/pulse. Layer-resolved profiling was achieved by moving horizontally from the left to the right side of the parchment sample in ~50-μm steps. The vertical axis was probed by removing 3–5 μm layers to obtain interference-free SF analysis. Then, the SF spectrum of each layer was recorded (Δλ = 70 nm) using an Aminco Bowman Series 2 luminescent spectrometer (Thermo Electron Corp., Madison, WI) equipped with a surface analysis accessory. The excitation and emission bandwidths were 4 nm. The angle between the excitation beam and the sample surface was 76°.

Results

LIBS and ICP measurements

It has been recognized that the elemental composition of parchment may be indicative of its age.1 The possibility of applying LIBS for the rapid identification or verification of historical parchments is of interest. Applying ANOVA, the authors found that each element (Ca, Na, K, Fe, and Mn), when probed on the parchment surface, acted as a simple marker for the fast identification of the parchment’s age. Good linear correlation was obtained between the mean LIBS and ICP values. This is essential, because LIBS measures the surface elemental concentration while ICP provides the bulk concentrations.

Figure 1 - Parchment sample grouping according to discriminant analysis. Discriminant analysis based on three elements (Ca, Fe, and Mg) calculating Euclidian distance is able to distinguish between the various parchment groups.

Moreover, discriminant analysis based on three elements (Ca, Fe, and Mg) calculating Euclidian distance was performed. The ratio of the LIBS integral intensity to the background was used, and the resulting discriminant groups are shown in Figure 1. A clear distinction between the two groups of parchment (modern and historical) can be observed.

Identification of the animal type is often required for the parchment restoration procedure. The intensity ratio of Mg/Cu allows for statistically significant differentiation at p = 0.05 between the animal types (calf, goat, and sheep) from which the parchment is made.

Depth profile measurements

The LIBS method allows for depth profile measurements, where the signals provide relative concentrations of the elements of interest as a function of probing depth. Two factors contribute to the actual depth profile of a given element: 1) the core composition, which is most likely related to the origin of the animal skin, and 2) the material added to the parchment during its processing and later occasional contamination. The discrimination between these two components is essential for restoration and preservation. No established method for parchment matrices was previously published. Therefore, the authors developed an experimental procedure for this purpose, by washing the samples in water. The washed samples are referred to in the following as the “clean matrix.”

Figure 2 - Depth profile of original parchment samples and the corresponding clean matrices. Differentiation between core composition (bulk) and added material/contamination (surface) is possible.

Some results of the parchment depth profiling are presented in Figure 2. As expected, the concentration of the checked elements in the original samples is significantly higher than in the clean matrix. In particular, the surface elemental content is much higher than in the core. Elemental distribution in a clean matrix is much more homogeneous than that of the original samples. The calcium content in historical samples is much higher than in the modern samples. This correlates with the authors’ surface measurements and previously published work.1 The experimental curves of depth profiling are nonsymmetrical due to matter ablation from the opposite side of the examined sample during the experiment. The measurements of magnesium depth profile indicate that it may be correlated to the animal type from which the parchment is made. Future work based on a larger set of parchment samples of known animal type is required in order to statistically validate the authors’ conclusion.

LBA and examination of parchment’s layered structure

Parchment has a multilayered structure; therefore, depth-resolved measurements are crucial in understanding the nature of fluorophores in each distinct layer. The layer-resolved study of SF was performed on both modern and historical samples.

Figure 3 - Parchment layered structure examination in a) modern and b) historical samples probed by synchronous fluorescence. Each spectrum provides the change in peak intensity and position at different continuously increasing depths.

Spectral characteristics of the bulk fluorescence are strongly affected by the variation in the layer’s thicknesses. The measured bulk fluorescence intensity is a mixture of fluorescence signals from the present layer and a few layers below. Each spectrum provides the change in peak intensity and position at different continuously increasing depths. The SF spectrum shows a shift in peak positions with a simultaneous fluorescence intensity change as a function of depth (Figure 3a and b). The contribution of agerelated substances to total fluorescence decreased rapidly with thickness. When sampling depth is continuously increased up to a half-thickness, the main fluorescence peak further decreases. At this point, the majority of the fluorescence signal is the superposition of native chromophores localized within these layers with only a negligible portion contributing to the real origin of the parchment sample being aged. The fluorescence spectra taken from a few bottom layers actually look like the opposite sample side. This means that the fluorescent emission of the dominant fluorophore in the last few layers probed from the grain side is identical to that of the flesh side. The information derived from the depthresolved spectra indicates that the parchment has a layered structure when the dominant fluorophore in the upper layer (grain side) is different from that in the lower layer (flesh side). Most importantly, it possesses the ability to quantify the contribution of each chromophore in each given layer to the whole picture.

Conclusion

LIBS elemental analysis of parchment samples was carried out. A rapid distinction between modern and historical samples, based on the LIBS data and/or discriminant analysis, was obtained. Animal type recognition was also possible. LIBS results were in agreement with the ICP data.

SF coupled with LBA was successfully applied for the first time to parchment analysis. The fluorescence spectrum is very sensitive to the conformational changes in the parchment environment and deterioration stage, and is a helpful technique for the examination of parchment conditions.

Reference

  1. Microanalysis of Parchment; Larsen, R., Ed.; Archetype: London, 2002.

The authors are with the Schulich Faculty of Chemistry, Technion-Israel Institute of Technology, Haifa 32000, Israel; tel.: +972 48292816; fax: +972 48295703; e-mail: chr21bv@tx.technion.ac.il. Significant parts of this material were presented as a poster at Pittcon® 2007. The authors thank Dr. Irena Rabin for supplying parchment samples.

Comments