Busy Times with FT-IR Spectroscopy

In the last few months we were rather quiet as we had a truly busy time. But now things have calmed down and we can present some of our recent findings.

In February, we conducted FT-IR spectroscopy. This method measures molecular vibrations induced by a specific amount of energy. On the basis of the resulting spectrum we can determine the chemical components and thereby the molecular decay, or age, of the material. It has previously been applied to organic material in archaeology, such as timber and straw, and recently also to both rag paper and wood pulp paper.

There exist several techniques in IR spectroscopy, including MIR (Mid Infrared) spectroscopy External Reflection, which works contactless with a relative broad measuring area but a relatively lower accuracy, and Attenuated Total Reflection, which is performed by softly pressing a crystal on the surface of the sample. Accuracy is higher, but a certain pressure is applied making this technique problematic for very fragile samples. For our “Life of paper” project, we used MIR spectroscopy.

In a first step, we conducted FT-IR spectroscopy on dated material, i.e. printed books and dated manuscripts. The results were used to create a model which we can use to date previously undated material. In a second step we conducted FT-IR spectroscopy on undated manuscripts and smaller pieces of reused paper without watermarks, such as slips of notes and wrappers for copies of charters. For many of these note slips and wrappers the Icelandic manuscript collector Árni Magnússon reused paper.

FT-IR spectroscopy with a Bruker Alpha II on an Icelandic paper manuscript from the 17th century (c) Silvia Hufnagel

The first version of the model predicted the age of the paper used in Iceland to +/- 36 years, giving us a time span of 72 years. Using different algorithms, the prediction rate went down to +/- 21 years, which is a great improvement. But while these results are great for a first step and for testing the method, they are not entirely satisfactory since methods from the humanities, such as palaeography, give similar and in some cases even more precise datings. We do hope, though, that more statistical analysis of the model will lead to even better results.