Characterization of archeological oak (Quercus sp.

) with mid and near infrared

spectroscopy

Anna Sandak 1, Jakub Sandak 1, Magdalena Zborowska 2, Wlodzimierz Prądzyński 2, Martino Negri 1

1 Trees and Timber Institute, IVALSA/CNR, San Michele All’Adige, Italy

2 Institute of Chemical Wood Technology, Poznan University of Life Sciences, Poznan, Poland

Abstract
Archeological wood, as most of natural materials, is slowly decomposing on the archeological site due to various
biological factors. In a consequence of such degradation significant changes within physical, mechanical and
chemical properties are occurring. Rapid and accurate estimation of the degradation level is extremely important
as it influences the selection of optimal restoration and conservation procedures.
The goal of this research was to verify the usefulness of infrared (FT-ATR and FT-NIR) techniques for
monitoring of the ageing process in archeological wood. The important advantage of the infrared spectroscopy is
its accuracy, simplicity, and ability of performing very high number of tests without needs of any destruction to
the workpiece. The methods explored in this research might be a novel tool assisting experts in evaluation of the
degradation state of archeological and historical wooden materials. Examples of infrared application in to
qualitative and quantitative evaluation have been presented. By linking both infrared techniques with novel
signal processing methods it is possible to improve the knowledge and interpretation of spectra.
1. INTRODUCTION
Wood is considered as a very friendly and useful raw material used extensively for various purposes
from ancient times. Archeological wood, as most of natural materials, is slowly decomposing on the
archeological site due to various biological factors. In a consequence of such degradation significant
changes within physical, mechanical and chemical properties are occurring. Rapid and accurate
estimation of the degradation level is extremely important as it influences the selection of optimal
restoration and conservation procedures.
Various analytical techniques might be applied for estimation of the archeological wood condition and
degradation stage. Chemical methods are very accurate and allow detailed interpretation of the
degradation process. An important limitation for use of chemical methods is however its destructive
nature, as well as long time duration of the test and its relatively high costs.
Recent developments in the fields of optics and electronics opened new possibilities for nondestructive
measurements of various physical and chemical properties of materials. Therefore, a tendency for
application of various analytical techniques, such as infrared spectroscopy into measurement of the
ancient wooden artifacts is observed nowadays [5]. The advantage, beside of the nondestructive
character, is the simplicity of measurement and possibility of numerous repetitions what significantly
increases the reliability and accuracy of data. Most of the state-of-the-art methods allow measurement
directly on the object in the field, without the necessity of sampling or sample transportation.
Instrumental methods, even possessing plenty of advantages, are rather rarely applied in the
conservation laboratories due to lack of the dedicated (to archeological wood) calibrations or simply
by limited trust to the novel equipment. In some cases the results provided by the equipment are not
directly related to the degradation degree of wood and some further interpretation is necessary. It is
important therefore to adopt the available instrumentation techniques for better serving a conservation
tasks and correlate results provided by the instrument to the real state of the historical object.

Infrared spectroscopy, in both near and mid ranges, is well known technique with a great potential for
chemical characterization of materials. It is very useful tool for identifying various inorganic and
organic compounds, base on their selective absorption of radiation in the infrared region. However
until recently, the sample preparation to be measured with infrared spectrometers has been rather
complicated; for example solving the sample or preparation of potassium-bromide (KBr) pellets for
sample analysis.
Fourier Transform Attenuated Total Reflectance (FT-MIR-ATR) is relatively new advancement of
traditional mid infrared spectroscopy. It uses a phenomenon of absorbing infrared energy during
reflection from the measured surface. This technique allows measurement of the wood powder or even
wooden block surface without time consuming KBr sample preparation. It is expected that by applying
FT-MIR-ATR in to area of archeometry, a new tool for estimation of the lignin, cellulose and
hemicelluloses content might be developed. On the other hand Fourier Transform Near Infrared (FT-
NIR) spectroscopy offers even simpler measurement. Unfortunately the signal-to-noise ratio is poorer
that this of FT-MIR-ATR and interpretation of spectra more problematical.

The goal of this research was therefore to verify the usefulness of FT-MIR-ATR and FT-NIR
techniques for monitoring of the ageing process in archeological wood. The performance of the both
methods would be compared to each other and to well-established experimental reference methods.

2. MATERIALS AND METHODS

2.1. Experimental samples
Five oak (Quercus sp.) pieces of the archeological wood have been used as experimental samples.
Each sample has been collected from the waterlogged site in the area of Poland. The details regarding
geographical origin and waterlogging are summarized in table 1. A piece of contemporary wood of the
same species has been also included in analyses for reference purposes. A small part of each wooden
block has been milled with the laboratory mill (Pulverisette 15, Fritsch GmbH), to transform the
structure into powder (fraction <0,5mm).

Table 1. Characteristics of the experimental materials

Sample code Origin Dating Waterlogging period, years
Q1 Biskupin VIII BC 2700
Q2 Bronisze (Warsaw) III AD 1700
Q3 Lednica X-XI AD 1000
Q4 Ostrow Tumski (Poznan) IX-X AD 1000
Q5 Szczecin XIII AD 700
Q6 Golabki XX AD 0

2.2. Chemical analysis

Chemical analyses have been performed on the experimental samples in order to generate reference
data for calibration and validation of the infrared methods. The analysis included determination of the
main chemical components concentration in the wood samples. Quantity of cellulose has been
determined according to the Seifert procedure (by using acetylacetone-dioxane-hydrochloric acid) and
lignin was conducted according to the TAPPI standard T 222 om-06 [3].

2.3. FT-NIR measurement

All experimental samples were measured by using Fourier transform near infrared spectrometer
VECTOR 22-N produced by Bruker Optics GmbH. The spectral range measured was between
4000 cm-1 and 12000 cm-1. The spectral resolution of the spectrophotometer was 8 cm-1, each spectrum
has been computed as an average of 32 successive measurements in order to minimize the
measurement error. Five separate measurements have been performed on each sample.

2.4. FT-IR-ATR measurement

All experimental samples have been also measured by using Fourier transform attenuated total
reflectance spectrometer Alpha produced by Bruker Optics GmbH. The spectral range measured was
between 4000 cm-1 and 600 cm-1. The spectral resolution of the spectrophotometer was 4 cm-1, each
spectrum has been computed as an average of 32 successive measurements in order to minimize the
measurement error. Five measurements have been performed on each sample. It has to be mentioned,
that all measurements for both techniques has been performed in climatic chamber (20°C and 65%RH)
to minimize influence of temperature and humidity of environment.

2.5. Data processing

OPUS 6.5 and National Instruments LabView 8.6 software packages have been used for signal
processing and data analyzes. Signal preprocessing included computation of derivatives, normalization
and smoothing. Derivatives were calculated according to the Savitzky-Golay algorithm. For the
mathematical management of the spectra and then for the evaluation of the results both Principal
Component Analysis (PCA) and Partial Least Square (PLS) matching techniques were applied. For 2D
correlation the algorithms based on the Noda [2] works has been implemented.

3. RESULTS AND DISCUSSION

Average spectra scanned from the archeological waterlogged oak samples by means of FT-NIR and
FT-MIR-ATR are presented in figure 1 and figure 2 respectively. It is clear that the degradation of the
waterlogged woods is evident on the spectra; however more detailed investigations are necessary to
point the effects of waterlogging. Table 2 presents a collection of the knowledge related to
interpretation of the spectra based on the literature data [4, 1].

Table 2 band assignments for interpretations of infrared spectra

Wavenumber
-1
cm band assigment
1 7000 OH stretching first overtone amorphous regions in cellulose
2 6718 OH stretching first overtone semi-crystalline regions in cellulose
3 6450 OH stretching first overtone crystalline regions in cellulose
4 6287 OH stretching first overtone crystalline regions in cellulose
5 5980 CH stretching first overtone aromatic skeletal due to lignin
6 5883 CH stretching first overtone
FT-NIR

7 5800 CH stretching first overtone furanose/pyranose due to hemicellulose

8 5587 CH stretching first overtone semi- or crystalline regions in cellulose
9 5464 OH stretching + CO stretching semi- or crystalline regions in cellulose
10 5219 OH stretching + OH deformation water
11 4890–4620 OH stretching + CH deformation cellulose
12 4404 CH2 stretching + CH2 deformation
13 4280 CH stretching + CH deformation semi- or crystalline regions in cellulose
14 4198 CH def. second overtone holocellulose
15 1724 CO unconjugated groups in lignin and carboxylic acid ester hemicelluloses
16 1587 Conjugated CO
17 1500 Aromatic skeletal; Benzene ring vibration in lignin
18 1450 CH deformation in lignin and carbohydrates;
19 1450 CH3, CH2, benzene ring vibration in lignin
20 1417 CH deformation in lignin and carbohydrates
21 1417 Aromatic skeletal vibrations combined with CH in plane deformation
FT-MIR-ATR

22 1363 CH deformation in cellulose and hemicelluloses;

23 1363 CH bending vibration in cellulose and hemicelluloses
24 1319 CH vibration in cellulose; CO in syringyl derivatives
25 1226 Syringyl ring; CO stretching lignin and xylan
26 1151 COC vibration in cellulose and hemicelluloses
27 1116 Aromatic skeletal; CC stretch;
28 1116 OH association band in cellulose and hemicelluloses
29 1024 CO stretch in cellulose and hemicelluloses;
30 1024 CO of primary alcohol
31 892 CH deformation in cellulose;
32 892 C group frequency in cellulose and hemicellulose

The most significant differences in the spectra can be observed for lignin (5980 cm-1, 1500 cm-1,
1450 cm-1, 1226 cm-1), hemicelluloses (5800 cm-1, 1724 cm-1, 1363 cm-1, 1151 cm-1) and amorphous
region in cellulose (7000 cm-1). No much difference on the spectra has been however observed in the
crystalline and semi-crystalline regions of cellulose (6718 cm-1, 6450 cm-1, 5587 cm-1, 5464 cm-1,
1024 cm-1)
Interpretation of results obtained with FT-NIR and FT-MIR-ATR measurements of the waterlogged
archeological oaks are similar to those reported by Tsuchikawa et al. [4]. They have concluded that the
amorphous, and partially the semi-crystalline, regions of cellulose, hemicellulose and lignin changes
during ageing. Conversely, the crystalline region in cellulose seems not to be affected by ageing.

0,000100

0,000080 A
0,000060
A B
0,000040
2nd der. of absorbance

0,000020

0,000000

-0,000020
Q1
Q1
-0,000040 Q2
Q2
Q3
-0,000060 Q3
Q4
-0,000080 Q5
Q6
C Q4
Q5
-0,000100 Q6
7500 7250 7000 6750 6500 6250 6000 5750 5500 5250 5000 4750 4500 4250 4000
-1 7500 7400 7300 7200 7100 7000 6900 6800 6700 6600 6500 6400 6300 6200 6100
wavenumber (cm )
wavenumber (cm-1)

B C

Q1 Q1
Q2 Q2
Q3 Q3
Q4 Q4
Q5 Q5
Q6 Q6

6100 6000 5900 5800 5700 5600 5600 5500 5400 5300 5200 5100 5000 4900 4800 4700 4600 4500 4400 4300 4200
wavenumber (cm -1) wavenumber (cm -1)

Figure 1 NIR spectra of waterlogged archeological oaks

2,0
Q1
Q2
Min-max normalization of absorbance

Q3

1,5
Q4
Q5
B
Q6

1,0
A

0,5

0,0
4000 3750 3500 3250 3000 2750 2500 2250 2000 1750 1500 1250 1000 750
wavenumber (cm-1)

Q1
Q2
A Q1
Q2
B
Q3 Q3
Q4 Q4
Q5 Q5
Q6 Q6

3700 3450 3200 2950 2700 1750 1500 1250 1000 750
wavenumber (cm-1) wavenumber (cm-1)

Figure 2 MIR-ATR spectra of waterlogged archeological oaks

 NIR MIR-ATR

Q3
Q2
Q6
Q4
Q5

Q4 Q1
Q1 Q5

Q6 Q3 Q2

Figure 3 Discrimination of the archeological oak groups with Principal Components Analysis,
Note: NIR; 1st derivative, vector normalization, 13 smoothing points, 3 factors, region 9000 to 4242 cm-1 and
MIR-ATR; vector normalization, 2 factors, region 1800 to 750 cm-1

rank 6, 1st derivative + vector normalization, 17 smoothing points, rank 6, 1st derivative + vector normalization, 17 smoothing points,
-1 -1
range 3607 – 3278 and 1968.2 to 984 cm range 3935 to 3606, 3279 to 2623, 1968 to 984 cm
45 70
y = 0,985x + 0,476 y = 0,998x + 0,057
R2 = 0,996 65 R2 = 0,999
RMSECV = 0,405 RMSECV = 0,366

40 60
estimated cellulose concentration (%)
estimated lignin concentration (%)

55

35 50

45

30 40

35

25 30

MIR-ATR lignin 25
MIR-ATR cellulose
20 20
20 25 30 35 40 45 20 25 30 35 40 45 50 55 60 65 70
reference lignin concentration (%) reference cellulose concentration (%)

rank 7, min-max normalization, rank 8, 1st derivative + vector normalization, 17 smoothing points,
range 10402 to 8797, 6402 to 4000 cm-1 range 8801 to 7197, 6402 to 4000 cm -1
45 70
y = 0,985x + 0,476 y = 0,998x + 0,057
R2 = 0,996 R2 = 0,999
65
RMSECV = 0,201 RMSECV = 0,289

40 60
estimated cellulose concentration (%)
estimated lignin concentration (%)

55

35 50

45

30 40

35

25 30

NIR lignin 25
NIR cellulose
20 20
20 25 30 35 40 45 20 25 30 35 40 45 50 55 60 65 70
reference lignin concentration (%) reference cellulose concentration (%)

Figure 4 Partial Least Squares estimation of the lignin and cellulose content in archeological waterlogged oak
woods by FT-NIR and FT-MIR-ATR spectroscopes
Other implementations of the FT-NIR and FT-MIR-ATR measurements, and its links to the reference
analysis, might be related to qualitative and quantitative evaluation of the samples. Figure 3 presents
an example of the discrimination of the samples of different degradation levels and/or provenances by
means of Principal Components Analysis. Very clear separation has been obtained by both FT-NIR
and FT-MIR-ATR techniques. Further improvement of this method (identification test) might be
utilized for rapid recognition of both degradation and origin of samples.

synchnous asynchnous
a)

b)

c)

Figure 5 2D spectral correlation between NIR and MIR-ATR spectra of waterlogged archeological oaks;
whole spectral range a), zoom to the significant ranges b), top view c)
Figure 4 shows an example of the infrared spectroscopy’s potential for estimation of the chemical
composition of the waterlogged samples. After processing spectra with Partial Least Squares
algorithm it was possible to determine the calibration models to be used for further prediction in the
unknown samples.

The last example of the spectra’ processing is 2D spectral correlation. The method developed by Noda
[2] bases on the computation of the auto-correlation or cross-correlation within/between several
spectra. Typically it is applied for visualization of the changes in to spectra due to so called
disturbances. These include temperature change, moisture change, concentration change, etc. In this
research however the spectral band has been considered as a disturbing factor and such evaluation is
called heterospectral. The specially designed software computed cross-correlation between FT-NIR
and FT-MIR-ATR spectra scanned from the archeological woods. The resulting charts are presented in
figure 5. Two diagrams are usually generated with 2D spectral correlation technique: synchronous and
asynchronous relations. The synchronous spectrum represents simultaneous or coincidental changes to
the spectrum. On the other hand asynchronous spectrum represents sequential or successive (but not
co-incidental) changes to the spectrum. The vertical axis of each 3D chart is an indicator of the
importance of the correlation between both spectral wavenumbers (NIR and MIR). It is clear the
highest values have been computed for all the wavenumbers belonging to the fingerprint region. The
magnified views are presented in figure 5 b and c.

The interpretation of the 2D spectral correlation is not straightforward, but if successful it can be very
useful for improving the spectrum interpretation by linking of unknown band assignments of one
technique to well established assignments of the second technique. Figure 6 is an example of the
knowledge building with the support of 2D spectral correlation; the known NIR band related to
hemicelluloses (#7, as in table 2) can be linked to the following bands discovered by MIR analysis
(#29, #28, #26, #25, #20, #18, as in table 2).

29 28 26 25 20 18

Figure 6 Example for interpretation of the 2D spectral correlation between NIR and MIR-ATR; hemicelluloses

4. CONCLUSIONS
The important advantage of the infrared spectroscopy is its accuracy, simplicity, and ability of
performing very high number of tests without needs of any destruction to the workpiece. The methods
explored in this research might be a novel tool assisting experts in evaluation of the degradation state
of archeological and historical wooden materials. Examples of infrared application in to qualitative
and quantitative evaluation have been presented. By linking both infrared techniques with novel signal
processing methods it is possible to improve the knowledge and interpretation of spectra.

Acknowledgments
The studies reported here are funded from the financial resources granted by the State Committee for
Scientific Research of Poland (KBN) within the framework of the N N309 135135 project.

References
1. Mohebby, B. (2005): Attenuated total reflection infrared spectroscopy of white-rot decayed beech
wood. International Biodeterioration and Biodegradation 55:247–251
2. Noda, I., Ozaki, Y. (2005): “Two-dimensional correlation spectroscopy” John Willey & Sons Ltd,
Chichester
3. TAPPI standard T 222 om-06
4. Tsuchikawa, S., Yonenobu, H., Siesler, H.W. (2005): Near-infrared spectroscopic observation of the
ageing process in archaeological wood using a deuterium exchange method. Analyst 130:379–384
5. Yonenobu, H., Tsuchikawa, S., Haraguchi, Y., Oda, H., Nakagawa, T. (2003) Non-destructive NIR
spectra measurement of dendrochronologically dated antique wood. Proceedings NIR-2003 - 11th Int'l
Conference