**2. Methods**

134 Multivariate Analysis in Management, Engineering and the Sciences

applications, e.g. humification in the course of composting.

their preservation to conditions that prevented or delayed deterioration.

time under different environmental conditions.

incineration ash and slag leads to carbonation [7, 8]. With respect to organic matter humic substances are built up resulting in a stable organic fraction with low turnover rates. These natural processes that come along with material ageing were adopted for technical

On the one hand natural processes serve as a model for anthropogenic activities with regard to the closed loop of material recycling, especially in the field of organic substances [9]. On the other hand every endeavour is made to prevent or retard the natural ageing, deterioration and degradation process of materials and to maintain a constant quality of products by adequate measures. There are several options to achieve this objective: modification of biogenic materials, treatment of the surface and application of chemical substances against microbial deterioration and ageing by abiotic factors. Although abiotic factors play a relevant role for ageing and deterioration of organic materials, biological processes dominate. Inorganic materials are primarily affected by chemical and physical attacks, but some specialised microbial communities are capable of promoting the ageing process of inorganic components. The environmental milieu plays a crucial role as it determines both biological and chemical reactions. Historical and archaeological finds owe

This study reports on natural ageing and degradation processes of organic matter and ageing of inorganic materials over weeks, years, decades and centuries. The questions to be answered focus on two main aspects: the environmental impact by ageing and deterioration of organic and inorganic matter and the proof of resistance of organic materials against biological degradation which is a main concern in material sciences to maintain the quality of products [10-12]. Ageing and deterioration can be described by many parameters. They focus on chemical and physical changes of the material by which the process is paralleled. In some cases, especially for product control, a single parameter might be sufficient to verify the ageing of materials [13]. For an overall characterisation of the state of deterioration those analytical methods are advantageous that provide a "fingerprint" of the material. FT-IR spectroscopy and STA were applied to reveal material characteristics and their changes over

FT-IR spectroscopy is based on the interaction of infrared radiation with matter. Infrared radiation provides the energy for molecule vibrations that become visible as absorption bands. The plot of wavenumbers (energy) within a defined range vs. band intensities results in the spectrum. Infrared spectra describe materials by the unique pattern and provide information on material chemistry. Band intensities depend on the concentration of the compound and the molar decadic absorption coefficient that is reflected in the spectrum and on individual properties of the functional group. Most molecules are infrared active and represented by diverse bands due to different types of molecule vibrations. They are characterised by a typical energy level and are therefore found at defined wavenumbers. The molecule skeleton and other functional groups influence the band position and can cause a band shift. Whereas pure substances show distinct bands that can be attributed to functional groups, complex materials feature broad and overlapping bands that are often not assignable. However, the material shows a "fingerprint". The information of underlying

#### **2.1. FT-IR spectroscopy and simultaneous thermal analysis**

FT-IR spectra of landfill samples were recorded by a Bruker Alpha® (Bruker, Germany) instrument in the mid infrared area (4000 cm-1 to 400 cm-1) in the attenuated total reflection mode (ATR). For the milled lignocellulosic materials the ATR-FT-IR spectra were collected by a Bruker Vertex® with a Pike MIRacle™ ATR device in the wavenumber range from 4000 cm-1 to 600 cm-1, at 4 cm-1 resolution averaging 32 scans. The milled sample was homogenised and directly applied on the ATR reflection module with a diamond crystal providing a measuring area of approximately 4 mm² and a pressure applicator. Twenty four scans per spectrum were collected at a resolution of 4 cm-1 and corrected against ambient air as background. The average of four spectra (maximum deviation of the four spectra from the average spectrum < 5%) was vector normalised prior to multivariate data analysis. Spectra treatment and data evaluation were carried out using the OPUS software.

Thermal analyses were carried out with a STA 409 CD Skimmer instrument (Netzsch GmbH) in an Al2O3 pan with the following combustion parameters: temperature range 30 – 950 °C, heating rate 10 °C·min-1, gas flow 150 ml·min-1 (80% He and 20% O2), sample amount 16.00 mg. The pyrolysis of wood powder was carried out at different temperatures (200, 300, 350, and 600 °C) under oxygen-free conditions (100% He) with a heating rate of 10 °C·min-1 and isothermal treatment for 20 min. After oxidative combustion of the pyrolysed wood

powder the enthalpy was calculated by integration of the area below the heat flow curve and a horizontal baseline from 30 to 650 °C, starting at 30 °C.

Ageing and Deterioration of Materials in the Environment – Application of Multivariate Data Analysis 137

samples outside the limits do not belong to the model. They are located in the quadrant "neither - nor". The 5% significance level means that 95% of the samples in the corresponding quadrants truly belong to the defined classes. New samples in these

Partial Least Squares-Discriminant Analysis (PLS-DA) is based on PLS regression to model the differences between classes. For the separation of two classes the PLS algorithm is used

Methods of the "calibration" group allow models to be developed for parameter prediction if the parameter is adequately reflected by the collected data, in this study by the spectral and thermal patterns. Contrary to classification models by which class assignment according to defined properties is performed, the prediction model provides distinct values of the parameter in question. Prediction models focus on the determination of dependent Yvariables for new samples that were characterised by independent X-variables. Based on an established validated X-Y model the Y-variable can be derived from X-variables. Due to this relationship only X-measurements are necessary. This procedure can be advantageous if expensive and time-consuming methods for the determination of Y-variables are replaced

Degradation of organic matter is a natural necessary process in the environment. On the one hand degradability is an inherent property of materials that depends on both chemistry and structure and on the other hand it is mainly influenced by environmental conditions that determine the velocity of this process. The balance between synthesis, transformation and degradation is a criterion of sustainability. Organic matter in mixed waste consists of native biomolecules, modified biomolecules or organic substances that are exclusively based on chemical syntheses. Modification means chemical and physical diversification. Modification of biomolecules, e.g. wood or cellulose is necessary to enhance the stability against microbial attacks. For biological degradation chemical or physical modification represents a barrier to some degree. Wood modification for instance such as acetylation and thermal treatment can increase the lifetime of a product, but pose a problem for degradation [22-24]. Microbial degradation of synthetics hardly takes place. Ageing by oxygen and UV exposure in the forefront can contribute to a certain bioavailability of molecule moieties. Due to the heterogeneity of materials and conditions in old landfills and dumps the turnover and the emissions are hardly predictable. This fact and the long aftercare phase led to the European "multi-barrier" concept that provides the pre-treatment of municipal solid waste prior to landfilling. This procedure ensures extensive degradation within a limited period of time under controlled aerobic conditions. Nevertheless, landfill sites from the past are still a current topic. Investigations of the solid waste focus on the assessment of the current stage, and the potential of future emissions can consequently be derived. Ageing of the landfill also comprises the attenuation of hazardous substances in the leachate that are released due

quadrants are therefore identified as members of the classes.

**3. Application of multivariate data analysis** 

**3.1. Deterioration of organic matter in dumps and landfills** 

by superior methods.

with the dummy variable (e.g. -1/+1) to distinguish the two defined groups.

For data evaluation the heat flow profiles and the temperature resolved curve of the CO2 ion current extracted from the mass spectrum were used. For data evaluation the integrated software PROTEUS was used.

The sample sets that were subjected to multivariate data analysis are mentioned in the text. Depending on the material and the question to be answered different multivariate evaluation methods were applied using The Unscrambler® 9.2 and 10.0 respectively (Camo®).
