Scrutinizing compost properties and their impact on methane oxidation efficiency

Waste Management 31 (2011) 871–883 Contents lists available at ScienceDirect Waste Management journal homepage: www.elsevier.com/locate/wasman Scrutinizing compost properties and their impact on methane oxidation efficiency Marion Huber-Humer ⇑, Johannes Tintner, Katharina Böhm, Peter Lechner BOKU – University of Natural Resources and Life Sciences Vienna, Institute of Waste Management, Muthgasse 107/3rd Floor, A-1190 Vienna, Austria a r t i c l e i n f o a b s t r a c t Article history: Methane emissions from active or closed landfills can be reduced by means of microbial methane oxi- Received 31 May 2010 dation enhanced by properly designed landfill covers and engineered biocovers. Composts produced Accepted 27 September 2010 using different waste materials have already been proven to support methane oxidation, and may rep- Available online 29 October 2010 resent a low-cost alternative to other suitable substrates such as sandy or humic-rich soils, which are frequently not available in sufficient amounts or are too costly. In the present study a data set of 30 different compost materials (different age and input materials) and mixtures, as well as seven soils and mineral substrates were tested to assess methane oxidation rate under similar conditions in a lab- oratory column set-up. Multivariate data analysis (discriminant analysis) was applied to predict the influence of 21 different parameters (chemical, maturation and physical) on methane oxidation rate in a PLS-DA model. The results show that bulk density, total nutrient content (nitrogen and phospho- rus), as well as the quantity and quality (with respect to maturity) of organic matter determined meth- ane oxidation rate in this data set. The model explained 50% of the data variation, indicating how characterisation of oxidation rate by single, even diverse conventional parameters was limited. Thus for the first time, Fourier Transform Infrared (FTIR) spectroscopy was applied to a series of samples to better determine the characteristics of methane-oxidising materials. The initial data obtained in this study appear to be most promising. The prediction of specific methane oxidation rate of a potential bio- cover material from FTIR spectra and multivariate data analyses is a target to be focused on in the future. Ó 2010 Elsevier Ltd. All rights reserved. 1. Introduction and chemical soil conditions (Boeckx and Van Cleemput, 1996; Hanson and Hanson, 1996; Scheutz et al., 2009a). However, the Annual global methane emissions from landfills are estimated majority of these factors are influenced and widely controlled by to be in the range of 500–800 Mt CO2-eq, representing the highest the construction of optimally designed and dimensioned biocovers, source of greenhouse gases within the waste sector (Bogner et al., and by choosing a suitable material for biocover application. 2007). These methane emissions from active or closed sanitary The selection of suitable materials represents one of the key- landfills, waste dumps or old landfill sites can be reduced by means issues in constructing biotic methane oxidation systems. In gen- of methane oxidation enhanced in properly designed landfill cov- eral, well-textured, porous substrates should be used, since they ers and engineered biocovers. The greater the cover area, the lower provide sufficient porosity for gas exchange, facilitating the pene- the specific methane load per m2 (provided that homogeneous tration of oxygen from the atmosphere and methane supply from landfill gas distribution occurs). This normally leads to overall waste, both of which are crucial to microbial methane oxidation higher emission mitigation efficiency, since lower landfill gas processes. Sandy media generally feature good oxidation perfor- fluxes can be more easily contained by the microbially-mediated mance (Whalen et al., 1990; Powelson et al., 2006), being further oxidation process, and the oxygen supply from the atmosphere optimized by amendment and admixture of organic matter or into the cover is more effective. The microbial methane oxidation nutrients (Kightley et al., 1995; De Visscher and Van Cleemput, process is moreover influenced by numerous other factors, such 2003). Considerably higher oxidation rates are reported for engi- as temperature, moisture content, and the prevailing physical neered, manufactured media such as inert expanded clay material (Gebert and Groengroeft, 2006), compost/expanded polystyrene pellet mixture (Powelson et al., 2006), or sand-compost perlite mixture (Philopoulos et al., 2009) used in biofilters, and small ⇑ Corresponding author. Tel.: +43 1 318 99 00; fax: +43 1 318 99 00350. bio-windows, respectively. Artificially designed and engineered E-mail addresses: [email protected] (M. Huber-Humer), [email protected] (J. Tintner), [email protected] (K. Böhm), media can provide more favorable homogeneous, controlled [email protected] (P. Lechner). physical and bio-chemical properties. However, in the construction 0956-053X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.wasman.2010.09.023 872 M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 of methane oxidation systems capping the whole landfill 2. Materials and methods surface (biocovers), or large areas of the latter, such as extended bio-windows, huge amounts of suitable substrates are needed, Over the past 10 years, numerous composts, mechanically and and availability or costs incurred frequently represent limiting biologically treated waste materials (mbt-material), soils and min- factors. eral substrates have been investigated by our institute in various Thus, the recovery of materials for use in biocover construction studies aimed at assessing methane oxidation. The materials have focuses largely on waste products, by-products and residues, been tested under similar conditions in column set-ups. The data including sewage sludge (Borjesson et al., 1998), composts (Humer set has been compiled focusing on routine chemical, physical and and Lechner, 1999, 2001; Barlaz et al., 2004; Wilshusen et al., maturation parameters. Finally, thirty different waste composts 2004b; Bogner et al., 2005; Mor et al., 2006; Stern et al., 2007; (including mbt-materials comparable to compost produced using Tanthachoon et al., 2008; Abichou et al., 2009; Philopoulos et al., municipal solid waste) in addition to seven soil and sandy sub- 2009), compost mixtures (Jugnia et al., 2008; Perdikea et al., strates were investigated. Moreover, FTIR spectroscopy was carried 2008) or even mechanically and biologically treated waste (Einola out as described below on still available samples (n = 16). et al., 2007, 2008; He et al., 2008), also featuring a proven high methane oxidation potential. Moreover, compost materials con- 2.1. Material characterisation tribute not only towards attenuating methane emissions from landfills but also several volatile organic compounds in landfill Materials were analysed for (1) chemical parameters, including gas (Scheutz et al., 2009b). pH-value, conductivity, ammonium, nitrate, nitrite, organic con- To summarize, numerous different compost materials and mix- tent (loss on ignition = LOI), total organic carbon (TOC), total nitro- tures have been investigated for use as potential support medium gen (Ntot), and optionally in some substrates total phosphorous for methane oxidation in recent studies. However, these materials (Ptot) and sulphate; (2) physical parameters, such as moisture con- feature a large variety of properties and thus exhibit a wide range tent (MC), water holding capacity (WHC), bulk density, air-filled of oxidation rates, likely due to the marked variations in their gen- pore volume (PV) at actual MC within the column test set, particle esis, the production process and applied input materials. Moreover, size distribution; and (3) maturity parameters, such as humic acids under unfavourable conditions, unsuitable compost material may (HA) in some materials, and respiration activity over 7 days (RA7), even lead to the production of methane rather than oxidizing it applied particularly to compost and waste materials. (Barlaz et al., 2004; Scheutz et al., 2009b), and may be susceptible Standard parameters were analysed in compliance with the to the formation of exopolymeric substances (EPS) (Streese and Austrian Compost Ordinance (2001) using a DX-120 Ion-chromato- Stegmann, 2003; Wilshusen et al., 2004a) and subsequent clogging graph (Dionex, Sunnyvale, CAÒ), and a spectrophotometer (HIT- of pores. The definition of a suitable support media for methane ACHI Ltd., Tokyo, JapanÒ). pH-value was determined in a 1:10 oxidation purposes is rather complex, with an even greater chal- eluate with de-ionised water. Loss on ignition (LOI) was deter- lenge being posed by characterisation and standard definition of mined by combustion in a muffle furnace at 550 °C. Total carbon heterogeneous materials such as composts from waste materials (TC) and total inorganic carbon (TIC) from which TOC was calcu- (municipal solid waste, bio-waste, yard waste, sewage sludge, etc.). lated, and nitrogen (Ntot) were determined by combustion in a To date, several specific features have proven to be fundamental Vario MaxÒ CNS-analyzer. Sulphate was eluted (50 g fresh sample in demonstrating the effective suitability of compost as support in 200 ml calcium chloride-dihydrate dilution, 1 h shaking) and media for methane oxidation. Critical factors are represented by determined using a DX-120 Ion-chromatograph. Particle size dis- the maturity and stability of organic matter (Huber-Humer et al., tribution was measured according to sieving method of DIN ISO 2009; Scheutz et al., 2009b), in addition to the long-term stability 3310-1 (1992). Respiration activity test was carried out according of the texture, and porosity (Kightley et al., 1995; Huber-Humer to Austrian Standards Institute (2004) using a Sapromat (Sapromat et al., 2008; Yuan et al., 2009), both of which are influenced by E, Voith SulzerÒ, Ravensburg, Germany) and extending the dura- the conditions of organic matter. However, conclusions reported tion of the test from 4 to 7 days in order to minimize the impact to date on the suitability of compost and other methane-oxidising of possible ‘‘lag-phases” by microorganisms. The respiration activ- materials have frequently been derived from single comparative ity test determines the amount of oxygen per g dry matter (mg O2/ investigations, thus being of a mainly intuitive nature and highly g DM) within a time period, used to aerobically degrade organic case-specific. substance, reflecting the basal respiration of a material. Humic acid The purpose of this paper is to outline and discuss the main (HA) analysis was carried out according to Gerzabek et al. (1993) properties and routine parameters which may be of use in defining based on alkaline extraction with 0.1 M sodium pyrophosphate and assessing the suitability of compost materials for biocover con- solution (pH = 10.5) and precipitation with HCl (37%). Based on struction, and in supporting material selection. Based on a data set gravimetric determination humic acid contents were calculated comprising 30 different compost materials tested for methane oxi- from photometrically-measured optical densities and referred to dation capacity under comparable environmental conditions in a organic dry matter (Meissl et al., 2007). The method applied in laboratory column set-up, and applying multivariate data analyses, analysis of air-filled pore volume (PV) at actual MC within the col- a first attempt was made to objectify the characterisation of com- umn-set up (see below) is described in Huber-Humer (2004). post materials and to define the main material parameters capable of influencing the achievement of high oxidation efficiency. More- over, initial investigations were performed using Fourier Trans- 2.2. Determination of methane oxidation capacity form Infrared (FTIR) spectroscopy to characterise compost material as support media in methane oxidation by means of a Oxidation capacity of the different materials was investigated in more comprehensive approach. FTIR spectroscopy has already counter-gradient soil columns with defined air supply at the top been applied previously in waste characterisation (Chen, 2003; and a controlled methane influx at the bottom. In contrast to Smidt et al., 2005; Smidt and Schwanninger, 2005; Smidt and small-dimensioned batch experiments, column-tests allowing a Meissl, 2007; Smidt et al., 2008) and in the prediction of parame- higher mass input are better suited for use in the presence of het- ters involved in respiration activity of bio-waste composts and erogeneous materials such as compost. Experiments were carried municipal solid waste (Meissl et al., 2007, 2008; Böhm et al., out at standard temperature (20 ± 1 °C) in a dark, climate chamber 2010a, b). with gas-tight, transparent plastic columns made of acrylic glass  M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 873 (height 90 cm, inner diameter 19 and 10 cm, respectively) accord- sub-optimal field conditions. This high threshold was set in order ing to Huber-Humer et al. (2009). The materials to be tested were to provide satisfactory material quality for an efficient oxidation introduced up to a height of 60–70 cm on top of a perforated PVC performance, which of course is impacted by varying environmen- plate for a more homogeneous gas distribution and potential water tal factors and cover (design) conditions in the field. drainage, and lightly compacted every 10–20 cm with a hand Initially ANOVA was performed to detect any significant differ- tamper at standardized conditions (18–19 kPa at a moisture con- ences between the two groups for chemical, maturation and phys- tent corresponding to 50–60% of the WHC). Underneath this plate ical variables. Interactions between parameters are not considered pure methane was continuously added at rates of approx. in this analysis. Type I error (a) was set at 0.05. 230 ± 20 l CH4/m2d. Methane input flow was controlled by means To classify materials according to methane oxidation capacity a of needle valves and measured by digital flow measuring devices discriminant analysis (PLS-DA) was performed. The discriminant (Aalborg Mass Flow Meter GFM17) and digital mass flow controller analysis was carried out using the program Unscrambler 9.2Ò (Brooks Mass Flow Controller 5850 TR), respectively. Air inflow (Camo Software AS, http://www.camo.com). The result of a PLS- from the top was set at a methane to oxygen ratio of about 1:2.5 DA model is a classification plot in which all samples are projected (flowing rate about 50–60 ml air/min) regulated by means of nee- along a projection line according to the a priori classification. In the dle valves and measured with rotameters. Prior to entering the loadings plot the measured variables (chemical, maturation and headspace above the substrate, the air was circulated through a physical parameters) are arranged to correspond with the influ- water-filled washing bottle to enhance moisture addition. ence exerted on the predicted classification. Variables farthest Each material was tested in a minimum of duplicate assays. The from the centre indicate a substantial influence on the model. By main test run periods varied between 60 and 90 days, with means of a discriminant analysis it can be checked, how strong is exception of one material (MSWC1) tested only 26 days, and three the influence of a single parameter on methane oxidation rate. materials (MSWC 2, MSWC 3, SSC6) with a longer test run of A contingency table was also created using methane oxidation 187 days. Based on the established influx and outflux rates, and rate of variables on the one hand, and ‘‘optical designation” (mean- methane concentrations present in the input and output of test ing empirical, intuitive evaluation based on author’s experience) of columns (measured using an infra-red methane detector (mea- compost materials on the other. The table comprised the number surement precision: 0.1% v/v), and a Flame-Ionisation-Detector of scores per combination and the percentage obtained. {2-test (microFID, PE Photovac lower detection limit of 0.5 ppmv) at CH4 was performed to test the significance of the table (a = 0.05). The concentrations in the outlet flux <0.2% v/v, respectively) the meth- strength of association was measured by means of Crameŕs V ane oxidation rates were calculated using a methane mass balance (Cramér, 1946). ANOVA and the contingency table were carried (Huber-Humer et al., 2009). The deviation in mass balance of dupli- out using SAS 9.1Ò. By means of a contingency table the influence cate column assays ranged between 1.7% and 3.0%. Results were of one categorical variable (the optical designation based on expressed as relative (%) methane oxidation rates (efficiency) asso- author’s experience) on a second categorical variable (methane ciated to methane loads applied. oxidation rate) can be tested. Moreover, simple excel-based linear, polynomial (square) and 2.3. Fourier Transform Infrared (FT-IR) spectroscopy exponential regression functions were applied to test the correla- tion between different parameters and methane oxidation Fourier Transform Infrared (FT-IR) spectroscopy is a laboratory efficiency. method commonly applied in numerous industrial and research fields, providing in general comprehensive information on the chemistry of waste and compost materials (Smidt and Meissl, 3. Results and discussion 2007). The spectrum gives information on the chemical composi- tion of the sample, rather like a chemical ‘‘fingerprint” (Smith, 3.1. Evaluation and multivariate data analysis of conventional 1999). Infrared spectroscopy has been proven to be an appropriate parameters and their impact on oxidation rate tool for process control and quality control of final composts (Smidt and Meissl, 2007). In Table 1 the materials are described by the methane oxidation In the present study mid infrared spectroscopy (MIR) using the rate obtained in the column set-ups (median value of oxidation attenuated total reflection (ATR) technique was applied to investi- efficiency at steady state), as well as organic matter content gate the fingerprint of available samples of compost, waste and (LOI), total nitrogen, bulk density, water holding capacity (WHC), mineral materials in this study. ATR spectra were collected with electrical conductivity (EC), respiration activity (RA7) and humic a Bruker AlphaÒ in the wavenumber range from 4000 to acids (HA). In Figs. 1 and 2 the conventional chemical, maturation 400 cm 1. The milled sample is directly applied on the ATR reflec- and physical parameters are summarized in boxplots. In spite of tion module containing a diamond crystal and pressed by a pres- the fact that approx. 18 compost-based materials displayed meth- sure applicator. The measured area was of approx. 4 mm2. ane oxidation efficiencies of almost 100%, however, considerably Twenty-four scans per spectrum were collected at a resolution of marked differences in the various parameters of these compost 4 cm 1 and a zero-filling factor of 2 and corrected against ambient substrates were revealed. In general, conductivity (see also de- air as background. Three spectra per sample were averaged for data tailed discussion below), sulphate, nitrate, Ptot, Ntot (TN), organic analysis. The maximum deviation of the three spectra from the matter content (LOI), TOC and WHC tend to be higher in group average spectrum was lower than 5%. ‘‘>95% oxidation efficiency”, while higher RA7 and bulk density val- ues (show also a higher data variability) are rather attributed to 2.4. Statistical and multivariate data analysis group ‘‘<95% oxidation efficiency”. pH, NH4 and humic acid content (HA) show no distinct trend, however, NH4 values are slightly high- The data set was classified in two groups, one comprising all er in group ‘‘<95% oxidation efficiency”, particularly including materials oxidizing at least 95% of the methane supplied and a sec- some outliers with high contents (see discussion below). ond including all other materials. The 95% limit was set following Two compost materials (and one mixture) featured singular re- the testing of materials under optimized moisture content, tem- sults in this data set: a sewage sludge compost (SSC3 – and conse- perature and oxygen supply conditions in a laboratory set-up, quently the mixture of this compost with sand (MIX 3)) showing implicating a marked curtailment of oxidation efficiency under high initial ammonium (1200 ppm DM) and nitrite concentrations 874 M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 Table 1 Median methane oxidation efficiency (at steady state), organic matter content (LOI), maturation and some physical parameters (selected parameters identified in outer ellipse of Fig. 4) of investigated materials (MSWC = municipal solid waste compost; SS = sewage sludge; SSC = sewage sludge compost; LOI = loss on ignition, WHC = water holding capacity; EC = electrical conductivity; RA7 = respiration activity, oDM = organic dry matter; HA = humic acids; na = not analysed); methane load = 230 ± 20 l CH4/m2d. Remarks Abbr. Median LOI Ntot Bulk WHC EC RA7 HA ox-rate density Unit % of load % DM kg/l % DM mS/cm mg O2/g oDM % oDM Well mature MSWC MSWC1 95 37 na 0.94 125 3.3 na na Relative fresh MSWC MSWC2 52 29 na 0.97 110 2.9 9.8 na Mature MSWC MSWC3 100 24 na 1.06 104 3.8 4.5 na Mature MSWC MSWC4 64 24 na 1.04 105 2.3 4.8 na Mature MSWC sieved to <15 mm MSWC5 100 25 0.85 1.02 89 2.4 6.8 6.8 Mature MSWC sieved to particle size <15 mm MSWC6 100 28 1 0.90 90 3.4 3.1 7.3 Quite mature MSWC sieved to <20 mm MSWC7 100 29 0.9 0.84 75 2.9 9.5 7.1 Quite mature MSWC sieved to <20 mm MSWC8 100 29 0.9 0.84 75 2.9 1.8 8.5 Mechanically–biologically treated MSW <20 mm WM-L 95 34 1.4 0.85 83 2.9 15.9 na 15y old MSW sieved <20 mm WM 45 6 0.16 1.50 40 0.6 2.4 na WM mixed with 30%v/v shredded wood and yard waste WM30 60 14 0.22 1.00 50 0.6 11.4 na WM mixed with 50%v/v shredded wood and yard waste WM50 65 18 0.23 0.90 65 0.6 16.9 na Well composted SSC sieved to <20 mm SSC1 100 28 na 0.84 124 0.9 6.9 5.8 Quite mature SSC sieved to <20 mm SSC2 100 28 na 0.90 123 1.0 2.7 6.1 50% w/w SS and 50% w/w wood chips not SSC3 35 30 1.3 0.83 122 1.5 6.2 9.9 well composted, sieved <20 mm 50% w/w SS and 50% w/w wood chips mature,<20 mm SSC4 100 26 1.15 0.90 121 0.7 1.9 6.0 50% w/w SS and 50% w/w wood chips well composted SSC5 100 26 1.1 0.85 109 1.2 1.2 9.0 50% w/w SS and 50% w/w wood chips well composted SSC6 100 30 1.2 0.82 120 2.6 1.7 7.4 50% w/w SS and 50% w/w wood chips well composted SSC7 100 29 1.25 0.83 129 1.2 1.1 11.3 50% w/w SS and 50% w/w wood chips well composted SSC8 100 27 1.2 0.82 109 3.0 1.6 11.7 Mature SSC SSC-O1 100 24 1.3 0.88 70 2.2 1.4 na Mature SSC SSC-N 99 31 1.2 0.76 110 2.1 1.0 na Immature bio-waste compost and wooden yardwaste BIO-COMP 0 45 1.3 0.57 147 1.4 14.5 na Mix 70% w/w SSC1 and 30% w/w silica sand MIX1 85 16 na 0.99 73 0.6 4.8 na Mix 70% w/w SSC2 and 30% w/w silica sand MIX2 75 16 na 1.10 67 0.7 1.9 na Mix 60% w/w SSC3 and 40% w/w SOI-2 MIX3 37 17 0.71 1.03 76 1.0 2.5 8.8 Mix 60% w/w SSC4 and 40% w/w SOI-2 MIX4 90 14 0.51 1.14 79 0.5 0.8 6.6 Mix 60% w/w MSWC5 and 40% w/w SOI-2 MIX5 100 16 0.5 1.20 60 1.6 1.7 6.8 Mix 70% w/w MSWC7 and 30% w/w sand MIX6 100 15 0.49 0.99 60 1.8 3.0 8.4 Mix 70% w/w MSWC and 30% w/w sand MIX7 100 17 0.64 0.99 60 1.7 5.2 7.6 Soil, mineral materials Excavated soil SOI-1 42 7 na 1.39 35 0.9 0.2 2.7 Garden soil SOI-2 98 6 0.23 1.39 39 0.1 0.2 5.3 Sandy, loamy landfill cover soil SOI-MD 43 3.6 0.04 1.26 31 0.3 0.3 na Top-soil from a landfill SOI-F 37 5.2 0.1 1.35 41 0.1 na na Fine, silty sand SA-F 18 1.1 <0.01 1.48 28 0.1 na na Topsoil SOI-Z 10 5.8 0.2 1.24 48 0.1 na na Loamy soil LOAM-Z 3 2.9 <0.01 1.41 37 0.1 na na (130 ppm), giving rise to a median methane oxidation rate of only tion and retention processes for CH4 within the column set-up. 35% (46 l CH4/m2d), and a slightly higher rate by the mixture MIX 3 This compost featured the highest organic matter content (LOI), (37%). Nitrite was only detectable in these materials within the and consequently the highest TOC and water holding capacity investigated data set. NH4 is assumed to be a strong inhibitor of (147% DM), in addition to one of the highest respiration activities methane oxidation at higher levels (Bronson and Mosier, 1994), (14.5 mg O2/g oDM) within the data set. Concentrations of inor- whilst nitrite is toxic for microorganisms at lower levels ganic nitrogen compounds were low. It is assumed that the imma- (Mancinelli, 1995). Further studies (Boeckx and Van Cleemput, ture organic matter and the turnover of the readily degradable 1996; Boeckx et al., 1996; De Visscher et al., 2001) have indicated carbon compounds (indicated by the high RA7) affected the low that the inhibition by ammonium is not always produced as a di- methane oxidation rate. rect result of actual concentrations, but rather of the nitrification The data illustrated in Table 1 and Figs. 1 and 2 indicate the pos- rate or N-turnover. A high nitrogen turnover was detected in this sibility of achieving good methane oxidation performance in di- material, with compost samples excavated from the columns after verse materials with a marked variety of properties, vice versa a test run of 35 days (data provided in Huber-Humer, 2004) indi- revealing how data obtained for comparable substrate parameters cating a clear shift between nitrogen compounds. The NH4 concen- do not categorically imply similar oxidation capacities. Hence, the tration dropped from 1200 ppm DM to 33–44 ppm, while nitrate definition of single – even diverse – parameters would not appear concentration increased from 240 ppm DM to approx. 1830 ppm to be sufficient to provide a comprehensive description of the suit- DM after the experiment. Most probably the nitrification process ability of a biocover or biofilter material. had suppressed methane oxidation in this set-up. The single parameter statistical analysis by means of ANOVA Another material tested, namely a bio-waste compost (BIO- revealed a significant difference between the two groups in data COMP), was distinguished by the median methane oxidation rate obtained for the following parameters with regard to methane oxi- of zero. This compost exhibited no continuous methane oxidation dation rate (a = 0.05): bulk density, moisture content (MC), water over a test run of 62 days. Several peaks in methane removal of holding capacity (WHC), electrical conductivity, total nitrogen about 10% (about 24 l CH4/m2d) were detected however on a spo- (Ntot), and loss on ignition (LOI). The higher the bulk density radic basis, likely attributable to interim and short-term adsorp- was, the lower the methane oxidation revealed. It was vice versa  M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 875 Fig. 1. Boxplots of the median oxidation rate, conductivity (mS/cm), pH, sulfate (ppm DM), ammonium (ppm DM), nitrate (ppm DM), Ptot (% DM), Ntot (=TN, % DM), LOI (% DM), TOC (% DM), RA7 (respiration activity) (mg O2/g DM), HA (humic acids) (% oDM), bulk density (kg/l), MC (% w/w wet matter) and WHC (% DM). for the other parameters MC, WHC, EC, Ntot and LOI. However, in- In order to ascertain the specific influence of various parameters ter-parameter interactions are not taken into consideration in this on oxidation rate and to identify the main variables in a complex, type of analysis. For this purpose more samples would be inferring system (such as methane oxidation behaviour in hetero- necessary. geneous materials), multivariate data analysis was performed. In 876 M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 50 100 100 80 80 airfilled PV 40 25 - 6.3 60 60 >25 30 40 40 20 20 20 SOI-1 SSC6 WM50 SSC5 SSC8 10 0 WM30 0 100 100 100 SA-F 80 80 80 2 - 0.63 60 60 60 6.3 - 2 <0.63 40 40 40 20 20 20 SA-F SA-F 0 0 0 >95% of <95% of >95% of <95% of >95% of <95% of methane methane methane methane methane methane oxidized oxidized oxidized oxidized oxidized oxidized groups groups groups Fig. 2. Boxplots of the air-filled PV (pore volume) (% v/v) determined at effective MC in the column-set up, and the particle size distribution fractions (fraction in mm in % w/w). Fig. 3 the result of the discriminant analysis is shown. The classifi- study-specific materials, thus implying that material criteria were cation of the samples according to the measured methane oxida- often evaluated intuitively. tion rate divides the data set into two classes (x-axis). The In Fig. 4 the loadings plot of the PLS-discriminant analysis illus- predicted values according to the model (based on the other trated in Fig. 3 is shown. Fig. 4 shows how the first principal com- parameters) are presented on the y-axis. The discriminant analysis ponent (PC) discriminated on the basis of physical parameters shown in Figs. 3 and 4 provides a good picture of factors capable of (bulk density) on the one hand, and to organic matter content, influencing methane oxidation efficiency, even though the classifi- and total carbon and nutrient contents, respectively, on the other. cation by the model is not completely satisfactory (R2 = 0.59). This The farer the parameter points are situated from the PC 1-line, the might be attributed to the availability of a limited data set repre- higher is their impact on the methane oxidation rate. The second senting a scarcity of variations, or to the fact that some important PC (PC 2) distinguished organic matter quality, however, producing variables exerting a strong impact on the oxidation capacity were a relatively low impact on model prediction (PC 2 accounts for only not considered in this study. However, typical material parameters 9% of data variation). All other parameters scattering within the in- taken into account in studies performed to date as relevant criteria ner ellipse have obviously no relevant impact on methane oxida- for methane oxidation (e.g. Humer and Lechner, 1999, 2001; tion efficiency in this data set. Unremarkably in view of the Streese and Stegmann, 2003; Mor et al., 2006; Tanthachoon findings obtained in the methane oxidation literature mentioned et al., 2008; Huber-Humer et al., 2009; Philopoulos et al., 2009; above, these data suggest that the physical properties, quantity Scheutz et al., 2009b) were included in the statistical evaluation. and quality of organic matter (such as maturation parameters) play Naturally, the experiences derived from respective studies were a crucial role in determining methane oxidation. mainly based on single comparative investigations of site- or A comparison of data obtained in the discriminant analysis with results by ANOVA clearly reveal how both methods identified bulk density, water holding capacity (WHC), electrical conductivity, to- tal nitrogen (Ntot) and loss on ignition (LOI) as the main critical factors capable of influencing oxidation rate. In this respect, electri- cal conductivity exerted an indirect influence in our data set, and should rather be regarded as a surrogate parameter e.g., for unrav- elling materials with high or low organic matter content (R2 between LOI and conductivity = 0.59). It may have represented an artefact, since the classification group ‘‘<95% oxidation rate” com- prised mainly mineral and soil materials and mixtures (compost with sand or soil), subsequently featuring lower conductivity com- pared to pure compost materials, producing a considerable impact on data variation. Literature reports reveal how conductivity may exert a small influence on the oxidation performance within a wide range, particularly in view of the high tolerance of methano- Fig. 3. Discriminant analysis (n1 = 17, n2 = 20) based on 21 chemical, maturation trophic bacteria to varying salt concentrations, and conductivity and physical variables (corr = 0.59). values, respectively (Figueroa, 1993). Conductivity has often been  M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 877 1 methane oxidation rate Ptot bulk density WHC EC PC2 0 LOI Ntot TOC RA7 humic acids -1 -1 0 1 PC1 Fig. 4. 2D scatter loadings plot of the PLS-discriminant analysis (PC1 accounts for 41%, PC 2 for 9% of data variation), WHC, water holding capacity; EC, electrical conductivity; LOI, loss on ignition; TOC, total organic carbon; Ntot, total nitrogen; and Ptot, total phosphorus. measured along with other factors in biocover and biofilter oDM. Material MSWC7, a municipal solid waste composted for materials (e.g., (Streese and Stegmann, 2003; Mor et al., 2006; 34 weeks (4 weeks in a closed system with forced aeration and then Tanthachoon et al., 2008; Philopoulos et al., 2009); however, little 30 weeks in windrows with periodical turn-over), achieved a med- information is available on the effect of electrical conductivity on ian methane oxidation rate of 100% (=250 l CH4/m2d) at an initial oxidation rate and has not been discussed thoroughly to date. Only RA7 of 9.5 mg O2/g oDM, slightly higher than the guiding value in Gebert et al. (2003) a more detailed discussion on the impact of (<8 mg O2/g oDM) recommended by Huber-Humer et al. (2009). conductivity on oxidation rate in a biofilter filled with crushed ex- However, despite the somewhat higher RA7, another indicator of panded clay can be found, reporting how salt concentrations the stability and maturity of organic matter, humic acid content, resulting in conductivity-values >6 mS/cm lead to a clear decrease reached values (HA = 7.1% oDM) towards the middle of those pre- in methane consumption. This reported value is much higher than sented within this dataset. As the determination of HA content is those values observed in the MSW and SS composts investigated a rather time-consuming and expensive extraction procedure, this here, which are in the same order of magnitude (taking into ac- parameter was not investigated for all samples, and consequently count the different elution ratios and periods applied in the diverse only few data are available (n = 19). The highest HA content studies) as conductivity values reported for different kitchen and (11.7% oDM) was measured in material SSC8, a mature sewage garden waste composts performing well in methane oxidation sludge compost (50% dewatered sewage sludge, and 50% wood (Mor et al., 2006), as well as yard waste compost tested in Streese chips windrow-composted over 2 years) which exhibited a 100% and Stegmann (2003). methane oxidation rate (=220 l CH4/m2d) and also showed a low Other parameters scattered throughout the outer ellipse of the initial RA7 (1.6 mg O2/g oDM). The lowest HA-content (2.7% oDM) loading plot shown in Fig. 4, and producing a greater influence was determined in a soil material (SOI-1) displaying a median on the model, are discussed in more detail below. methane oxidation rate of 42% (=105 l CH4/m2d) and at an initial RA7 of 0.2 mg O2/g oDM. Compared to other studies, HA content 3.1.1. Maturity of organic matter (RA7 and HA content) of these waste composts and materials were rather low. Meissl As assessed by PLS-discriminant analysis, the impact of respira- et al. (2007) reported HA values between 5% and 40% oDM in bio- tion activity on model prediction seemed relatively low, indicating waste composts. According to Lechner and Smidt (2003) composts that this parameter did not strongly influence the different oxida- derived from kitchen and yard waste feature in general consider- tion rates of investigated substrates. Mor et al. (2006) evaluated ably higher HA contents than composts made from sewage sludge the relationship between respiration rate (measured CO2 produc- (10–19% oDM). The lowest HA contents were exhibited by MSW- tion in batch tests at 22 °C) and methane oxidation, detecting no composts, and mechanically–biologically pretreated wastes, significant relation either. It is undeniable that in immature com- respectively, typically not exceeding 12% oDM (Tintner et al., post displaying high respiration activity, and thus a higher O2 2010), corresponding to the data obtained in the present study. demand, decomposing microorganisms may compete for oxygen No correlation was observed between HA content and RA7 within with methanotrophic bacteria, and consequently oxidation rates the tested data set (n = 19, linear R2 = 0.17, and somewhat higher could be limited. However, no significant correlations between as exponential regression R2 = 0.35). According to the loading plot basal respiration and oxidation rate have been reported in litera- (Fig. 4) HA content seems to have a slightly higher impact on oxida- ture to date or in the present study (R2 = 0.19). As can be seen in tion rate than RA7. To the authors’ knowledge, HA-content and its Fig. 5 several compost materials performed well in methane oxida- relationship with methane oxidation capacity have not been re- tion despite their higher respiration activity. Material WM-L, a mu- ported in literature to date; thus, this parameter cannot be dis- nicipal solid waste treated both mechanically and subsequently cussed in a comprehensive manner. Furthermore, since the over a 4 week period aerobically–biologically and then extensively analytical method used to determine HA content is particularly windrow-composted (without forced aeration) for another nine time-consuming and expensive, this parameter is not suited as a months, exhibited a median methane oxidation rate of 95% routine method for practical use. However, by applying FTIR- (=210 l CH4/m2d) at a relatively high initial RA7 of 15.9 mg O2/g spectroscopy in association with multivariate data analyses, in 878 M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 100 18 17 90 16 15 80 ox-rate % of load 14 median ox-rate (% of CH 4 load) RA7 13 70 RA7 (mg O2 / g oDM) 12 60 11 10 50 9 suggested guide value 8 40 7 6 30 5 4 20 3 10 2 1 0 0 M 5 SS 1 SO N SS 2 SS 4 M 6 SS 5 SS 6 SS X 7 M 4 S 7 C 8 M 1 W X2 BI S 3 M 1 SO D -C 3 P W -2 M 1 SW 3 SW 5 SW 6 SW 7 SS 8 SW 0 SW 0 W 4 2 SO WM M L C C C C C C SS SC -O IX IX IX IX IX O SC I- C M C M C M C C C C M M5 M M3 - M - I-M I M C I I SW O M M Fig. 5. Median methane oxidation rates and corresponding respiration activity of tested materials; methane load = 230 ± 20 l CH4/m2d; the horizontal line represents suggested guide value for respiration activity within 7 days for compost materials (Huber-Humer et al., 2009). particular partial least squares regression, HA-content can be pre- to mineralization of readily degradable carbon compounds, as dicted from the compost spectra as already described in literature TOC obtained for the excavated material after the test (data in (Meissl et al., 2007; Böhm, 2009). Huber-Humer, 2004) did not confirm this trend. The change in res- The majority of suitable materials featuring >95% oxidation rate piration activity was more likely due to nitrifying activity during exhibited a lower RA7 than the suggested guiding value (Fig. 5). the experiment: it is assumed that the higher initial RA7 value Nevertheless, as discussed above, in some cases the negative influ- was mainly caused by nitrification of the high initial ammonium ence produced by a high RA7 may be outcompeted by other factors concentrations in the material taking place in the Sapromat, rather or even by interactions between these influencing parameters. For than by readily degradable carbon substances. The latter findings example the materials WM-L and MSWC7 (discussed above) were indicate that RA7 (oxygen consumption) reflects diverse biodegra- both characterized by a low bulk density (0.85 and 0.84 kg/l, dation processes in a material capable of exerting a potential im- respectively), which is positive for the gas permeability and hence pact on oxidation rate, and that RA7 does not only describe the for the oxygen supply. In this study it was invariably tried to pro- carbon turnover processes. Thus, interpretation of RA7 data is vide optimal conditions through provision of a hyperstoichiometric rather complex and cannot be directly compared to data obtained amount of oxygen for the methane oxidation process at a ratio from batch tests measuring CO2 formation, as often applied in pre- of 1:2.5 (methane:oxygen), although under steady state condi- vious studies (Mor et al., 2006; Einola et al., 2008). tions a ratio of 1:2 is sufficient for complete methane turnover (Mancinelli, 1995). Hence, it could be speculated that for both 3.1.2. Quantity of organic matter (LOI), total nutrient contents, bulk endogenous, still actively decomposing bacteria, and methane-oxi- density and water holding capacity (WHC) dising bacteria the oxygen supply in the compost material was suf- As indicated in Fig. 4 the organic matter content produces a ficient, leading to no oxygen competition. Thus, even in the clear impact on oxidation performance. Spatial grouping with the presence of an increased RA7, high oxidation rates can be achieved parameters WHC and total nutrients in the loading plot lead us at optimized oxygen supply. However, since the above conditions to assume that interactions occur between these parameters. are not prevailing in the field, and to achieve conservative results Literature reports provide extensive proof of the effect pro- with regard to methane oxidation already shortly after biocover duced by organic matter content on physical parameters in soil application, care should be taken to use mature compost materials (Scheffer and Schachtschabel, 1984), and it determines the physi- with an initially low respiration activity. cal properties in bio-waste materials (Kraft, 2001). In the data set On the other hand, few compost materials with sufficiently low examined organic matter content (LOI) had a clear impact on bulk RA7 did not perform well with regard to methane oxidation (see density (R2 = 0.84 significant at a = 0.05, shown in Fig. 6), and cor- Fig. 5). In these materials several factors and parameters achieved related positively with water holding capacity (WHC: linear sub-optimal values, thus outcompeting a suitable RA7. For in- R2 = 0.77, exponential R2 = 0.89). No distinct influence of organic stance, material WM, a 15 year old municipal solid waste that matter was found on air-filled pore volume (R2 = 0.29), which is was excavated, mechanically prepared and biologically treated also affected by additional variables, particularly prevailing mois- over a period of several weeks, showed a low RA7 (2.4 mg O2/g ture conditions (MC). Both MC and air-filled PV scatter throughout oDM), but also low organic matter and nutrient content, as well the inner ellipse of the loading plot illustrated in Fig. 4, producing as a comparable high bulk density (1.5 kg/l), all factors capable of no relevant influence on oxidation efficiency in this data set (in this potentially negative impact on oxidation efficiency. context it must be considered that MC and the compaction rate Material SSC3 (already discussed above) displayed low oxida- (which again have an impact on the air-filled pore volume) were tion efficiency due to unfavorable nitrogen turnover processes tried to keep as similar as possible in each column-set up, which within the test columns, negatively impacting oxidation perfor- may hamper an evident impact on the oxidation efficiency). mance. However, respiration activity in SSC3 also clearly dropped Applied simple linear statistic showed no direct relation (linear during the test run (6.2–2.7 mg O2/g oDM) and cannot be assigned R2 = 0.14) between LOI and oxidation efficiency, polynomial  M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 879 50 45 organic matter content (LOI) (%DM) 40 y = -43.258x + 64.902 35 R2 = 0.8411 30 25 20 15 10 5 0 0.4 0.6 0.8 1 1.2 1.4 1.6 bulk density (kg/l) Fig. 6. Correlation between organic matter content (LOI) and bulk density; R2 significant at a level of significance a = 0.05. 100 1.6 90 1.4 80 median ox-rate (% of CH 4 load) 1.2 70 bulk density (kg/l) 1 60 50 impact by high NH4, 0.8 NO2 or RA7 40 0.6 30 0.4 20 ox-rate % of load bulk density kg/l 0.2 10 0 0 BIO-COMP SSC1 SSC2 SSC4 SSC5 SSC6 SSC7 SSC8 SSC-O1 SOI-2 SOI-1 SSC3 SSC-N SOI-MD SA-F SOI-F SOI-Z MIX5 MIX6 MIX7 MIX4 MIX1 MIX2 MIX3 LOAM-Z MSWC3 MSWC5 MSWC6 MSWC7 MSWC8 MSWC1 WM50 MSWC4 WM30 MSWC2 WM-L WM Fig. 7. Median methane oxidation rate and corresponding bulk density of tested materials; in the grey ellipse materials are marked where other unsuitable parameters like high NH4, NO2 or RA7 affected the oxidation rate more strongly than bulk density. (square) function expressed a somewhat better correlation single parameters, thus rendering the general description of suit- (R2 = 0.54). able biocover materials extremely complex. The values obtained for bulk densities varied between 1.5 kg/l Moreover, it is an acknowledged fact that organic matter con- in material WM (old MSW sieved to <20 mm) and fine, silty sand tent largely affects the nutrient status in a substrate. Linear regres- (SA-F), and 0.82 kg/l in sewage sludge composts mixed with wood sion analysis performed between organic matter content (LOI) and chips (SSC6 and SSC8) in the test-sets where all materials were Ntot showed a R2 = 0.88, and between LOI and Ptot R2 = 0.65, packed into the columns in a standardised way at the same com- respectively, in the data set examined. Although the influence of paction rate (18–19 kPA). The material-specific relation between inorganic nitrogen, particularly NH4, on oxidation rate is discussed oxidation efficiency and corresponding bulk density is shown in in detail in the literature (Boeckx et al., 1996; De Visscher et al., Fig. 7. 2001; Bodelier and Laanbroek, 2004; Scheutz et al., 2009a), only As can be seen, three materials examined (marked with the grey a few studies have addressed the impact of Ntot (Wilshusen ellipse) do not follow the slight trend in the graph. In these mate- et al., 2004b; Mor et al., 2006; Huber-Humer et al., 2009). rials other parameters exerted a higher impact on oxidation rate. In Methanotrophic bacteria have a relatively high nitrogen de- SSC3 and MIX 3 NH4 and nitrite concentrations were high, in mate- mand (Scheutz et al., 2009a). In a landfill cover inorganic nitrogen rial BIO-COMP it was assumed that the high RA7 was implicated. compounds can be biologically and chemically transformed, con- On the other hand, soil SOI-2 with higher bulk density featured a sumed by plants and microorganisms or leached out. Hence, in good oxidation rate (98% = 245 l CH4/m2d). These findings once landfill environments with a high ratio of methane (=carbon again underlined the marked interferences occurring between source) to nitrogen, limitation of inorganic nitrogen may occur 880 M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 (Bodelier and Laanbroek, 2004). Thus, for field application it is specific surface area increases with higher organic matter content important that cover materials provide long-term nitrogen supply in a substrate. According to Scheffer and Schachtschabel (1984) to support methane oxidation. This may be achieved by the release sand provides a SSA <0.1 m2/g, clay 5–400 m2/g, soils usually of nutrients from slowly mineralising organic matter, and organic 5–500 m2/g, and humified organic matter 800–1000 m2/g. Thus, nitrogen, respectively. However, usually only approx. 5–15% of mature compost materials afford extensive specific surface areas, the mainly organically bound nitrogen pool is plant-available and since microbial activity generally occurs on the surfaces of (and assumed to be also available for microorganisms) in the first the particles, use of these materials may potentially enhance years after application (Amlinger et al., 2003). The C/N ratio of the methane oxidation. Moreover, SSA may be of use providing a more material, which shall be low, seems to be a determining criteria for comprehensive assessment of suitability, indicating an impact on the availability of nitrogen for methanotrophic microorganisms in physical (porosity, mass transport), chemical (nutrient supply, a cover material (Bodelier and Laanbroek, 2004; Scheutz et al., cation exchange capacity), and biological issues (settlement area 2009a). In the data set studied here Ntot ranged between <0.01% for microbes). However, the measurement of SSA is quite time- DM (detection limit, in the fine sand SA-F and the loamy soil consuming and sophisticated (Arnepalli et al., 2008), and practica- LOAM-Z) and 1.4% DM (material WM-L). C/N-ratios ranged be- ble routine methods are currently not available. Future research tween 9.5 (SOI-Z) and 42 (WM50 as an outlier) with a median C/ must be undertaken to focus on these issues. N-value of 14. Ptot varied between 0.16% DM in a soil (SOI-2) A simple optical designation (intuitive material evaluation exhibiting good oxidation rate and 3.06% DM (SSC 7) in sewage based on authors’ experiences) was also applied to the materials sludge compost. In general, composts derived from sewage sludge investigated in this data set. The contingency table of the two vari- show clearly higher Ntot and Ptot amounts compared to MSW ables methane oxidation rate and optical designation of compost is composts, due to their formation process (accumulation of nutri- given in Table 2. The hypothesis of no dependency of the v2-test ents in sludge during waste water treatment). However, in all com- was rejected (a = 0.05), implying that the use of optical designation post-based materials the nutrient status was regarded as being driven by intuitive classification of mature compost should provide sufficient to achieve good methane oxidation rate. a proper indication of the suitability of a material and the expected Last but not least, probably one of the most important parame- methane oxidation rate. The strength of this association is rela- ters provided by organic matter in a material and capable of tively low, however, with value obtained for Crameŕs V corre- enhancing methane oxidation is constituted by internal specific sponding to 0.72. surface area (SSA), i.e., the surface area to volume or mass ratio. In biofilter operation SSA is regarded as a determining criterion 3.2. Preliminary approach to determine methane oxidation rate and underlying mass exchange and biological reaction (Amanullah impacting material properties by FTIR-spectroscopy et al., 1999; Streese and Stegmann, 2003). Higher specific surface areas clearly improve the removal of diverse compounds from To demonstrate the different spectral characteristics infrared the gas flow introduced (Amanullah et al., 1999). However, in liter- spectra of different composts and sand are illustrated in Fig. 8. Sev- ature reporting on methane oxidation and biocover performance, eral ‘‘indicator” bands are marked in the spectra. The aliphatic this parameter has been rather neglected to date, and was indeed methylene bands at 2920 and 2850 cm 1 represent the skeleton not directly investigated in our data set. It is a well-known fact that of many biomolecules. The amide II band at approx. 1540 cm 1, the (aromatic) amine band at 1320 cm 1, the band at 1260– 1240 cm 1 (C–O vibration of carboxylic acids and C–N of amide Table 2 III) can be assigned to readily degradable organic matter. The band Contingency table of the variables methane oxidation rate (grouped according to at 1510 cm 1 represents lignin. The band at 1640 cm 1 cannot be discriminant analysis) and optical designation of mature compost. unequivocally assigned due to overlapping of aromatic C@C vibra- >95 % methane oxidation <95 % methane oxidation tions, functional group vibrations such as C@O (carboxylates and Mature compost 18 (49 %) 0 (0 %) amides), C@C (alkenes) and OH from water, all contributing to this Other material 6 (16 %) 13 (35 %) strong band. Bonded and non-bonded hydroxyl groups and water (O–H stretch) are reflected by the band at about 3400 cm 1. 0.18 1430 0.16 1030 3400 875 0.14 3620 0.12 3690 2920 2850 1780 1640 absorbance 0.1 2520 mature MSW- 0.08 compost 0.06 mature SS- 0.04 compost fresh biowaste 0.02 compost 1510 sand 1540 1320 1260 0 3900 3400 2900 2400 1900 1400 900 400 wavenumbers (cm -1) Fig. 8. FTIR spectra of different biocover materials (mature MSW-compost MSWC8, mature SS-compost SSC-O1, fresh bio-waste compost BIO-COMP, and sand SA-F) providing different methane oxidation capacities.  M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 881 Furthermore, inorganic components of clay minerals (1030 cm 1) compounds like NO3, NH4 and NO2 showed no distinct influence and carbonates (2520, 1780, 1430 and 875 cm 1) can be found in on the oxidation efficiency, although, some compost materials with all samples. The bands at 3690 and 3620 cm 1 can be attributed high NH4 and NO2 contents expressed clearly low oxidation effi- to kaolin. It is clearly visible that the band heights of the different ciency. Since it is well known from the literature that high NH4 materials in Fig. 8 differ considerably. Due to this fact different contents can inhibit methane oxidation, these nitrogen parameters properties of the materials can be indicated by the spectral charac- should be still kept in view for material selection. teristics. Various authors have demonstrated that several chemical However, no statistically significant correlations were obtained parameters such as TOC, C, N, available N (Malley et al., 2005; between these single parameters and oxidation efficiency in com- Michel et al., 2006; Huang et al., 2007), respiration activity (Meissl post-based materials using linear, polynomial and exponential et al., 2007) and humic acids (Meissl et al., 2007, 2008) of composts regression functions. Even with multivariate data analysis only and properties such as plant suppression (Michel et al., 2006) and 50% of the data variation was explainable in this study. Thus, the phytotoxicity (Ludwig et al., 2006) can be defined using FTIR spec- classification of single conventional parameters capable of affect- troscopy and multivariate data analysis. Thus, crucial parameters ing suitability alone seems to be not distinctive. In this regard, determining methane oxidation capacities can be derived from the interaction of organic matter with other criteria may constitute FTIR spectra. As already shown in Böhm et al. (2010b) conventional a key-issue. For instance, organic matter produces an impact on parameters of composts like Ntot, TOC, LOI, as well as quality physical properties, particularly bulk density and water holding parameters HA and RA can be reliably predicted from the FTIR- capacity, and also on chemical parameters. These parameters inter- spectra. act making their single effect on oxidation efficiency and are diffi- The methane oxidation rates of the different biocover materials cult to discern. Furthermore, a higher organic matter content illustrated in Fig. 8 are summarized in Table 1. Since for the param- improves the specific surface area which seems not only to be eter ‘‘methane oxidation efficiency” the development of a proven implicated in mass transfer but also an essential factor in the model is under way, the prediction of the methane oxidation rate establishment of methanotrophs and their reactivity volume. How- using FTIR spectroscopy and multivariate data analyses is still a ever, the determination of this parameter is very sophisticated, and target to be reached. these effects were not investigated, being beyond the scope of the For this calibration more samples have to be investigated pro- present study. viding the methane oxidation capacities together with the infrared Composts are heterogeneous materials with variable genesis spectra. Currently only 16 samples are available; thus, no reliable and a very specific variety of different properties. Obviously, their correlation data can be provided. interactive and highly complex effects on methane oxidation can- not be adequately defined by means of conventional parameters to date. Conclusions reached as to their suitability have mainly been 4. Conclusions and outlook based on single comparative investigations, frequently of an intui- tive and highly study-specific nature. Moreover, compost proper- Waste compost materials or compost mixtures are often consid- ties may vary over time due to further mineralization and ered for use in biocover, biofilter or bio-window construction, since degradation processes. In general, their hardly predictable proper- their suitability to enhance methane oxidation processes has al- ties may limit practical application to date. ready been proven, and in practice, other proper materials, such The use of FTIR spectroscopy may contribute towards providing as sandy or humic-rich soils, are often not available in sufficient a more comprehensive characterisation of compost substrates, as amounts, or are too costly. Furthermore, options should be made well as other materials, and a better evaluation of substrates for available with regard to compost materials not suited for use in methane oxidation purposes. In this study a preliminary attempt agriculture. These materials are often produced directly on or close has been made to test the applicability of this method in assessing to a landfill, thus minimizing transportation distances. parameters affecting methane oxidation performance. Although to Previous experiences and findings in the literature have shown date only a small data pool has been tested, the data obtained indi- that the maturity of a compost substrate, indicating the stability of cate – in accordance with literature results in which the use of FTIR organic matter, and physical properties such as bulk density, and spectroscopy by means of regression models has been successfully texture, may affect the suitability of materials to facilitate methane applied to predict other waste properties and features – that this oxidation in biocovers. The results from this study indicate that approach may represent a promising tool for the characterisation bulk density, the organic matter (LOI), TOC and nutrient (Ptot and selection of methane oxidation substrates in the future. Stud- and Ntot) content, water holding capacity and electrical conductiv- ies undertaken in the future should focus on the creation of a reli- ity had an relevant impact on the oxidation efficiency of the inves- able, sufficiently large data pool for use in specific prediction tigated materials, and may be helpful pre-screening parameters for model generation. However, in the meantime we have to rely compost material selection. Respiration activity and humic acid mainly on time-consuming column tests and/or batch tests to as- content, both describing the maturity of composts, also seem to sess methane oxidation capacity and material suitability. Further- be essential, although their impact could not definitely be attrib- more, maturation parameters in compost materials are particularly uted to the materials’ oxidation efficiencies. The parameter RA7 re- difficult to analyze, their determination is time-consuming, inter- flects the total oxygen demand of a substrate, including not only national standard protocols for a consistent measurement proce- the carbon respiration but also the oxygen demand for further pro- dure are not yet available, and the comparison of these data and cesses, e.g., nitrogen turnover. Thus, this parameter can strongly material suitability is virtually impossible on an international level. interfere with other material properties, as well as environmental In this regard therefore, the new approach presented here may be factors such as the overall oxygen supply to the cover system. particularly promising. The conventional determination of the humic acid content is quite laborious, therefore this parameter may not become a routine cri- Acknowledgements terion for material selection in practice, as long as it can not be re- placed by other, more practicable methods like the prediction via The authors would like to thank Reinhold Ottner for the FTIR spectra. measurement of the FTIR spectra. Moreover, we thank Oliver The texture (in this study investigated as particle size distribu- Gamperling for providing data from his Master thesis to be tion and air-filled pore volume at actual MC), and nitrogen included in the data set. 882 M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 References Huang, G., Han, L., Liu, X., 2007. Rapid estimation of the composition of animal manure compost by near infrared reflectance spectroscopy. J. Near Infrared Spec. 15, 387–394. Abichou, T., Mahieu, K., Yuan, L., Chanton, J., Hater, G., 2009. Effects of compost Huber-Humer, M., 2004. Abatement of landfill methane emissions by microbial biocovers on gas flow and methane oxidation in a landfill cover. Waste Manage. oxidation in biocovers made of compost. University of Natural Resources and 29, 1595–1601. Applied Life Sciences Vienna, Vienna. Amanullah, M., Farooq, S., Viswanathan, S., 1999. Modeling and simulation of a Huber-Humer, M., Gebert, J., Hilger, H., 2008. Biotic systems to mitigate landfill biofilter. Ind. Eng. Chem. Res. 38, 2765–2774. methane emissions. Waste Manage. Res. 26, 33–46. Amlinger, F., Götz, B., Dreher, P., Geszti, J., Weissteiner, Ch., 2003. Nitrogen in bio- Huber-Humer, M., Röder, S., Lechner, P., 2009. Approaches to assess biocover waste and yard waste compost: dynamics of mobilisation and availability – a performance on landfills. Waste Manage. 29, 2092–2104. review. European Journal of Soil Biology 39, 107–116. Humer, M., Lechner, P., 1999. Alternative approach to the elimination of greenhouse Arnepalli, D.N., Shanthakumar, S., Hanumantha Rao, B., Singh, D.N., 2008. gases from old landfills. Waste Manage, Res. 17, 443–452. Comparison of methods for determining specific-surface area of fine-grained Humer, M., Lechner, P. 2001. Design of a Landfill Cover Layer to Enhance Methane soils. Geotech. Geol. Eng. 26, 121–132. Oxidation; Results From a Two Year Field Investigation. ‘‘SARDINIA 2001 – Austrian Standards Institute, 2004. OENORM S2027-1, Stability parameters Eighth International Waste Management and Landfill Symposium‘‘, Cagliari, describing the biological reactivity of mechanically biologically pretreated Sardinia, Italy, CISA. residual wastes – part 1: respiration activity (AT4). Jugnia, L.B., Cabral, A.R., Greer, C.W., 2008. Biotic methane oxidation within an Barlaz, M.A., Green, R.B., Chanton, J.P., Goldsmith, C.D., Hater, G.R., 2004. Evaluation instrumented experimental landfill cover. Ecol. Eng. 33, 102–109. of a biologically active cover for mitigation of landfill gas emissions. Environ. Kightley, D., Nedwell, D.B., Cooper, M., 1995. Capacity for methane oxidation in Sci. Technol. 38, 4891–4899. landfill cover soils measured in laboratory-scale soil microcosms. Appl. Eviron. Bodelier, P.L.E., Laanbroek, H.J., 2004. Nitrogen as a regulatory factor of methane Microb. 61, 592–601. oxidation in soils and sediments. FEMS Microbiol. Ecol. 47, 265–277. Kraft, E. 2001. Charakterisierung von Bioabfällen – Voraussetzung für die Boeckx, P., Van Cleemput, O., 1996. Methane oxidation in a neutral landfill cover Optimierung des Kompostprozesses. Schriftenreihe Manuskripte zur soil: influence of moisture content, temperature, nitrogen-turnover. J. Environ. Abfallwirtschaft. W. Bidlingmaier. Berlin, Germany, Rhombos Verlag. Band 5. Qual. 25, 178–183. Lechner, P., Smidt, E. 2003. Humic Acids, a parameter for compost quality – a new Boeckx, P., Van Cleemput, O., Villaralvo, I., 1996. Methane emission from a landfill approach for a sustainable soil. ‘‘SARDINIA 2003” – Ninth International Waste and the methane oxidising capacity of its covering soil. Soil Biol. Biochem. 28, Management and Landfill Symposium, S. Margherita di Pula, Cagliari, Italy, 1397–1405. CISA. Bogner, J., Abdelrafie Ahmed, M., Diaz, C., Faaij, A., Gao, Q., Hashimoto, S., Ludwig, B., Schmilewski, G., Terhoeven-Urselmans, T., 2006. Use of near infrared Mareckova, K., Pipatti, R., Zhang, T. (2007) Waste management – contribution spectroscopy to predict chemical parameters and phytotoxicity of peats and of working group III to the fourth assessment report of the intergovernmental growing media. Sci. Hortic-Amsterdam 109, 86–91. panel on climate change. Climate Change, Mitigation. Malley, D.F., McClure, C., Martin, P.D., Buckley, K., McCaughey, W.P., 2005. Bogner, J., Spokas, K., Chanton, J., Powelson, D., Fleiger, J., Abichou, T. 2005. Modeling Compositional analysis of cattle manure during composting using a field- landfill methane emissions from biocovers: a combined theoretical – empirical portable near-infrared spectrometer. Commun. Soil Sci. Plan 36, 455–475. approach. ‘‘SARDINIA 2005 – Tenth International Waste Management and Mancinelli, R.L., 1995. The regulation of methane oxidation in soil. Annu. Rev. Landfill Symposium‘‘, Cagliari, Sardinia, Italy, CISA. Microbiol. 49, 581–605. Böhm, K., 2009. Compost quality determination using infrared spectroscopy and Meissl, K., Smidt, E., Schwanninger, M., 2007. Prediction of humic acid content and multivariate data analysis. University of Natural Resources and Applied Life respiration activity of biogenic waste by means of Fourier transform infrared Sciences Vienna, Vienna. (FTIR) spectra and partial least squares regression (PLS-R) models. Talanta 72, Böhm, K., Smidt, E., Binner, E., Schwanninger, M., Tintner, J., Lechner, P., 2010a. 791–799. Determination of MBT-waste reactivity – an infrared spectroscopic and Meissl, K., Smidt, E., Schwanninger, M., Tintner, J., 2008. Determination of humic multivariate statistical approach to identify and avoid failures of biological acids content in composts by means of near – and mid-infrared spectroscopy tests. Waste Manage. 30, 583–590. and partial least squares regression models. Appl. Spectrosc. 62, 873–880. Böhm, K., Smidt, E., Binner, E., Tintner J., Lechner P., 2010b. Prediction of various Michel, K., Bruns, C., Terhoeven-Urselmans, T., Kleikamp, B., Ludwig, B., 2006. quality parameters of biowaste composts by means of fourier transform Determination of chemical and biological properties of composts using near infrared (FTIR) spectroscopy, Book of abstracts of the 7th International ORBIT infrared spectroscopy. J. Near Infrared Spec. 14, 251–259. 2010 Conference, Heraklion, Crete, June 29 – july 3 2010, p. 104. Mor, S., De Visscher, A., Ravindra, K., Dahiya, R.P., Chandra, A., Van Cleemput, O., Borjesson, G., Sundh, I., Tunlid, A., Svensson, B.H., 1998. Methane oxidation in 2006. Induction of enhanced methane oxidation in compost: Temperature and landfill cover soils, as revealed by potential oxidation measurements and moisture response. Waste Manage. 26, 381–388. phospholipid fatty acid analyses. Soil Biol. Biochem. 30, 1423–1433. Perdikea, K., Mehrotra, A.K., Hettiaratchi, J.P.A., 2008. Study of thin biocovers (TBC) Bronson, K.F., Mosier, A.R., 1994. Suppression of methane oxidation in aerobic soil for oxidizing uncaptured methane emissions in bioreactor landfills. Waste by nitrogen fertilizers, nitrification inhibitors, and urease inhibitors. Biol. Fert. Manage. 28, 1364–1374. Soils 17, 263–268. Philopoulos, A., Ruck, J., McCartney, D., Felske, C., 2009. A laboratory-scale Chen, Y., 2003. Nuclear magnetic resonance, infra-red and pyrolysis: application of comparison of compost and sand – compost – perlite as methane-oxidizing spectroscopic methodologies to maturity determination of composts. Compost biofilter media. Waste Manage. Res. 27, 138–146. Sci. Util. 11, 152–168. Powelson, D.K., Chanton, J., Abichou, T., Morales, J., 2006. Methane oxidation in Cramér, H. 1946. Mathematical methods of statistics, Princeton University Press. water-spreading and compost biofilters. Waste Manage. Res. 24, 528–536. De Visscher, A., Schippers, M., Van Cleemput, O., 2001. Short-term kinetic response Scheffer, F., Schachtschabel, P., 1984. Lehrbuch der Bodenkunde. Enke-Verlag, of enhanced methane oxidation in landfill cover soils to environmental factors. Stuttgart, Germany. Biol. Fert. Soils 33, 231–237. Scheutz, C., Kjeldsen, P., Bogner, J.E., De Visscher, A., Gebert, J., Hilger, H.A., Huber- De Visscher, A., Van Cleemput, O., 2003. Simulation model for gas diffusion and Humer, M., Spokas, K., 2009a. Microbial methane oxidation processes and methane oxidation in landfill cover soils. Waste Manage. 23, 581–591. technologies for mitigation of landfill gas emissions. Waste Manage. Res. 27, DIN ISO 3310-1 1992. Test sieves; technical requirements and testing; test sieves of 409–455. metal wire cloth. Scheutz, C., Pedersen, G.B., Costa, G., Kjeldsen, P., 2009b. Biodegradation of methane Einola, J.K.M., Kettunen, R.H., Rintala, J.A., 2007. Responses of methane oxidation to and halocarbons in simulated landfill biocover systems containing compost temperature and water content in cover soil of a boreal landfill. Soil Biol. materials. J. Environ. Qual. 38, 1363–1371. Biochem. 39, 1156–1164. Smidt, E., Eckhardt, K.U., Lechner, P., Schulten, H.R., Leinweber, P., 2005. Einola, J.K.M., Sormunen, K.M., Rintala, J.A., 2008. Methane oxidation in a boreal Characterization of different decomposition stages of biowaste using FT-IR climate in an experimental landfill cover composed from mechanically– spectroscopy and pyrolysis-field ionization mass spectrometry. Biodegradation biologically treated waste. Sci. Total Environ. 407, 67–83. 16, 67–79. Figueroa, R.A. 1993. Methane oxidation in landfill top soils. ‘‘SARDINIA 1993” – Smidt, E., Meissl, K., 2007. The applicability of Fourier transform infrared (FT-IR) Fourth International Waste Management and Landfill Symposium, Cagliari, spectroscopy in waste management. Waste manage. 27, 268–276. Sardinia, Italy, CISA. Smidt, E., Meissl, K., Schwanninger, M., Lechner, P., 2008. Classification of waste Gebert, J., Groengroeft, A., 2006. Passive landfill gas emission - Influence of materials using Fourier transform infrared spectroscopy and soft independent atmospheric pressure and implications for the operation of methane-oxidising modeling of class analogy. Waste Manage. 28, 1699–1710. biofilters. Waste Manage. 26, 245–251. Smidt, E., Schwanninger, M., 2005. Characterization of waste materials using FT-IR Gebert, J., Groengroeft, A., Miehlich, G., 2003. Kinetics of microbial landfill methane spectroscopy – process monitoring and quality assessment. Spectrosc. Lett. 38, oxidation in biofilters. Waste Manage. 23, 609–619. 247–270. Gerzabek, M.H., Danneberg, O., Kandeler, E. 1993. Bestimmung des Smith, B. 1999. Infrared Spectral Interpretation. London, New York, Washington DC, Humifizierungsgrades. Bodenbiologische Arbeitsmethoden. F. Schinner, R. CRC Press, Boca Raton. Öhlinger, E. Kandeler and R. Margesin. Wien, Springer Verlag, pp. 107–109. Stern, J.C., Chanton, J., Abichou, T., Powelson, D., Yuan, L., Escoriza, S., Bogner, J., Hanson, R.S., Hanson, T.E., 1996. Methanotrophic bacteria. Microbiol. Rev. 60, 439– 2007. Use of a biologically active cover to reduce landfill methane emissions 471. and enhance methane oxidation. Waste Manage. 27, 1248–1258. He, R., Ruan, A., Jiang, C., Shen, D.s., 2008. Responses of oxidation rate and microbial Streese, J., Stegmann, R., 2003. Microbial oxidation of methane from old landfills in communities to methane in simulated landfill cover soil microcosms. Bioresour. biofilters. Waste Manage. 23, 573–580. Technol. 99, 7192–7199.  M. Huber-Humer et al. / Waste Management 31 (2011) 871–883 883 Tanthachoon, N., Chiemchaisri, C., Chiemchaisri, W., Tudsri, S., Kumar, S., 2008. Wilshusen, J.H., Hettiaratchi, J.P.A., Stein, V.B., 2004b. Long-term behavior of Methane oxidation in compost-based landfill cover with vegetation during wet passively aerated compost methanotrophic biofilter columns. Waste Manage. and dry conditions in the tropics. J. Air Waste Manage. 58, 603–612. 24, 643–653. Tintner, J., Smidt, E., Böhm, K., Binner, E., 2010, submitted, Investigations of Yuan, L., Abichou, T., Chanton, J., Powelson, D.K., De Visscher, A., 2009. Long-term biological processes in Austrian MBT plants. Waste Manage. numerical simulation of methane transport and oxidation in compost biofilter. Whalen, S.C., Reeburgh, W.S., Sandbeck, K.A., 1990. Rapid methane oxidation in a Practice periodical of hazardous, toxic, and radioactive. Waste Manage. 13, 196– landfill cover soil. Appl. Eviron. Microb. 56, 3405–3411. 202. Wilshusen, J.H., Hettiaratchi, J.P.A., De Visscher, A., Saint-Fort, R., 2004a. Methane oxidation and formation of EPS in compost: effect of oxygen concentration. Environ. Pollut. 129, 305–314.