Figure 2 presents the overall oxidation efficiencies of methane, the methane concentrations in the raw biogas and the moisture contents in the control and enriched columns, for all the campaigns with loading rates 1 (22 gCH4.m− 2.d− 1) and 2 (44 gCH4.m− 2.d− 1).
The first 40 days of the experiments (with a loading rate of ≅22 gCH4.m− 2.d− 1) seem to have been sufficient for a suitable acclimatation of the methanotrophic population. Even the first campaigns with loading rate 1 showed methane removal efficiencies equivalent to those found after several months of monitoring. Figure 2 also shows that the methane oxidation efficiencies were greater in the enriched column (p < 0.01) for all the campaigns and with both loading rates (p < 0.01). For the campaigns with loading rate 1, the mean methane oxidation efficiencies were 98.4% (SD = ± 1.4) and 89.5% (SD = ± 2.6); and for the campaigns with loading rate 2, the mean methane oxidation efficiencies were 92.6% (SD = ± 1.4) and 82.3% (SD = ± 2.6), for the enriched and control columns, respectively. In fact, landfill soils with higher organic matter contents are associated with relatively high methane oxidation efficiencies [14, 20, 28].
As mentioned above, substrates rich in organic matter are an important source of carbon and nutrients for the metabolic kinetics. Table 2 shows the C/N ratios and pH values of the packing materials used in the control and enriched columns.
Table 2
Initial and final characteristics of the packing materials used in the biofilters
|
Column
|
|
Control
|
Enriched(1)
|
Parameter
|
Initial
|
Final
|
Initial
|
Final
|
C/N
|
14.4
|
19.7
|
9.4
|
10.6
|
pH
|
5.5
|
5.7
|
5.7
|
5.8
|
(1) According to Table 1, this composition only refers to the top 15 cm of the enriched bed. |
In general, low C/N values are associated with elevated methane oxidation rates due to the greater nutrient presence [18]. Taking the compost prepared by Bernal et al. [50] with seven organic waste mixtures (including sewage sludge) as the parameter, a C/N ratio ≤ 12 indicates a mature compost, and hence adequate for the organic amendment of soil. In the present study, there was a reduction in the proportion of N in relation to C from the start to the finish of the experiment in both columns, reflecting nutrient consumption during methane oxidation (Table 2). The difference was considerably greater in the control column (C/N from 14.4 to 19.7), well above the reference value (12). Since the entire carbon and nitrogen contents of the two biofilters evaluated were originally present in the landfill cover soil and SWWTP and there was no nutrient replacement during the experiment, the values showed evidence of a greater nitrogen deficit in the control column since the start of the experiment, which became more and more accentuated throughout the 13 months the biofilters were operated. In the case of the enriched columns, the greater nutrient demand by the methanotrophic community [18] was compensated by the higher N content present in the enriched portion of the enriched bed, which corroborated with the significantly higher methane oxidation efficiencies in this biofilter as compared to the control column (p < 0.01, Fig. 2). It is thus clear that when dealing with organic packing materials, the nutrient content of the support medium must be frequently monitored [51] so as to guarantee their replacement during the operational phase of the biofilter and maintain an adequate C/N ratio for the requirements of the methanotrophs.
Suitable pH values for methanotrophic activity are close to neutrality, but the pH-range in the biofilters can vary considerably depending on the characteristics of the methanotrophic population, the type of packing material, the environmental conditions etc. [18, 19, 33, 51, 52]. On observing the slightly acid pH values reported in Table 2, it appears that this parameter did not influence the methane oxidation efficiency. For practically identical pH conditions in the two columns, the oxidation efficiencies in the enriched bed were close to 100% in all the campaigns of loading rate 1. In fact, although tolerant of pH variations, it is fundamental for methanotrophs that the pH values of the medium remain constant throughout their biological activity [25, 33] as occurred in the present study.
Although both biofilters were watered in the same way throughout the monitoring period, the enriched column showed a higher moisture content throughout almost all the campaigns (Fig. 2), showing a greater water holding capacity (WHC) in this biofilter (at least in the upper portion of the bed, with its higher organic matter content). In fact, in addition to a greater water holding capacity, substrates with higher organic matter contents manage to retain the moisture content of the medium for a longer time [14]. Almost all the moisture contents were within the range of ≅50–55% (Fig. 2) for both columns and both loading rates. There is no consensus in the literature concerning an optimum moisture range for methane oxidation [6, 33, 34] and these values can vary a lot as a function of the substrate characteristics [19, 53, 54]. However, the moisture values obtained corroborated with the reference range of 50–60% for most of the organic supports [55], and the slightly higher moisture contents obtained in the enriched column reflect greater methane oxidation microbial activity, since water is one of the CH4 oxidation products. Even though regularly trying to maintain the moisture content, this parameter varied considerably throughout the biofilter monitoring period, mainly due to the climatic conditions which dictated a greater or lesser drying of the upper portion of the filter beds. Nevertheless, considering that this was a field experiment under real conditions, the fluctuations in moisture content observed during the months of monitoring were acceptable. It is important to mention that the variation in moisture content during the campaigns did not significantly affect methane oxidation (p = 0.54) and the parameter of moisture did not show a significant coefficient on the model generated, demonstrating that the moisture control carried out during the experiment was efficient.
Duplication of the biogas inlet loading rate caused the mean methane oxidation efficiencies to fall 6.0% in the enriched column and 8.0% in the control column. The relationship between methane removal efficiency and the inlet loading rate and inlet CH4 concentration during biotreatment processes has been the subject of several studies [56–59], An increase in loading rate (or gas flow rate) generally increases the velocity with which the gas passes through the filter bed, thus reducing the residence time of the contaminant in the biofilter. At first, an increase in loading rate increases the transfer rate of the substrate (contaminant) present in the gaseous phase (bulk) to the biofilm, where the microorganisms grow and degrade the contaminant. However, as the biogas loading rate or even the contaminant concentration in the bulk continue increasing, there is a tendency for the methane conversion rate to decrease, since the mass transfer of this compound to the biofilm becomes the rate limiting step of the process. In addition, the mass transfer of methane at the gas-liquid interface is also affected by the low solubility of methane in the aqueous medium [23, 51, 53, 58, 60]. Thus, even though the maximum loading rates reported in the literature (and hence the biotreatment efficiency) may vary considerably as a function of the system characteristics and operational conditions [6, 57, 61, 62], reduced loading rates (or elevated residence times) can favour methane oxidation in bio-based cover systems [18, 35], as observed in the present study. The parameter “loading rate” was also significant for methane oxidation efficiency, and for loading rate 1 the mean efficiencies were indeed statistically higher than for loading rate 2, for both the enriched and control columns (Table 3).
Table 3
Contrasts calculated by Tukey’s test at 5% significance between the mean oxidation efficiencies of the biofilters and the loading rates
Loading rates
|
Biofilters
|
Efgox (average)
|
2
|
Control
|
82.3d
|
1
|
Control
|
89.5c
|
2
|
Enriched
|
92.6b
|
1
|
Enriched
|
98.4a
|
Means followed by the same letter do not differ significantly according to Tukey’s test at 5% significance. |
The methane concentrations in the raw biogas throughout the 25 campaigns of the experiment varied from 39.9 to 43.6% for loading rate 1 and from 39.3 to 43.1% for loading rate 2 (Fig. 2). The concentrations of the principal components of the LFG can vary temporally and spatially as a function of parameters such as the moisture content, temperature, pH value, nutrient availability in the waste mass, landfilling conditions, waste composition, decomposition phase of the residue and even the climatic conditions [63, 64, 65]. However, this fluctuation in the CH4 concentrations was not sufficiently high to interfere with the biofilter operations (control and enriched), since no relationship was observed between the inlet CH4 concentration and its oxidation efficiency, independent of the column and loading rate. In fact the CH4 concentration did not present a statistically significant coefficient in the linear models generated (p = 0.51). Thus, under the conditions of the present study, the variations in loading rate presented a much greater influence on the methane oxidation efficiency than the variations in methane concentration.
For all the campaigns with both loading rates 1 and 2, Fig. 3 presents the overall oxidation efficiencies of methane on the control and enriched columns as a function of the packing materials (Tsoil) and environmental (Tenvir) temperatures.
T1 and T2: Thermometers installed, respectively, 40 and 10 cm from the filter bed surface.
C: Control column E: Enriched column
Independent of the loading rate under evaluation, the temperatures of the two filter beds at 40 cm (T1) and 10 cm (T2) from the surface showed temperatures above the environmental temperatures in all the monitoring campaigns. In fact, the biological oxidation of methane is an exothermic reaction, liberating ≅211 kcal per mol of oxidized methane [25, 59]. According to Fig. 3, when loading rate 2 was applied to the biofilters and the methanotrophic activity increased, higher temperatures were registered on the enriched column at 10 cm from the surface, where there was a greater amount of organic matter provided by the SWWTP. In addition, the heat of the raw biogas supplied to the two columns as from the vertical drain added to the exothermicity of the system, which can lead to a considerable increase in the filter bed temperature [6]. Many studies have indicated a strong correlation between oxidation efficiency and the temperature of the medium [19, 24, 26, 35, 36]. Most of these studies reported an increase in methane removal efficiency with increase in temperature of the medium, and, similarly, a reduction in removal efficiency with decrease in temperature, mentioning temperatures in the range from 25–35ºC to obtain optimal methane oxidation efficiency in different soil types [19, 59]. However, such behaviour cannot be considered as a general rule. Depending on the conditions of the study, one can obtain methane bio-oxidation efficiencies above 80%, even at lower temperatures [37, 38]. In the present study, as shown in Fig. 3, even at filter bed temperatures below 20ºC, the methanotrophs maintained their biological activity, obtaining methane removal efficiencies equivalent to those registered at temperatures above 25ºC. The heat liberated during methanotrophic activity or even that provided by the raw biogas (as mentioned above), allow for a certain “independence” of the system in relation to eventual falls in temperature of the filter beds or even oscillations in the environmental temperature [25]. Also, according to Gebert et al. [13], the methanotrophic population can adapt to variations in the environmental temperature, including shifting the optimal temperature range for methane oxidation, guaranteeing the performance of the biofilter even under low environmental temperature conditions. In the present study, even using an ample environmental temperature range (≅11 to 30°C), all the overall oxidation efficiencies of methane on the enriched columns were above 97% for loading rate 1 and 90% for loading rate 2, as evidence that the temperature did not constitute a limiting parameter in this biological process. It was verified statistically that neither of the temperatures evaluated (T1 and T2) on the two biofilters or even the environmental temperature, presented significant coefficients for the oxidation efficiency prediction models.
In general, the variables that significantly influenced the methane oxidation efficiency were the differences between the packing materials (enriched and control beds) and the loading rates (Fig. 4). These two predictive variables explained about 92% of the behaviour related to Efgox (R2adj = 0.92) for the three regression models evaluated. Thus for the models T1, T2 and T environmental, the determination coefficients adjusted by the degrees of liberty (R2adj) presented no difference in value.
Figures 5 and 6 present, for loading rates 1 and 2 respectively, the vertical oxidation methane profiles (Efvox) and the ratio between the carbon dioxide and methane concentrations (CO2:CH4) throughout the filter beds. The values plotted are the means of the campaigns for loading rate 1 (13 campaigns) and 2 (12 campaigns).
The determination of the ratio between the carbon dioxide and methane concentrations (CO2:CH4 ratio) allows one to identify the depth of the methane oxidation zone within the filter bed. In the case of an upward flow biofilter, the CO2:CH4 ratio tends to increase from bottom to top as a function of the gradual oxidation of methane to carbon dioxide. The analysis of the methane oxidation profile presupposes, for example, that all the CH4 is converted to CO2, that microbe respiration is negligible (all the CO2 measured coming exclusively from the oxidation of CH4), that the size of the methanotrophic community is stable and that the system operates under steady state conditions [15,18,66].
Figures 5 and 6 revealed that the principal methane oxidation zones occurred in the upper ≅23 cm of the control bed and upper ≅40 cm of the enriched bed for loading rate 1; and in the upper ≅23 cm of both columns for loading rate 2. Thus oxidation was more significant in a larger portion of the enriched bed as compared to the control bed. In addition to supplying the soil with nutrients, the addition of organic matter to the cover layer increased the porosity of the medium [67,68], improving the availability of oxygen to the methanotrophs and making CH4 circulation through the filter bed easier [14, 15, 20, 29]. The greater porosity of the medium guarantees a longer residence time of the gases in the packing material, corroborating with an increase in the methane oxidation rate in the biofilter [25]. Although the parameter “porosity” was not determined in the present study, the addition of SWWTP to the cover soil resulted in a coarser medium (Fig. 7), making the penetration of atmospheric air easier. Although the SWWTP was only added to the upper 15 cm of the bed, even the lower layers of the enriched column showed a slightly superior performance as compared to the control column in terms of methane oxidation capacity. Since moistening was always done at the top of the packing materials, part of the organic matter concentrated at the top of the column was carried down to the deeper zones of the enriched bed, improving the conditions for methanotrophic activity in this region of the bed.