Mechanisms of distinct activated carbon and biochar amendment effects on petroleum vapour biofiltration in soil

Khaled M. Bushnaf ab, George Mangse b, Paola Meynet b, Russell J. Davenport b, Olaf A. Cirpka c and David Werner *b
aDepartment of Earth and Environmental Sciences, El-mergab University, Khoms, Libya
bSchool of Engineering, Newcastle University, Newcastle upon Tyne, NE1 7RU, England, UK. E-mail: david.werner@ncl.ac.uk; Fax: +44 (0)191 208 6502; Tel: +44 (0)191 208 5099
cCenter for Applied Geoscience, University of Tübingen, 72074 Tübingen, Germany

Received 8th July 2017 , Accepted 23rd August 2017

First published on 1st September 2017


We studied the effects of two percent by weight activated carbon versus biochar amendments in 93 cm long sand columns on the biofiltration of petroleum vapours released by a non-aqueous phase liquid (NAPL) source. Activated carbon greatly enhanced, whereas biochar slightly reduced, the biofiltration of volatile petroleum hydrocarbons (VPHs) over 430 days. Sorbent amendment benefitted the VPH biofiltration by retarding breakthrough during the biodegradation lag phase. Subsequently, sorbent amendment briefly reduced the mineralization of petroleum hydrocarbons by limiting their bioavailability. During the last and longest study period, when conditions became less supportive of microbial growth, because of inorganic nutrient scarcity, the sorbents again improved the pollution attenuation by preventing the degrading microorganisms from being overloaded with VPHs. A 16S rRNA gene based analysis showed sorbent amendment effects on soil microbial communities. Nocardioidaceae benefitted the most from petroleum hydrocarbons in activated carbon amended soil, whereas Pseudomonadacea predominated in unamended soil. Whilst the degrading microorganisms were overloaded with VPHs in the unamended soil, the reduced mobility and bioavailability of VPHs in the activated carbon amended soil led to the emergence of communities with higher specific substrate affinity, which removed bioavailable VPHs effectively at low concentrations. A numerical pollutant fate model reproduced these experimental observations by considering sorption effects on the pollutant migration and bioavailability for growth of VPH degrading biomass, which is limited by a maximum soil biomass carrying capacity. Activated carbon was a much stronger sorbent for VPHs than biochar, which explained the diverging effects of the two sorbents in this study.



Environmental significance

Understanding sorption and biodegradation interlinkages is fundamentally important for understanding the fate of pollutants in the environment. Here we present for the first time a mechanistic study of long-term activated carbon and biochar amendment effects on the fate of biodegradable pollutants in soil. We use chemical analysis, microbial community characterization by next generation gene sequencing, and numerical modelling to discuss the effects of activated carbon and biochar amendments on the biodegradation of volatile petroleum hydrocarbons in soil by dynamic microbial communities. Our case study demonstrates contrasting effects of sorption during different phases of the microbial growth cycle, and sorbent effects on the soil microbial community composition. Our scientific insights are applicable to the design of sorbent-based soil remediation and other biofiltration designs.

Introduction

Vast amounts of oils and fuels are used worldwide, and petroleum spillages are among the most frequently reported incidents of soil and water pollution.1 Petroleum hydrocarbon vapours migrate from spilled liquids through the vadose zone towards indoor air, outdoor air, and groundwater, posing a significant exposure pathway in contaminated land usage scenarios.2 The biofiltration of volatile petroleum hydrocarbons (VPHs) in soil provides a barrier to contaminant spreading and can considerably reduce exposure risks.3–6

Sorption and biodegradation are key processes contributing to this natural VPH attenuation in soil,6 but potential interactions between sorption and biodegradation are manifold and difficult to predict.

Fig. 1a and b illustrate how an enhanced sorption capacity (expressed as linear solid–water distribution coefficient Kd) may affect the natural attenuation of a biodegradable petroleum vapour that is transported by gas diffusion over a certain distance in unsaturated soil. The predictions are based on the common assumptions that VPH degradation occurs in the soil water and can be described by first-order kinetics.3,7 If these assumptions are valid, enhanced sorption (i.e., a larger Kd-value) will (i) delay VPH migration through a layer of soil and (ii) reduce the peak concentration of VPH if the source is small and finite, but (iii) the cumulative VPH mass breaking through the soil layer over the lifetime of the source will eventually reach the same level, that is, it is independent of Kd. In reality, the dynamics of soil biomass,8,9 growth limitations,10 microbial ecology,11,12 and metabolic regulation at the cell level13,14 complicate this picture greatly, and, depending on the dominating mechanisms, enhanced sorption may potentially reduce or enhance cumulative VPH mass breakthrough.


image file: c7em00309a-f1.tif
Fig. 1 Schematic illustration of the predicted sorbent amendment effects on (a) the VPH concentration in soil; and (b) the VPH attenuation in soil; and (c) an illustration of the column study set-up. The qualitative illustrations of the Kd effects on VPH concentrations and attenuation in soil are based on the assumption of a finite source of one biodegradable compound passing through the column by gas-phase diffusion with first-order rate biodegradation occurring only in the soil pore water phase.

Understanding complex trade-offs between sorption and biodegradation is broadly relevant to pollutant fate predictions in the environment, and especially important for sorbent-based in situ soil remediation, which seeks to reduce exposure risks at contaminated sites.15–20 We performed long-term column studies, amending activated carbon or biochar to construction sand exposed to biodegradable VPHs released from a finite NAPL source to compare the sorbent amendment effects on the attenuation of typical VPHs. Our study aim was to derive the scientific understanding required for predicting coinciding sorption and biodegradation effects on the VPH fate in the environment, so that these scientific principles can then be practically applied to the design of enhanced petroleum vapour biofiltration in soil with the addition of carbonaceous sorbent materials such as activated carbon or biochar. We tested the hypothesis that the sorbent amendments would not just delay the pollutant breakthrough, but more lastingly enhance the petroleum vapour biofiltration.

Materials and methods

Gravelly sand (soil), biochar (BC) and activated carbon (AC)

For a porous medium, we chose gravelly construction sand (89.5% by weight sand with 0.063 to 2 mm particle size). The total organic carbon (TOC) content was 1.6 ± 0.1%, the total carbon content was 4.1 ± 0.1%, and the pH measured in 0.01 M CaCl2 was 7.43. A commercial biochar, produced from wood chips by fast pyrolysis at high temperature (800 °C) in a fixed bed reactor, was obtained from Environmental Power International, EPI (Wiltshire, UK), and used in this study due to its large surface area. This biochar was ground to a particle size <163 μm and had a TOC content of 85 ± 2% and an alkaline pH of 9.25 ± 0.16. Type F400 activated carbon (Chemviron Carbon Ltd, Lancashire, UK) was also ground to a particle size <163 μm and had a TOC content of 73 ± 1% and a pH of 8.73 ± 0.10. Biochar and activated carbon surface areas were measured by CO2 adsorption using an Intelligent Gravimetric Analyzer (IGA) (Hiden Isochema Ltd., Warrington, UK) and determined to be 928 m2 g−1 and 1012 m2 g−1, respectively.

Chemicals

Twelve major constituents of gasoline or kerosene combined in typical ratios3 represented a well-defined non-aqueous phase liquid (NAPL) source, prepared from high purity chemicals obtained from Sigma-Aldrich (Dorset, UK).21 The NAPL source of each column consisted of five straight-chain alkanes (0.62 g pentane, 1.13 g hexane, 1.11 g octane, 2.02 g decane, and 0.89 g dodecane), four branched and/or cyclic alkanes (1.10 g methylcyclopentane, 1.83 g methylcyclohexane, 1.24 g cyclohexane, and 2.07 g isooctane), and three monoaromatic compounds (0.69 g toluene, 0.85 g m-xylene, and 1.05 g 1,2,4-trimethylbenzene). These petroleum hydrocarbon compound classes vary in polarity and biodegradability.22–24 Sulfur hexafluoride (SF6) (Sigma-Aldrich, Dorset, UK) was used as a conservative tracer.

Column experiments

Three horizontally placed glass columns of 120 cm length and 7.8 cm internal diameter (Fig. 1c) were homogeneously packed over a length of 93 cm with the gravelly sand (soil), gravelly sand with 2% dry weight (d.w.) biochar (soil & BC), and gravelly sand with 2% d.w. activated carbon (soil & AC), respectively. TOC contents were 1.6 ± 0.1, 3.0 ± 0.1, and 2.7 ± 0.3% d.w., total porosities 0.55 ± 0.02, 0.60 ± 0.03 and 0.59 ± 0.02 cm3 cm−3, and initial volumetric water contents 0.11 ± 0.01, 0.12 ± 0.02, and 0.11 ± 0.01 cm3 cm−3, for the soil, soil & BC and soil & AC columns, respectively. Thus, in agreement with previous work,25 the carbonaceous amendments moderately enhanced soil macroporosity. After packing, the columns were left undisturbed for five days to monitor background soil respiration, and then, nominally on day 0, a vial containing 20 mL VPH mixture was tightly connected to a curved glass tube with 1.1 cm internal diameter at one end of the columns using a Teflon-lined rubber seal. The other column ends were purged into a fume hood with a water-saturated airflow rate of 5 ± 1 mL min−1 to strip the petroleum vapours and simulate a near-zero concentration boundary condition. Each column was equipped with seven sampling ports positioned at 15 cm distances, and the ports were sealed with GC septa (Thermogreen LB-2, Supelco, Bellefonte, USA). Soil air samples were withdrawn through these ports with gas tight syringes for analysis. An initial SF6 tracer test was conducted to determine the tortuosity of the gas-filled pore space.26 After connecting the sources to the columns, VPH and CO2 and O2 concentrations in soil air were monitored weekly at each sampling port (21 in total) for the duration of the experiment. On day 430, the soil was removed from the columns in three equal sections (namely the ‘VPH source side’ between ports 1 and 3, the ‘Middle’ between ports 3 and 5, and the ‘Atmospheric boundary side’ between ports 5 and 7 in Fig. 1c) and soil subsamples were immediately extracted with 40[thin space (1/6-em)]:[thin space (1/6-em)]60 v/v dichloromethane/pentane for residual VPH quantification. Inorganic nutrient contents and total and organic carbon contents were also determined.

Batch experiments

A series of replicated batch microcosm experiments with the three column soil types and appropriate sterile controls were performed to complement the column studies and measure VPH sorption coefficients, VPH biodegradation rates, and nutrient availability effects on VPH biodegradation and soil microbiology. Batch study details are provided in the ESI, and followed protocols described by Bushnaf et al.21

Soil microbiology analysis

Total bacterial cell numbers were determined by mixing 5 μL of sample stored in an ethanol/PBS mixture with 995 μL of filtered-sterile phosphate buffer saline (PBS, Oxoid), as previously described.15 Microbial communities were investigated by high-throughput sequencing. DNA was extracted in duplicate from 500 mg of soil (wet weight), using the FastDNA Spin kit (MP Biomedicals, UK). Amplicon libraries were prepared from each extracted DNA sample by PCR amplifying the V4–V5 regions of the 16S rRNA gene.27 Details are reported in the ESI. The individual amplicon libraries were pooled together in equimolar amounts, and sequenced on a Roche 454 GS FLX+ System by NewGene Ltd (International Centre for Life, Newcastle upon Tyne, UK), for most samples, and (following withdrawal of that platform by the supplier) on an Ion Torrent Personal Genome Machine (School of Engineering, Newcastle University, UK), for the samples from one of the batch studies. Further details, including those about bioinformatics, are provided in the ESI.

Analytical protocols

VPH analysis was performed with an Agilent HP-7890 gas chromatograph using a flame ionization detector (Agilent Technologies, Palo Alto, USA), and CO2 and O2 and tracer SF6 analysis was performed on a Fisons 8060 gas chromatograph linked to a Fisons MD800 mass spectrometer.21 More details of analytical protocols, including those for inorganic nutrient analysis, are provided in the ESI.

Modelling

To interpret the column study results, an earlier petroleum vapour fate model16 was improved by (i) considering VPH intraparticle diffusion in biochar and activated carbon, which is particularly important for simulating activated carbon amendment effects, and (ii) accounting for logistic growth with a maximum soil biomass, which provides greater flexibility as compared to the earlier model version,16 since any limiting factor, including nitrogen availability, can potentially be used to predict the soil biomass carrying capacity. Like its previous version,16 the model considers the effect of changes in source composition on VPH concentrations in soil air above the NAPL source using Raoult's law, the effect of air–water and air–solid partitioning on VPH migration from the NAPL source through the soil towards the atmospheric boundary side of the columns, the biodegradation of VPHs dissolved in soil pore water, and the growth and decay of VPH degrading biomass. Detailed modelling assumptions, equations, and boundary conditions are provided in the ESI.

Results and discussion

Sorbent amendment effects on VPH sorption

Despite their similar BET surface areas, the biochar was a weaker VPH sorbent than the activated carbon (see distribution coefficients Kd in Table S1 in the ESI). Amendment of gravelly sand with 2% biochar enhanced the sorption of straight-chained, branched, and cyclic alkanes by approximately a factor of 2 while the sorption of monoaromatic compounds was increased by up to a factor of 10, confirming preferential monoaromatic hydrocarbon sorption by this biochar.21 The addition of 2% activated carbon to gravelly sand much more substantially enhanced the sorption of all VPHs, by up to three orders of magnitude. Because the biochar and activated carbon had similar total surface areas, the observed differences are likely explained by different pore network structures and surface chemistries. The higher pH of the biochar for example suggests a more basic nature, which may hinder the adsorption of hydrophobic alkanes. In this context, it is important to note that sorption performance varies greatly for different biochar types,20 meaning that other biochar types may be able to bind VPHs as strongly as activated carbon.

Inorganic nutrient availability and VPH sorption effects on petroleum vapour biodegradation in batches

When added at low concentrations (0.002 mL of VPH mixture per 15 g of wet soil), all VPHs were biodegradable, as indicated by the apparent first-order biodegradation rates measured in the batch studies (Table S1 and Fig. S1 in the ESI). A biodegradation lag phase was noted for some VPHs such as isooctane (Fig. S1b in the ESI). When added at higher concentrations (0.03 mL of VPH mixture per 15 g of wet soil), VPH biodegradation was hindered by insufficient inorganic nutrient availability (Fig. 2). In agreement with previous studies,28 provision of inorganic nitrogen and phosphorus to achieve a VPH-C to N to P ratio of 100[thin space (1/6-em)]:[thin space (1/6-em)]10[thin space (1/6-em)]:[thin space (1/6-em)]1 significantly accelerated VPH biodegradation by intrinsic soil microorganisms (t-test, two-tailed, p < 0.01), while provision of only nitrogen or phosphorus had lesser benefits, which were not statistically significant (t-test, two-tailed, p > 0.05, blue bars in Fig. 2). Inorganic nutrient amendment effects on VPH biodegradation in the soil & AC batches were not statistically significant in any of the treatments (t-test, two-tailed, p > 0.05, green bars in Fig. 2). Apparently, strong sorption by activated carbon became the VPH biodegradation rate controlling factor when N and P were sufficiently available. Without inorganic nutrient additions, the extent of VPH biodegradation was not significantly different in soil and soil & AC batches (t-test, two-tailed, p > 0.05), showing that both sorption and inorganic nutrient scarcity may limit VPH biodegradation in the geomedia investigated. Observations in soil & BC batches (red bars in Fig. 2) fell in between the soil and soil & AC trends, as would be expected based on the respective Kd values (Table S1 in the ESI).
image file: c7em00309a-f2.tif
Fig. 2 VPH mass residuals after fifteen days in live soils with and without 2% biochar (BC) or activated carbon (AC) amendment relative to sterile controls. Nitrogen (N) and/or phosphorus (P) amendments were added based on a mass ratio of 100[thin space (1/6-em)]:[thin space (1/6-em)]10[thin space (1/6-em)]:[thin space (1/6-em)]1 for VPH-C to NH4+-N to H2PO4-P.

Activated carbon and biochar amendment effects on the petroleum vapour attenuation in the column study

For the petroleum vapour migration exposure pathway, only those VPHs which emanate from the soil can potentially reach receptors via their presence in indoor and outdoor air. We therefore use the cumulative VPH-C emanation over the 430 day period as a measure of the incomplete VPH attenuation (Fig. 3a), and by this measure, the activated carbon amended soil column most effectively, and the biochar amended soil column least effectively, attenuated the VPHs released by the NAPL source. The cumulative VPH-C emanation from the soil & AC column was at all times much lower than those from the soil and soil & BC columns. While the total VPH-C emanation was higher for the soil & biochar column than for the unamended soil column (3.18 versus 2.56 g by day 430, respectively), which could be partially due to the somewhat greater air-filled porosity of the soil & biochar column, emanation of monoaromatic VPHs, which bind strongly to the biochar used in this study, was lower for the soil & biochar as compared to the unamended soil column (0.005 versus 0.020 g monoaromatic VPH-C by day 430, respectively). The cumulative CO2-C emanation (Fig. 3b) was initially substantially lower for the soil & AC as compared to the other two columns, but then rose more steadily to reach and surpass the flux from the other columns. Since background soil respiration measured before the connection of the NAPL source was very low in this study, the VPH-C to CO2-C ratio relates on a carbon-normalized basis the amount of VPH which was able to pass unaltered through the column (i.e. was not attenuated), to the amount of VPH which has been mineralized into CO2. By day 430, the ratio of the cumulative VPH-C to CO2-C emanation from the columns was only 1[thin space (1/6-em)]:[thin space (1/6-em)]77 for the soil & AC column, showing substantial VPH attenuation, as compared to 1[thin space (1/6-em)]:[thin space (1/6-em)]1.3 and 1[thin space (1/6-em)]:[thin space (1/6-em)]0.8 for the soil and soil & BC columns, respectively.
image file: c7em00309a-f3.tif
Fig. 3 Measured, cumulative (a) VPH carbon and (b) CO2 carbon emanation from the soil columns, shown as diamond symbols. Solid lines show the model predictions for the soil (blue), soil & BC (red) and soil & AC (green) columns, with optimized microbiological parameter values.

Notable characteristics of the petroleum vapour and CO2 breakthrough curves measured in the column study

VPH and CO2 breakthrough curves measured at port 4 in the middle of the three columns are compared in Fig. 4. To reduce the number of plots, the twelve VPH compounds have been aggregated into the petroleum vapour compound classes of straight-chain alkanes (Fig. 4a), branched/cyclic alkanes (Fig. 4b), and monoaromatic hydrocarbons (Fig. 4c), while CO2-C data provide context (Fig. 4d). Characteristic features distinguished the measured VPH breakthrough curves in soil and soil & BC from those anticipated by a first-order rate model (Fig. 1a): instead of a monotonic increase and then decline of the VPH concentrations, an initial spike in the measured VPH concentrations was followed by a temporary dip around days 5 to 10, especially notable for the straight-chain alkanes, which was followed by a secondary maximum around days 20 to 30, before the concentrations of the more volatile VPHs again declined due to the exhaustion of their finite NAPL source (Fig. 4a–c). These features confirm the three VPH biodegradation phases tentatively identified in our previous, short-term study:16 (i) a biodegradation lag phase; (ii) a brief period of intensive VPH biodegradation; followed by (iii) less optimal conditions for VPH biodegradation.
image file: c7em00309a-f4.tif
Fig. 4 Carbon-content normalized VPH and CO2 concentrations measured in the middle of the columns for the sum of the (a) straight-chain alkanes, (b) branched/cyclic alkanes, (c) monoaromatic hydrocarbons, and (d) CO2. Insets show the early data for the first two months. Solid lines show the model predictions for the soil (blue), soil & BC (red) and soil & AC (green) columns, with fitted microbiological parameter values.

Mass balances, total cell counts and nutrient availability at the end of the column study

On day 430, the remaining NAPL sources consisted of 99% of the three least volatile compounds, 1,2,4-TMB (0.2–5%), decane (11–31%) and dodecane (56–89%), showing that the other nine VPHs had been largely exhausted from the source. The soil & AC column had the lowest NAPL source residual (Table 1), which can be explained by the enhanced sorption of VPHs by the activated carbon fostering VPH volatilization from the NAPL. VPH residuals within the columns on day 430 were highest for the soil & AC column, due to the strongest VPH retention in this column. However, the VPH residual is too small to alone explain the reduced VPH-C emanation from the soil & AC column at the atmospheric boundary side (Table 1). The VPH residual in the soil & AC column consisted of 84% of isooctane, which is one of the more recalcitrant VPH compounds.7 Poor extractability of VPHs from activated carbon can therefore not explain the absence of more substantial VPH residuals, since it would have affected isooctane and other alkanes similarly. Instead, an apparent increase in the soil carbon content, especially notable in the soil & AC column, but also in the two other columns, can potentially account for the unknown gaps in the VPH-C mass balances (Table 1). Increased soil carbon content can be due to VPH metabolite formation, humification of biomolecules from VPH degrading biomass, or precipitation of CO2-C produced from VPH degradation as carbonates.
Table 1 Volatile Petroleum Hydrocarbon (VPH) carbon mass balance on day 430. Initially, 12.5 g VPH-C was present in each source
Column Source (g) Retained as VPH in soil (g) Emanated as VPH (g) Emanated as CO2 (g) Change in biomass Ca (g) Total accounted (g) Change in total soil C (g)
a Calculated assuming a carbon content of 10−13 g per cell.
Soil 1.15 ± 0.09 0.0 2.56 3.32 0.080 ± 0.026 7.11 +10 ± 15
Soil & 2% BC 1.36 ± 0.10 0.01 ± 0.01 3.18 2.63 0.037 ± 0.023 7.22 +16 ± 20
Soil & 2% AC 0.65 ± 0.03 0.40 ± 0.05 0.045 3.45 0.063 ± 0.019 4.61 +51 ± 28


Inorganic nutrient availability in the columns was reduced on day 430 as compared to the initial conditions (Table 2) and was generally lowest on the NAPL source side of the columns, and especially low in the soil & AC column, where CO2 production was highest towards the end of the experiments. Ammonium-N, which is more readily assimilated than nitrate-N, was the most depleted inorganic nutrient (>90%) as compared to the initial conditions. Inorganic nutrient scarcity is thus the likely cause of suboptimal VPH biodegradation conditions in the later phase of the column study.

Table 2 Water extractable inorganic soil nutrients and total soil cell count data
Sample ID Ammonium (μg NH4+-N per g) Nitrate (μg NO3-N per g) Phosphate (μg H2PO4-P per g) Total cell count (cells per g)
Before the column experiment
Soil 2.9 ± 0.2 8.2 ± 2.0 0.8 ± 0.07 2.7 ± 0.2 × 107
Soil & 2% biochar 5.8 ± 0.5 5.5 ± 0.57 0.2 ± 0.02 7.8 ± 0.2 × 107
Soil & 2% activated carbon 6.1 ± 1.2 6.5 ± 1.0 0.3 ± 0.03 3.9 ± 0.7 × 107
[thin space (1/6-em)]
At the end of the column experiment
Soil, VPH source side 0.18 ± 0.04 4.0 ± 0.4 0.2 ± 0.02 21.4 ± 1.3 × 107
Soil, middle 0.18 ± 0.05 6.8 ± 0.3 0.4 ± 0.03 11.5 ± 6.6 × 107
Soil, zero conc. Boundary side 0.24 ± 0.03 6.5 ± 0.5 0.7 ± 0.03 18.6 ± 4.1 × 107
Soil & 2% biochar, VPH source side 0.06 ± 0.003 1.1 ± 0.5 0.07 ± 0.01 20.2 ± 1.6 × 107
Soil & 2% biochar, middle 0.05 ± 0.01 2.1 ± 0.1 0.12 ± 0.02 13.9 ± 1.1 × 107
Soil & 2% biochar, zero conc. Boundary side 0.04 ± 0.01 2.2 ± 0.1 0.13 ± 0.03 11.2 ± 4.6 × 107
Soil & 2% activated carbon, VPH source side 0.04 ± 0.01 0.0 ± 0.0 0.17 ± 0.02 18.0 ± 4.6 × 107
Soil & 2% activated carbon, middle 0.05 ± 0.01 0.18 ± 0.01 0.16 ± 0.01 12.5 ± 1.1 × 107
Soil & 2% activated carbon, zero conc. Boundary side 0.05 ± 0.01 2.4 ± 0.02 0.12 ± 0.02 17.9 ± 9.1 × 107


Total cell counts were up to a factor of 8 higher at the end of the experiment as compared to the initial conditions (Table 2), showing a lasting enhancement in the soil microbial biomass following VPH exposure. Amongst locations (source, middle and atmospheric sides of the columns) and soil types (soil, soil & BC, and soil & AC) there was at most a factor of 2 variability in the total cell counts, suggesting that the extent of this long-term biomass growth was not strongly altered by the sorbent amendment or distance from the source.

Petroleum vapour addition, and biochar or activated carbon amendment effects on the soil microbial community composition

After quality filtering, 148[thin space (1/6-em)]882 and 365[thin space (1/6-em)]166 sequences, for batch and column samples respectively, were recovered and represented by 4568 OTUs for the batch samples and 12[thin space (1/6-em)]265 OTUs for the column samples at 97% similarity (i.e. distinguished at the species level). The average Bray–Curtis similarity matrix (log transformed relative abundance data) was mapped onto a 2D ordination space to reveal VPH and sorbent-amendment as well as spatial effects (Fig. 5). Short-term effects of VPH exposure at high concentrations became apparent on day 6 in the batch studies (Fig. 5a), where sorbent amendment strongly shaped bacterial and archaeal community-structure (ANOSIM based on Bray–Curtis similarity, global R sorbent amendment effect = 0.655, p < 0.001). Long-term VPH exposure effects on microbial communities became apparent on day 430 in the column study (Fig. 5b), with a significant effect of sorbent amendment on the bacterial and archaeal community structure (ANOSIM, global R = 0.647, p < 0.01).
image file: c7em00309a-f5.tif
Fig. 5 Non-metric multi-dimensional scaling (NMDS) analysis based on Bray–Curtis similarities at the species level (log transformed) for the sorbent amended and unamended soil samples (duplicates), in the (a) batch study on day 6, with nutrients (WN) and without nutrients (WON) and (b) column study on day 430.

Relative OTU abundance rankings compared to the control soils at time zero were used to further evaluate the nature of the bacterial and archaeal community composition changes, induced by the VPH exposure (Fig. 6 and Tables S2–S4 in the ESI). An improved abundance ranking of an OTU provides evidence for better ecological competitiveness of this OTU under the altered conditions. Not surprisingly, several OTUs from families known to include petroleum hydrocarbon degraders29–39 substantially enhanced their ranking following VPH exposure (Fig. 6a). In batch studies with soil, members of the families Nocardiaceae (i.e. from the genus Rhodococcus) and Pseudomonadaceae (i.e. Pseudomonas umsongensis) gained most significantly following straight-chain alkane addition, while the Pseudomonadaceae benefitted more strongly than the Nocardiaceae from branched and cyclic alkane addition. Probably due to toxic effects at high monoaromatic hydrocarbon concentration, only small ranked abundance changes were observed following monoaromatic hydrocarbon addition.


image file: c7em00309a-f6.tif
Fig. 6 Illustration of relative abundance rank changes following the addition of 0.03 mL VPHs as (a) compound classes or (b–d) a complete mixture of the twelve VPHs in the batch study for (a and b) soil, (c) soil and biochar and (d) soil and activated carbon.

The microbial community response to VPH addition was affected by the sorbent amendments (Fig. 6b–d). While some OTUs, such as one from the genus Acidovorax, benefitted from the addition of the twelve VPH mixture with and without the sorbent amendments, an OTU from the genus Achromobacter was more successful in the soil and soil & BC than in the soil & AC batches, OTUs from the Pseudomonadaceae family were most successful in the unamended soil batches, an OTU from the genus Hydrogenophaga was most successful in the soil & BC batches, and an OTU from the Nocardioidaceae family most successful in the soil & AC batches (Fig. 6b–d and Table S3 in the ESI). Competitive interactions between groups of microorganisms with hydrocarbon degradation potential are often governed by environmental parameters: Actinobacteria (to which the Nocardioidaceae belong) tend to be most successful in soils with low carbon substrate availability, and Proteobacteria (to which the Pseudomonadaceae belong) most successful in soils with high carbon substrate availability.40 Furthermore, stable isotope labeling to identify nitrogen-incorporating bacteria in petroleum-contaminated Arctic soils showed that, within the major families of Actinobacteria, Nocardioidaceae and Microbacteriaceae were more effective than Nocardiaceae at assimilating ammonium.41 Combining these insights with the current study results it can be speculated that Nocardioidaceae have especially high affinity for carbon substrates and inorganic nitrogen, and thus gain ecological competitiveness in the soil & AC environment, where the bioavailable VPH concentrations (Fig. 4) and inorganic nitrogen concentrations (Table 2) were lower than those in the unamended soil.

On day 430 in the column study, when VPH concentrations had fallen to very low levels, some of these sorbent-related short-term trends were still notable (Table S4 in the ESI). An OTU from the Nocardioidaceae family showed the strongest relative rank abundance increase near the VPH source in the soil & AC column, while an OTU from the Pseudomonadaceae family had significantly increased relative abundance ranks in the soil and soil & BC column, but not in the soil & AC column. An OTU from the genus Hydrogenophaga again gained most consistently in the soil & BC column.

Modelling results

Fig. 3 and 4 compare measured with simulated VPH concentrations aggregated for the different compound classes. Model parameter values are compiled in the ESI (Tables S5–S8). Model predictions for the soil column based on the independently measured batch study data, and some parameter values from the literature (as indicated in the ESI), already captured the characteristic features of the measured breakthrough curves (early peak, temporary dip, and secondary peak, see Fig. S2 in the ESI) certainly much better than a first-order rate model (Fig. 1a). Quantitatively, straight-chain alkane concentrations were well predicted, branched and cyclic alkane concentrations were under-predicted, and monoaromatics concentrations over-predicted, based on the independent model parameter calibration. An improved fit for the soil column data (Fig. 3 & 4, blue lines) was obtained with the following four changes to the microbiological modelling parameters: (i) increasing the lag phase for the branched and cyclic alkanes from 3 to 6 days; (ii) reducing the second order growth rates for the branched and cyclic alkanes by a factor of 3.3, whilst increasing the rates for the monoaromatic hydrocarbons by a factor of 1.5; (iii) reducing the soil biomass carrying capacity by a factor of 1.9; and (iv) increasing the biomass decay rate, which was obtained from the literature,16 by a factor of 2. Fig. S3 in the ESI shows the predictions for the soil & BC and soil & AC columns when only the amendment effects on the VPH sorption (Kd values in Table S1 in the ESI) were considered, i.e. assuming that the microbiological parameters fitted for the soil column study also applied in the soil & BC and soil & AC columns. These predictions already aligned well with the measured data, indicating that the main amendment effects on the petroleum vapour fate relate to VPH sorption, and resulting effects on the VPH mobility and bioavailability. A further improved overall data fit (shown as red and green lines in Fig. 3 & 4) could, however, be obtained, with second order growth rate adaptations for each soil type (as detailed in Table S8), in particular by increasing second-order growth rates for the soil & AC column by up to a factor of 10. Higher second-order growth rates for the activated carbon amended soil as compared to the unamended soil would imply that VPH degrading microorganisms in the soil & AC column had higher specific substrate affinity42,43 than those predominating in the unamended soil, which aligns with the observed soil microbial community composition changes discussed above. The remaining discrepancies between measured data and model simulations generally also show a slightly overestimated VPH removal for the initial experimental period, when the VPH loading of the columns was high, and a slightly underestimated VPH removal from about day 120 onwards (Fig. 3a), which also suggest that the soil microbial communities can adapt and be effective at removing VPHs when these compounds are present at low concentrations. Temporal changes in soil microbial community composition with impacts on second-order growth rates are not currently considered by the model.

Conclusions

Carefully designed sorbent amendments can stabilize the pollutant biofiltration process in soil by preventing rapid pollutant breakthrough during periods of biological inefficiency (i.e. during the lag phase), and by spreading and smoothing out pollutant concentration peaks which may otherwise overload the soil's pollutant removal capacity (i.e. when inorganic nutrient scarcity limits further growth of pollutant degrading biomass). An activated carbon amended, near-surface soil layer could thus be a beneficial pollution control measure at VPH contaminated sites with irregular pollutant releases, for example above a polluted aquifer with a fluctuating water table.44 Similarly, biochar or activated carbon amendments can enhance the robustness of other biofiltration designs such as stormwater biofiltration in retention ponds and soak-aways.45 The suggestion that Nocardioidaceae predominate VPH degradation in the presence of sorbents like activated carbon, which strongly bind VPHs, while Pseudomonadaceae predominate the VPH degradation at higher bioavailable concentration levels is of broad interest to researchers studying the microbial ecology of petroleum biodegradation in soil.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

George Mangse was supported by an Overseas Scholarship from the Petroleum Technology Development Fund. A Fellowship award of the Alexander von Humboldt Foundation to David Werner enabled collaboration between Newcastle University and the Eberhard Karls University of Tübingen. We thank Ian Head and Neil Gray for helpful comments on the interpretation of the microbiological data.

References

  1. Environment Agency, Oil pollution, http://www.environment-agency.gov.uk/research/library/position/41233.aspx, accessed 06.05, 2010.
  2. R. De Búrca, H. Morgan, J. Jeffries and N. Earl, Soil Guideline Values for toluene in soil Report Science Report SC050021/toluene SGV, 2009 Search PubMed .
  3. G. Pasteris, D. Werner, K. Kaufmann and P. Hohener, Environ. Sci. Technol., 2002, 36, 30–39 CrossRef CAS PubMed .
  4. N. Dakhel, G. Pasteris, D. Werner and P. Höhener, Environ. Sci. Technol., 2003, 37, 2127–2133 CrossRef CAS PubMed .
  5. H. Luo, P. R. Dahlen, P. C. Johnson and T. Peargin, Environ. Sci. Technol., 2013, 47, 1977–1984 CrossRef CAS PubMed .
  6. T. H. Wiedemeier, H. S. Rifai, C. J. Newell and J. T. Wilson, Natural attenuation of fuels and chlorinated solvents in the subsurface, Wiley, New York, 1999 Search PubMed .
  7. P. Höhener, C. Duwig, G. Pasteris, K. Kaufmann, N. Dakhel and H. Harms, J. Contam. Hydrol., 2003, 66, 93–115 CrossRef .
  8. D. W. Ostendorf, T. H. Schoenberg, E. S. Hinlein and S. C. Long, Environ. Sci. Technol., 2007, 41, 2343–2349 CrossRef CAS PubMed .
  9. N. C. Choi, J. W. Choi, S. B. Kim and D. J. Kim, Appl. Microbiol. Biotechnol., 2008, 81, 135–141 CrossRef CAS PubMed .
  10. P. Höhener, N. Dakhel, M. Christophersen, M. Broholm and P. Kjeldsen, J. Contam. Hydrol., 2006, 88, 337–358 CrossRef PubMed .
  11. A. Elazhari-Ali, A. K. Singh, R. J. Davenport, I. M. Head and D. Werner, Environ. Pollut., 2013, 173, 125–132 CrossRef CAS PubMed .
  12. N. L. Capiro, M. L. B. Da Silva, B. P. Stafford, W. G. Rixey and P. J. J. Alvarez, Environ. Microbiol., 2008, 10, 2236–2244 CrossRef CAS PubMed .
  13. M. K. Chang, T. C. Voice and C. S. Criddle, Biotechnol. Bioeng., 1993, 41, 1057–1065 CrossRef CAS PubMed .
  14. M. A. Dinamarca, A. Ruiz-Manzano and F. Rojo, J. Bacteriol., 2002, 184, 3785–3793 CrossRef CAS PubMed .
  15. P. Meynet, S. E. Hale, R. J. Davenport, G. Cornelissen, G. D. Breedveld and D. Werner, Environ. Sci. Technol., 2012, 46, 5057–5066 CrossRef CAS PubMed .
  16. P. Meynet, E. Moliterni, R. J. Davenport, W. T. Sloan, J. V. Camacho and D. Werner, Soil Biol. Biochem., 2014, 68, 20–30 CrossRef CAS .
  17. L. Jakob, T. Hartnik, T. Henriksen, M. Elmquist, R. C. Brändli, S. E. Hale and G. Cornelissen, Chemosphere, 2012, 88, 699–705 CrossRef CAS PubMed .
  18. U. Ogbonnaya and K. Semple, Agronomy, 2013, 3, 349 CrossRef CAS .
  19. J. Tang, W. Zhu, R. Kookana and A. Katayama, J. Biosci. Bioeng., 2013, 116, 653–659 CrossRef CAS PubMed .
  20. S. E. Hale and D. Werner, in Biochar for Environmental Management: Science, Technology and Implementation, ed. J. Lehmann and S. Joseph, Routledge, Milton Park, UK, 2nd edn, 2015, pp. 625–654 Search PubMed .
  21. K. M. Bushnaf, S. Puricelli, S. Saponaro and D. Werner, J. Contam. Hydrol., 2011, 126, 208–215 CrossRef CAS PubMed .
  22. C. J. Englert, E. J. Kenzie and J. Dragun, Bioremediation of Petroleum Products in Soil, Lewis publisher, Chelsea, MI 48118, 1993 Search PubMed .
  23. F. Solano-Serena, R. Marchal, M. Ropars, J. M. Lebeault and J. P. Vandecasteele, J. Appl. Microbiol., 1999, 86, 1008–1016 CrossRef CAS PubMed .
  24. K. H. Baker and D. S. Herson, Bioremediation, McGraw-Hill, New York, 1994 Search PubMed .
  25. H. M. S. K. Herath, M. Camps-Arbestain and M. Hedley, Geoderma, 2013, 209–210, 188–197 CrossRef CAS .
  26. D. Werner and P. Höhener, Environ. Sci. Technol., 2002, 36, 1592–1599 CrossRef CAS PubMed .
  27. L. Wilhelm, G. A. Singer, F. Christina, T. J. Battin and K. Besemer, ISME J., 2013, 7, 1651–1660 CrossRef CAS PubMed .
  28. R. M. Atlas and R. Bartha, Adv. Microb. Ecol., 1992, 12, 287–338 CAS .
  29. S. J. Assinder and P. A. Williams, Adv. Microb. Physiol., 1990, 31, 1–69 CrossRef CAS PubMed .
  30. F. Chaillan, A. Le Fleche, E. Bury, Y. Phantavong, P. Grimont, A. Saliot and J. Oudot, Res. Microbiol., 2004, 45, 629–634 Search PubMed .
  31. M. Deng, J. Li, F. Liang, M. Yi, X. Xu, J. Yuan, J. Peng, C. Wu and J. Wang, Mar. Pollut. Bull., 2014, 83, 79–86 CrossRef CAS PubMed .
  32. A. Fahy, T. J. McGenity, K. N. Timmis and A. S. Ball, FEMS Microbiol. Ecol., 2006, 58, 260–270 CrossRef CAS PubMed .
  33. G. D. Gojgic-Cvijovic, J. S. Milic, T. M. Solevic, V. P. Beskoski, M. V. Ilic, L. S. Diokic, T. M. Narancic and M. M. Vrvic, Environ. Pollut., 2011, 159, 616–622 CrossRef PubMed .
  34. N. Hamamura and D. J. Arp, FEMS Microbiol. Lett., 2000, 186, 21–26 CrossRef CAS PubMed .
  35. N. Hamamura, S. H. Olson, D. M. Ward and W. P. Inskeep, Appl. Environ. Microbiol., 2006, 72, 6316–6324 CrossRef CAS PubMed .
  36. S. Meyer, R. Moser, A. Neef, U. Stahl and P. Kampfer, Microbiology, 1999, 145, 1731–1741 CrossRef CAS PubMed .
  37. S. Mukherjee, N. K. Bardolui, S. Karim, V. V. Patnaik, R. K. Nandy and P. K. Bag, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., 2010, 45, 1048–1053 CrossRef CAS PubMed .
  38. N. J. Palleroni, D. H. Peiper and E. R. B. Moore, in Handbook of Hydrocarbon and Lipid Microbiology, ed. K. N. Timmis, Springer-Verlag, Berlin, Heidelberg, 2010, pp. 1787–1798 Search PubMed .
  39. R. E. Parales, J. L. Ditty and C. S. Harwood, Appl. Environ. Microbiol., 2000, 66, 4098–4104 CrossRef CAS PubMed .
  40. T. H. Bell, E. Yergeau, C. Maynard, D. Juck, L. G. Whyte and C. W. Greer, ISME J., 2013, 7, 1200–1210 CrossRef CAS PubMed .
  41. T. H. Bell, E. Yergeau, C. Martineau, D. Juck, L. G. Whyte and C. W. Greer, Appl. Environ. Microbiol., 2011, 77, 4163–4171 CrossRef CAS PubMed .
  42. F. P. Healey, Microb. Ecol., 1980, 5, 281–286 CrossRef CAS PubMed .
  43. K. Kovárová-Kovar and T. Egli, Microbiol. Mol. Biol. Rev., 1998, 62, 646–666 Search PubMed .
  44. D. Werner and P. Höhener, The influence of water table fluctuations on the volatilization of contaminants from groundwater, Groundwater Quality: Natural and Enhanced Restoration of Groundwater Pollution (Proceedings of the Groundwater Quality 2001 Conference held at Sheffield, UK, June 2001), Sheffield, UK, 2002, IAHS Publ. No. 275. Search PubMed .
  45. B. A. Ulrich, E. A. Im, D. Werner and C. P. Higgins, Environ. Sci. Technol., 2015, 49, 6222–6230 CrossRef CAS PubMed .

Footnotes

Electronic supplementary information (ESI) available. See DOI: 10.1039/c7em00309a
These authors contributed equally to the work.

This journal is © The Royal Society of Chemistry 2017