Thermal treatment and leaching of biochar alleviates plant growth inhibition from mobile organic compounds

Study species

The herbaceous species Lolium multiflorum Lam. (annual ryegrass: hereafter ryegrass) and Trifolium repens L. (white clover: hereafter clover) are commonly used as temperate forage crops, for erosion control (Ledgard & Steele, 1992), and are early colonizers of temperate sites following fire events (Keeley et al., 1981; Milberg, 1995). Ryegrass is a fast growing, nutrient-demanding annual dicot that can tolerate a range of soil conditions. Clover is a slow-growing perennial monocot that has low nutrient requirements as its associated root symbionts can biologically fix atmospheric nitrogen (Ledgard & Steele, 1992). Clover in previous studies has shown increased nitrogen fixation in soils amended with biochar (Quilliam, DeLuca & Jones, 2013). Because of this contrast in life histories, resource use, and growth forms, we expected that changes to soil properties would have differential effects on growth and performance in these two species. Because larger effects of biochar on growth have been reported for annuals over perennials (Biederman & Harpole, 2013), and since the strongest negative effects in experiment 1 were detected in L. multiflorum, it was selected to test if biochar post-treatments could enhance plant responses in experiment 2.

Experiment 1: effects of amount, type, and application method

Plants were grown in 11.33 cm² surface area growth containers (Ray Leach SC-10 cone-tainers; Stuewe and Sons, Tangent, OR, USA) 21 cm depth, 164 mL volume, with a fiberglass screen placed at the base of the container to reduce soil loss in soil amended with three sawdust-derived biochars (Table 1): 1) sugar maple sawdust pyrolyzed in a batch unit at 475 °C (MB) (∼1.5 h total pyrolysis time), 2) spruce sawdust pyrolyzed in a batch unit at 422 °C (SB) (∼1.5 h total pyrolysis time), and 3) 80% sugar maple, 20% yellow birch sawdust pyrolyzed in a flow-through pyrolysis unit at 575 °C (MFT) (∼0.5 h total pyrolysis time). The batch unit was purged with nitrogen gas, and the flow-through unit consisted of a long feed-screw that resulted in tight packing of feedstock to limit oxygen access during pyrolysis. Biochars were applied at 5, 10, 20, and 50 t/ha either mixed into the soil or applied to the surface; the control treatment consisted of soil without biochar. These biochars were selected for the present study as they are expected to be used in industrial scale applications and their effects on temperate plant and soil functioning have been the subject of recent study (Sackett et al., 2015; Noyce et al., 2015; Noyce et al., 2016, Mitchell et al., 2015). Five replicates were used for the biochar treatments and 10 replicates for were used for the control treatments (to increase statistical power for control vs treatment contrasts). In total, 4 addition rates (5, 10, 20, and 50 t/ha) × 3 biochars (MB, SB, MFT) × 2 application methods (mixed, top-dressed) × 5 replicates), + (10 control plants (no biochar, field soil) × 2 species (clover, ryegrass), resulted in n = 260 plants. Biochar use is expected to target acidic, coarse-textured soils of low nutrient status. Accordingly, the soil used in the experiment was a drystric cambisolic field soil (or drystric brunisol based on the Canadian System of Soil Classification (1998)), that was sandy-loam in texture (Table 2), collected in the summer of 2011 from the uppermost mineral layer (< 10 cm) of a managed forest at Haliburton Forest and Wildlife Reserve, Haliburton, ON, Canada (45°15′N, 78°34′W), and re-used from control treatments in a previous experiment. All plants were supplemented with 0.082 g of slow-release NPK fertilizer representing relatively high inputs found in disturbed temperate herbaceous systems (Garbutt & Bazzaz, 1987): addition rates are 130 kg/ha (N), 5 kg/ha (P), 48 kg/ha (K), respectively. Germination was initiated by placing seeds on greenhouse benches in water-saturated vermiculite. One seedling was transplanted into each prepared treatment pot five days after germination. Fertilizer was applied following planting. A randomized block design was employed using five blocks in which plants were randomly placed, with each treatment combination was represented once in each block. Supplemental lighting was applied to maintain a 16 h photoperiod. The experimental growth period was 28 days (December 20, 2012–January 17, 2013) and 41 days (December 20, 2012–January 30, 2013) for ryegrass and clover, respectively. Average daily greenhouse temperature throughout the experiment was 16.5 °C with average daily maximums and minimums of 20 and 16 °C, respectively. All plants were top watered semi-daily (daily to every other day) to saturation and allowed to drain.

Experiment 2: modulation of effects through leaching and heating treatments

To test the effect of pre-application biochar heating and leaching on plant performance, ryegrass was grown in pots (as above) with additions of MFT biochar processed to remove mobile compounds before additions. Because, there was no difference in plant performance between biochars in experiment 1, suggesting similar compounds responsible for the observed phytotoxicity, only MFT was selected for experiment 2 since it was most available at the time of experimental set-up. Two leaching treatments and three heating treatments were applied to biochars before being applied as both mixed and top-dressings at a dose of 10 t/ha (n = 8). We leached biochar with 1:1 (v:v) mixture of biochar to de-ionized water for either 0.5 h (WW-0.5) or 24 h (WW-24) on an oscillating table. Biochar-water slurries were then suction filtered using Whatman #4 filter paper, and the filtrate captured. Filtrate from the leachings was collected and 5.0 mL of filtrate was applied as a separate treatment to plants (n = 8) one day following planting (Leach-0.5, Leach-24). Leached biochars were then rinsed with 400 mL of de-ionized water, suction filtered, and applied to their respective treatments. Biochars were placed in a convection oven to be heat treated for 24 h at 50 (Heat-50), 100 (Heat-100), and 150 °C (Heat-150), cooled, and added to soil. To assess the effect of leaching and heating, un-processed biochar treatments (mixed and top-dressing) were also established (n = 8), as well as a control treatment containing no biochar (n = 10). In total, 2 leaching treatments + 3 heating treatments + 1 un-processed biochar × 2 application methods × 8 replicates + 2 leachate additions × 8 replicates + 10 controls, resulted in n = 122 plants. Soil used in the experiment was the same field soil as used in experiment 1 (Table 2). Experimental growth period was 28 days (January 28–February 26, 2013). Average daily greenhouse temperature was 18 °C, with daily average maximum and minimum temperatures of 21 and 15 °C, respectively. All plants were top watered semi-daily to saturation.

Biochar and soil characterization

Biochar was characterized (Table 1) as follows: electrical conductivity (EC) and pH were measured in a 1:20 (w:v) biochar to H₂O solution using pH and EC probes (Rajkovich et al., 2012). Particle size distribution was determined from dry sieving following the ASTM D2862 method (ASTM, 1999). Total moisture content, ash content, volatile matter, and organic matter content were determined by oven drying at 105 °C, by muffle furnace combustion at 750 °C, by heating at 950 °C in sealed containers for 7 min, and through loss on ignition for 0.5 h in a muffle furnace at 500 °C by following ASTMD1762-84 (ASTM, 2007). Total carbon and nitrogen were quantified using combustion analysis using an Elementar VarioMax (Elementar Analysensteme GmbH, Hanau, Germany). Total P and cations (K⁺, Ca⁺, and Mg²⁺) for MB and SB were obtained using a sulphuric acid digest and Mehlich III extraction, respectively, then analyzed by ICP-OES (SPECTRO Analytical Instruments GmbH, Kleve, Germany) (see Sackett et al., 2015).

Soil characterization (Table 2) was conducted at the Agriculture and Food Laboratory, University of Guelph, Guelph, ON.

Plant growth measurements

For both experiments, above- and below-ground plant biomass and leaf area were measured at the end of the experiments. Because leaf area indicates plant carbon gain, nutrient uptake, and physiological performance, and has demonstrated responsiveness to biochar additions (Graber et al., 2010), we measured leaf area using a Li-3100C leaf area scanner with a resolution of 1 mm² (LiCor Biosciences, Lincoln, Nebraska, USA). Below-ground biomass was separated from soil and biochar media by dry sieving followed by gentle washing of roots with water. Above-ground and below-ground biomass, including scanned leaves, was dried for 48 h at 65 °C and weighed. Since increased nutrient uptake, nitrogenase activity, and nodulation has been reported in legumes amended with biochar (Rondon et al., 2007; Quilliam, DeLuca & Jones, 2013), before drying we counted nodules on clover roots under a dissection microscope. Only nodules with leghemoglobin were scored (visible red colour).

SPME–GC-MS of volatile and leachable compounds

We followed Rombolà et al. (2015) for qualitative analysis of SB and MB solid char and their water leachates, with minor modifications. Briefly, for headspace (HS)-SPME of solid biochars, 0.5 grams of sample was placed in a 10 mL headspace vial and spiked with 2.0 mL of 1 ppm O-eugenol as an internal standard. HS vials were placed on a heating plate at 150 °C for 30 min while a Carboxen-PDMS fiber was inserted into the top 1 mL of the headspace vial on a Agilent 7890A GC system. Following heating the fiber was inserted into the injector and analytes thermally desorbed at 250 °C for 10 min prior to direct injection. GC analysis was carried out using a DB-WAX column (20 m length, 0.20 μm width, 0.10 mm i.d) following the thermal program in Rombolà et al. (2015): 80 °C for 5 min, then 10 °C/min to 250 °C for a total run time of 28 min. MS analysis was carried out on a 5975C Agilent MS with peak identification conducted using the NIST reference spectral library. Detection was made under electron ionization at 60 m/z and an acquisition of 1 scan/s. Hits with over 70% probability and relative match factors higher than 800 were considered; peaks of internal standards were confirmed.

Direct injection (DI)-SPME of leachates was performed on 3 mL of dionized water leachates (following washing method used for MFT) with 1.0 mL of 2 M KH₂PO₄ buffer. This sample was spiked with 2.5 mL of 1 ppm O-eugenol, and 2.5 mL of 5 ppm 2-ethyl butyric acid as internal standards, and then placed into 10 mL headspace vials. The Carboxen-PDMS fiber was directly inserted into the solution under magnetic stirring for 30 min at 250 °C. GC-MS analysis was performed as above. Quantification of analytes was not attempted due to the un-reliability of SPME in producing accurate calibration curves with internal standards, a problem also noted in Spokas et al. (2011). All GC-MS analysis were conducted at the Teaching and Research in Analytical Chemistry and Environmental Sciences (TRACES) Centre, University of Toronto, Scarborough (1065 Military Trail, Toronto, ON, M1C 14A). SPME-GC-MS was conducted on SB and MB only since MFT was no longer available at the time of quantification.

Statistical analysis

Both experiments had treatment structures that consist partially of a full set of factorial combinations, have multiple controls, and test other treatments of interest (i.e. have an augmented factorial design); we therefore used multiple contrasts to test our hypotheses (following Schaarschmidt & Vaas, 2009). This allows the pair-wise comparison of both individual treatments (e.g., MFT biochars vs control) and groups of treatments (e.g, all biochars at 5 t/ha vs control). The test statistic for each contrast was calculated from: $$\text{T}=\left(\Sigma c_i{\widehat{u}}_i\right)/\left(\sigma \Sigma c_i/n_i\right)$$

Where û_i is the mean estimate derived from the linear model described in Bretz, Hothorn & Westfall (2002), and c_i is the contrast coefficient, and σ represents the residual error. Contrast coefficients are shown in Tables S1–S3. Simultaneous confidence intervals (95%) were derived from: $$\left[\Sigma c_i{\widehat{u}}_i\pm \left(q_1-\alpha ,M,R\right)\alpha \left(\sqrt{\Sigma c_i^2/n_i}\right)\right]$$

Where the critical value, (q₁ − α, M, R), is calculated from the multivariate t-distribution with dimension M and correlation matrix R (see Bretz, Genz & Hothorn, 2001) for critical value determination. Simultaneous confidence intervals were used to display the magnitude and direction of mean comparisons. Analyses were done using the “multcomp” package (Bretz, Hothorn & Westfall, 2002; Hothorn et al., 2008) in the statistical software R version 3.1.0 (R Core Team, 2012). Many biochar studies have complex treatment structures and treatments of interest: we expect wide use of multiple contrast tests in future biochar research as an effective and informative analysis tool, we have therefore included the R script as a supplementary file for use as a reference.

To test the effect of biochar addition rate on plant performance traits we performed linear regressions with above- and below-ground biomass and leaf area as response variables and with biochar dosage as the independent variable. We log-transformed response variables to represent effect sizes metrics reported in recent meta-analyses of plant growth responses to biochar (Liu et al., 2013; Biederman & Harpole, 2013; Thomas & Gale, 2015) following: RR = ln (B/C), where RR is the response ratio metric, B is the mean biomass of biochar treated plants, and C is the mean biomass of control plants without biochar. Regressions were performed for each of the three biochars used in the experiment and were not separated by application method (top-dressing vs complete mixtures). Analyses were conducted in the statistical software R version 3.1.0 (R Core Team, 2012).

Experiment 1: negative and null effects on plant growth regardless of biochar type, dosage, or application method

Application of biochar resulted in null or negative biomass responses in ryegrass and clover regardless of biochar type, dosage, or application method. The average of the three mean responses of all treatments using MFT, SB, and MB revealed decreased ryegrass aboveground biomass relative to controls by 25%, leaf area by 17%, and belowground biomass by 35% (Fig. 1; Table 3). All three individual biochar types negatively impacted all growth traits examined in ryegrass (Fig. 1), but had no significant effect on clover growth traits (Table 3). Nodulation in clover did not differ across treatments (Table S2). All performance traits in ryegrass, but not clover, were inhibited by biochar at all dosages and this inhibition decreased with biochar addition rate. Linear regressions detected significant negative relationships of above- and below-ground biomass and leaf area with increasing biochar addition rate for all three biochars used in almost all comparisons (Fig. 2). The mean of all the biochar treatments at 50 t/ha showed reduced ryegrass aboveground biomass by 34%, leaf area by 28%, and belowground biomass by 43% (Fig. 1; Table 3). Mixing biochar directly into the soil was significantly less inhibiting than direct surface applications for all growth parameters in ryegrass (Fig. 1; Table 3).

Experiment 2: heating and leaching biochar prior to application alleviates growth inhibition in ryegrass

Ryegrass grown in biochar either water leached or convection heated prior to application performed significantly better than ryegrass grown in un-processed biochar (Fig. 3). Un-processed biochar reduced ryegrass aboveground biomass by 24% (Fig. 3A), leaf area by 22%, and belowground biomass by 28% (Fig. 3B; Table 4). There was no difference between mixed and top-dressed applications of the un-processed biochar. Post-treating the biochar by water leaching for 24 h or heat treatment ≥ 100 °C significantly increased aboveground biomass as compared to un-processed biochar (Fig. 3A). Biochar that was water-leached for 24 h increased aboveground biomass by 34%, and belowground biomass by 22% compared to un-processed biochar (Table 4). Convection heating of biochar for 24 h at 100 and 150 °C increased leaf area by 20 and 23%, respectively (Fig. 3; Table 4). However, aboveground biomass and leaf-area of ryegrass grown with any processed biochars did not differ significantly from ryegrass grown in control (without biochar) soil (Table 4).

The water-based filtrate from biochar leaching applied immediately following transplantation significantly decreased growth of ryegrass as compared to plants grown in control soil (Table 4). Filtrate applied from the 24 h leaching most strongly inhibited growth, decreasing aboveground biomass by 40%, leaf area by 27%, and belowground biomass by 36% (Fig. 4; Table 4). Filtrate from the 30-min water leaching was less inhibiting, decreasing leaf area by 18%. The significant decrease in ryegrass growth traits due to the application of filtrate was similar in magnitude to the application of un-processed biochar (Fig. 4; Table 3).

Mobile organic compounds detected with SPME-GC-MS

SPME-GC-MS of solid chars and leachates revealed primarily short-chained carboxylic acids, phenols, hydrocarbons; with notable differences between chars (Fig. 5). HS-SPME of MB qualified acetone, acetonitrile, acetic acid, 2-ethylbutyric acid, and butyric acid (Fig. 5A). HS-SPME of SB revealed acetic acid, benzene, decane, methyl isocyanide, ethylbenzene, dodecane, and 6-methyoxybenzofuran (Fig. 5B). DI-SPME of MB leachates revealed acetic acid, butyric acid, valeric acid, phenol, 2,4-di-tert-butylphenol and benzoic acid (Fig. 5C). DI-SPME of SB leachates similarly detected acetic acid, butyric acid, 2,4-di-tert-butylphenol and benzoic acid; however, oxime, and 5-methyoxybenzofuran were also detected (Fig. 5D).

Leached compounds from biochar inhibit plant growth

Effluent from biochars leached with de-ionized water decreased ryegrass growth by up to 40% (Table 4), similar to the decreases in growth shown by treatments with the un-processed biochar. This result confirms the growth-inhibitory nature of the aqueous fraction of pyrolysis-produced bio-oils and is consistent with others that have found inhibitory effects from biochar-leached chemical compounds on plant germination (Rogovska et al., 2012; Buss & Mašek, 2014; Kołtowski & Oleszczuk, 2015) and microbial communities (Lehmann et al., 2011). Here we report inhibition in a variety of performance traits, rather than germination, from less intense leachings than in (Rogovska et al., 2012), and from more typical neutral to basic chars (pH = 6.5–7.4) than the acidic char studied by Buss & Mašek (2014) (pH = 3.64). Since we observed a positive relationship between leaching duration and compound concentration, we recommend that future applications of biochar use pyrolysis post-treatments that leach for longer durations. We watered our plants semi-daily to saturation, potentially leaching compounds quicker than would occur naturally. Phytotoxicity to biochar in areas of low rainfall, or otherwise under growth conditions with low watering frequency might indeed be greater.

Feedstock characteristics and the conditions of pyrolysis strongly influence the type and amount of mobile organic compounds from biochars (Quilliam et al., 2013; Hale et al., 2012), and subsequently, the toxicity of leachates is likely to also be related to system conditions, in particular soil characteristics and microbial composition. Spokas et al. (2011) report lower concentrations of volatile organic compounds in biochars produced from open-pit, kiln, slow, and steam activated pyrolysis compared to those produced via gasification and fast pyrolysis. Mobile organic compounds in aqueous leachates from poultry litter, but not corn stalk biochars, resulted in germination inhibition of cress seeds (Lepidium sativum) in Rombolà et al. (2015). Indeed, spruce and maple charcoals had different volatile matter content and compositions of mobile organic compounds in our study. The effects of mobile organic compounds on soil microorganisms will further influence plant responses (Dutta et al., 2016), with recent research showing shifts in microbial communities from biochar are influenced by soil properties (Noyce et al., 2015). Soil physical and chemical properties such as porosity, aeration, leachability, moisture content, and pH are also likely to play a critical role in determining hormetic effects of leached mobile organic compounds from biochar, and should be the focus of much further study.

Qualification of phytotoxic compounds and the potential for hormesis

The increasingly reported null and neutral responses to biochar necessitates quick and reliable analysis of the compounds responsible. SPME-GC-MS promises to serve this role and is becoming the standard for qualification and even semi-quantification of volatile and leachable organic compounds in biochar (Spokas et al., 2011; Buss et al., 2015; Rombolà et al., 2015). We detected similar low molecular weight compounds to Rombolà et al. (2015) and Buss et al. (2015) with slight differences in GC-MS phases and procedures. However, we were unable to analyze the other maple biochar produced in a different pyrolysis unit and slightly different feedstock composition for mobile compounds, which potentially limits the array of compounds detected, as well as inference on the influence of pyrolysis conditions on the formation of mobile organic compounds. Buss et al. (2015) report concentrations of phenols and organic acids from moderate temperature pyrolysis of softwood chars in concentrations slightly lower (∼22.4 mgL⁻¹ and ∼52.2 mgL⁻¹, respectively) than concentrations that have inhibited root elongation in prior studies (Lynch, 1980; Feng et al., 1996).

Pyroligneous acids, or “wood vinegars,” have promoted plant growth and establishment in several studies (Kadota, Hirano & Imizu, 2002; Du et al., 1998). Acetic acid is the primary component of the chemical fraction in wood vinegars which have a long history of use as plant growth stimulators (Mu, Uehara & Furuno, 2004; Mungkunkamchao et al., 2013). Mu, Uehara & Furuno (2003) showed strong hormetic responses in seed germination of four species: Lactuca sativa, Rorippa nasturtium-aquaticum, Cryptotaenia japonica, Chrysanthemum corronarium. Strong inhibition of germination and radical growth was observed at high concentrations, while dilutions stimulated germination and growth, with obvious species-specific responses. Root development was increased, but not shoot development, when diluted wood vinegar was added to Pyrus pyrifolia cuttings (Kadota, Hirano & Imizu, 2002). Similarly, radicle length was increased when aqueous extracts of one of several charcoals was added to corn seeds in Rogovska et al. (2012). Accurate quantification of volatile and leachable organic compounds in biochar is necessary to determine the type of hormetic plant responses; however, current SPME-GC-MS techniques offer only semi-quantification with noted difficulties (Spokas et al., 2011; Dutta et al., 2016). Germination tests could be utilized to evaluate hormetic effects from mobile organic compounds in biochar: principally, to identify compounds responsible, possible synergistic effects, and species-specific toxicity thresholds.