Comparative characterization of biochars produced at three selected pyrolysis temperatures from common woody and herbaceous waste streams

Raw material selection

Six biochars were prepared using pine sawdust and pea straw as feedstocks. Sawdust was obtained from pine (Pinus radiata), grown in several plantations ranging between Eurobin (Victoria- 36°38′18.9″S 146°51′06.2″E) to Tumut (New South Wales, Australia - 35°18′58.6″S 148°13′51.7″E) plantations. Pea straw (Pisum sativum) was acquired from a wholesaler (Peninsula Hay) situated in the Mornington region of South East Victoria (38°24′12.9″S 144°58′34.6″E). In this region the pea plant is used to fix nitrogen in pastures and later harvested for use as feed or mulch.

Pyrolysis of raw materials

Biochar was produced by tightly packing 400 g of a single feedstock into a 1 L internal volume (Radius–7 cm; Height–6.5 cm) stainless steel cylindrical vessel with a spring clamped lid which exerted a small downward force strong enough to prevent atmospheric exchange yet still allow evolved gases to escape under positive pressure. No inert gases were employed as oxygen was prevented from entering the vessel by the lid, any remaining oxygen existing in the vessel was either exhausted during heating or expelled through expansion during the temperature ramping process. Therefore, inside the vessel was considered an oxygen limited environment. The vessels were then placed in a furnace and the temperature ramped at 8.3 °C/min to a respective 350 °C, 500 °C or 750 °C, followed by a 1 h dwell time. These temperatures were selected as a gradient and are spread across the upper and lower as well as median thresholds for slow pyrolysis. After pyrolysis, each vessel was placed in the draft of a fume hood to allow an hour to cool before opening, to prevent ignition. The above process was carried out four times. All biochars were passed through a 1 mm sieve, homogenized and stored in polypropylene containers under standard lab conditions until analysis. Biochars were coded P (Pine Sawdust) and S (Pea Straw), and temperature groups (P350, P500, P750, S350, S500 and S750).

Characterization of biochars

Statistics

Descriptive statistics, Two-way factorial ANOVA and Pearson Correlation were carried out on IBM’s SPSS Statistics 22 package (Armonk, NY, USA). Two- way ANOVA was the key tool in verifying significant trends between biochar parameters governed by pyrolysis temperature and differences between feedstock type. Univariate factorial analysis allowed the identification of the main effect responsible for any trends observed, differentiated as temperature, feedstock or interaction. Significant results are displayed in the format (F_1,6 = X, p < 0.05), where the p value is alongside the F value. F-crit values can be ascertained using F tables and the subscript numbers, the first being the degrees of freedom followed by the number of sample groups for that parameter. Pearson Correlation results are displayed as follows (R = X, n = X, p = X), where R is the correlation factor, n denotes the number of groups of samples sampled and the last figure corresponds to the p value.

Chemical and physical characterization

Temperature was found to be the main effect influencing yield (F_2,18 = 12.1, p < 0.05), with yields decreasing at higher pyrolysis temperatures for both, straw and pine feedstocks. Straw exhibited greater yields than pine at both 500 °C and 750 °C (25% and 23% compared to 21% and 20% respectively), while at 350 °C pine (34%) produced a higher yield than straw (29%). This is likely due to the interplay of temperature and the two feedstocks, which differ in water, lignin, cellulose and hemicellulose composition. Feedstock and temperature have been identified as the most influential parameters in the decomposition of woody and herbaceous biomasses to produce biochar (Zhao et al., 2013; Benavente et al., 2018). This is due to the variation in each feedstock with respect to their content of the biopolymers; hemicellulose, cellulose and lignin, all of which degrade at different temperature ranges (Qian et al., 2015; Kambo & Dutta, 2015; Yeo et al., 2017). Comparatively the rigid structure required by trees results in a higher proportion of lignin in softwoods than in herbaceous grasses, while grasses are more cellulosic (Das & Sarmah, 2015; Azargohar et al., 2013). Hemicellulose is the easiest degraded of the three major components, with complete degradation starting at 330 °C (Yeo et al., 2017; Buss et al., 2015). The greatest proportion of cellulose degradation occurs at temperatures above 427 °C, though degradation can begin at lower temperatures, generating much volatile matter as carbon-oxygen and carbon-hydrogen bonds are broken (Buss et al., 2015). Lignin is the most recalcitrant of these three major components, with complete degradation evident only after temperatures exceeding 607 °C. This is due to lignin’s structure consisting of multiple ether linkages and functional groups such as hydroxyl and methoxy (Yeo et al., 2017).

A difference in surface morphology was observed using SEM in the form of cellular structure between the feedstocks, as well as, increased fracturing of structure with increased pyrolysis temperature (Fig. 1). The cells seen in pine biochars were longer and more cylindrical than the short cuboid cells noted in straw based biochars, and the pores visible on the surface of all biochars were similar to those reported in previous literature (Shaaban et al., 2014).

Bulk density was similar across feedstocks, ranging between 0.12–0.17 g/cm³. Pine at 350 °C had the lowest bulk density, however at 750 °C both biochar types were matched in density (Table 2). This would have implications when biochars are used as soil conditioner. Thus, bulk density is of utmost importance to rainfall infiltration. Moreover, a decrease in bulk density would have ramifications, increasing soil porosity and soil aeration, and, potentially leading to a positive effect on microbial respiration.

Surface area increased in both feedstock types with higher pyrolysis temperature and large surface area differences were observed between feedstock types (Table 2). Feedstock (F_1,11 = 529.0, p < 0.05), temperature (F_2,11 = 471.6, p < 0.05) and their interactions (F_2,11 = 132.5, p < 0.05) were significant factors with respects to biochar surface area and this is consistent with similar studies, carried out in other lignocellulosic wastes (Das & Sarmah, 2015). Pine biochars had higher surface areas than straw biochars in the 500 °C (278.0 ± 4. cm²/g) and 750 °C (397.3 ± 4.1 cm²/g) experiments, however the depressed values at 350 °C (16.3 ± 5.2 cm²/g) were suggested to be due to the lower temperatures resulting in the underdevelopment of pores (Abdel-Fattah et al., 2015) and clogging of pores with tars which could not volatilize (Das & Sarmah, 2015). These results fit within the range expressed by in Table 1 for surface area.

Moisture levels ranged between 1.5 and 4.0% in biochars produced, an increase in moisture was observed with higher pyrolysis temperature (Table 2). Temperature and feedstock were each found to hold significant effects over biochar moisture levels (F_1,6 = 9.6, p < 0.05 and F_1,6 = 169.4, p < 0.05, respectively). Biochars prepared at 750 °C had the highest moisture levels, particularly P750, it is suggested that this is absorbed from the atmosphere due to the higher surface area of the material.

Temperature was found to be a main effect for VM content (F_2,6 = 42.9, p < 0.05), decreasing from 350 °C biochars to 750 °C biochars (Table 2). This is consistent with other studies (Table 1) as VM loss occurs as outgassing volatiles (Al-Wabel et al., 2013). Feedstock was also identified as a main effect (F_1,6 = 31.4, p < 0.05), exhibiting higher average VM in straw biochars than pine biochars. An interaction for both factors was present F_2,6 = 8.3, p < 0.05 suggesting an interplay between these two factors. Volatile matter is of importance to explain the microbial and plant responses following biochar addition to the soil, although this is a poorly understood interaction, due to the large amount of individual volatile compounds present in biochars (Spokas et al., 2011).

For ash, feedstock was determined to be a main effect (F_1,6 = 219.7, p < 0.05), such that ash content was significantly higher in straw biochar than pine. Ash fractions increased with higher temperature (Table 2), which was demonstrated to be a main effect by univariate factorial ANOVA (F_2,6 = 6.6, p < 0.05). This demonstrates temperature’s role in forming the ash fraction, compared to feedstocks role in defining the fraction available for maximum ash formation. Further, a statistically significant interaction between feedstock and temperature was found (F_2,6 = 5.3, p < 0.05). Ash represents the largely inorganic fraction that cannot be volatized or degraded by combustion, including potassium (K), calcium (Ca) magnesium (Mg), carbonates and heavy metals (Hmid et al., 2014). Ash compounds hinder the formation of aromatic structures that contribute greatly to fixed carbon content.

Fixed carbon was found to be higher in pine biochars than in straw biochars (Table 2). Feedstock was found to be the main effect for this difference in fixed carbon between biochars (F_1,2 = 36.9, p < 0.05) and is supported by literature (Zhao et al., 2013). FC values were slightly higher than those seen in literature (Table 1). Fixed carbon values were in agreement with those of the thermostability index. In general, higher values of these would be indicative of a longer residence time of biochar in soil.

Thermal stability increased with pyrolysis temperature for all biochars, lower thermal stability was noted for straw biochars (Table 2). This suggests that pine biochars will be more recalcitrant in the environment than straw biochars (Das, Sarmah & Bhattacharyya, 2016).

A reduction in oxygen and hydrogen containing functional groups, primarily carbonyl (1,690–1,700 cm⁻¹) and carboxyl groups (1,690–1,760 and 1,210–1,320 cm⁻¹) between 350 °C and 500 °C, was observed through FTIR (Figs. 2 and 3). In both feedstock types, FTIR suggests that at 750 °C biochars are namely comprised of C-C bonds and that most of the other functional groups and volatile components had been lost, this is similar to trends expressed in by results found for other biochars produced from woody feedstocks (Alburquerque et al., 2013; Abdel-Fattah et al., 2015). In all biochars, H and O values were comparable to literature, though slightly lower values (Table 2). Hydrogen content decreased with higher pyrolysis temperature and a difference was observed between feedstocks, with hydrogen in straw derived biochars being lower than in pine biochars (Table 2). Both feedstock and temperature significantly influenced hydrogen content (F_2,2 = 250.4, p < 0.05 and F_1,2 = 22.9, p < 0.05 respectively).

Nitrogen was not affected by higher pyrolysis temperature in straw biochars, however N increased with pyrolysis temperature for pine biochars (Table 2). Feedstock was the main factor affecting nitrogen content (F_1,2 = 24.0, p < 0.05). Similarly, feedstock was the main factor influencing percentage sulphur remaining in biochars (F_1,2 = 57.3, p < 0.05). Sulphur and nitrogen values for all biochars compared well to literature values seen in Table 1.

H:C ratio is a measure often used to discern the degree of aromatization in biochars as increases in carbon are inversely related to hydrogen through polymerization, dehydration and volatization. In these experiments H:C decreased in both biochar types with higher pyrolysis temperature, highlighting temperature as a main effect (F_2,2 = 393.1, p < 0.05) and suggesting an increase in aromatization (Table 2). The loss of H is indicative of water and surface acid functional group loss, such as hydroxyl (OH) and carboxyl (COOH) through volatilization. Higher pyrolysis temperatures results in a greater loss of VM, oxygen and hydrogen as due to depolymerisation of biopolymers and the carbonisation of the feedstock to a more recalcitrant form through decarboxylation, dehydration, de-carbonylation, de-methylation, condensation and aromatisation reactions (Das & Sarmah, 2015; Heitkötter & Marschner, 2015).

Inverse relationships were observed between moisture and H:C (R =  − 0.912, n = 6, p = 0.006); bulk density and H:C (R =  − 0.827, n = 6, p = 0.021); and bulk density and yield (R =  − 0.851, n = 6, p = 0.016). The relationships between bulk density, yield, and H:C can all be understood through mass loss. Hydrogen is lost through dehydration while percentage carbon content increases through condensation and graphitization, affecting H:C. Yield decreases in proportion to H:C via mass loss, whereas bulk density increases due to the formation of graphite like structures. These are temperature dependent relationships, though the initial feedstock does play a major role in their resilience to thermal degradation. It is intuitive that moisture levels decrease from feedstock to biochar through the loss of water as steam due to the elevated temperatures used in pyrolysis. However, the increase in moisture in finished biochars as a function of temperature is surmised to be due to the hygroscopic effect exerted by their high surface area. This is further supported by the correlations seen in BET surface area relating in a negative manner to yield (R =  − 0.824, n = 6, p = 0.022), O: C (R =  − 0.976, n = 6, p < 0.001) and VM (R =  − 0.964, n = 6, p = 0.001), all of which are characteristics which decrease with higher pyrolysis temperature. This relationship highlights the impact outgassing of VM has in the formation of pores and hence a higher surface area (Das & Sarmah, 2015).

Electrical conductivity (EC) and pH increased with higher pyrolysis temperature in all cases. Both pH and EC were found to be higher in straw biochars than in pine biochars (Fig. 4). Temperature (pH: F_2,6 = 1706.8, p < 0.05; and EC: F_2,6 = 179.5, p < 0.05), feedstock (pH: F_1,6 = 4621.7, p < 0.05; and EC: F_1,6 = 279.4, p < 0.05) and their interactions (pH: F_2,6 = 73.5, p < 0.05; and EC: F_2,6 = 103.0, p < 0.05) were found to play a significant role in pH and EC values. This suggests the differences within feedstock groups were due to pyrolysis temperature, the different values across temperature groups were due to feedstock and the extent of the difference was due to the interaction of these two factors. Increases in EC are the result of a gain in the amount of ions present through the increase in ash fraction (Alburquerque et al., 2013). Similarly, it is well-established that increasing pyrolysis temperature tends to favour the alkalinity of biochars (Yuan, Xu & Zhang, 2011). This is partly due to an increase in inorganic carbonates (Yuan, Xu & Zhang, 2011).

CEC denotes the ability of biochar to bind cations (Hmid et al., 2014), and was found to increase with higher pyrolysis temperatures and was higher in straw derived biochars (Fig. 4). Feedstock was the main factor influencing the differences between biochars (F_1,2 = 115.9, p < 0.05).

Strong positive Pearson correlations were observed between pH, ash, and CEC creating a grouping of correlated parameters (p < 0.05). These correlations were in line with expected increases caused by pyrolysis temperature and therefore increased ash content, which result in increased pH through the concentration of ions such as K, Ca, Mg and carbonates (Kuzyakov, Bogomolova & Glaser, 2014; Heitkötter & Marschner, 2015). These increases in pH elevated the biochars CEC as this parameter is pH dependent (Abdel-Fattah et al., 2015). In addition, significant positive correlations regarding sulphur (R = 0.858, n = 6, p = 0.014) and nitrogen (R = 0.868, N = 6, p = 0.013) with ash were obtained.

Contaminant analysis

Six of the eight heavy metals tested were detected in the biochars, with mercury and cadmium being below the limit of quantitation (<LQ) (2mg/kg) in all samples (Table 3). Lead, nickel and zinc were found to be below IBI guidelines in all biochars as illustrated by (Table 3). Similarly, arsenic, chromium and copper were below guideline levels in biochars produced from straw feedstocks.

Biochars produced from pine had elevated levels of arsenic, copper and chromium that were 1–2 orders of magnitude above IBI limits in all pine biochars excepting chromium at 750 °C. The presence of these three heavy metals at elevated concentrations was due to raw pine being milled alongside treated pine, in turn contaminating the feedstock with chromated copper arsenate. Chromated copper arsenate is the most commonly used wood preservative and treated woods typically contain chromium, copper and arsenic concentrations within the range 2.6–9.8 mg g⁻¹, 5.3–19.0 mg g⁻¹ and 5.2–16.3 mg g⁻¹ respectively (Jones & Quilliam, 2014). It is likely, in Victoria and elsewhere, that timber from construction waste or wood sourced from forests, would have significant concentrations of chromated copper arsenate, which would impact the quality of the biochar. If these heavy metals are allowed to enter into the environment, through biochar application, their excessive availability could have detrimental toxic effects on local wildlife or crops (Alburquerque et al., 2013; Freddo, Cai & Reid, 2012). Typically, metals measured and detected increase in concentration with higher pyrolysis temperature as found in previous work (Benavente et al., 2018). However, there are exceptions. It is likely that arsenic could be lost as volatile arsenic trioxide (As₂O₃) during pyrolysis ((Jones & Quilliam, 2014); (Helsen & Van den Bulck, 2000) as it has been demonstrated that 15–24% of arsenic concentration can be lost during incineration at 400 –800 °C (Yan et al., 2008). While it is possible for a fraction of arsenic to be liberated as the volatile arsenic trioxide in the presence of oxygen, it does not account for the magnitude of arsenic loss seen in this study within an oxygen limited environment at high temperature. A dissimilar trend with temperature was found for copper and chromate concentrations in pine and straw biochars. This result could be due to differences in the partition of heavy metals between the biochar and the tar. Differences in the molecular structure of the feedstocks and how contaminants are distributed in the matrix could have contributed to these results (Farrell, Rangott & Krull, 2013). Further, it has been demonstrated that with respects to total heavy metals the result can be dependent on the extraction method (Enders & Lehmann, 2012). Hence the term pseudo-total heavy metals must be applied, as this designates that it is the maximum that can be extracted for the method used, which is often determined by availability of lab equipment and access to materials and reagents (Beesley, Moreno-Jiménez & Gomez-Eyles, 2010). In this study the remainder of the heavy metals tended to be higher in concentration in straw derived biochars than in pine. Levels for Zn were comparable to those seen for similar feedstocks for both straw and pine biochars, the same is true of Pb in pine biochars (Srinivasan & Sarmah, 2015; Zhao et al., 2013), demonstrating the small pool of analogue studies from which to draw literature comparisons in this area. In Australia or in the state of Victoria, there are currently no specific biochar application guidelines and it is necessary to rely on IBI guidelines for heavy metal limits (Table 3). Using these criteria all straw biochars qualify as soil enhancers and all pine biochars would be unsuitable for land application due to elevated arsenic concentrations (International Biochar Initiative, 2013). It has been demonstrated that elevated levels of arsenic, chromium and copper have a negligible effect on the determination of other parameters such as EC and pH (Jones & Quilliam, 2014). Elevated metals should not interfere with ultimate analysis, proximate analysis or surface area determination.

The sixteen USEPA PAHs measured in the study were below the limit of detection (0.5 mg/kg) in most biochars with the exceptions being S500 and P350. S500 was found to contain a single PAH, 1.2 mg/kg naphthalene. Comparatively, P350 contained a total of 6 PAHs (Table 4), with 7,12-dimethylbenz(a)anthracene concentration dominating at 35.0 mg/kg. . Out of the 6 biochars created, the PAH level in one biochar (P350) is above the suggested IBI PAH limit (6 mg/kg) (International Biochar Initiative, 2013). PAHs are generated during biochar production through incomplete combustion of biomass. PAH concentration in biochars is feedstock-dependent (Wang, Wang & Herath, 2017). Naphtalene is usually the major hazardous PAH (Wang, Wang & Herath, 2017). In agreement with our results, it is well documented that PAH concentration in biochars diminishes with the temperature of pyrolysis (Wang et al., 2018). Oleszczuk, Jośko & Kuśmierz (2013) demonstrated that in environmentally relevant applications to soil (10% rate of addition), biochar’s PAH content could inhibit root growth for Lepidium sativum up to 92% compared to controls. Root growth inhibition started at concentrations of 5%. Further a significant relationship was established between total PAHs, leached from biochars using water, and the mortality of a test planktonic crustacean (Daphnia magna). These demonstrate the threat biochars containing PAHs could pose to the environment and agricultural lands if application rates appropriate to each biochar are not determined.