2.1 Biochar production

Pellets were produced with a diameter of 6 mm on a roller wheel mill (WK230, EverTec, Groß-Zimmern, Germany) from straw (Jumbo/Coop, Basel, Switzerland, 5.9% ash content, cf. S1 Table) and softwood (Allspan Spanverarbeitung GmbH, Karlsruhe, Germany, 0.4% ash content), respectively. Three batches each of straw and wood pellets were produced at different points in time from the same biomass (first batch: biochars produced at 400–600 °C in 50°-steps, second batch: 620–800 °C in 20°-steps, third batch: replicate biochars produced at 600 and 700 °C). Smaller temperature increments starting at 600 °C were chosen, as this is the temperature range of most commercial pyrolysis plants and the greatest changes in carbon speciation were expected [23]. The biomass composition is shown in S1 Table. Experimental pyrolysis was performed with a PYREKA research pyrolysis unit (Pyreg GmbH, Dörth, Germany), a continuously operated auger reactor [22] adjusted to a residence time of 10 min. This setup is described in detail in Hagemann et al. [22]. Feeding rate was kept constant for each batch of feedstock and was in the range 0.4–0.7 kg h^-1. The reactor was purged with 2 L min^-1 N₂. Biochars were collected for 30–45 min to achieve 50–150 g per sample. After changing the pyrolysis temperature during continuous feeding of biomass into the reactor, biochar produced during the subsequent 30 min was discarded; when the input of a new biomass was started, the production of the first 45 min was discarded. We labeled the biochars beginning with the feedstock (i.e., W for wood, S for straw) followed by the pyrolysis temperature (e.g., a biochar produced at 600 °C from straw pellets is denoted as S600). If the temperature indication is followed by the capital letters A-C (e.g., W600A), it is a replicate of the biochar production (intra-day precision). The replicates were produced during an ongoing continuous pyrolysis, each with a sampling interval of 30 min.

2.2 Biochar characterization

Elemental analysis (CHN) was performed according to DIN 51732. The SEC was determined while the ground biochar (< 0.2 mm) was subjected to a pressure of 10 kN between two electrodes of the “Black Gauß I” device, which equals 30 MPa. The apparatus and procedure are described in detail elsewhere [19]. Both methods are compliant with the analytical guidelines of the European Biochar Certificate [24]. The ash content needed to express the BC_HyPy content on a dry and ash-free (daf) basis was quantified according to DIN 51719 (550 °C). Oxygen (O) was determined on a vario EL-cube (elementar, Langenselbold, Germany).

For HyPy, 100−200 mg of biochar sample were loaded with a Mo catalyst using an aqueous/methanol (80%/20%) 0.2 M solution of ammonium dioxydithiomolybdate [(NH₄)₂MoO₂S₂]. Catalyst weight was ~ 10% of the sample weight. The catalyst-loaded biochar was dried (110 °C, 24 h) and samples were placed in quartz tubes (20 mm long), sealed with a sintered disc at the base, and placed in the HyPy reactor. The samples were heated at a rate of 300 °C min^-1 from 50 to 250 °C (i.e., within 40 seconds), then heated at 8 °C min^-1 from 250 °C until the final temperature of 550 °C, which was held then for 2 min under a hydrogen pressure of 15 MPa. More details are described elsewhere [11]. A controlled constant hydrogen sweep-gas flow of 5 L min^-1 in the reactor, measured at ambient temperature and pressure, ensured that the labile products were quickly removed from the samples. The mass and carbon content of the HyPy residue were quantified.

4.1 Biochar properties determined by feedstock type, preparation, and pyrolysis conditions

Pyrolysis leads to the volatilization of low molecular weight carbonaceous compounds rich in O and H, resulting in an increase in carbon content and a decrease in H/C molar ratio of the solid product [12,21], which was confirmed in the present study. However, when using straw, a high-ash biomass (5.9%, S1 Table), the carbon content of the resulting biochars decreased at temperatures above 550 °C because of the non-proportional accumulation of mineral matter [12], which was not observed for the biochars made from low-ash wood (0.4%).

While low-temperature biochar is an electrical insulator, high-temperature biochar is electrically conductive due to the presence of conjugated π-electrons in its aromatic carbon structure. The degree of graphitization and aromaticity strongly influence conductivity, as a higher proportion of sp²-hybridized carbon enhances charge transport. Biochars produced at higher pyrolysis temperatures exhibit increased conductivity due to larger graphitic ordering and π-electron delocalization as the result of a higher degree of polycondensation [25,26]. X-ray diffraction and ¹³C nuclear magnetic resonance spectroscopy revealed that aromaticity and the degree of polycondensation increase with pyrolysis temperature, which is further indicated by the reduction of the H/C molar ratio [21,23,27–29]. The data on the SEC presented here fit well into the state of knowledge: at higher pyrolysis temperatures, biochars present a higher degree of polycondensation, which results in higher electrical conductivity. It should be emphasized that the increase occurred exponentially. In the parameter range investigated here (up to 800 °C and H/C = 0.1), a linear increase in the logarithmic SEC can be observed without any signs of saturation. For a better mechanistic understanding, future studies should include even higher pyrolysis temperatures and the measurement of reference materials such as conductive carbon black and defined nano carbon species.

It was unexpected that the conductivity of straw biochar was systematically higher than that of wood biochar in the temperature range 400–700 °C, despite the higher ash content of straw. Ash content negatively impacts the SEC of biochar, as demonstrated by both artificial mixtures (data not shown) and intrinsic variations [30], where higher ash fractions consistently lead to lower SEC when biochars are produced at similar temperatures from the same feedstock. At the same time, the presence of ash-forming substances also influences the speciation of pyrogenic carbon compounds through catalytic effects [31] and could therefore lead to an increase in SEC, depending on the composition of the ash. To investigate this in greater detail, a database with more than two different biomass and ash compositions would be required, which was beyond the scope of this study.

The increase in the BC_HyPy content with increasing pyrolysis temperature and decreasing H/C molar ratios fits into the context of the literature presented above and confirms the previous HyPy studies with lab-produced biochars [7,12,13,15]. Remarkably, the data show a saturation in BC_HyPy content when pyrolysis temperatures exceeded 600 °C and 680 °C for wood and straw, respectively, with BC_HyPy > 90% of TC. This is in line with the sigmoidal-like progression of BC_HyPy observed by McBeath and colleagues in pyrolysis experiments conducted at 300–900 °C [12]. Further experiments with higher pyrolysis temperatures and/or longer residence times should investigate in more detail if BC_HyPy plateaus > 90% TC or if (virtually) all carbon in biochar can be BC_HyPy. The latter can be expected but is not observed so far, which may indicate an artefact or contamination in the analytical HyPy.

Howell and colleagues [15] suggested a limit of 75 wt% BC_HyPy in biochar, which was exceeded by several samples in this study, with up to 90.6 wt% BC_HyPy, representing as much as 99.0% of its total carbon content (W700C, Table 1). Interestingly, Howell and colleagues produced biochar from woody biomass at temperatures of up to 800 °C and up to 10 min holding time (+ 100 °C/min heating rate), which at first glance would appear to be comparable to the conditions used in this study, as carbon speciation and aromaticity in particular are controlled by feedstock selection, (maximum) pyrolysis temperature, and the residence time in the pyrolysis setup [27]. While the BC_HyPy content given as the percentage of the total biochar weight was lower in straw than in wood biochar, which is due to the higher ash content of the straw feedstock, feedstock selection had no consistent impact on the BC_HyPy when expressed as a ratio to the TC content of the biochar.

McBeath and colleagues quantified BC_HyPy in biochars produced at up to 900 °C from a broad range of biomasses, which covered ash content of 0.1–39.8% and suggested that higher ash content, and specifically the content of amorphous silica, may inhibit the formation of polycondensated structures and thus reduce BC_HyPy. They performed pyrolysis in batches of 20–200 g of biomass in a muffle furnace flushed with nitrogen and controlled the temperature in the biomass bed and held the desired temperature for 1 h. Their data on BC_HyPy of biochars from pine wood and corn stover is in good agreement to the data on biochar from softwood and straw presented in this study, respectively (Fig 2a and 2b). For biochars produced at 500 °C or less, the present study showed higher BC_HyPy. In our study, the pyrolysis temperature was measured on the reactor wall. We observed selectively that the pyrolytic system no longer had to be actively heated when performing pyrolysis at 400–500 °C and that in some cases, temperatures above the set pyrolysis temperature were measured as the pyrolysis process was obviously exothermic in this temperature range. These observations were not systematically documented but are generally in line with literature [32] and measured temperatures did not deviate more than 5–10% from the set pyrolysis temperature between 400–500 °C. Still, higher BC_HyPy content compared to McBeath et al. in the range of 400–500 °C might be the result of actually higher pyrolysis temperatures due to exothermal reactions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Comparison of the content of BC_HyPy from biochar obtained from a continuously operating auger reactor (this study) with similar studies: McBeath et al. [12] pyrolyzed batches of 20-200 g of pre-dried pine wood (a) and corn stover (b) under nitrogen flow in a muffle furnace.

H/C molar ratio was not available in this study. Howell et al. (c) performed thermal treatment of pulverized woody biomass (<425 µm) in a thermogravimetric analyzer at 300-800 °C for 1-10 minutes under flow of air or nitrogen as an inert gas; only data points with H/C molar ratio < 0.7 were included.

https://doi.org/10.1371/journal.pone.0330206.g002

Howell and colleagues used pulverized biomass sieved to < 0.425 mm, while considerably larger biomass pellets were used in the present study. They had a diameter of 6 mm and a length of approximately 5–10 mm (S1 Fig). The biomass particle geometry affects the heating rate and gas exchange during pyrolysis and thus impacts biochar carbon speciation, which may explain the different results [33,34]. Moreover, Howell et al. performed thermal treatment in a thermogravimetric analyzer using only 100 mg biomass under inert gas or oxygen, which reduces secondary pyrolysis reactions [15]. The latter are known to result in highly aromatized carbon species, as demonstrated in industrial pyrolysis devices [35]. The resulting biochars consistently had lower BC_HyPy contents at similar H/C molar ratios compared to the biochars produced in our study (Fig 2c). A high (>90 wt% daf) BC_HyPy content was only achieved by Howell et al. when some type of gasification was performed (thermal treatment under the supply of oxygen that is not sufficient for full oxidation) [15,36]. This highlights the need to conduct analyses of “real-life” industrial biochars. The use of a pilot plant in the present study was a compromise between practice-oriented, industrial-like pyrolysis conditions and the possibility of testing a range of pyrolysis temperatures under otherwise constant conditions.

4.2 Prospects and limits of using BC_HyPy in multi-pool decay models

The IPCC suggested estimating biochar persistence for national greenhouse gas inventories via the pyrolysis temperature [37], which is (supposedly) simple and, above all, inexpensive. Also, our data (Fig 2a) could be interpreted in this way. However, the reality is more complex: In our experiments on PYREKA, the temperature was the single difference in pyrolysis conditions, whereas other factors impacting biochar properties, such as reactor design, particle size of biomass, residence time of solid and gaseous pyrolysis products, and residual oxygen concentration in the reactor, were constant [38]. As such, our findings to this end should not be considered universally applicable and cannot be directly extrapolated to comparisons of biochars derived from different feedstocks or subjected to varying production processes. In practice, different reactor designs and a wide range of pyrolysis conditions impact biochar properties despite the general consensus that pyrolysis temperature is the most important pyrolysis process determining biochar properties [15,21,27,33,36,39]. Moreover, determining pyrolysis temperature in practice is often challenging to impossible due to moving parts in most reactors and the challenge of establishing ideal heat transfer between thermocouples and the biomass [15,40]. Also, the heterogeneity of biochar (e.g., with regard to PAH contents) suggests considerable variability of temperature distribution within an industrial pyrolysis reactor [41]. Thus, pyrolysis temperature should not be used for the parametrization of decay models for individual biochars. Instead, biochar decay models must be parameterized by analytical data of the produced biochar. Robust and, in the best case, simple methods are needed to enable high-throughput analysis of biochar persistence.

Initially, HyPy was not designed to quantify a persistent biochar carbon fraction. Instead, its purpose was to isolate black carbon from organic impurities in environmental samples. The HyPy residue is an operationally defined, thermally stable carbon fraction (H/C < 0.5, > 7-ring polycondensed clusters), which is quantified with high reproducibility and precision (triplicate measurements typically within ±2% variation coefficient). As there is strong consensus in the literature that highly condensed aromatic clusters in biochar can be considered persistent (but not inert) [42–44], BC_HyPy could be one factor to parameterize multi-pool decay models. Still, HyPy does not measure the actual size and speciation of these aromatic clusters but provides an operationally defined threshold measure for the degree of poly-condensation. To study the speciation of the BC_HyPy fraction, X-ray diffraction [23,27] can be employed and used for correlations with other biochar properties.

Electric conductivity could be an indicator for the speciation of BC_HyPy. SEC was in the range of 10¹–10³ for biochars produced at 680–800 °C, which obviously differ in carbon speciation based on previous research [27], despite showing plateauing BC_HyPy at >90% TC. To better understand the role of SEC as an indicator of biochar persistence, further investigations are necessary to determine whether and to what extent SEC could also be influenced by parameters other than carbon speciation, e.g., the ash content. Solid state electric conductivity may be one of the easiest methods in biochar persistence analysis to perform with little experimental equipment required.

Non-BC_HyPy may include more or less alkanes, heterocyclic aromatic compounds, PAHs with up to seven aromatic rings, and their alkylated counterparts, as revealed by GC-MS of the compounds volatilized during HyPy. It may also include some larger aromatic molecules that cannot be quantified with conventional GC/MS [8]. Its composition depends on feedstock, pyrolysis temperature, and further biochar production conditions [7,9,45]. Understanding the stability and fate of non-BC_HyPy in the environment is needed to quantify the time-dependent carbon sequestration of the less-persistent fraction of a given biochar [46]. It would allow to distinguish between the persistent, semi-persistent, and labile fractions of biochar and to derive a time-dependent carbon-sink accounting curve for the total biochar carbon applied to soil or materials.

Both SEC and BC_HyPy correlate well with the H/C molar ratio, which is currently used to approximate a stable carbon fraction in individual biochars based on incubation-derived data on biochar persistence [2–4]. Other studies correlate biochar persistence with its O/C molar ratio [15,47]. However, the O content of biochar is usually calculated after quantifying C, H, N, S, and ash content with insufficient precision [48], while the direct measurement is not standardized yet [24]. Thus, the determination of O content in praxis is less reliable than most other biochar properties.

Another approach to identify a persistent carbon fraction in biochar is the analysis of macerals (organic minerals) in the carbonaceous material according to guidelines of the International Committee for Coal and Organic Petrology – ICCP [49–52]. Here, light microscopy is used to identify structures in an embedded and polished biochar sample. The reflectance of visible light is then determined microscopically according to ISO 7404−5 (vitrinite reflectance) to quantify the content of the maceral inertinite, which is considered the most recalcitrant maceral. However, the name inertinite does not mean that this maceral is inert, but that it is far less reactive than others [53]. Future research should compare the different persistence proxies, including elemental analysis (H/C and O/C molar ratios), HyPy, SEC, vitrinite reflectance and other methods, preferably on industrial biochars. There is an urgent need to consolidate the findings from physico-chemical characterization of biochar [47,54], from controlled incubation experiments [5,44] and from field trials [55,56] into a unified understanding of biochar persistence in the environment, as the current data and their interpretation are still perceived as contradictory.