Biochar prepared at different pyrolysis temperatures affects urea-nitrogen immobilization and N₂O emissions in paddy fields

Biochar preparation and characterization

Maize straw was collected from an experimental field at Shenyang Agricultural University, Shenyang, Liaoning Province, China. After oven-drying (85 °C, 24 h), the straw was cut into small pieces (<2 mm) and stored in sealed plastic bags. The straw samples were then transferred to a rectangular porcelain container (150 × 100 × 50 mm) and placed in a muffle furnace for pyrolysis at a heating rate of 15 °C min⁻¹. Temperature was first raised to 200 °C and then to a final temperature of 400 or 700 °C, maintained for 1 h (Gai et al., 2014). To minimize the oxygen content in the reaction, the container was filled with straw and tightly sealed. The biochar was cooled and then stored at room temperature until analysis and experimentation. The Brunauer-Emmett-Teller surface area and pore volume of biochar were determined using a V-Sorb 4800P surface area and porosimetry analyzer (Gold APP Instrument Corporation China, Beijing, China). Basic properties of biochar samples obtained at each different temperature are shown in Table 1. Each sample was analyzed in triplicate.

Site description and experimental soil

Paddy soil was collected in an experimental field (41°50′N, 123°24′E) managed by the Rice Institute of Shenyang Agricultural University. The site experiences a typical semi-humid temperate continental monsoon climate, with a mean annual temperature and precipitation of 8.3 °C and 500 mm, respectively, and 183 frost-free days. The accumulated temperature (>10 °C) is 3,300–3,400 °C. Mean annual precipitation is concentrated, with slight variation in mean annual temperature (Sui et al., 2016). Temperature and precipitation data were recorded during the rice growing season (June to October 2016, Fig. 1).

The soil was taken from the surface layer (0–20 cm) in April 2016. The soil was classified as Histosols according to the United States Department of Agriculture (USDA) soil taxonomy. The basic properties of the soil prior to the experiment were as follows: initial pH = 6.8 (1:2.5, water/soil, w/v) (HANNA HI2221, Italy), bulk density = 1.46 g cm⁻³, total N = 1.87 g kg⁻¹, and total C = 15.39 g kg ⁻¹. The soil were air-dried, passed through a 2-mm nylon sieve, and mixed thoroughly with biochar at different rates before use.

Experimental design

The experiment followed a 2 × 3 × 3 factorial completely randomized design. There were two pyrolysis temperatures (400 and 700 °C; B400 and B700, respectively), three biochar application rates (0, 0.7, and 2.1%, w/w; equivalent to 0, 15, and 45 t ha⁻¹; C0, C0.7, and C2.1, respectively), and three N fertilizer levels (0, 168, and 210 kg N ha⁻¹; N0, N168, and N210, respectively). Each treatment group included three PVC pots (30 cm diameter ×25 cm height), giving a total of 54 pots. A total of 14 kg of soil with or without biochar was packed into each pot at a depth of 20 cm.

The rice (Oryza sativa L.) japonica variety ‘Shennong 265’ was cultivated in 2016. Two seedlings at the three-leaf stage were transplanted from the nursery bed to each pot on 29 May 2016. In the experiment, 36% total urea was applied before transplanting (¹⁵N-labeled urea as base fertilizer), with 24% at the active tillering stage (unlabeled urea as tillering fertilizer) and 40% at the ear primordial stage (unlabeled urea as panicle fertilizer). The ¹⁵N-labeled urea (10.18 atom% ¹⁵N abundance) was provided by Shanghai Research Institute of Chemical Industry in Shanghai, China. In addition, all treatments received the same amounts of phosphorus (615 kg P₂O₅ ha⁻¹ as triple super phosphate) and potassium (200 kg K₂O ha ⁻¹ as potassium chloride) as base fertilizers. Fertilizers were dissolved in water and then added into pots. All pots were watered regularly to maintain flooded conditions, with water level 5 cm above the soil surface (except for aeration at the top-tillering stage to control effective tillering). Pots were kept outdoors. To reduce the effects of precipitation, a mobile steel-framed plastic canopy (800 cm long ×300 cm wide ×200 cm high) was used.

Plant sampling and analysis

At maturity (14 October, 2016), all plants were harvested and separated into grains and straw then oven-dried to a constant weight at 70 °C for 48 h and weighted to determine total yield. Grain moisture was determined using a hand-held moisture tester after drying (John Deere, Moline, IL, USA), and grain yield was estimated with a 14.5% moisture content.

Dry plant samples (grain and straw) were ground and sieved (0.15 mm) to analyze total N and ¹⁵N content (% in atoms) by isotope ratio mass spectrometry. ¹⁵N analyses were performed using elementar ISO prime 100 (Isoprime Ltd., Germany). Stable nuclides and the natural abundance differ from the atom% excess of the element. The background value (atom %) was subtracted from the experimental value to give the atom% excess. The natural ¹⁵N abundances of the plants and soil were estimated by averaging the values of all experimental treatments, respectively. Plant N uptake, N use efficiency, and the percentage of plant N derived from urea fertilizer were calculated using the following equations: (1)${\displaystyle \text{Plant N content}=\text{Total N concentration in dry biomass}\times \text{weight of dry biomass}}$ (2)${\displaystyle \text{Ndff}=\frac{Af\text{\%}-Acf\text{\%}}{Au\text{\%}-Acf\text{\%}}\times \text{TN(plantorsoil)}}$ (3)${\displaystyle \text{Ndfs}=\text{TN}-\text{Ndff}}$ where Ndff is the N in the plant or soil derived from ¹⁵N-labeled urea fertilizer (mg pot⁻¹) and Ndfs is the N in the plant derived from the soil (mg pot⁻¹), TN is the total N content in the plant or soil (mg pot⁻¹), and Au, Acf, and Af are the ¹⁵N abundance in the ¹⁵N-labeled urea fertilizer (10.18 atom%), natural ¹⁵N abundance in the plant or soil, and total ¹⁵N abundance in the plant or soil, respectively.

The recovery of ¹⁵N-labeled urea in the plant tissue or percentage retained in the soil was derived at the harvest stage using the following equation (Bronson et al., 2000): (4)${\displaystyle \mathrm{REN}\left(\text{\%}\right)=\frac{Ndff}{F}\times 100}$ where F is the amount of ¹⁵N-labeled urea applied (mg pot⁻¹).

Soil sampling and analysis

Three soil samples were obtained using a hand-operated core sampler (inner diameter = 3.5 cm, 20 cm deep) from each pot after harvest (October 2016). Soil samples were sealed in plastic bags and maintained on ice in an insulated box then transported to the laboratory where they were stored at −20  °C until use. Each soil sample was divided into two parts, one for analysis of water content and another for physico-chemical analyses. Soil water content was determined after oven-drying ca. A total of 5 g of samples at 105 °C for 48 h. Before analysis of other soil properties, samples were ground using a Wiley millto and passed through a 2-mm sieve to remove root detritus. Subsamples were further passed through a 0.15-mm sieve for determination of total N using an Elementar Variomax CNS Analyzer (German Elementar Company, 2003). Soil inorganic N (NH₄⁺-N) was extracted with 2.0 M KCl solution (Prayogo et al., 2013), filtered by Whatman No. 1 filter paper, and quantitated with a Continuous Flow-injection Analyzer AA3 (SEAL Analytical, Inc., Norderstedt, Germany). Bulk density was determined using a cylinder (100 cm³) with additional samples collected at a 0–10 cm soil depth. Soil porosity was calculated as follows: (5)${\displaystyle \mathrm{TP}=\left(1-\frac{{\rho }_b}{{\rho }_p}\right)\times 100}$ (6)${\displaystyle \mathrm{CPP}=\frac{FC}{WMC}\times BD}$ (7)${\displaystyle \mathrm{AFP}=\mathrm{TP}-\mathrm{CPP}}$ where TP is the total porosity (%), ρ_b is the bulk density (g cm⁻³) and ρ_p is the soil specific gravity (g cm⁻³) in Eq. (5); CPP is the soil capillary porosity (%), FC is the field capacity (%), and WMC is the wilting moisture content (%) in Eq. (6); AFP is the soil air filled porosity (%) in Eq. (7).

Gas sampling and analysis

N₂O emission fluxes were measured across the rice growing season (June to October) using static opaque chambers (Wang et al., 2011). The size of the chambers (32 cm diameter × 70 or 120 cm height) was adapted to rice growth, with 70 cm height before the heading stage and 120 cm height after heading. The chambers were also wrapped in aluminum foil to reduce internal temperature changes, and equipped with circulating fans to ensure complete gas mixing during gas sampling. During the growing season, gas fluxes were measured approximately every two weeks then once more after each fertilization or water control practice. This measurement frequency was adjusted to capture the period of most active N loss in more detail.

Chambers were placed on stainless steel pedestals during the period of each flux measurement. The edge of the stainless steel pedestals had a groove filled with water to seal the gas chamber during gas collection. Gas samples were collected using a 50 mL air-tight syringe (Singla & Inubushi, 2014) at 15-min intervals (0, 15, 30 and 45 min) after the chambers were closed. N₂O flux measurements were conducted between 8 and 10 a.m. (Zou et al., 2005). N₂O concentrations were analyzed using a gas chromatograph (Agilent 7890A; Agilent Technologies, Santa Clara, CA, USA) and hourly emissions of N₂O were determined from the slope of the mixing ratio change with four sequential samples. Quality checks were applied and N₂O flux measurements were discarded if the r² of the linear regression of the fluxes was <0.90. Cumulative emissions of N₂O during the growing season were calculated using trapezoidal integration to interpolate fluxes between successive sampling days (Millar et al., 2018).

The N fertilizer-induced N₂O emission factor was calculated by the difference in cumulative N₂O-N emission during the rice growing season between treatments with or without N fertilization, divided by the fertilizer N applied (Eq. (8)): (8)${\displaystyle \text{EF}\left(\text{\%}\right)=\frac{\text{Cumulative}{\text{N}}_2\text{O}-\text{N}\left(\text{fertiliztion}\right)-\text{cumulative}{\text{N}}_2\text{O}-\text{N}\left(\text{unfertilized}\text{control}\right)}{\text{Fertilizer}N\text{applied}}\times 100}$ where EF is the N fertilizer-induced N₂O emission factor.

Statistical analysis

Treatment effects were assessed by three-way analysis of variance (ANOVA) using SPSS Statistics 18.0 (IBM, Somers, NY, USA), and significance was expressed at P <0.05. All data are expressed as the mean ± standard error (n = 3). Multiple comparisons of means were based on Fisher’s least significant difference (LSD) test at a 5% significance level unless stated otherwise.

Soil physico-chemical properties

A significant decrease in soil bulk density was observed with increasing biochar rate, irrespective of pyrolysis temperature and N level. A high biochar rate significantly increased soil porosity to 50.43–52.54% under low-temperature biochar treatment (B400) with or without N fertilization; however, under high-temperature biochar treatment (B700), this increasing effect was only observed without N fertilization. Biochar had no significant effects on soil capillary porosity or air-filled porosity between the different treatments of pyrolysis temperature or biochar rate (Table 2).

Under high-N treatment (N210), soil NH₄⁺-N decreased slightly with increasing biochar rate, irrespective of pyrolysis temperature. The lowest soil NH₄⁺-N concentration was observed at 3.79 µg g⁻¹ under B400 and 3.71 µg g⁻¹ under B700 (Fig. 2). Biochar rate had no significant effects on soil NH₄⁺-N.

¹⁵N recovery in rice-soil system

A higher biochar rate resulted in lower N uptake from ¹⁵N-labeled urea in rice, irrespective of pyrolysis temperature and N level. Under B400, a low application rate of biochar resulted in significantly higher recovery of ¹⁵N in rice compared with a high biochar rate; however, under B700, no significant difference was observed in ¹⁵N recovery between the two biochar rates. Following N fertilization alone at 210 kg ha⁻¹ (C0), N uptake percentage from ¹⁵N-labeled urea reached 15.80 and 15.30% under B400 and B700, respectively. Compared with non-biochar treatment, a high application rate of B400 decreased N uptake from ¹⁵N-labeled urea by 55.03 and 44.03% under N210 and N168, respectively; by contrast, B700 caused a smaller decrease in N210 (44.03%) compare with N168 (55.52%). Similarly, N uptake from the soil also decreased with increasing biochar rate (Table 3).

Following biochar application at a high rate, the recovery rate of ¹⁵N-labeled urea in rice was 7.06–7.11 and 6.78–8.56% under B400 and B700, respectively. Compared with non-biochar treatment, significant decreases were observed in the ¹⁵N recovery rate across the two pyrolysis temperatures, irrespective of N level (Table 3). The residual soil ¹⁵N rate obtained with a high biochar rate was increased compared with low-biochar and non-biochar treatments, under both B400 and B700; in particular, the difference reached a significant level under N168 (P < 0.05). Overall, the highest ¹⁵N recovery in the plant-soil system was obtained with biochar application under B400 (540.56 mg pot⁻¹) and B700 (500.77 mg pot⁻¹). The corresponding N recovery rates in the plant-soil system under these conditions were 33.58 and 31.10%, ranking highest among all treatments, although there were no significant differences (Table 4).

A large proportion of ¹⁵N in the rice-soil system was presumably lost, and lowest ¹⁵N loss rates were found under B400 (66.42%) and B700 (68.90%) following co-application of 2.1% biochar and 168 kg ha⁻¹ N fertilizer. Under B400, a smaller ¹⁵N loss rate was observed following N fertilization at 168 kg ha⁻¹ compared with 210 kg ha ⁻¹, irrespective of biochar rate. No significant difference between the two pyrolysis temperatures was observed in terms of ¹⁵N loss rate. Concerning biochar rate alone, a high application rate resulted in a lower N loss rate (71.82 and 72.94%) than a low application rate (78.00 and 79.69%) under both B400 and B700. In the presence of 168 kg ha⁻¹ N fertilizer, a high biochar rate of B700 significantly decreased the ¹⁵N content in rice plants compared with non-biochar treatment (P < 0.05). Both B400 and B700 enhanced soil ¹⁵N retention at the 2.1% biochar rate, while B400 retained a larger soil ¹⁵N value compared with B700 (Table 4).

N₂O emissions from paddy soil

N₂O emissions from the soil were significantly affected by pyrolysis temperature, biochar rate, N level, and their interactions (Table 5). N₂O emissions fluxes initially peaked on day 2 then decreased on day 20 after transplanting, except under co-application of 2.1% biochar and 168 kg ha⁻¹ N fertilizer. Under B400, co-application of 2.1% biochar and 210 kg ha⁻¹ N fertilizer resulted in wide fluctuations in N₂O emissions after water control and then a slight decrease after topdressing with N, while under B700 a sharp decrease was observed (Fig. 3). Cumulative N₂O emissions ranged from 0.75 to 3.75 kg ha⁻¹ and significantly decreased with increasing biochar rate, regardless of pyrolysis temperature and fertilizer level. The application of 2.1% biochar without N fertilizer resulted in the least cumulative N₂O emissions under both B400 and B700. Meanwhile, 2.1% biochar treatment caused a notable decrease in cumulative N₂O emissions compared with 0.7% biochar treatment (P < 0.05; Fig. 4).

The N fertilizer-induced N₂O emission factor, calculated as the percentage of total N supplied through urea, ranged from 0.09 to 0.77%. The lowest emission factor was obtained under B400 with 2.1% biochar plus 210 kg ha⁻¹ N fertilizer, while the highest was observed under B700 with 0.7% biochar plus 168 kg ha⁻¹ N fertilizer. The average emission factor under high-biochar treatment (C2.1) was 0.14%, which was significantly lower than that under all other biochar treatments.

Rice biomass and yield

The aboveground biomass (including straw and grain) of rice plants at harvest was significantly higher under high-N treatment alone compared with all other treatments (P < 0.05). However, biochar alone had no significant effects on rice biomass, irrespective of pyrolysis temperature. Co-application of 2.1% biochar (B400) and a high N level decreased the aboveground biomass by 21.58% compared with non-biochar treatment, whereas a low N level caused a smaller decrease of 10.15%. Under B700, the rice biomass under high-N treatment alone was significantly higher than that under co-application of 2.1% biochar and 168 kg ha⁻¹ N fertilizer. Moreover, under 2.1% biochar (B700), high-N treatment caused an increase in biomass of 19.07 g pot⁻¹ compared with low-N treatment. There were no significant differences in rice biomass between biochar treatments at different pyrolysis temperatures (Table 5).

Grain yield was markedly higher under 0.7% biochar treatment compared with 2.1% biochar treatment, but only at a low pyrolysis temperature. The average yield with 0.7% biochar was 62.64 and 62.37 g pot⁻¹ under B400 and B700, respectively. Three-way ANOVA revealed the significant effects of biochar rate and N level on rice biomass and grain yield (Table 5).

Biochar application reduces N loss in the rice-soil system

Following biochar application, fertilizer N can be lost as NH₄⁺-N; however, the amount of NH₄⁺-N loss is much smaller than NO₃⁻-N (e.g., by one order of magnitude), and therefore, the NO₃⁻-N content in paddy soil is very low and barely detectable (Cheng et al., 2017) . In the present study, we found that soil NH₄⁺-N concentrations occurred at relatively low levels across treatments (3.71–6.07 µg g⁻¹), whereas Riaz et al. (2017) observed the lowest extractable NH₄⁺-N concentration (0.51 µg g⁻¹) in the 2% biochar-treated soil in laboratory microcosms. Since we adopted single extractions with 2 M KCl solution, the results of soil NH₄⁺-N might be underestimated. Interestingly, the soil NH₄⁺-N concentration did not respond to increasing rate of biochar application alone (N0); nonetheless, a small decrease was observed with increasing biochar application following 210 kg ha⁻¹ N fertilization (N210), irrespective of pyrolysis temperature (Fig. 2). Ding et al. (2010) also observed a slight decrease in the cumulative loss of NH₄⁺-N in the 0–20 cm surface layer of sandy silt soil after bamboo charcoal application, which was attributed to the NH₄⁺-N sorption capacity of biochar. In our study, we observed a significantly lower soil NH₄⁺-N concentration under high-temperature biochar treatment (B700) than low-temperature biochar treatment (B400) with 0.7% biochar application only (Fig. 2). The variation in the NH₄⁺-N sorption ability of biochar is thought to be due to (1) difference in the ash content between biochar samples; and (2) difference in the amount of particular functional groups on the biochar surface (Zheng et al., 2013). In addition, biochar often have anionic sites, which can increase the ability of soil to adsorb and reserve NH₄⁺ (Sohi et al., 2010).

Another pathway of N loss is gas emission (e.g., N₂O), which is a major problem encountered in paddy fields (IPCC, 2014). Throughout our study period (i.e., the rice growing season), biochar treatments caused notable decreases in cumulative N₂O emissions compared with non-biochar treatment (Fig. 4). This result corroborates previous studies that biochar application to agriculture soil can efficiently mitigate N₂O emissions (Dicke et al., 2015, Hagemann et al., 2017; Sun et al., 2017). Interestingly, there were no spikes of N₂O emissions after N fertilization events in our study, which might be related to the gas sampling at wide time intervals. However, a recent study also found that corn straw-derived biochar application to alkaline clay soil did not affect N₂O emissions (Wu et al., 2018). Consensus has yet to be reached on how and why biochar reduces N₂O emissions across different plant-soil systems (Cayuela et al., 2014).

In our study, N₂O emissions were higher under B400 than B700 treatment (Fig. 3). The possible mechanisms underpinning the higher N₂O emissions after application of low-temperature biochar (400 ° C) are as follows: (1) The N immobilization and NH₄⁺ sorption ability of B400 was enhanced, making it a better substrate for NH₃ oxidation and accumulation of NO₂⁻, a majority substrate of N₂O (Clough et al., 2013); (2) B400 contained a higher N content, which could stimulate growth of nitrifying and denitrifying bacteria, indirectly contributing to N₂O emissions; (3) B400 had a lower C/N ratio, resulting in faster N cycling (Ali, Kim & Inubushi, 2015). Cayuela et al. (2014) and Harter et al. (2016) suggested that lower N₂O emissions following biochar application were due to microbial reduction of N₂O to N₂ via nosZ gene-containing microorganisms. In the current study, we did not verify nosZ gene abundance throughout the N₂O monitoring period. Further studies are therefore needed to quantify the nosZ gene under different biochar treatments and thereby determine the microbiological mechanisms of N₂O emission reduction.

As previously reported, biochar may alter the soil N content (Clough et al., 2013). However, our hypothesis that low pyrolysis temperature indirectly affects soil N retention for plant uptake was not supported by the experimental results, as no differences were found in the recovery of ¹⁵N in rice plants between B400 and B700 treatments. Zhou et al. (2017) revealed that biochar prepared with mixed materials at different temperatures caused an increase in both residual soil ¹⁵N and subsequent plant ¹⁵N uptake in an agricultural field in Canada (Zhou et al., 2017). These results suggest that biochar immobilizes soil nutrients onto its surface, thereby increasing soil ¹⁵N concentrations at higher application rates through increased porosity and surface area (Atkinson, Fitzgerald & Hipps, 2010; Liang et al., 2006).

In our study, the pyrolysis temperature of biochar did not significantly affect the recovery rate of ¹⁵N in rice or the residual soil ¹⁵N, yet biochar application did decrease N₂O emissions in paddy soil. We assume that a large portion of N loss was emitted as N₂ from denitrification (Dong et al., 2015). The relationship between residual soil ¹⁵N and the pyrolysis temperature of biochar observed here might be explained by the fact that high-temperature biochar favors soil fungal and bacterial colonization, in turn enhancing gaseous N losses and decreasing N retention (Gul et al., 2015; Nguyen et al., 2016). Complementary experiments further suggested that applying a low rate of high-temperature biochar (>450 °C) resulted in more correlations between microbial taxa, with a large number of microorganisms appearing to influence soil N retention (Nelissen et al., 2014). The higher residual soil ¹⁵N content observed under B400 treatment was therefore not simply the function of a single factor, and further analysis of long-term biochar application in the field is therefore required to determine the agricultural-environmental win-win benefits and improve crop yield. The effects of biochar on soil microbial biomass N content, microbial activity, and N fixation processes of key factors of N fixation process in various soils also need to be studied. Consequently, the use of straw-derived biochar remains challenging, requiring an accurate pyrolysis temperature and application standards. Further research is therefore needed to support the universal use of straw-derived biochar in agriculture.

Biochar application decreases rice productivity in short-term pot experiments

In the present study, the rice biomass response to biochar varied with pyrolysis temperature, biochar application rate, and N fertilizer level, with biochar application worsening plant growth. Surprisingly, treatment with 2.1% biochar, regardless of N fertilizer level, resulted in a ca. 13.35% decrease in rice biomass under B400 compared with non-biochar treatment (Table 5). This could be attributable to the higher contents of polycyclic aromatic hydrocarbons and volatile organic compounds in biochar prepared at lower pyrolysis temperature, which may cause bio-toxic effects (Zhu et al., 2018). Our results are therefore contrary to those of Zhao et al. (2014), who suggested that rice straw biochar applied at 9 t ha⁻¹ had a positive effect on rice/wheat growth. These growth improvements could be explained by bioavailable NH₄⁺-N and NO₃⁻-N levels after application of straw biochar (Sui et al., 2016; Dong et al., 2015). Rajkovich (2010) found a small increase or no change in aboveground biomass following soil application with varying rates of feedstock biochar obtained at different pyrolysis temperatures, whereas Dao et al. (2013) observed a three-fold increase in aboveground biomass after application of 80 t ha⁻¹ biochar compared with the non-biochar control. Overall, therefore, the plant biomass response to biochar depends not only on the characteristics of the biochar, the application rate and crop species, but also on the experimental set-up and original soil conditions (Biederman & Harpole, 2013; Chan et al., 2008; Lehmann et al., 2003).

There is a direct relationship between crop biomass and grain yield (Sakamoto et al., 2006). The results obtained from our pot experiment suggest that co-application of biochar with N fertilizer could markedly decrease rice yield under B400, but not B700, compared with the non-biochar control (Table 5). Our finding lends support to Jian et al. (2018) that laboratory-produced biochar at higher pyrolysis temperature contained lower bioavailable concentrations of polycyclic aromatic hydrocarbons and did not release volatile organic compounds at room temperatures. Wheat-straw biochar (12 t ha⁻¹) was previously found to have no significant effect on rice production in the first season (Xie et al., 2013). However, Zhang et al. (2012a) revealed that application of wheat straw biochar (20 t ha⁻¹) resulted in a 10% increase in rice yield in the first cycle and more than 9.5% increase in the second cycle. Such sustainable yield-increasing effects of biochar were also found in other experimental studies on crops. Together, these data suggest that pyrolysis temperature has only a small effect compared with biochar application rate on the short-term yield potential in paddy soil.