Rice (Oryza sativa L) plantation affects the stability of biochar in paddy soil

Conversion of rice straw into biochar for soil amendment appears to be a promising method to increase long-term carbon sequestration and reduce greenhouse gas (GHG) emissions. The stability of biochar in paddy soil, which is the major determining factor of carbon sequestration effect, depends mainly on soil properties and plant functions. However, the influence of plants on biochar stability in paddy soil remains unclear. In this study, bulk and surface characteristics of the biochars incubated without rice plants were compared with those incubated with rice plants using a suite of analytical techniques. Results showed that although rice plants had no significant influence on the bulk characteristics and decomposition rates of the biochar, the surface oxidation of biochar particles was enhanced by rice plants. Using 13C labeling we observed that rice plants could significantly increase carbon incorporation from biochar into soil microbial biomass. About 0.047% of the carbon in biochar was incorporated into the rice plants during the whole rice growing cycle. These results inferred that root exudates and transportation of biochar particles into rice plants might decrease the stability of biochar in paddy soil. Impact of plants should be considered when predicting carbon sequestration potential of biochar in soil systems.

biotic oxidation. This suggested that abiotic oxidation was very important for the initial oxidation of biochar. Interactions between biochar and microbes are known to occur within a short time period after application to soil, thus influencing the biochar stability in soil 14 . Through a six-month laboratory incubation study, Santos et al. 15 suggested that biochar carbon could be utilized by all bacteria groups, especially gram ( + ) bacteria. Using compound specific isotopic analysis of phospholipid fatty acids, Farrell et al. 16 found that soil microorganisms were able to rapidly take up and metabolize small proportions (3%) of biochar-carbon. In addition, for the direct utilization of biochar carbon within soil microbial communities, microbial degradation of biochar has been shown to occur through co-metabolism. Hamer et al. 17 reported that glucose addition accelerated mineralization of maize biochar by 58%, rye biochar by 72% and wood biochar by 115% via microbial co-metabolism compared to the treatment without glucose addition.
While abiotic and microbial degradation of biochar has been shown to significantly influence biochar stability, the effect of plants on biochar stability is not known. The interaction of plants with the soil and microbes, however, are known to occur through plant roots. Rooting systems provide a source of organic carbon to soil microbes, which enhance microbial biomass, leading to a promotion of microbial co-metabolism that can increase biochar degradation 17 . Roots can also regulate the microbial community structures and activities 18 by changing soil physical and chemical characteristics (e.g. water and pH) and by exuding organic substances. Therefore, plants are able to directly influence biochar decomposition rates through abiotic or indirectly influence biochar through microbial oxidation. Radial leakage of oxygen from the roots 19 may enhance oxidation of biochar while organic molecules released from the rice root can adsorb either onto the mineral layers of the biochar or directly to the biochar surfaces 14 . Therefore the adsorption onto biochar contributes to soil organic carbon stabilization and, in turn, plant roots can impact on biochar stability in a soil ecosystem. Especially in paddy fields, the impact of rice plants on biochar stability is expected to occur because paddy soils cycle through saturation and non-saturated states. However, the effects of rice plants on biochar mineralization and characteristics in paddy soil ecosystem remain largely unexplored. Yet, biochar is regularly added to paddy soils because rice-straw derived (RS) biochar applications have proven to be an effective means to reduce greenhouse gas (GHG) emissions, increase carbon sinks and increase crop yields in paddy soils 1,[20][21][22] .
The objective of this study was to test the effect of rice plants on the decomposition of RS biochar in paddy soil. We hypothesized that rice plants would stimulate RS biochar degradation via roots exudates and microbial oxidation.

Results
Decomposition and characteristics of biochar without rice plants. During the pre-incubation period, the initial CO 2 efflux rates from the soil were very low. However, the CO 2 efflux rates increased to a maximum after 12 days of incubation when the rates reached 22.23 mg m −2 h −1 in the soil without biochar amendment (SO) treatment and 34.56 mg m −2 h −1 in the biochar amendment (SC) treatment (Fig. S1). After 12 days of pre-incubation, the δ 13 C-CO 2 evolved from the biochar amendment had a significantly higher signature than that of the control (Fig. 1). As calculated by equation 2, the BC mineralization rate was 2.45 × 10 −2 μg CO 2 -C g −1 biochar h −1 ,or 2.45 × 10 −6 % h −1 .
Elemental composition. The C and O compositions of the biochar sampled from the SC treatments after pre-incubation and, at the jointing, heading, and maturing stages of rice growth showed that the total C content increased from 76.28% to 87.20% during the first 3 stages, but decreased during the maturing stage (Table S1). In contrast, O content decreased from the pre-incubation stage until the heading stage and then increased at the maturing stage. As a result, the O/C ratio decreased from 0.24 to 0.08 during the first 3 stages but then increased to 0.13 at the maturing stage.
Fourier-transform infrared spectra (FTIR) analysis. Infrared-spectra showed that the original biochar, prior to any treatment, contained carbonyl, hydroxyl and aliphatic C−H groups. After pre-incubation, bands arising from aromatic C-H out of plane vibrations (700-900 cm −1 ) became more apparent in the biochar samples (Fig. 2). Other functional groups, including the aromatic C = C groups (1600 cm −1 ) kept constant during all rice cultivation stages.
Decomposition and characteristics of biochar during the various rice growth stages. Decomposition of biochar. During the rice cultivation stages, CO 2 efflux rates in the soil were significantly greater in the presence of rice plants than without rice plants, irrespective of biochar amendment. Biochar amendments had no significant impact on the CO 2 efflux rates in soils during the whole period of rice growth (Fig. S1). During the 4 rice cultivation stages, the CO 2 efflux rate was constant except during the filling stage when for both treatments with rice plants, the CO 2 efflux rate was smaller than all other stages.
The δ 13 C−CO 2 signature of the evolved CO 2 from biochar decomposition ranged between -19.13‰ and -15.64‰ (Fig. 1). Based on these isotopic CO 2 signatures, the amount of CO 2 derived from biochar was determined for the biochar amended treatments at the heading and the maturing stages of the rice growth cycle. Compared to the heading stage, the biochar mineralization rate in the maturing stage, increased when no rice plants were present while the biochar mineralization rate in the biochar treatment with rice plants decreased (Table S2). This resulted in similar biochar mineralization rates during the maturing stage. At the heading and maturing stage, the presence of rice plants did not significantly impact on the biochar mineralization rates.
Elemental composition. The C and O compositions of the biochar in the presence of rice plants at the jointing, heading and maturing stages of rice growth showed that the total C content, increased from 85.96% to 86.36% during the first 2 stages of rice growth, but decreased during the maturing stage (Table  S1). In contrast, O content decreased from the jointing stage until the heading stage and then increased at the maturing stage. Therefore, O/C ratio increased from the heading stage until the maturing stage. However, the presence of plants did not significantly impact on the total C content, O content and O/C ratio of biochars sampled from the biochar treatment without (SC) and with (SCR) rice plants during the various rice growing stages.
Thermogravimetric (TG) analysis. The overall weight loss when temperatures were increased to 1000 °C amounted to 25.37% for the original biochar but only 18.27% for the SC treated biochar at the end of the maturing stage and 18.62% for the SCR treated biochar (Fig. 3). The derivative of the thermogravimetric (Dr-TG) analysis curves demonstrated the non-isothermal decomposition of biochar. The greatest rate of weight loss as indicated by Dr-TG analysis curves occurred in the temperature range of 30-60 °C due to moisture losses. When the temperature was greater than 100 °C, the original biochar showed a maximum weight loss rate at 934 °C. Biochar samples from both the SC and SCR treatments had much lower weight loss rates above 750 °C than the original biochar. Biochar samples from both the SC and SCR treatments showed a maximum weight loss rate at 610-620 °C.
FTIR analysis. During the whole rice cultivation stage, in the presence of rice plants, functional groups in the biochar samples such as carbonyl, hydroxyl and aromatic C=C (1600 cm −1 ) groups did not show any significant changes over time (Fig. 2). The presence of rice plants did not significantly impact on the functional groups of the biochar samples compared to the samples from the treatments without rice plants.
X-ray Photoelectron Spectroscopy (XPS) analysis. The biochar particles and powder after being incubated with rice plants contained significantly more O atoms, as shown by the O1s peak area, than the original biochar and biochar being incubated without rice plants (Fig. S3).
The bulk properties of biochar (biochar powder, Table 1) showed the same trend in elemental contents of C and O as the results of elemental analysis (Table S1), specifically, a small increase in C and a decrease in O after being incubated. The O/C ratio of the original biochar of the interior (biochar powder, Table 1) was similar to the O/C ratio of the surfaces (biochar particles, Table 1). The presence of rice plants did not show a significant difference in the O/C ratio of biochar powder after incubation. However, after incubation with rice plant, the O/C ratio of biochar surfaces increased by 72% of the value of original biochar and 19% of the biochar incubated without rice plants (biochar particles, Table 1). These increases were mainly due to the formation of phenolic C and carboxyl C on the biochar surfaces ( Table 2).

Incorporation of biochar 13 C into microbial biomass and rice plants.
During the tillering, heading and maturing stages of rice growth, the presence of rice plants did not significantly impact on the microbial biomass carbon (MBC) for the biochar amended soils (SC and SCR) (Fig. S2). The isotopic signatures of dissolved organic carbon (DOC) and the sum of DOC and MBC at the maturing stage of rice cultivation in Fig. 4 showed that more 13 C labeled carbon from the biochar (0.0016% of 13 C input) was incorporated into the microbial biomass when rice plants were present (SCR) than when they were not (SC) (−0.006% of 13 C input).
At the end of the maturing stage, the incorporation of 13 C labeled biochar carbon in the above ground biomass was 0.15% and in the below ground biomass was 0.45% of total 13 C input (Table 3). In total, 0.047% of the carbon in the biochar was incorporated into rice plants during the whole rice cultivation cycle.

Discussion
The total thermo-gravimetrical weight loss (25.37%, Fig. 3) of the original biochar was lower than that of the lignin (54.7%) and black carbon-rich soils from natural forests (29.55%) 23,24 . Recent studies have shown that TG could be used to estimate the labile fraction of carbon in biochar and evaluate the thermal stability of soil organic matter by using the fraction of volatile matter with respect to the sum of volatile matter content and fixed C 24,25 . Therefore, the considerably lower total weight loss in the TG analysis indicated that the biochar used in this study had a relatively high thermal stability.
It has been suggested from the elemental composition that a biochar material containing lower O/C ratios would be more stable than a biochar material with higher O/C ratios 26 . In this experiment, the O/C ratio of the biochars significantly decreased after pre-incubation (Table S1). This suggested that the stability of biochar might be increased after incubation in soils. Mao, et al. 27 showed that biochar residues in Terra preta soils were mainly composed of about 6 fused aromatic rings and these biochar residues represented a particularly stable form of soil organic matter. The FTIR in this experiment demonstrated that the intensity of aromatic C-H out-of-plane vibrations (700-900 cm −1 ) of the biochars increased after the pre-incubation period and the aromatic C=C (1600 cm −1 ) functional groups kept constant during the whole incubation period (Fig. 2). This also suggested that the stability of biochar had increased after the pre-incubation period. On the other hand, the total weight loss of the biochar after incubation in soil   (SC and SCR treatment) was significantly lower than that of the original biochar. This suggested that the thermal stability of the biochar increased by incubation in soil. Therefore, it could be inferred that biochar after pre-incubation in soil was very stable. It has been assumed that plants may directly influence biochar decomposition rates by abotic oxidation or indirectly by microbial oxidation. However, in this study rice plants did not have a significant effect on the biochar decomposition rates (Table S2). The method using isotopically labeled CO 2 has some limitation when it comes to calculating biochar decomposition rates. In particular, biochar could be converted into some forms of organic matter that may be excluded from the calculations. In this experiment, more carbon from biochar was converted to microbial biomass carbon when rice plants were present than when they were not. Kuzyakov, et al. 28 reported that 0.3%0.95% of the 14 C in biochar was incorporated into microbial biomass and 17% into polysaccharides. These parts of carbon that biochar converted into were not included in the calculations using isotopically labeled CO 2 .
As shown in the FTIR and TG analysis, no significant differences in the functional groups and total weight loss were observed between the biochar samples collected from SC and SCR treatments (Fig. 2, Fig. 3). These results suggested that rice plants had no significant effects on the biochar bulk characteristics. However, the increase of atomic O/C ratios of biochar particles in the presence of rice plants (Table 1) suggested that rice plants enhanced the surface oxidation of the biochar. The differences between surface and bulk properties of the biochars that were exposed to rice plants highlighted the fact that higher oxidative states on the surface region of the biochar particles occurred in comparison to the interior 29,30 . Such differences were also observed in this study whereby significant increases in the surface oxygen functional groups of the biochar appeared both in SC and SCR treatments compared to the entire particles (Table 2). Surface oxidation of biochar led to an increase in phenolic groups 13 . In this study, the presence of -COR groups of the biochar significantly increased when biochar was incubated in the presence of rice plants. In addition, LeCroy, et al. 31 reported that biochar exposed to the plant-soil environment reportedly showed more surface oxidation in the form of increased proportions of -COR,-C=O and -COOR groups. Increased surface oxidation of the biochar in the presence of rice plants might have resulted from the rice exudates and radial leakage of oxygen from the roots 19 . Oxygen being the most common electron acceptor, may have oxidised biochar through abotic reactions boosted by a high density of electron donating π-electrons. A short-term incubation conducted by Cheng, et al. 13   at 30 °C with microbial inoculation showed that there were no changes in the measured indicators of surface oxidation, which suggested that surface oxidation might happen as a function of abiotic oxidation, such as chemisorption of oxygen (O) by aging processes 32 . On the other hand, rice exudates, such as low-molecular weight organic compounds (free exudates), high-molecular weight gelatinous materials (mucilage) and sloughed-off cells, could be adsorbed directly onto the biochar surfaces 14 . This would provide more available carbon for soil microorganisms to increase the co-metabolic decomposition of the biochar. Despite that further research is needed to clarify this mechanism, the increase in the biochar surface oxidation of the biochar with rice plants indicated that biochar stability in the soil was influenced by plants in the long term. Microbes can utilize biochar through two pathways: directly as a carbon source; and indirectly through co-metabolism 17,33 . Kuzyakov, et al. 34 indicated that biochar was mainly decomposed through co-metabolism and that biochar was negligible as a microbial C source. In this study, only 0.0016% of carbon from the biochar was incorporated into the soil microbial biomass in the presence of rice plants. The incorporation of the small amount of carbon from biochar into microbial biomass therefore confirms the low microbial availability of biochar as a substrate. However, the amount of carbon from the biochar that was incorporated into the soil microbial biomass in the presence of rice plants was significantly greater than that without the rice plants (−0.006% of 13 C input, calculated using data from Fig. 4). The enhancement of biochar incorporation into the soil microbial biomass under rice plants may be explained by the effect of root exudates on the microbial co-metabolism in the soil. Root exudates include ions, free oxygen, enzymes, mucilage and a diverse array of carbon-containing primary and secondary metabolites 35,36 . These exudates provide a source of organic carbon for soil microbes, leading to a promotion of microbial co-metabolism that can increase the biochar degradation 17 or enabling some microbes to actively degrade the recalcitrant C compounds with their extracellular enzymes 37 . It is reported that arbuscular mycorrhizal fungi (AMF) could form associations with plant roots through root exudates 38 and play a pivotal role in facilitating biochar degradation by altering it physically and chemically 39 .
This study has shown a direct incorporation of biochar carbon into rice plants. While only 0.047% of the carbon in the original biochar was incorporated into the rice plant in this study, a large amount of biochar carbon may be lost by being converted to CO 2 through the metabolism of rice plants in the paddy fields. Especially in China there are many paddy fields. Therefore, the incorporation of biochar carbon into rice plants will greatly decrease the carbon sequestration effect of biochar in a paddy soil ecosystem. Studies have shown that labile components of biochar are susceptible to microbial decomposition. Some metabolites of microorganisms such as acetate and butyrate could also be metabolized by the roots 40 . Therefore, carbon in biochar is not only directly taken up by rice plants but also indirectly utilized by the rice plant through microbial decomposition. Furthermore, it is reported that the transport of biochar particles increased with decreasing particle sizes and pyrolysis temperature 41 . Carbon nanoparticles could be taken up by plant roots, penetrate through vascular systems and then be transported to the plant shoots 42,43 . Therefore, we may speculate that rice roots are able to take up some biochar nanoparticles in paddy soil and then transport them into plant shoots. Although the transportation mechanisms of biochar carbon incorporation into rice plants are still not clear, results of this study implied that the role of biochar in carbon sequestration in the soil ecosystem would be greatly overestimated in the long term without the consideration of the impacts of plants. Therefore, the impact of plants should be considered when predicting biochar stability in the soil ecosystem and the role of biochar in the global carbon cycle.

Methods
Soil characteristics and production of 13 C labeled rice-straw biochar. A clay loam soil was collected from a paddy field in Shangyu City, Zhejiang Province, China. In this paddy field, rice plants had been grown under normal rice growing conditions by local farmers. The soil was sampled by taking composite samples in the plowing layer (0-20 cm) of the paddy field and then transported to the agricultural experiment station of Zhejiang University, China where it was air-dried and sieved through a 2-mm mesh screen.
In order to quantify biochar decomposition rates, 13 C labelled rice straw was generated 44 . The biochar was produced by slowly pyrolysing 13 C labelled rice straw in a GDL-1500X tubular furnace (Kejin, Hefei, China). The labelled rice straw was heated at a rate of 5-10 °C/min, followed by a residence time of 3 h at 500 °C under a N 2 gas flow of 1 L/min to ensure an oxygen-free atmosphere. The percentages of 13 C atoms in the soils and biochars were analyzed by δ 13 C isotope ratio mass spectrometry (ThermoFinnigan/ mat 253). The pH of the soils and biochars were measured in a 1:5 (w/ v) and 1:20 (w/ v) water solution (Cheng et al., 2006). Total C, H, N contents were determined using an elemental analyzer (Flash EA1112, Thermo Finnigan, Italy). Ash contents in percent of the biochar were quantified according to the American Society for Testing and Materials (ASTM) D1762-84. Total oxygen contents were calculated by subtracting the C, H and N contents that were water and ash free basis, assuming that biochar and soils were only composed of C, N, H, and O. Detailed physicochemical characteristics of the soils and biochars are summarized in Table 4.
Experimental setup. Before conducting the experiment, the paddy soil was pre-incubated with (SC treatment) and without 13  light under continuous flooded conditions in plastic buckets (D b = 11 cm, D t = 18.5 cm, H = 10 cm) in order to decompose labile components within the biochar. The Soil-Biochar mixtures were created by incorporating 2.5% (w/w) 13 C labeled RS biochar into the pre-incubated soil and non-amended soil was used as a control (n = 3 replicates of each).
The pre-incubated soils with and without biochar were used to create 2 experimental treatments (n = 3 replicates of each). These were: Soil with rice plants (SR), and Soil-Biochar mixture with rice plants (SCR). The pre-incubated soils that did not have rice planted in them were used as the Control treatment. All the soil samples were incubated at 20-30 °C with natural light for a complete rice growing season of 105 days. These soils were kept under continuous flooded conditions, and drained 3 weeks prior to the harvest. In order to inhibit the growth of algae, the buckets were shaded with dark foils during the whole incubation period.
Two rice seedlings of the commercially grown variety Oryza sativa L., cv. Xiushui 134, were transplanted into each bucket for the treatments with rice plants (SR and SCR). In order to ensure sufficient nutrients for the rice plants to grow, 250 ml of nutrient solution that contained 0.22 g kg −1 N (NH 4 NO 3 ), 0.20 g kg −1 P 2 O 5 (NaH 2 PO 4 ·2H 2 O), 0.20 g kg −1 K (K 2 SO 4 ) and trace elements, were uniformly mixed with the soils, two weeks before transplanting, and 100 ml of nutrient solution was applied three more times during the whole tillering stage of the rice cultivation stage. Nutrient solutions were also applied to Control soils without rice plants. Enough additional buckets were prepared to enable replicate (n = 3) samples to be destructively sampled at the pre-incubation, tillering, heading and maturing stages of the rice cultivation.
Carbon dioxide measurements. The CO 2 flux measurement was determined by the closed chamber method at four growth stages of rice cultivation. These growth stages were the jointing, heading, filling and maturing stages. In order to exclude the influence of photosynthesis, CO 2 was sampled at night (the sampling lasted for one hour each time). The concentration of CO 2 in the gas samples was analyzed using a gas chromatograph (GC-14B, Shimadzu, Japan) 20 and stable carbon isotopic ratios were measured using ThermoFinnigan/mat253 to estimate the fraction of respired CO 2 that was derived from biochar. Biochar characteristics. Biochar particles (diameters from 150 μm to 300 μm) were randomly picked from the soil of SC and SCR treatments at pre-incubation, jointing, heading and maturing stages using tweezers and an anatomical lens (45 × , SZ61, Olympus) 30 . Elemental compositions of the biochars were conducted by an elemental analyzer (Flash EA1112, Thermo Finnigan, Italy). FTIR absorption spectra were collected from wavelengths 4000 to 400 cm −1 . In order to prepare biochar samples for FTIR analyses, 1 mg of the samples had been finely ground, sieved through a 150 μm sieve, and mixed with 200 mg KBr. Carbon chemical functional groups were assigned to wave numbers according to reference 21 .
TG and XPS analyses were only conducted on biochar samples collected, from the maturing stage, after 105 days. TG was operated using a AT DSCQ1000 (USA) under a nitrogen atmosphere. The biochar samples, with an initial weight of 10 mg, were placed inside an aluminium oxide crucible. The temperature through the reaction was increased at a heating rate of 10 °C min −1 . Dr-TG analysis curves were determined.
In order to investigate the surface characteristics of biochar particles, XPS analysis was conducted using a VG ESCALAB MARK II (England), which used a focused monochromatic Mg Kα X-ray (1253 6 eV) source for excitation and a spherical section analyzer. The spectra of the biochar particles were assumed to represent the properties of the particle surface. Finely-ground biochar powder, whose spectra were assumed to represent both exterior and interior properties of the biochar were also measured. The high resolution XPS spectra of C1 S and O1 S were used to quantify C and O forms on the surface and in the interior of biochar particles. The deconvolution of the C1 S and O1 S spectra were determined using a non-linear least squares curve fitting program with a Gaussian-Lorentzian mix function and Shirley background subtraction. Details of quantification were described by Cheng et al. 13 .
Soil microbial and plant biomass analyses. Soil samples were collected during the growth season of the rice at 30 days (jointing), 65 days (heading) and 105 days (maturing). Soil MBC was determined on fresh soil samples by the chloroform fumigation-extraction method 45 . Above ground and below ground biomass of the rice plants were collected from each bucket just before harvest, air-dried, ground and sieved through a 0.3 mm mesh screen to measure total carbon with an elementary analyzer (vario MAX,  Germany). To quantify the incorporation of 13 C labeled carbon from biochar into the microbial and rice plants biomass, the 13 C signature of soil microbial and above ground and below ground biomass carbon at the maturing stage of rice growth was determined by mass spectrometer (ThermoFinnigan/mat253).
Calculations and statistical analyses. CO 2 emission fluxes were calculated using the equation: F = ρ × h × dc/dt × 273/T, where F is the CO 2 emission flux, ρ is the density of CO 2 at temperature T, h is the height of the closed chamber, dc/dt is the change of the CO 2 concentration per unit in the closed chamber, and T is the Fahrenheit temperature of the closed chamber. Isotopic ratios were expressed in standard delta notation as δ 13 CO 2 in units of per mil (%) relative to the Vienna Pee Dee Belemnite (VPDB) standard. Isotopic ratios are expressed as Atom% 13 C after being labeled, which is calculated as follow (R Standard = 11237.2 ± 90 × 10 −6 ): Where Atom% 13 C treatment is the 13 C atom percentage of different types of carbon with biochar amendment, Atom% 13 C control is the 13 C atom percentage without biochar amendment, Atom% 13 C biochar is the 13 C atom percentage of biochar. All data were reported as means and standard deviation (SD) of the means. Analysis of variance and the least significant difference (LSD) tests were used to determine the significant differences between treatments with and without rice plants at confidence level of P < 0.05.