High-Temperature Hay Biochar Application into Soil Increases N2O Fluxes

Biochar has been proposed as an amendment that can improve soil conditions, increase harvest yield, and reduce N losses through NO3− leaching and N2O emissions. We conducted an experiment to test the hay biochar mitigation effect on N2O emissions depending on its production temperature. The pot experiment consisted of the soil amendment with three different production temperature biochars (300 °C, 550 °C, 850 °C) alone and in combination with three different organic fertilizers (cattle slurry, slurry digestate, vinasse), in growth chamber conditions. The effects of biochar and fertilizer were both significant, but the interaction biochar:fertilizer was not. The amendment with the three fertilizer types and with the highest production temperature biochar resulted in significantly higher cumulative N2O fluxes. Biochar did not show a mitigation effect on N2O emissions when applied with organic fertilizer. Cumulative emissions were higher with biochar addition, with increasing emissions for increasing biochar production temperature. Our results support the idea that biochar cannot be considered as a universal tool for the reduction of N2O emissions.


Introduction
The urge to reduce emissions of greenhouse trace gasses (GHG) is generally acknowledged [1], with special focus on nitrous oxide (N 2 O) due to its high global warming potential [2], and for being the main ozone-depleting substance [3]. Anthropogenic sources of N 2 O account for almost 40% of global emissions, of which, agriculture represents 67%-80% [4]. Nitrogen fertilization, both organic and inorganic, is the main factor explaining the contribution of agriculture to N 2 O emissions [4,5]. N 2 O emissions from the soil are the product of two main processes: nitrification and denitrification [6,7]. Soil water content, soil temperature, soil pH, soil ammonium and nitrate content and carbon (C) availability are some of the key factors that regulate the prevalence of one of these processes as well as its importance.
Organic fertilization is the main fate of manure slurry [8], a by-product of livestock farms rich in NH 4 + , but it has been found to produce higher N 2 O emissions than synthetic fertilizer for the same amount of applied nitrogen [9,10]. Slurry digestate is the by-product of anaerobic digestion of slurry, also used for fertilization [11], and can have a higher NH 4 + content than the original slurry [12].
Digestion of slurry reduces emissions of CH 4 [13] but does not significantly reduce N 2 O emissions [14]. Vinasse is usually the by-product of sugar and ethanol production and it has also been reported to have higher N 2 O emissions than inorganic fertilizers [15,16]. Table 1. Mixture for the 0-10 cm depth layer for each of the treatments and amount of applied biochar and/or fertilizer (g m −2 ). BC300: torrefied granules at 300 • C, BC550: BC300 pyrolyzed at 550 • C, BC850: BC300 pyrolyzed at 850 • C, CS: cattle slurry, SD: slurry digestate, VN: vinasse. The soil used in the experiment was a sandy loam (73% sand, 22% silt, 5% clay) excavated from the A horizon of a permanent grassland on Dystric Endostagnic Glossic Retisol (Colluvic) (FAO World Reference Base). Soil pH KCl was 4.46, and N tot , C tot , P tot , K tot , Ca tot , and Mg tot concentrations were 0.7, 12.4, 0.214, 0.715, 0.538, and 1.48 g kg −1 , respectively.

Biochar
Hay from a permanent grassland dominated by reed canary grass (Phalaris arundinacea L.), cut at the seed ripening phase, was pressed into granules of 10-20 mm in length and 7 mm diameter. These granules were torrefied in the Fraunhofer Institute in Germany (Fraunhofer Institute for Environmental, Safety, and Energy Technology, Sulzbach-Rosenberg, Germany) at a temperature of 300 • C. Torrefaction is a mild pyrolysis, with lower temperatures, but also in the absence of oxygen [51,52]. These torrefied granules (hereafter BC300) were afterward pyrolyzed at 550 • C or 850 • C (hereafter BC550 and BC850, respectively) in the Lithuanian Energy Institute (Lietuvos Energetikos Institutas, Kaunas, Lithuania). The pyrolysis was performed with a slow pyrolysis reactor under laboratory conditions. A container filled with 22 L of the torrefied hay pellets was inserted into the furnace. Before heating, N 2 , at a rate of at 15 L min −1 , was blown through the bottom of the furnace. The furnace was kept at the desired temperature (550 or 850 • C) for one hour. After the pyrolysis, the resulting biochar was left to cool in a closed device for 24 h. The pH for the different biochars ranged from 6.8 for BC300 to 11.8 for BC850 and the C/N ratio from 19.0 for BC300 to 26.6 for BC850 (Table 2). Further details of the different biochar properties can be found in Raave et al. [53].

Fertilizers
The fertilizers used in the experiment were cattle slurry (CS), yeast industry residue vinasse (VN), and cattle slurry digestate (SD). The pH of the three fertilizers was 4.5, 7.5, and 7.9 for vinasse, cattle slurry, and slurry digestate respectively (Table 3), and the NH 4 + -N content ranged between 0.214 for cattle slurry and 0.27 (%) for vinasse, with no detectable presence of NO 3 − -N. The combination of the amount of fertilizer applied (Table 1) and the dry matter content in each fertilizer results in the total amount of dry matter applied (cattle slurry: 29,275 kg/m 2 ; slurry digestate: 19,425 kg/m 2 ; vinasse: 11,986 kg/m 2 ).

Cover Crop
The vegetation cover consisted of ryegrass (Lolium perenne L.). Into each pot, 100 ryegrass seeds were sown, after filling the pots with the substrate and watering to field capacity. During the experiment, the growth room air temperature was kept at 17 • C and air relative humidity at 60%, with a 13:11 h light:dark cycle. All pots were watered with the same amount of water three times a week, at a rate of 125 mL pot −1 during the first growing period (0-30 days) and 150 mL pot −1 during the second growing period (30-60 days). The water amount was calculated to avoid leaching and adjusted according to plant biomass size. Plants were harvested at the end of each growing period. The week before harvesting, watering was increased by 25 mL pot −1 .

Flux Measurements
Flux measurements were carried out by the static closed chamber method [54] between June and August of 2017. The PVC chambers had the same dimensions as the pots, with 110 mm diameter and 30 cm height, and a total volume of 2.4 L. They were designed to fit inside the neck of the pots, achieving a hermetic sealing. The chambers had two holes, one for the sample extraction pipe, and the second for the temperature sensor. Trace gas measurements were carried out weekly, always the next day after watering to assure that the pots had a similar moisture level during all measurement dates.
The gas concentration in the collected air was determined using the Shimadzu GC-2014 gas chromatography system (Shimadzu Corporation, Kyoto, Japan) equipped with an electron capture detector (ECD) and a flame ionization detector (FID). The system is based on the automated gas chromatographic system described by Loftfield et al. [56] and is located in the Department of Geography of the Institute of Ecology and Earth Sciences at the University of Tartu (Estonia).
The N 2 O flux was calculated from the slope of the least-squares linear regression of the N 2 O concentrations versus time [57,58], using the equation: V is the volume of the chamber headspace (m 3 ), and A is the surface entailed by the chamber (m 2 ). Thus, dC/dt is the rate of change in concentration with time (ppm(v) s −1 ). Three replicates for each of the treatments were sampled. The adjusted R 2 value of the linear regression is used to check the quality of the chamber measurement, and if necessary, one of the observations is discarded, using the remaining three for the linear fit [58].
The cumulative flux was calculated as the time-integration of the total daily fluxes, after gap-filling by linear interpolation between measurement points [59].

Fourier Transformed Infrared Spectrophotometer (FTIR) Analysis
Fourier transformed infrared spectrophotometer (FTIR) analyses of the different biochars was carried out using the Thermo-Nicolet iS10 Fourier transformed infrared spectrophotometer (FTIR, Thermo Fisher Scientific, Waltham, MA, USA), with a 4 cm −1 resolution over the range from 4000 to 400 cm −1 , and 32 scans per analysis. The resulting spectra were corrected against the ambient air spectrum as background, applying an automated baseline correction.

Statistical Analysis
Statistical analyses were carried out using R programming language [60]. The fulfillment of test assumptions was checked prior to analysis. A three-way analysis of variance (ANOVA) was performed to study the effect of biochar, fertilizer, and number of days and the interaction between them. A post-hoc Tukey's HSD (honest significant difference) analysis was performed after ANOVA to discriminate between groups for each treatment factor, with the package 'agricolae' in R [61]. The Pearson correlation coefficient was used for correlation analysis. Correlation analysis for fertilizer properties was performed after weighing, according to the application rate to compensate for the different amount of fertilizer applied, and for the differences in dry matter content. Cumulative fluxes are represented by box-and-whisker plots according the conventions described by Tukey [62].

Results
The biochar and fertilizer amendment significantly increased N 2 O emissions (Table 4, Figure 1). The effect of both factors (ω 2 = 0.039 and ω 2 = 0.037, for biochar and fertilizer) was similar. Among organic Agronomy 2020, 10, 109 6 of 17 fertilizers, the three studied fertilizers (CS, SD, and VN) had a significant impact on N 2 O emissions ( Figure 2). N 2 O emissions between the three organic fertilizer treatments were not significantly different (p > 0.05) (Figure 1). Table 4. Results of the three-way analysis of variance (ANOVA) and effect size (ω 2 ) testing the effects of biochar production temperature (BC), applied fertilizer type (FERT), days since beginning of the application of fertilizer (TIME), and the interactions between these factors on N 2 O fluxes for the full duration of the experiment. † Stars indicate significance: ***: 0.001; **: 0.01; *: 0.05; ns: not significant.  Figure 2). N2O emissions between the three organic fertilizer treatments were not significantly different (p > 0.05) ( Figure 1). Table 4. Results of the three-way analysis of variance (ANOVA) and effect size (ω 2 ) testing the effects of biochar production temperature (BC), applied fertilizer type (FERT), days since beginning of the application of fertilizer (TIME), and the interactions between these factors on N2O fluxes for the full duration of the experiment. † Stars indicate significance: ***: 0.001; **: 0.01; *: 0.05; ns: not significant.  The effect of biochar on N 2 O emissions was influenced by its production temperature and increased with it (Figure 1), but only fluxes from BC850 differed significantly from the control. Emissions from BC850 were also significantly higher than those of BC300 (p < 0.05), but not from those of BC550 ( Figure 3). The effect of biochar on N2O emissions was influenced by its production temperature and increased with it (Figure 1), but only fluxes from BC850 differed significantly from the control. Emissions from BC850 were also significantly higher than those of BC300 (p < 0.05), but not from those of BC550 (Figure 3). Biochar mixing with organic fertilizer significantly increased N2O emissions (p < 0.001, for nonpaired t-test) compared to the fertilizer treatment alone (Figure 1). That result is opposite to our hypothesis as we expected that biochar addition to organic fertilizer would reduce N2O emissions.

Sum of Squares
N2O emissions were significantly influenced by the number of days passed after BC and fertilizer incorporation into soil (factor time) (Table 4, Figure 4). The differences between treatments were statistically significant (p < 0.05) only for the first two sampling dates (days 3 and 9 after the application of the organic fertilizers and biochar) of the duration of the experiment (40 days) ( Figure  4). After 9 days, N2O emissions decreased and stabilized in all treatments (including control).
The total cumulative emission for the peak period (20 days) was the greatest from the treatments Biochar mixing with organic fertilizer significantly increased N 2 O emissions (p < 0.001, for non-paired t-test) compared to the fertilizer treatment alone (Figure 1). That result is opposite to our hypothesis as we expected that biochar addition to organic fertilizer would reduce N 2 O emissions. N 2 O emissions were significantly influenced by the number of days passed after BC and fertilizer incorporation into soil (factor time) (Table 4, Figure 4). The differences between treatments were statistically significant (p < 0.05) only for the first two sampling dates (days 3 and 9 after the application of the organic fertilizers and biochar) of the duration of the experiment (40 days) (Figure 4). After 9 days, N 2 O emissions decreased and stabilized in all treatments (including control).
Agronomy 2019, 9, x FOR PEER REVIEW 8 of 16 studied period (Table 6), showing a significant fit for both Ctot (p < 0.05, slope = 0.25) and Ntot (p < 0.05, slope = 7.04) ( Figure 5).   The total cumulative emission for the peak period (20 days) was the greatest from the treatments where organic fertilizers had been mixed with biochar (13.485 mg N/ha), followed by treatments where only organic fertilizer (7.528 mg N/ha) and biochar (5.105 mg N/ha) were applied and the smallest flux was from the control treatment (-0.357 mg N/ha). Among the combination of organic fertilizer and biochar treatments, the greatest total cumulative emission was from treatments where BC850 had been mixed with the fertilizer (19.862 mg N/ha).
Final pH values ranged between 5.08 and 6.09 for all substrate combinations for both treatments (Table 5), while final total carbon content (%) ranged between 1.10 and 1.33 and total nitrogen content (%) between 0.052 and 0.073 among the different treatments (Supplementary Table S1). No significant correlation was found between total emitted N 2 O with the properties of the organic fertilizers (pH KCl , NH 4 + -N, NO 3 − -N, P, K, Ca, Mg) nor with the properties of biochar (pH, ash content, neutralization ability, surface area, cumulative pore volume) or nutrient content (N tot , C tot , P tot , K tot , Ca tot , Mg tot ). N 2 O flux in BC treatments was influenced by the release of nutrients from biochar. A linear model was fit between nitrous oxide fluxes and both released C tot and N tot in the studied period (Table 6), showing a significant fit for both C tot (p < 0.05, slope = 0.25) and N tot (p < 0.05, slope = 7.04) ( Figure 5).    Nutrient uptake by the plant cover was negligible in the first two weeks of the experiment (when the highest N 2 O emissions as well as the biggest differences between treatments were monitored), as the plants had not emerged yet or were in the sprout or seedlings stages. Total final nitrogen uptake by the plant cover was significantly different (p < 0.05) between fertilizers (control, CS, SD, and VN), but no difference (p > 0.05) was found in nitrogen uptake between biochars for the different fertilizers (control, CS, SD, and VN) ( (Supplementary Table S2). Regarding P tot , K tot , Ca tot , and Mg tot , uptake by the crop cover between biochar treatments for the different fertilizer treatments showed significant differences (p < 0.05) only for total phosphorus [53].
The pyrolyzation process resulted in the breaking of the cellulose, hemicellulose, and lignin functional groups (1500-1000 cm −1 region), present in BC300 but not in BC850, as shown by FTIR analysis (Supplementary Figure S1) [63][64][65][66], with intermediate absorbance peaks for BC550, especially for lignin (1593 and 1032 cm −1 ). An increase in aromatic groups (region 1000-400 cm −1 ) was observed with increasing pyrolysis temperature, especially for BC850, with BC550 also showing intermediate peaks. Principal component analysis (PCA) of the FTIR spectra ( Supplementary Figures S2 and S3) shows that BC850 is differentiated from BC300 and BC550 on differences of aromatic, aliphatic, and hydroxyl groups (first component, 70.6% of variability explained), while differences in cellulose, hemicellulose, and lignin content differentiate the three different production temperature biochars (first and second principal components).

Discussion
Organic fertilization significantly increased N 2 O emissions as a consequence of the large amount of NH 4 + -N being incorporated into the soil with the addition of fertilizer. Although, according to pH values, conditions were mostly favorable for denitrification in the soil, some nitrification spots would have allowed oxidation of the ammonium from the organic fertilizers into nitrates, especially in the initial period. Emissions were not influenced by the lower pH or the higher content of K, Ca, and Mg in vinasse. It is well known that biochar can affect the water-holding capacity of the substrates [22], but because of the small relative weight of biochar amendment (0.272% of total soil weight; 0.725% for the 0-10 cm layer), no significant effect on soil moisture compared to control, nor on differences in soil moisture between treatments, is expected. Moisture was also assumed to be not significantly different between treatments, as no significant effect on soil water-holding capacity was expected, watering was the same for all treatments, no leaching or negligible leaching was monitored, uptake by plants was not different [53], and, as the relative amount of amended biochar was small and the biochar pellets were mixed, albedo of the soil should not be affected, therefore not affecting evaporation.
Contrary to our hypothesis, amendment with biochar did not reduce N 2 O emissions. Biochar addition into soil resulted in higher emissions of N 2 O (p < 0.05). Moreover, emissions were higher with higher temperature biochar. The ineffectiveness of the studied biochar in reducing N 2 O emissions could be explained by its low C/N values (Table 2), as the C/N values of the studied biochars were much lower compared to those commonly reported in the literature [21], even for biochar produced from herbaceous residues [23]. Previous research has documented that N-rich biochars with low C/N had no mitigation effect on N 2 O emissions [21]. A meta-analysis by Cayuela et al. [23] found that biochars with a low C/N ratio (<30) did not reduce N 2 O emissions. Spokas and Reicosky [31] found that, from 16 different studied biochars, only biochar with high N content, high labile organic matter, and a low C/N ratio, increased N 2 O emissions, with only one of the 16 different biochars, with a C/N ratio of 19.4, resulted in higher N 2 O emissions on all soils, while the other biochar generating higher emissions in different soil types presented C/N ratios between 21.85 and 207.5. Clough et al. [46] showed that biochar can increase N 2 O emissions, if soil moisture is favorable for denitrification, but also when large amounts of C and N are released from biochar. This is in line with our results, as the biggest losses of C and N were found for BC850, from which the highest emissions were also measured. Although BC850 presented the lowest initial N content (Table 2), and extractable carbon and nitrogen were higher for lower production temperature biochar (Supplementary Table S1), the higher release of nutrients measured with higher temperature biochar could be a consequence of the pH of the soil-biochar interface being closer to the optimal for ammonification, increasing the mineralization rate of biochar. The idea that BC's enhancing effect on N 2 O emissions comes from C and N release is supported by the positive significant relation found from the linear fit between N 2 O flux and the released C and N amount. Released carbon becoming available would promote denitrifiers' growth, and reduce O 2 availability in the substrates, indirectly promoting denitrification too [14]. Available C is a key factor controlling denitrification [67][68][69], which was the predominant process involved in N 2 O production in the substrates, as a consequence of soil pH during the experiment and soil moisture, as the measurements were carried out the day after watering. The pH values at the end of the experiment for the substrate mixes were in the range between 5.08 and 6.09 (Table 5), which are below the optimum for nitrification (7.8-8.9) [70]. Denitrification is less sensitive to pH than nitrification, and, at lower values of pH, the ratio N 2 O:N 2 increases [71], supporting that denitrification was the predominant process involved in the N 2 O emissions. Moreover, Hütsch et al. [72] reported the biggest losses due to denitrification, mainly N 2 O, for pH values of 5.2 and 5.9 for sandy and loamy soils respectively, which are almost coincident with the final pH values of the substrates tested in this experiment. It has also been reported that fast increases in soil pH can cause a sharp increase in denitrification potential, due to the solubilization of organic matter [73]. It has been documented that denitrifying bacteria recover faster when soils are rewetted after dry periods [74], and that rewetting after dry periods increases the concentration of dissolved organic carbon (DOC) [75] and C availability [76,77]. This would explain why the emissions peaked after the initial watering of the dry soil, and why the biggest differences among treatments were measured during the first two weeks. Firestone and Davidson [6] established that large increases in N 2 O emissions, like those observed after rewetting, are more typical for denitrification than for nitrification, which supports our assumption that conditions in the substrates were more favorable for denitrification, and that denitrification was the main reason for the observed emissions. According to Petersen [78], denitrification in manure slurries is promoted by the readily assimilable carbon. This would explain why biochar amendment does not reduce emissions with organic fertilization. In fact, organic carbon input can promote anaerobic conditions, even at low water-filled pore space (WFPS) values, increasing denitrification potential and N 2 O emissions [79].
Higher losses of C with the highest production temperature biochar would explain why this biochar presented the highest emissions.
The higher pH from higher pyrolysis temperature biochars could have also contributed to the differences in N 2 O fluxes, for higher rates of denitrification are expected under slightly alkaline than under acid conditions [80]. Finally, higher phosphorus content in higher production temperature biochar could have also promoted higher N 2 O emissions, as P content promotes N 2 O emissions from both nitrification and denitrification processes [81].
An alternative mechanism for explaining the differences in emissions after application of the different biochars could be a combination of biochar-mediated soil organic matter mineralization enhancement, in parallel with NH 4 + sorption by the biochar which, in turn, reduces the nitrification potential. Although biochar has been reported to have opposed effects on mineralization [46], meta-analyses show a general increase in mineralization activity in biochar-amended soils [22,23], an increase that is usually attributed to the priming effect of biochar, by stimulating microorganisms to mineralize recalcitrant soil organic matter in response to the C input [21,41,46]. This would lead to an increase of nitrification activity [21], due to the NH 4 + becoming available as a result of the ammonification process. Ammonification is not so sensitive to low pH as nitrification [82,83]. In parallel, NH 4 + sorption by biochar would decrease nitrification activity, explaining the differences in the final N 2 O emissions from different treatments. As sorption takes place in the surface of biochar, it will be conditioned mostly by the biochar pH and not by the pH of the substrates. The different pH and, consequently, NH 4 + sorption capacity of the different biochars would result in different NH 4 + availability and, therefore, nitrification activity. Higher temperature biochar presents lower NH 4 + sorption potential [84,85], leading to more available NH 4 + in soil for nitrification. Although ammonium sorption capacity increases with increasing pH [84,85], higher pH values will cause a shift in the ammonium-ammonia equilibrium towards ammonia [86]. Wang et al. [86,87] observed that adjusting the pH of the solution to 7 resulted in increased NH 4 + sorption. Assuming that the three biochars enhanced mineralization, then BC300, with its pH close to 7 and higher content of negatively charged functional groups present in cellulose, hemicellulose, and lignin with lower temperatures, as shown in the FTIR analysis (Supplementary Figure S1), would have been the one with the highest NH 4 + sorption capacity, followed by BC550. However, the low levels of organic N and P in the soil would limit ammonification, and the low pH of the substrates would limit nitrification, therefore seeming unlikely that the combination of ammonification and NH 4 + sorption could cause the measured differences in the N 2 O fluxes. Even for biochar and fertilizer combined amendment, where N was not limiting due to the addition of NH 4 + -N, as the interaction between biochar and fertilizer was not significant, this mechanism of combined ammonification and nitrification did not play a significant role, with the low pH for nitrification being the most likely limiting factor.

Conclusions
Organic fertilizers increased N 2 O emissions in comparison to the control as expected, but the increase was not different among fertilizers. Biochar amendment did not mitigate the increase in N 2 O emissions that resulted from organic fertilization, as biochar itself produced higher emissions. No significant interaction between biochar and fertilizer treatments was found. Biochar addition also increased emissions, with this difference being significant only for higher temperature production biochar. Our findings support the idea that biochar cannot be considered as a universal tool for the reduction of GHG emissions, and that each biochar, depending on its physical and chemical properties, can have a different effect. A low C/N ratio and a high N content in biochar can lead to higher emissions in biochar-amended soils. This study highlights the need for more research, focusing on different feedstocks and production methods of biochar, to provide a better understanding and allow future modeling of the mechanisms relating biochar properties and N 2 O emissions. This is particularly important for biochars produced from non-woody source materials, and especially with low C/N ratios, more likely to have an unexpected and undesired effect on GHG emissions.

Supplementary Materials:
The following are available online at http://www.mdpi.com/2073-4395/10/1/109/s1, Figure S1: Fourier-transform infrared spectroscopy (FTIR) spectrum of the studied biochars (300 • , 550 • , and 850 • ) with identified functional groups. Figure S2: Biplot of the first two components of the principal component analysis (PCA) of the Fourier-transform infrared spectroscopy (FTIR) spectrum values for the three production temperature biochars (300 • , 550 • , and 850 • ). Figure S3: Scatterplot of wavenumber values and eigenvalues for the first two principal components of the principal component analysis (PCA) of the Fourier-transform infrared spectroscopy (FTIR) spectrum data. Table S1: Extractable Nitrogen and Carbon of the different production temperature biochar after 120 h. Table S2