Alkaline Wood Ash, Turbulence, and Traps with Excess of Sulfuric Acid Do Not Strip Completely the Ammonia off an Agro-waste Digestate

The present study combined two nutrient management strategies to improve the marketability of a waste-derived fertilizer: (a) isolation of ammoniacal nitrogen and (b) preparation of a bulk soil amendment. The wood fly ash with low content of pollutants was added to an agrowaste anaerobic digestate as alkaline stabilizer, which promoted the volatilization of ammonia and adsorption processes, and as nutrient supplement. The 39.71 ± 1.44 g blend was incubated for 60 hours at 20°C and 100 rpm in a closed chamber (250-mL Schott Duran bottle) with a 5.21 ± 0.10 mL sulfuric acid trap of 10 different concentrations (0.11, 0.21, 0.32, 0.43, 0.54, 0.64, 0.75, 0.86, 0.96, and 1.07 mol/L). For analytical purposes, the sulfuric acid, water-soluble, and water-insoluble fractions of the blend were isolated after the incubation. The 1.07 mol/L sulfuric acid solution contained 23.69 ± 5.72 % more of ammonical nitrogen than the 0.11 mol/L solutions. However, in all cases the amount of nitrogen in the H2SO4 compartment was lower than the one in the water-soluble and water-insoluble fractions. Only the 15.52 ± 2.13 % of the nitrogen accounted after the incubation was found in the H2SO4 trap. The bottleneck of the NH3 stripping process was the rate of mass transfer at the interface between the blended fertilizer and the headspace of the closed chamber. The organic phosphorus was more susceptible to be adsorbed during the alkaline treatment with non-intrusive acidification than the nitrogen and carbon. Activation of the ash as adsorbent before mixing with the digestate should improve the properties of the blend as slow release fertilizer, since more nutrients would end in the water-insoluble fraction.


Introduction
The commercialization of waste-derived fertilizers is constrained by the level of contaminants [1]. It might be easier to isolate the plant nutrients contained in the residues and sell them as conventional fertilizers [2]. The addition of ashes to organic manures, such as anaerobic digestates, could be proposed for: (a) improving the properties of these organic amendments as fertilizers; (b) reducing the greenhouse emission and phosphorus leaching associated to the management and use of these materials; (c) manufacturing of a granular fertilizers; and (d) increasing the pH to promote the volatilization of ammonia and subsequent capturing the NH3 in a sulfuric acid trap. The ammonium sulfate could be sold as liquid fertilizer (40-60% (NH4)2SO4 solution) or as solid after crystallization. On the other hand, the use of clean materials with low content of pollutants, such as biomass ash and agro-industrial digestate, enables the end-of-waste status and the marketability of the blend as bulk soil amendment [3][4][5][6][7][8][9].
Mixing wood ash and agro-waste digestate to get an approximate carbon, nitrogen, phosphorus ratio (C/N/P) of 40/2/1 could enhance the use efficiency of these elements upon spreading the blend in the soil [10,11]. Given the high pH of the blend ash-digestate, this material could be used as liming agent, after the removal of the ammoniacal nitrogen (NH4 + -N) [12]. This nutrient management strategy might offer better results than the preparation of the blend under acid conditions to promote the adsorption and to minimize the loss of NH4 + -N [13]. to a 4.5% (w/v) blend of biochar and cattle slurry mitigated less the NH3 emissions than just applying the 0.3 mL H2SO4 (98%) to 50 g of cattle slurry to reach a pH 5.5. The stepwise mechanism of acidification, dehydration, and adsorption or flocculation is widely used to improve the management of anaerobic digestates [6,14]. The present study analyzed how the H2SO4 non-invasive acidification) affects the alkaline stabilization of a blend of Wood Fly Ash (WFA) and Post-Harvest Vegetable Waste Digestate (PVWD) in terms of: (a) NH4 + -N recovery, (b) N, C, and P availability and, (c) overall C/N/P. The severe H2SO4 non-invasive acidification was meant to decrease the amount of NH3 in the gas phase and create a greater gradient of concentration, which would be enough to overcome the mass transfer resistance at the layer between the blend WFA+PVWD and the headspace. The H + that remains in the blend WFA+PVWD due to the dissociation of WS NH4 + and subsequent volatilization of NH3, might promote dehydration and adsorption processes affecting the availability of nitrogen, carbon, and phosphorus.

Materials and Methods
For the preparation of the blend WFA+PVWD, the 10 mL of ultrapure milli-Q ® water was added to each gram of sample before blending (1:10). This improved the fluency of the waste streams and generated the Water-Soluble (WS) fraction of the blend WFA+PVWD. A detailed description of the components of the blend WFA+PVWD is shown in the  repetitions for each of the 10 H2SO4 non-intrusive acidifications) were prepared. This methodology also offered more realistic results about the potential of the H2SO4 noninvasive acidification to affect the composition of the blend WFA+PVWD. The way of conducting this experiment was based on a previous study [15] (Unpublished) and aimed that, by the end of the treatment, all the fractions of the blend were in equilibrium, including the H2SO4 fraction. It is important to mention that the stoichiometric amount required to capture all the nitrogen of the blend WFA+PVWD ( Table 1) would correspond to a 3.43-mL solution of 0.11 mol/L H2SO4.   A 3-µm filtration was required for the solid-liquid separation of the WS extract and the WI material. The three fractions of the blend were weighed with a precision balance and the calculation of the volumes of H2SO4 solution and WS extract was done assuming a density of 1 g/mL. The availability of an element was defined as the ratio of the WS form to the WI form. The average recovery effectiveness of an element was calculated as its final amount measured after the incubation divided by its initial amount before starting the incubation. The trap effectiveness of stripping the NH3 off the headspace was calculated as the ratio of the [H2SO4] NH4 + -N to the nitrogen which could not be found in any of the WS and WI fractions. The One-Way Analysis of Variance (ANOVA) was performed with Microsoft Excel (p<0.05) to decide whether the H2SO4 non-invasive acidification affected significantly the composition of the blend WFA+PVWD.

Fractionation of the blend after the incubation
The mass of the 5 mL H2SO4 aqueous solutions in the traps increased in agreement with the content in sulfuric acid (Figure 3a). The Table 1: Initial characterization, expressed in fresh basis, of a blend (n=4) of 0.51 ± 0.07 g WFA and 3.40 ± 0.41 g PVWD prepared in a closed chamber with a 4.39 ± 0.02 g trap of 0.11 mol/L H2SO4 aqueous solution. The amounts of milli-Q ® water added to WFA and the PVWD were 5.58 ± 0.02 g and 30.43 ± 0.05 g, respectively [17].
A volume of 17.11 ± 3.45 mL of WS extract was lost during the 60 hours incubation at 100 rpm and 20 °C and subsequent filtration of the blends (Figure 3b). According to the ANOVA test (p<0.05), there was not significant increase in the amount of WS fraction recovered when using H2SO4 solutions in the trap with greater concentration than 0.43 M. This effect would be explained by the neutralization of the surface negative charges of the colloids of the PVWD by adding a cationic surfactant or via intrusive acidification, which make feasible their dehydration and flocculation. Similarly, the losses of WI material were 0.13 ± 0.05 g and did not show dependence on the concentration of the H2SO4 solution in the trap (Figure 3c). The explanation for the constant losses could be the procedure followed to achieve the solidliquid separation. Some of the WI material would have remained stuck to the walls of the closed chamber and any weight gain due to the hydration of the ashes would have been lost during the drying at 105°C, before weighting the mass of the WI fraction recovered [13,14].

pH of sulfuric acid trap and the blend WFA+PVWD
Since the volume of the traps were 5.21 ± 0.10 mL (Figure 3a), the pH of the H2SO4 solutions were measured directly (at 23.38 ± 0.36 °C) and also after eleven times dilution, to ensure a better contact with the probe of the Mettler Toledo ® Seven CompactTM S220 pH/Ion meter. For the second set of measurements, the pH in the undiluted traps was calculated by increasing an order of magnitude on the concentration of the H + species determined in the eleven times diluted H2SO4 solutions. The first thing that needs to be highlighted is that the pH decreased during the incubation. This is opposite to what was expected since the absorption of NH3 should increase the pH of the H2SO4 solutions.
Understanding why the pH of the trap decreased is important to enhance the absorption and recovery of the NH3 in the headspace. It could be possible that the pH of the trap decreased because of the evaporation of the water and subsequent increase in the concentration of H + ions.
However, the losses of the mass of the traps were negligible. It should be noted that, even when the pH decreased during the incubation, greater values than the theoretical ones were obtained. For example, the 1.07 M H2SO4 solution should have a pH lower than zero (i.e. -0.03) before the incubation and the value measured was 0.74 ± 0.02. After the incubation, the calculated value of pH from the measurements in the eleven times diluted H2SO4 solutions (0.03 ± 0.06; Figure 4a) was lower than the values measured in the undiluted traps (0.45 ± 0.05; Figure 4a). Thereby, these calculated values of the pH could be considered more accurate than the values of the pH obtained from the direct measurements of the H2SO4 solutions after the incubation. Therefore, the H2SO4 solutions needed to be diluted for the measurement of the variations in the concentrations of the H + species due to the upper detection limit of the pH-meter. The pH of the blend WFA+PVWD was not affected by the level of non-invasive acidification (10.40 ± 0.46; Figure 4b).  (Table 1) and the milli-Q ® water used to prepare the blend (Figure 1). The initial mass of WI material was assumed the dry matter of the blend ( Table 1).

Speciation of nitrogen
The [H2SO4] NH4 + -N was the only nitrogen species that increased significantly (p<0.05) with respect to the initial characterization ( Table  1) Table  1). High [H2SO4] NH4 + -N was expected because of the low pH of the H2SO4 solutions in the trap (0.29 ± 0.24; Figure 4a) were able to absorb the NH3 available in the headspace. The effect of the non-intrusive acidification can be seen in Figure 5a, which shows increase (23.69 ± 5.72 %) in [H2SO4] NH4 + -N and decrease (41.08 ± 17.46 %) of WS NH4 + -N when the concentration of the H2SO4 solution in the trap was increased from 0.11 mol/L to 1.07 mol/L. Furthermore, Figure 3b shows the increase in the amount of WS fraction due to the presence of a H2SO4 trap with a concentration greater than 0.43 M. This significant dehydration of the WI fraction has a p-value of 0. 16 invasive acidification did not promote the adsorption of WS N as much as the HCl intrusive acidification.  A calibration procedure is typically required for measuring the concentration of NH3 in the air using H2SO4 solutions [18]. Nevertheless, in the present study, the trap effectiveness in all the conditions evaluated was 100% (i.e. all the NH3 released to the headspace was absorbed in the H2SO4 solutions in the trap) and what limited the depletion of the WS NH4 + -N, was the mass transfer resistance at the film between the fluidized blend and the gaseous phase. This constant value of trap effectiveness was calculated without considering the amount of sulfuric acid used, which ranged from 0.05 ± 0.00 g to 0.54 ± 0.01 g of H2SO4. Otherwise, there would be a difference of one order of magnitude between the effectiveness of the less concentrated and the most concentrated traps. The concentration of the H2SO4 trap did not affect the volatilization of NH3 from the blend WFA+PVWD. Similar amounts of [H2SO4] NH4 + -N were found in all the conditions evaluated (0.99 ± 0.38 mg; Figure  6b). Most of the nitrogen recovered was in the form of WI N and only the 15.52 ± 2.13% of the nitrogen accounted after the incubation was in the form of [H2SO4] NH4 + -N. The overall recovery effectiveness of nitrogen was 117% (Figure 6b), as more nitrogen was accounted after the incubation than in the initial characterization. Most of the volatilization of NH3 took place immediately after the blending and the absorption in the H2SO4 trap continued progressively during the course of the incubation, due to the mass transfer resistance [16]. Thereby, in processes with short contact time between the gas and the H2SO4 solution, it might be convenient to use bubbling systems to increase surface area between the two phases and thus the rate of transfer of NH3 [19].
As the WFA+PVWD blend had a pH of 10.40 ± 0.46 (Figure 4b), the 90 % of the WS NH4 + -N was in the form of NH3 [20]. The closed chamber continuously shaken at 100 rpm was not enough to enable the equilibrium between the three fractions of the blend WFA+PVWD. This setup was chosen for its low capital and operating cost but the use of an excess of H2SO4 in the trap could not be justified technically and economically. It might be possible to attain the depletion of the WS NH4 + -N in the blend WFA+PVWD using advanced equipment, which allow to operate at higher temperatures under vacuum conditions (e.g. 65 º C and 25.1 kPa; [7]) or perform hydraulic cavitation [21]. Another processing option would be to reduce the moisture content of the blend, for the production of the granular fertilizer. As the surface area of the dewatered material is greater than the fluidized blend WFA+PVWD, the emissions of ammonia increase [5,22,23].

Speciation of carbon
Similarly to nitrogen, most of the carbon was in the WI form ( Figure  7a).  Table 1). Therefore, the H2SO4 non-intrusive acidification did not promote the adsorption of the WS Corg onto the WFA, the release of CO2 or emission of volatile organic molecules. It is important to mention that, Ukwuani & Tao (2016) reported the flux of other compounds different from NH3, such as cyclohexene, towards the H2SO4 trap. In the present study, the concentration of the H2SO4 solution in the trap was not responsible of any phenomena which affected the distribution of carbon between the fractions of the blend WFA+PVWD and did not affect the recovery effectiveness of carbon after the incubation. The alkaline pH (Figure 4b) of the blend prevented the losses of carbon and the recovery effectiveness after the incubation was 115% (Figure 7b), which could be related to the fact that the blend was a sink of carbon.

Speciation of phosphorus
The concentration of WS PO4 3--P (3.75 ± 1.45 mg/kg; Figure 8a) and WI P (792.88 ± 218.92 mg/kg; Figure 8a) after the incubation was the same as before the incubation (5.42 ± 3.84 mg WS PO4 3--P/kg and 636.01 ± 140.53 mg WI P /kg; Table 1). Thus, most of the phosphorus was in the form of WI P, regardless the concentration of the H2SO4 solution in the trap. Both the fact that there were no losses of phosphorus via gaseous emissions in the studied conditions and the adsorption of the WS Porg could explain the 115 % of average recovery effectiveness (Figure 8b). Any change in the amount of adsorbed WS Porg would be accounted by the colorimetric analytical method (i.e. molybdenum blue reaction) followed. The reason is that the sulfuricperoxide digestion of the WI fraction led to the formation of WS PO4 3--P, which was measured with the segmented flow analysis. The availability of phosphorus went from 0.0083 ± 0.0021 mg WS PO4 3--P/mg WI P at the beginning of the incubation to 0.0048 ± 0.0015 mg WS PO4 3--P/mg WI P at the end. This tiny difference in the availability could prevent losses via leaching in an open system, for example, when applying phosphorus to land at a rate of 26 kg/ha [10].

Conclusions
Exceeding the H2SO4 non-intrusive acidification beyond the stoichiometric limit was not an efficient way of stripping the WS NH4 + -N off the blend WFA+PVWD. The depletion of the NH3 in the headspace of the closed chamber and the turbulence created with the 100 rpm rotary mixing did not sort out the bottleneck of NH3 transfer from the WFA+PVWD blend to the gas phase. Advanced processing conditions (e.g. vacuum thermal stripping and hydraulic cavitation) or dewatering of the blend are required to increase the rate of NH3 volatilization. Although a 10 times increase of the concentration of the solution in the trap resulted in 23.69 ± 5.72 % more [H2SO4] NH4 + -N, this only represented an increase of 3.57 ± 2.05 % in the overall fate of nitrogen in this fraction. The C/N/P:38.74 ± 17.56/2.38 ± 0.69/1 found after the incubation was the same as the intended C/N/P:40.55 ± 15.72/2.38 ± 0.80/1 (Table 1), as there was no losses during the treatment. The greatest share of these elements was found in the WI fraction and the WS Porg was more susceptible to be adsorbed under the studied conditions than the WS Norg and WS Corg. This controlledrelease fertilizer should minimize the greenhouse gas emissions, eutrophication of underground waters upon land application, and pollution swapping. Activating the WFA as adsorbent, for example via calcination at temperatures greater than 500 º C, could improve the valorization of PVWD by enhancing the retention of nutrients in the WI fraction of the blend. Funding sources