Activated Carbon Reduced Nitrate Loss from Agricultural Soil but Did Not Enhance Wheat Yields

Abstract

Wheat requires a high quantity of nitrogen to grow efficiently and produce a high number of nutritious grains (=high yield). The consequences of fertilizer use in uncontrolled conditions are well known, e.g., nitrogen leakage leading to impacts on ecosystems. One solution to reduce these impacts could be activated carbon, which is already used to treat wastewater. In this study, we assessed the efficiency of four activated carbon formulations applied to two agricultural soils in a column leaching test, a greenhouse pot experiment, and a field experiment. In the latter two experiments, wheat was grown with one dose of one selected activated carbon and several nitrogen fertilization conditions. The goal was to find an activated carbon that could stabilize nitrate while improving soil health and thus increase wheat yields. We showed that nitrogen leaching reduction (between 59% and 79% when significant in the column test) was dependent on the raw material used and the activation process. The controlled pot experiment demonstrated that wheat growth was dependent on nitrogen application (60 to 80% increase) and that the activated carbon addition did not enhance yields. Finally, field trials showed that the addition of 1% activated carbon did not result in higher wheat yields compared to those in the non-amended plots in both the absence and presence of nitrogen. In conclusion, although the activated carbon chosen is a strong nitrate-retaining agent, it does not deliver the expected yield gains, probably due to excessive retention, which prevents increased plant growth. More research is needed to improve activated carbon as a fertilizer.

1. Introduction

In Europe, wheat is the most commonly cultivated grain crop with 73% of the croplands used for cereals being dedicated to wheat [1]. Wheat is particularly important in France, as the country is the fifth highest producing and fourth highest exporting country in the world [2]. Wheat yields are highly dependent on chemical fertilizer, especially for nitrogen. And with the expected increase in the human population, yields will need to be further increased (by 60% compared to those in the 2000s [3]), as this higher need for food is combined with demographic pressure, and thus, cropland surfaces are not increasing (and could potentially decrease due to construction needs). Therefore, chemical fertilizer use in agriculture has greatly increased within the last decades, which has affected the environment. For instance, only a part of the applied N is consumed by plants [4], and the rest is emitted into the atmosphere as a greenhouse gas, contributing to global warming, or it is leached [5,6,7], polluting ground waters and participating in eutrophication [7,8]. Such pollution due to agriculture intensification has been recognized since the 1970s in the industrialized countries of North America and western–central Europe [9]. For instance, nitrogen has been recognized as the second most important chemical contaminant in surface and ground waters [7]. The excessive application of chemical N fertilizer also endangers soil organic matter (OM) reserves, as it promotes mineralization [3], which ultimately puts pressure on crop yields and contributes to greenhouse gas emissions and, ultimately, global warming. In addition, these chemical fertilizers are fossil based, and the elements need to be mined, which damages the environment. As the pools diminish, we will face a shortage in the upcoming decades [10]. Finally, with the rise in the price of chemicals over the last few years, fertilization has put pressure on farmers’ income. Moreover, since the 1990s, a slowing down of the increase in yields has been observed across Europe, and France has even been facing a small reduction in yields since 2000s [2]. Therefore, there is a need to find solutions to replace chemical fertilization, or at least to improve its use efficiency, and thus reduce its loss and its known consequences on ecosystems.

One of the solutions is to provide nutrients via organic amendments such as compost, manure, or biochar. Most of those amendments are rich in nutrients as well as OM, which improves the fertility of the soil. However, compost and manure are easily degradable and thus need to be applied frequently. On the contrary, biochar is a more stable product but may contain less nutrients. Biochar is produced from the pyrolysis of biomass, usually waste materials such as wood chips, crop straws, manure, sewage sludge, etc. It is characterized by a high carbon (C) content, high porosity, and high surface area [11]. Biochar has a great sorption capacity for cations but is less efficient for anions like NO₃⁻, which is the principal form through which N is lost via leaching [12,13]. To improve its sorption capacity, biochar can be modified/activated, yielding activated carbon (AC). Several studies have shown that AC has a great capacity to sorb nitrate through electrostatic interactions with functional groups or physical retention in pores in batch sorption tests [14,15,16,17]. Activated carbon is already used as a filter in the treatment of wastewater. However, as sorption occurs through electrostatic interaction and physical retention, the surface properties of the activated carbon have a great effect on its sorption capacities, and thus, the initial feedstock and activation methods affect nitrogen retention, as we demonstrated in our previous study [17]. But although the evaluation of AC sorption capacity for nitrate has been well evaluated in microcosm batch experiments, research on AC in real soil conditions is still scarce, and not only AC properties but also application doses and initial soil characteristics influence nutrient dynamics in the soil [18,19].

The objectives of our study were to evaluate the effects of a selected AC formulation (selection made from our previous results, [17]) on (i) leachate composition, (ii) microbial activity, and (iii) plant performance. Although studies have assessed those parameters separately, we evaluated them all together through a multiscale approach, which has not been performed yet. In this multiscale approach, we performed laboratory experiments in microcosms (column leaching test) and mesocosms (pot experiment), together with a field validation experiment (macrocosm), which is lacking in most of AC-related studies. In these three experiments, we examined associations between plant yield and soil health indicators to evaluate the effectiveness of AC as a replacement for chemical nitrogen fertilizer. Based on this multiscale approach, we hypothesized that (i) the application of AC reduces the loss of nitrate through leaching; (ii) the application of AC improves the activity of microorganisms, especially in relation to the N cycle; and (iii) the application of AC improves the yield of crops.

2. Materials and Methods

2.1. Soils and Amendments

Two French agriculture sites were studied, which were both located in the Centre-Val-de-Loire administrative Region (France). The first soil was located in Saugy (S) in the agricultural area called “Beauce”. The second site was located in Tournoisis (T) in the “Champagne-Berrichonne” agriculture area. Both soils have a clay–limestone texture. Soils were characterized for basic physico-chemical properties (Chambre d’Agriculture du Loiret, France). The results are shown in

Table 1. Soil chemical properties.

		Saugy	Tournoisis
Soil texture		Clay limestone	Clay limestone
Cation exchange capacity (meq/100 g)		18.2	34.7
Total nitrogen (g.kg−1)		2.16	2.27
Organic carbon (g.kg−1)		24.2	20.9
Organic matter (g.kg−1)		43.2	45.4
C/N		11.2	9.2
pH (H2O)		8.28	8.33
Olsen [P] (mg.kg−1)		55	43
Exchangeable cation concentrations	K (mg.kg−1)	416	437
	Ca (mg.kg−1)	8476	11,129
	Mg (mg.kg−1)	79	211
	Na (mg.kg−1)	6.68	16.3
Total element concentrations	Cu (mg.kg−1)	2.6	2
	Zn (mg.kg−1)	3	1.4
	Mn (mg.kg−1)	25.2	16.9
	Fe (mg.kg−1)	7.8	9.3

.

Based on the results of our previous batch sorption study [17], we have selected four AC (coded 5K, 2K, CS, L27) (Jacobi Carbons, France), presenting the best nitrate sorption capacity. The codes of the materials are based on their industrial names. Two of those AC were produced from a mineral feedstock (i.e., 5K and 2K), and the other two were from a vegetal feedstock (i.e., L27 and CS). They were obtained following either a physical activation (for 5K, 2K, and CS) or a chemical activation for L27. The materials have been characterized during our previous study [17]. Data are summarized in

Table 2. Physico-chemical properties of the activated carbons (n = 5). EC = electrical conductivity, Eh = redox potential, SSA = specific surface area, TPV = total pore volume, MPD = mean pore diameter.

Activated Carbon	pH	EC (µS.cm−1)	Eh (mV)	SSA (m2.g−1) *	TPV (cm3.g−1) *	MPD (nm)	C (%)	H (%)	N (%)
L27	1.9 ± 0.1	2680 ± 2	525 ± 1	1500	2.78	15	89 ± 5	2.10 ± 0.09	1.54 ± 0.16
5K	10.5 ± 0.0	122 ± 18	51 ± 7	1060	1.77	15	85 ± 1	0.04 ± 0.03	2.85 ± 0.46
2K	9.5 ± 0.0	5 ± 0	200 ± 4	1050	0.87	9	82 ± 2	0.36 ± 0.01	1.91 ± 0.79
CS	10.7 ± 0.0	349 ± 4	157 ± 2	1100	0.67	7	86 ± 4	0.59 ± 0.13	0.13 ± 0.02

* Results of one BET measurement.

.

2.2. Microcosm Laboratory Experiment: Column Leaching Test

From our previous experiment [17], four activated carbons with good sorption capacity toward nitrate were selected. But the effects can be different once in the soil. To select the most appropriate material to apply to the soils, a leaching test was performed. In this test, the four materials were applied to the two agricultural soils, and the leaching of nitrate was monitored. The goal of this experiment was to evaluate the influence of AC in the short to medium term (one trimester) on nitrate loss through leaching and on microbial activities, then to select the most promising material to be used in the mesocosm laboratory and macrocosm field studies, and last verify hypotheses 1 and 2 at the microscale.

For this leaching test, closed columns (50 mL volume) were filled with 40 g of soil (S or T), amended with either 5K, 2K, CS, or L27, at three doses (1%, 2%, and 5%, dw/dw). A control treatment, without amendment, was also prepared for each soil. In total, per soil, 65 columns were prepared (five replicates per treatment). The substrates were watered at field capacity and equilibrated for two weeks. After those two weeks, 25 mL of N solution, prepared from the commercial N solution 390 (fertilizer often used by the farmers) diluted in ultra-pure water to obtain a concentration of 50 N units, was applied to each column. The solution 390 contains 39 units of nitrogen per 100 L in the form of NO₃ (9.75 units), NH₄ (9.75 units) and urea (19.5 units). The concentration chosen corresponds to a N dose applied by farmers during one fertilization episode.

Leachates were collected at regular intervals, i.e., after 0, 14, 28, 42, 56, 84, and 105 days, by adding 25 mL of dH₂O, except for day 0, when the N solution was applied, and collecting the solution (25 mL) at the bottom of the column. The samples were used directly to measure pH and electrical conductivity (EC), using a multimeter (SevenExcellence, Mettler-Toledo, Columbus, OH, USA). Following, samples were filtered (0.45 µm, cellulose acetate membrane) and analyzed for anion (fluoride, bromide, chloride, nitrate, nitrite, phosphate, and sulfate) concentrations, using ionic chromatography (IC) (ThermoFisher, Courtaboeuf, France). The analysis was performed under 1.2 mL.min⁻¹ isocratic eluent (4.5 mM sodium carbonate/1.4 mM sodium bicarbonate) with a 31 mA suppressor current and an AS22 column (ThermoFisher, Courtaboeuf, France). Each run lasted for 13 min, compounds were identified based on their retention time, and concentrations were calculated using a 4-point curve (0–30 mg.L⁻¹), which was prepared by diluting commercial stock solutions. Only the anions with measured concentrations above limits, i.e., nitrate and phosphate, were kept and will be presented in this study. Data were expressed on a soil dry weight basis, and the cumulated leached amount was calculated for each column. Ammonium was also measured but concentrations were below detection limits.

After 105 days, soil samples from the control treatments (S and T without activated carbon amendment) and from the 5K amended treatments were sampled and analyzed for enzyme activities. The choice of those treatments was based on the fact that the material 5K was the most promising to reduce nitrate leaching through time in both soils. The enzymes measured were β-glucosidase, alkaline and acid phosphatases, hydrolysis, and urease (methods described in

Table 3. Brief description of the methods used to measure soils enzymatic activities.

Enzyme	Substrate	Buffer	Absorbance Wavelength	Reference
β-glucosidase	4-nitrophenyl-β-D-glucopyranoside (10 mM)	Citrate phosphate buffer (0.15 M, pH 4–5)	410 nm	[20,21]
Alkaline phosphatase	4-nitrophenyl phosphate disodium salt hexahydrate (5 mM)	Tris-HCl (0.1 M, pH 8)	410 nm	[21,22]
Acid phosphatase	4-nitrophenyl phosphate disodium salt hexahydrate (5 mM)	Sodium acetate (0.1 M, pH 5)	410 nm	[21,22]
Hydrolysis	Fluorescein diacetate (50 mM)	Potassium phosphate (60 mM, pH 7.6)	490 nm	[21,23]
Urease	Urea	Sodium acetate (50 mM, pH 5)	650 nm	[24]

). Such enzymes are pertinent indicators of soil health, regarding the C, N and P cycles, as well as the general microbial activity. In addition, nitrification potential was assessed [20]: 1.5 g of fresh soil (stored at 4 °C) was mixed with 10 mL of potassium phosphate buffer (pH 7.2, 1 mM) containing NH4Cl (1 mM) and incubated (105 rpm) for 144 h at 18 °C. Following centrifugation (14,000 rpm, 10 min), the supernatant was collected, and nitrite and nitrate concentrations were measured by IC (as described earlier).

2.3. Pot Culture Test

Based on the results of the column-leaching test, the material 5K was selected for the crop growth mesocosm test under greenhouse conditions. This pot experiment aimed to evaluate the effect of this material under controlled conditions not only on the leaching of nitrate but also on the loss of organic carbon, the health of the plants, and the microbial activities, and thus verify at the meso-scale the three hypotheses.

For each soil, five treatments were prepared: (i) control (S or T) in which no amendment or fertilization was added, (ii) fertilization (S+F or T+F), (iii) 1% (dw/dw) activated carbon (S+AC1 or T+AC1), (iv) 1% (dw/dw) activated carbon and fertilization (S+AC1+F or T+AC1+F), and (v) 0.5% (dw/dw) activated carbon and fertilization (S+AC0.5+F or T+AC0.5+F). After their preparation, substrates were put in 4 L pots (containing 4 kg of soil) and equilibrated for two weeks at field capacity. Ten pots were prepared per treatment, which were further split in two: half were left unvegetated, and the other half were sown with three seeds of wheat and subsequently further reduced to one plant per pot. The pots were placed in a controlled climatic chamber with the following parameters: 22 °C during the day and 16 °C during the night, 16 h of light, with a light intensity of 400 W. Fertilization (commercial N solution 390) was applied to the substrates of the concerned treatments three times during the experiment time course: after 34 days (50 N units), after 83 days (50 N units), and after 117 days (40 N units). The amount of N applied, and timing, was based on agriculture practices.

During the experiment, leachates were collected on three occasions: on the day of sowing (T0), after 48 days (T48) (after the first N application), and near the end of the experiment after 191 days (T191). Leachates were analyzed for pH, EC, and nitrate concentrations, as described in Section 2. In addition, total organic C content was measured using a TOC analyzer (Shimadzu TOC-L, Shimadzu France, Marne la Vallée, France).

After 212 days, wheat plants were harvested and evaluated for dry weight (after 72 h of drying at 50 °C), 1000-grain weight, and C and N contents (Flash 2000 (Thermo Fisher Scientific, France)).

Soils, collected at the beginning (T₀) and the end (T_F) of the experiment, were analyzed for basic chemical properties (CEC, OM content, and total C and N), as well as for microbial activities, through a Biolog^® plate (as described in Lebrun et al., 2021 [21]), and the measurement of diverse enzyme activities (hydrolysis, β-glucosidase, urease, and nitrification potential), using the same methods as described above.

2.4. Field Validation

To consolidate the results observed under controlled laboratory conditions and because climatic conditions, in addition to the pedologic properties, can influence AC effects on soil, we performed a field trial on each of the studied sites (Saugy and Tournoisis). In this experiment, we aimed to evaluate under real climatic conditions the influence of AC on crop yields, to validate hypothesis 3, and thus confirm the results from the controlled pot test.

On each site, 32 microplots (3 × 10 m) were set up as two sets of 16 microplots separated by 3 m. One of those sets was amended with 14 t.ha⁻¹ of the selected AC (5K), while the other set was left unamended. Inside each set, the microplots were placed in two rows, separated by 1 m, and each microplot line was separated by 0.3 m (Figure 1).

In total, four fertilization treatments were applied (four microplots in the unamended set, and four microplots in the amended set for each fertilization treatment) (Figure 1): (i) the recommended N dose (based on soil N residue and crop N requirement), which was noted XN, (ii) the recommended N dose lowered by 40 N units, which was noted X-40N, (iii) the recommended N dose supplemented by 40 N units, which was noted X+40N, and (iv) no fertilization, which was noted 0N. Those N fertilization treatments were split into three supplies on the Saugy soil, and into four supplies on the Tournoisis soil, following the regular practice of the farmer on the site.

After amendment application, wheat was sown, following the regular farmer management practices. Except for the N fertilization, all the other treatments were completed as required based on the farmer’s practices. At the end of the growing season, plants were harvested to evaluate yield and grain protein content (Chambre d’Agriculture du Centre-Val-de-Loire, Orléans, France).

2.5. Data Analysis

All the data were analyzed on the R software (version 4.4.2) [25]. In each case, soil and AC were analyzed separately, following the same procedure: the normality (Shapiro test) and homoscedasticity (Bartlett test or Fligner’s test) of the data were verified, and means were compared using either the ANOVA test (for normally distributed data) or the Kruskal–Wallis test (for non-normally distributed data), which was followed by a post hoc test (Tukey’s HSD test or Dunn’s test). In addition, in the column-leaching test, the activated carbon effect was evaluated, following the same process (except for the post hoc test). Difference was considered significant at p-value < 0.05.

3. Results

3.1. Leaching Column Test

3.1.1. Leachate Analysis

Cumulative nitrate leaching. After 105 days, the total amount of nitrate-N leached from the non-amended Saugy soil was 0.07 mg.g⁻¹ (Figure 2). The effect of AC type was highly significant (p < 0.001) regardless of the application rate. The materials CS and 2K had a significant effect only when applied at 5%, leading to a decrease by 76% and 59%, respectively, whereas the material L27 showed no significant effect at any application dose. The material 5K decreased cumulated leached nitrate-N content by 70% on average even when applied at 1% with no observed dose-dependent effect.

The amount of nitrate-N leached from the Tournoisis soil after 105 days was 0.06 mg.g⁻¹ (Figure 3). Overall, the type of activated carbon was very significant (p < 0.01). Only two materials significantly reduced nitrate-N leaching when applied on Tournoisis: CS decreased total leached nitrate-N content by 74% on average, at both application rates (2% and 5%), while 2K decreased it only when applied at 5% (a 62% decrease). Although the activated carbon 5K initially decreased nitrate-N leaching, it did not have any significant effect on the total amount of nitrate leached by the end of the experiment.

Cumulative phosphate leaching. After 105 days, almost no phosphate was leached from the non-amended soil of Saugy (0.002 mg.g⁻¹ soil) (Table S1). There was a highly significant effect (p < 0.001) of the activated carbon type on phosphate leaching. The materials CS, 2K, and 5K had no effect on the total amount of phosphate leached from Saugy, whatever the dose they were applied at, except for a slight increase observed with 5% CS. In contrast, the application of L27, even at 1%, induced a significant increase in phosphate leaching with a significant dose-dependent effect. In more detail, the total leached phosphate contents were 0.022 mg.g⁻¹ (L27 at 1%), 0.049 mg.g⁻¹ (L27 at 2%), and 0.126 mg.g⁻¹ (L27 at 5%).

In the Tournoisis soil, phosphate concentrations in the leachate of non-amended soil were below the detection limit (Table S1). The activated carbon type effect was highly significant (p < 0.001). Similar to Saugy, the addition of CS, 2K, and 5K had no effect on the total amount of phosphate leached from Tournoisis regardless of application rate. However, the application of L27, even at 1%, induced a significant increase in phosphate leaching with a dose-dependent effect. In more detail, total leached phosphate contents were 0.011 mg.g⁻¹ (L27 at 1%), 0.027 mg.g⁻¹ (L27 at 2%), and 0.117 mg.g⁻¹ (L27 at 5%).

pH and EC of the final leachates. The pH and EC of the final leachate of non-amended Saugy were 8.1 and 411 µS.cm⁻¹ (Table S2), respectively. There was no significant effect of the activated carbon type on leachate EC, while it had a significant impact (p < 0.001) on pH. The pH was only affected by the application of 5% CS, leading to a 0.3-unit increase, and 5K at 1% (+0.1 unit), while EC was not affected for all of the AC used.

For the Tournoisis soil, the pH and EC of the final leachate were 8.4 and 499 µS.cm⁻¹ (Table S2). No materials significantly affected pH and EC, except for the 0.3-unit decrease in pH following the addition of L27, whatever the concentration. The activated carbon type effect was highly significant (p < 0.001) for pH and non-significant for EC.

3.1.2. Soil Analysis

Enzyme activities were evaluated only in the controls and the soils amended with 5K, as it was the material showing the best decrease in nitrate leaching, even from the beginning of the leaching test, although it was not significant at the final time point for the Tournoisis soil. In Saugy soil, none of the evaluated enzymes was significantly affected by the application of the activated carbon 5K (

Table 4. Enzyme activities measured in the two soils (S = Saugy, T = Tournoisis), amended with the activated carbon 5K, at 0%, 1%, 2% and 5% (w/w), after 105 days of the leaching test. Letters indicate a significant difference between the activated carbon dose inside each soil (p < 0.05) (n = 5).

	β Glucosidase (nmol Glc.g−1 soil.min−1)	Alkaline Phosphatase (nmol PNP.g−1 soil.min−1)	Acid Phosphatase (nmol PNP.g−1 soil.min−1)	Hydrolysis (nmol FDA.g−1 soil.min−1)	Urease (nmol NH4+.g−1 soil.min−1)	Nitrification Potential (mg NO2−.kg−1 soil.h−1)
S	1.46 ± 0.20 a	0.04 ± 0.01 a	0.13 ± 0.01 a	11.24 ± 1.43 a	1.01 ± 0.13 a	0.88 ± 0.08 a
S+1%5K	0.67 ± 0.30 b	0.04 ± 0.00 a	0.12 ± 0.02 a	11.60 ± 1.15 a	0.78 ± 0.09 a	0.89 ± 0.22 a
S+2%5K	0.58 ± 0.07 b	0.03 ± 0.00 a	0.11 ± 0.00 a	8.84 ± 1.31 a	0.96 ± 0.11 a	0.88 ± 0.17 a
S+5%5K	0.36 ± 0.10 b	0.03 ± 0.00 a	0.11 ± 0.02 a	3.95 ± 0.97 b	0.76 ± 0.04 a	0.81 ± 0.20 a
T	1.81 ± 0.40 a	0.04 ± 0.00 a	0.15 ± 0.01 a	15.59 ± 0.78 a	0.75 ± 0.10 ab	1.63 ± 0.12 b
T+1%5K	0.06 ± 0.06 c	0.03 ± 0.00 a	0.14 ± 0.01 a	8.69 ± 0.36 b	0.66 ± 0.08 b	2.00 ± 0.13 ab
T+2%5K	0.36 ± 0.15 bc	0.04 ± 0.01 a	0.14 ± 0.01 a	4.41 ± 0.73 c	1.24 ± 0.17 a	2.38 ± 0.24 a
T+5%5K	0.64 ± 0.08 ab	0.04 ± 0.01 a	0.13 ± 0.00 a	3.11 ± 1.31 c	0.98 ± 0.14 ab	2.17 ± 0.17 ab

Glc = glucose; PNP = p-nitrophenol; FDA = fluorescein diacetate.

) except for the 75% decrease in hydrolyzing activity when 5K was applied at 5%.

In Tournoisis soil, three enzyme activities were significantly affected by the 5K amendment (Table 4): the β-glucosidase activity decreased by 97% and 80% with the application of 1% and 2% 5K, while the hydrolysis of FDA decreased by 44% (5K at 1%), 72% (5K at 2%), and 80% (5K at 5%). Nitrification potential increased by 46% but only with the application of 5K at 2%.

3.2. Pot Experiment

3.2.1. Leachate Analysis

Organic carbon. At the beginning of the experiment (T₀), the OC content in the leachate of the non-amended Saugy soil was 19 mg.L⁻¹ (

Table 5. Concentration (mg.L⁻¹) in organic C and nitrate (NO₃-N) in the leachates of the Saugy soil, differentially amended, collected at T₀, T₄₈ and T₁₉₁, in the pot experiment. S = Saugy; F = fertilization; AC1 = activated carbon (5K) applied at 1%; AC0.5 = activated carbon (5K) applied at 0.5%. Letters indicate significant differences (n = 5, p < 0.05) within each sampling time.

		Organic Carbon (mg.L−1)			NO3−N (mg.L−1)
		T0	T48	T191	T0	T48	T191
S	No	19.0 ± 0.9 a	5.8 ± 0.4 a	2.8 ± 0.5 b	19.15 ± 6.87 a	33.38 ± 2.31 ab	52.08 ± 13.62 ab
	Yes	/	6.1 ± 0.2 a	3.0 ± 0.6 b	/	1.26 ± 0.63 d	0.11 ± 0.01 e
S+F	No	/	7.0 ± 1.0 a	4.3 ± 0.4 a	/	43.83 ± 8.96 a	93.74 ± 9.94 a
	Yes	/	6.7 ± 0.4 a	2.3 ± 0.2 bc	/	8.75 ± 3.40 cd	0.11 ± 0.02 e
S+AC0.5+F	No	4.4 ± 0.4 b	1.8 ± 0.2 b	1.2 ± 0.1 c	15.26 ± 3.83 a	20.23 ± 2.70 bc	64.26 ± 7.00 ab
	Yes	/	2.0 ± 0.2 b	1.3 ± 0.1 c	/	8.79 ± 2.87 d	0.23 ± 0.08 cde
S+AC1	No	3.9 ± 0.4 b	1.9 ± 0.1 b	1.4 ± 0.1 c	14.16 ± 3.51 a	10.94 ± 2.32 cd	6.85 ± 3.16 bcd
	Yes	/	1.9 ± 0.1 b	1.3 ± 0.1 c	/	3.49 ± 1.14 d	0.12 ± 0.03 de
S+AC1+F	No	/	2.1 ± 0.1 b	1.2 ± 0.1 c	/	22.07 ± 3.49 ab	20.20 ± 3.21 abc
	Yes	/	1.9 ± 0.1 b	1.2 ± 0.1 c	/	7.82 ± 0.95 cd	0.12 ± 0.03 e

Values in italics indicate water concentrations in a “very good state” (10 mg.L⁻¹), and values in bold indicate water concentrations in a “good state” (50 mg.L⁻¹) based on the French Directive (Directive Cadre sur l’Eau (DCE) 2000/60/CE).

), and it decreased with the addition of the AC, without any dose effect, by an average of 78%. In the leachates collected on day 48, the OC content in the S treatment was 5.8 mg.L⁻¹. Only the presence of AC, whatever the application dose, decreased it by 67% on average, while the fertilization or presence of plants had no effect. After 191 days, the OC content in the leachates of the non-amended Saugy soil was 2.8 mg.L⁻¹ and the application of AC, whatever the dose, decreased it by 55% on average. Fertilization increased OC leaching (+53% compared to S) but only when applied to the non-amended soil, while in this case, plant presence reduced OC leaching (−47% compared to the non-vegetated S+F treatment).

In Tournoisis soil, the OC content in the initial leachate (T₀) of the non-amended soil (T) was 30.6 mg.L⁻¹, and the application of AC, whatever the dose, significantly reduced OC leaching by 61% on average (

Table 6. Concentration (mg.L⁻¹) in organic C and nitrate (NO₃-N) in the leachates of the Tournoisis soil, differentially amended, collected at T₀, T₄₈ and T₁₉₁, in the pot experiment. T = Tournoisis; F = fertilization; AC1 = activated carbon (5K) applied at 1%; AC0.5 = activated carbon (5K) applied at 0.5%. Letters indicate significant differences (n = 5, p < 0.05) within each sampling time.

		Organic Carbon (mg.L−1)			NO3−N (mg.L−1)
		T0	T48	T191	T0	T48	T191
T	No	30.6 ± 2.8 a	10.6 ± 0.6 a	5.2 ± 1.7 ab	5.82 ± 1.98 a	37.03 ± 7.80 a	61.12 ± 15.25 a
	Yes	/	11.1 ± 0.9 a	2.7 ± 0.4 abc	/	1.57 ± 0.62 e	8.52 ± 2.63 b
T+F	No	/	9.5 ± 0.3 a	5.8 ± 1.6 a	/	41.76 ± 5.40 a	98.31 ± 19.91 a
	Yes	/	10.5 ± 0.6 a	3.9 ± 2.1 abc	/	13.34 ± 1.72 bcd	5.12 ± 2.18 b
T+AC0.5+F	No	11.5 ± 0.9 b	2.6 ± 0.2 b	3.2 ± 0.6 abc	1.37 ± 0.10 b	20.59 ± 4.17 abcd	84.09 ± 11.28 a
	Yes	/	2.9 ± 0.2 b	1.3 ± 0.1 c	/	4.05 ± 0.88 de	8.62 ± 4.26 b
T+AC1	No	12.3 ± 1.3 b	3.5 ± 0.3 b	4.0± 1.1 abc	1.27 ± 0.31 b	18.60 ± 4.92 abc	65.77 ± 16.12 a
	Yes	/	2.8 ± 0.2 b	1.4 ± 0.1 c	/	0.37 ± 0.13 e	2.03 ± 0.82 b
T+AC1+F	No	/	3.1 ± 0.2 b	3.6 ± 0.4 abc	/	33.83 ± 5.28 ab	108.06 ± 19.99 a
	Yes	/	3.0 ± 0.1 b	2.9 ± 1.4 bc	/	4.06 ± 1.39 cde	7.67 ± 4.00 b

Values in italics indicate water concentrations in a “very good state” (10 mg.L⁻¹) and values in bold indicate water concentrations in a “good state” (50 mg.L⁻¹) based on the French Directive (Directive Cadre sur l’Eau (DCE) 2000/60/CE).

). In the second sampling (T₄₈), the OC content in the T treatment was 10.6 mg.L⁻¹ and only the presence of AC, whatever the dose, decreased it (−72% on average). In the last sampling (T₁₉₁), the OC leaching in T was 5.2 mg.L⁻¹, and significant decreases were only observed in the vegetated T+AC0.5+F (−75% compared to non-vegetated T) and vegetated T+AC1 (−73% compared to non-vegetated T) treatments.

Nitrate-N. In the soil of Saugy, nitrate-N concentration in the first leachate was 19.15 mg.L⁻¹, and AC had no effect (Table 5). In the first leachate, all concentrations were above the 10 mg.L⁻¹ threshold, indicating a very good underground water quality. On the second leachate at day 48, the nitrate-N concentration in the non-amended Saugy soil was 33.38 mg.L⁻¹. The addition of AC only significantly reduced nitrate-N leaching by 62% when applied at 1%, while fertilization increased nitrate-N leaching by 75% when applied to the treatment S+AC1. The presence of plants significantly reduced nitrate-N leaching in all cases except S+AC1. At this stage, all non-vegetated conditions had nitrate concentrations above the good state threshold. On the last leaching test, the nitrate-N concentration was 52.08 mg.L⁻¹ in S, and no effect of AC amendment and fertilization, alone or combined, was observed, while plant growth reduced nitrate-N leaching in all cases except in the case of S+AC1. However, it has to be noted that although the effects of AC were not significant, we can observe that compared to S, concentrations in nitrate-N in the treatment S+AC1 were reduced from above 50 mg.L⁻¹ to below 10 mg.L⁻¹, demonstrating that the application of AC to soil can change the underground water from a not good state into a very good state in regard to nitrate. Similarly, compared to fertilized soil (S+F), nitrate values passed from a not good state (>50 mg.L⁻¹) to a good state (>10 mg.L⁻¹). The presence of plants allows for obtaining nitrate concentrations that are considered a “very good state” (<10 mg.L⁻¹).

In the soil of Tournoisis, the first leachates of non-amended T treatment contained 5.82 mg.L⁻¹ nitrate-N on average (Table 6), and applying AC significantly reduced nitrate-N leaching by 75% on average. In these leachates, all concentrations were below the threshold considering underground water to be in a very good state. At T₄₈, the nitrate-N concentration in the leachate of T was 37.03 mg.L⁻¹, and the addition of AC or fertilizer had no effect, while plant growth decreased nitrate-N leaching in all cases, although it was not significant in the case of T+AC0.5+F. At this stage, even though only plants had a significant effect, and the reduced nitrate-N concentration was at a level considered to be a “good state”, the application of AC at 1% (in the presence of plants) allowed the reduction in nitrate-N to concentrations considered to be in a very good state. In the last leachates, collected at T₁₉₁, the nitrate-N concentration was 61.12 mg.L⁻¹ in T, and the only significant effect was found with plant growth, which reduced nitrate-N leaching in all cases. At the end of the experiment, all the non-vegetated conditions were in a “not good state” in regard to nitrate leaching, while all vegetated conditions were in a very good state. These results suggest that over time, activated carbon has much less effect on nitrate-N leachate in Tournoisis than in Saugy.

3.2.2. Plant Analysis

Biomass production. When grown on Saugy soil, without any treatment, wheat plants produced an average of 18.6 g of aerial biomass and 2.0 g of root biomass. Aerial biomass increased only with fertilizer amendments, by 76% on average (

Table 7. Dry weight and C and N contents in the wheat plants grown for 212 days on Saugy and Tournoisis soils, differentially amended. S = Saugy; T = Tournoisis; F = fertilization; AC1 = activated carbon (5K) applied at 1%; AC0.5 = activated carbon (5K) applied at 0.5%. Letters indicate significant differences (n = 5, p < 0.05) within each soil.

	Aerial Biomass (g)	Root Biomass (g)	1000-Grain Weight (g)	Carbon Content (%)	Nitrogen Content (%)
S	18.6 ± 0.45 b	2.0 ± 0.4 b	30.5 ± 2.1 b	40.5 ± 0.3 b	0.74 ± 0.10 a
S+F	33.2 ± 1.5 a	5.0 ± 1.3 a	28.5 ± 3.8 b	42.1 ± 0.4 a	0.71 ± 0.08 a
S+AC0.5+F	33.7 ± 1.3 a	3.7 ± 0.8 ab	40.7 ± 3.3 ab	42.6 ± 0.8 a	0.78 ± 0.08 a
S+AC1	19.7 ± 0.2 b	2.1 ± 0.4 ab	50.8 ± 5.3 a	38.7 ± 0.5 c	0.91 ± 0.06 a
S+AC1+F	31.4 ± 1.7 a	3.6 ± 0.7 ab	39.7 ± 6.4 ab	42.6 ± 0.2 a	0.71 ± 0.05 a
T	18.5 ± 0.8 b	0.7 ± 0.2 a	32.3 ± 5.9 a	39.7 ± 0.3 b	1.00 ± 0.08 a
T+F	29.4 ± 1.6 a	0.8 ± 0.2 a	33.5 ± 7.3 a	42.2 ± 0.1 a	0.93 ± 0.04 a
T+AC0.5+F	26.4 ± 2.5 a	1.0 ± 0.3 a	46.4 ± 7.9 a	41.2 ± 0.3 a	1.23 ± 0.15 a
T+AC1	15.7 ± 1.1 b	0.5 ± 0.2 a	28.1 ± 3.4 a	41.1 ± 0.3 ab	0.85 ± 0.04 a
T+AC1+F	25.1 ± 1.8 a	0.6 ± 0.2 a	33.8 ± 4.0 a	41.3 ± 0.9 a	0.89 ± 0.03 a

), while root biomass was only increased in the S+F treatment by 2.5 times compared to S. The weight of 1000 grains for treatment S was 30.5 g (Table 6), and it only increased by 67% with the application of 1% AC.

On Tournoisis soil, aerial and root biomasses were, respectively, 18.5 g and 0.7 g on the non-amended treatment (Table 7). The aerial biomass was increased only when fertilizer was applied, by 46% on average, while the root biomass was not affected. The 1000-grain weight was 32.3 g under treatment T and was not affected by any amendments.

Carbon and nitrogen content. When grown on the non-amended Saugy, the C and N contents of the aerial biomass were, respectively, 40.5% and 0.74%, on average (Table 7). The application of fertilizer increased C content by 5% on average, while the application of 1% AC alone decreased it by 4%. The N content was not affected by any of the treatments.

On the soil of Tournoisis, for plants grown on the non-amended treatment, the aerial C and N contents were 39.7% and 1.00%, respectively (Table 7). Only the application of fertilizer increased C content, by 5% on average, while no treatment affected the N content.

3.2.3. Soil Analysis

Soil chemical properties. The OM content and CEC were measured in the soil at the beginning (T₀) and the end (T_F) of the experiment. The Saugy soil contained 9.6% OM at the beginning of the experiment, and the application of AC had no effect (Table S3). At the end of the experiment, the OM content was 13.53% and the application of AC increased the OM content by 9% but only when applied alone at 1%, while the application of fertilizer on this treatment (S+AC1) decreased the OM content by 8%. Plant growth decreased the OM content in all cases except S+AC1+F. The CEC of non-amended Saugy was 8.8 cmol.kg⁻¹ at T₀, and the value was not affected by the application of AC. At T_F, the CEC was 8.2 cmol.kg⁻¹ and none of the treatments affected the CEC compared to S. Only the growth of the plant increased CEC in a few treatments; i.e., S+F and S+AC1.

In Tournoisis, the OM content was 11.3% at T₀, and the only significant increase was observed with the application of 0.5% AC (+8%) (Table S4). At the end of the experiment, the OM was 17.5% on non-amended T, and the application of AC and fertilizer alone had no effect, while the combination of 1% AC and fertilizer led to a 10% increase in the OM content. Plant growth decreased the OM content in all cases except T+AC0.5+F. The CEC value was 12.1 cmol.kg⁻¹ at T₀, and AC application had no effect. At the end of the experiment, the CEC was 11.8 cmol.kg⁻¹ on T, and none of the treatments had a significant effect.

Soil microbial activities. The Biolog^® test showed that at T₀, on the non-amended Saugy soil, the AWCD (average well color development) was 1.12 (Table S5) and the AC application had no effect, while at TF, the AWCD value was 1.53 and neither the AC nor the fertilizer had any effect when applied alone. However, the application of 0.5% AC combined with fertilization decreased AWCD by 31%, and plant growth had contrasting effects depending on the substrate: a 2-fold increase in AWCD on S, a 27% decrease in AWCD on S+AC1, and no effect on the other treatments. The richness value at T₀ was 27 and not affected by the AC amendment. At T_F, the richness was 26, and AC and fertilization had no effect when applied alone but induced a decrease when applied together: AC0.5+F led to a 19% decrease and AC1+F induced an 8% decrease. Plant growth increased richness in all cases except S+AC1 and S+AC1+F. The hydrolyzing activity was 37 nmol FDA.min⁻¹.g⁻¹ on S at T₀, and the addition of AC led to a 58% decrease in activity on average. At T_F, the hydrolyzing activity on S was 4.71 nmol FDA.min⁻¹.g⁻¹, and the addition of AC at 1% decreased it by 20% (with fertilization) to 31% (without fertilization), while fertilizer had no significant effect and plant growth either decreased (−19% on S), increased (+31% on S+F) or had no effect on the hydrolyzing activity. The β-glucosidase activity was 10.3 nmol PNP.min⁻¹.g⁻¹ on S at T₀, and AC application decreased it by 23% on average. At the end of the experiment, activity was 0.18 nmol PNP.min⁻¹.g⁻¹ and the addition of AC or fertilization alone had no effect, while the combination of 1% AC and fertilization led to a 2.4-fold increase in β-glucosidase activity. Plant growth either led to an increase (2-fold, on S), a decrease (−51%, on S+AC1+F), or no effect. The nitrification potential was only measured at the end of the experiment, and the value was 2.59 mg NO₂.h⁻¹.kg⁻¹ on S, and none of the treatments had a significant effect. Finally, urease activity was 0.16 µg NH₄.h⁻¹.kg⁻¹ on S at TF, and AC and fertilizer had no effect when applied alone. The combination of 1% AC and fertilization (S+AC1+F) led to a 62% decrease in urease activity compared to S, while plant growth only significantly affected urease activity in two cases: it induced an 89% decrease in S and a 3-fold increase in S+AC1+F.

On the Tournoisis soil, the AWCD of T₀ was 1.24 and the application of 1% AC decreased the AWCD value by 21% compared to T (Table S6). At the end of the experiment, the AWCD value was 0.60 on T and the application of AC1 alone increased it by 32%, while plant growth increased it in all cases. The richness was 28 at T₀, and AC application had no effect compared to the control treatment T. At T_F, the richness value of T was 18 and the application of 1% AC increased it by 28%, while the presence of plants increased it in all cases except T+AC1. The hydrolyzing activity was 34.9 nmol FDA.min⁻¹.g⁻¹ on the non-amended soil at T₀, and AC application led to a 63% decrease when applied at 0.5% and a 33% decrease when applied at 1%. At T_F, the activity was 4.01 nmol FDA.min⁻¹.g⁻¹ and the only significant effect observed was a 32% decrease in the presence of plants on T+AC1. The β-glucosidase activity was 11 nmol PNP.min⁻¹.g⁻¹ at T₀ on the non-amended T and decreased by 29% on average with the application of AC. At T_F, the activity was 0.29 nmol PNP.min⁻¹.g⁻¹, and the application of AC at 1% decreased it by 41% (without fertilization) to 79% (fertilization), while the treatment AC0.5+F decreased it by 41%. Fertilization induced a decrease in β-glucosidase (−65%) only when applied to T+AC1. Plant growth increased β-glucosidase activity by 3.5-fold but only in the treatment T+AC1+F. Finally, the nitrification potential and urease activity were 3.09 mg NO₂.h⁻¹.kg⁻¹ and 0.20 µg NH₄.h⁻¹.g⁻¹ on T, respectively, and none of the treatments had any effect.

3.3. Field Experiment

Wheat yield. On Saugy, the wheat yield was 34.3 qt.ha⁻¹ when no AC or N fertilizer was applied (Figure 4). The application of AC had no effect regardless of the N levels, whereas applying N increased yield, by almost 2-fold on average, with no difference between N doses applied.

On Tournoisis, the yield on the non-amended and non-fertilized condition was higher than on Saugy, at 45.73 qt.ha⁻¹ (Figure 4). In all cases, AC negatively affected yield across all treatments, while N application increased it by 82% on average.

Grain protein content. The protein content in the grains of the plants grown on the Saugy soil, non-amended and non-fertilized, was 11.88% on average (Figure 4). No effect of the AC and/or fertilization was observed.

On Tournoisis, the grain protein content on the control (no AC and no N fertilization) was 11.78% (Figure 4) and no effect of AC application was found, while the N fertilization significantly increased protein content by 34% on average without any significant difference between the N doses.

4. Discussion

4.1. The Application of Activated Carbon Stabilizes Carbon in the Soil, Potential for Carbon Storage

Activated carbon, being primarily composed of carbon, contributes to an increase in the soil’s OC content. In general, the addition of a high quantity of OC induces its mineralization, which is followed by its emission as CO₂ and its leaching toward underground water [26]. However, C coming from biochar materials was shown to be less leached than other organic amendments. For instance, Liu et al. [27] demonstrated that only 0.14 to 0.18% of the C coming from biochar was lost through leaching. Such observations corroborate our results as well as those of other studies [28,29,30,31]. The main reason for the reduced leaching of OC is the nature of the C coming from biochar materials, including AC, which is recalcitrant and resistant to microbial degradation [32]. This can explain the diminution of the activity of the β-glucosidase, which is an enzyme involved in the C cycle; such a decrease could also indicate a reduction in C mineralization and thus loss through emission. In addition, soluble C coming from the soil could be sorbed on biochar [28,31]. Such a decrease in carbon leaching shows the potential of AC for C long-term storage.

4.2. Activated Carbon and Plant Development Allows for a Reduction in the Leaching of Nitrate as Well as the Potential for a Reduction in Chemical Fertilizer Use

The reduction in nitrate leaching following AC amendment observed in the column and pot experiments was in accordance with previous studies [14,33,34,35]. The following principal explanations were given for such a reduction in nitrate leaching: (i) an immobilization of N in the microbial biomass [14], (ii) an increase in the microbial activity, in particular the denitrification [14,36], and (iii) a sorption/precipitation of N on the AC surface [33,34,35]. Based on the measures performed in this study, the last explanation seems to be the most influencing in the case. Indeed, the involvement of denitrification, although it was not measured, seems unlikely. The measure of the nitrification potential showed an increase following AC application, implying a potential higher generation of nitrate. Such high nitrate content could have been sorbed and/or precipitated on the AC surface, as it has been demonstrated in our previous study for the four AC tested here [17]. This seems to be particularly true for the soil of Tournoisis, in which the AC 5K increased nitrification potential and thus nitrate formation but reduced nitrate leaching, demonstrating an immobilization of the generated nitrate into the soil—either via direct sorption or precipitation on the AC surface or indirectly through the modification of the soil properties. However, we cannot rule out other mechanisms such as microbial immobilization, especially when comparing the efficiency of those materials in the batch sorption test and in the column leaching test. In the sorption test, the efficiency in sorbing nitrate decreased in the order L27 > 5K > 2K = CS, while L27 was the least efficient in the column test and 5K and CS were the best. This shows that (i) once in the soil, AC sorption properties are modified, and (ii) AC may have modified soil properties, influencing the mobility of nitrate. In addition, in several cases, plant development further reduced nitrate leaching, which can be due to an uptake of nitrate by plants or a modification of the soil properties through plant exudation. The data on the leaching of nitrate show that some treatments were able to reduce it to levels below the threshold of the nitrate concentration in ground water (25 mg.L⁻¹) in Europe [1]. This highlights that a plant cover, and/or the application of 1% AC, are a good solution to protect ground water quality.

4.3. Activated Carbon Diminishes the Functional Diversity of the Soil, Producing a Negative Effect on Microbes

In general, AC reduced the microbial functional diversity and decreased or did not affect enzyme activities related to C and N cycles. On the contrary, plants tended to increase those parameters. The hydrolyzing activity, a general indicator of soil biological activity [37], decreased in both pot experiment and column experiments in a dose-dependent manner. Such a result is in contradiction with previous findings [37,38]. This could be related to the release of toxic substances by AC [39]. However, this explanation seems unlikely as the other enzymes measured, such as urease, alkaline phosphatase, and acid phosphatase, were not affected in the pot experiment. The activity of the β-glucosidase, an enzyme involved in the C cycle, was generally decreased by the application of the AC, which is consistent with previous studies [39,40]. The other enzymes activities were in most cases not affected. In general, the application of organic amendments stimulates enzyme activities due to the addition of OM and nutrients to the soil [41,42]. However, the influence of biochar-based materials on soil enzyme activities showed contrasting effects, with stimulation [43,44], inhibition [39,42] or no effect [45]. Therefore, it is difficult to draw clear conclusions on the response of the microbial community to biochar material addition. The reduction in β-glucosidase observed in our study can be related to (i) the recalcitrance of the C present in the AC, which cannot be processed by microorganisms [46], and/or (ii) a co-localization of the C and the microorganisms inside the AC pores, as it serves as a habitat [47], which improves the C use efficiency. Therefore, the microorganisms requirement for OM mineralizing enzymes is reduced [48]. Another observation was the non-effect of AC on N enzymes in the pot experiment and a decrease in the column experiment. In those experiments, nitrate leaching had differing behavior. More precisely, nitrate leaching was reduced in the column test and not affected in the pot when only AC was added. Therefore, we can hypothesize that the reduction in N-related enzymes in the column test was related to a sorption of the nitrate by the AC, making it less available for microorganisms, whereas in the pot, nitrate did not seem to be immobilized by AC, and thus it was still accessible for microorganisms, which explains the non-effect in that case. Similarly, phosphatase activities were not affected in the column, which could be related to the fact that phosphate was neither increased nor decreased in that test. Ultimately, our results showed that AC amendment reduced C mineralization, which could lower C emission into the atmosphere and increase C storage [49] while having no effect, positive or negative, on the other element cycling enzymes. In particular, the non-augmentation of urease activity could indicate that N mineralization and thus emission was not increased, which is generally observed in agriculture following the application of OM [50].

4.4. Plant Growth Is Not Enhanced by Activated Carbon Addition

Our results did not show any significant improvement in crop yield with AC application except in the case of 1000-grain weight on the soil of Saugy amended with 1% AC. This is in contradiction with previous studies [14,51,52]. But other studies showed the non-effect of AC on crop yield [35]. The improvement of plant growth has been related to (i) an improvement of soil fertility and (ii) a reduction in pollutant availability. In our cases, the first explanation is more likely. However, although soil fertility was increased, i.e., an increase in OM and a reduction in nitrate leaching, biomass production was not improved. Such a non-effect can have several explanations: (i) the fertility improvement was not enough to lead to significantly higher plant growth in relation to the lower dose of AC applied (1%) compared to other studies (3–5%) [51,52]; (ii) the nitrate retained by the AC (as shown by the reduced concentration in the leachates) could have not been available for the plants; and/or (iii) the fertility of the soil was already optimal for plant growth. However, this last explanation seems unlikely, as plant growth increased in the fertilized treatment, which shows the dependency of growth on N. Furthermore, we can see that on the soil of Tournoisis, yields were increased more importantly by N fertilization than on Saugy. This shows the important dependency of the soil of Tournoisis to N fertilization and could explain why AC reduced yield in this particular soil with a sorption of nitrate on its surface. Our results also show that field studies are more contrasted and highly dependent on climatic conditions, which is not the case in controlled greenhouse experiments. Further studies need to be conducted to improve this material in order to bring benefits to crops. Finally, based on the field experiment, it seems that N fertilizer application could be reduced, as even the lowest dose of N improved yield, at similar levels as the regular N dose.

4.5. The Effect of Activated Carbon Is Modulated by the Activated Carbon and Soil Type

This study first showed the crucial influence of the feedstock and activation method on the soil property changes in response to AC amendment. Based on those two parameters, the resulting AC will have different physical and chemical properties, as revealed by the characterization of the materials used in this study. Those parameters, such as pH, surface area, and pore volume, will significantly influence the effect of AC on the soil. For instance, the higher surface area and pore volume of the L27 activated carbon will increase its sorption capacity toward nitrate, as shown in our previous study [17], but its highly acidic pH could lead to soil acidification. However, once mixed with the soil, this AC that showed the best potential toward nitrate immobilization did not meet expectations and did not significantly reduce nitrate leaching. In addition, it promoted phosphate leaching in relation to the chemical solution used for its activation. The activated carbon 5K, which was the second best in the sorption testing [17], was revealed to be the best to prevent nitrate leaching even at the lowest concentration. This material was not the one with the highest surface area, pH, or pore volume. Therefore, its efficiency must be related to other properties not assessed here or an interaction between them.

The second crucial parameter that will affect how the soil will respond to AC amendment is the initial soil characteristics. In our study, we observed higher responses in the soil of Saugy compared to the soil of Tournoisis. Several reviews and meta-analyses showed that applying organic amendments under tropical or arid climates had higher influences than under temperate climates [53,54] and that the largest effects were observed on sandy and loamy texture soils [54]. However, both soils came from the same region and had the same texture. Therefore, the higher response of the Saugy soil compared to Tournoisis soil must be related to other physico-chemical properties. For instance, the soil of Tournoisis had a higher CEC than the soil of Saugy and thus could retain more nutrients [55]. Moreover, Ca, Mg, and Na contents were higher in Tournoisis, showing higher initial fertility, as also confirmed by the higher yield in the control of Tournoisis compared to the control of Saugy. As Ca could indicate the presence of calcium carbonates, and thus the stabilization of OM with time, this could also explain the higher OC in Tournoisis and the reduced effect of the AC by the end of the experiment [56]. Finally, both soils came from different “agricultural regions” and had different physical characteristics. For instance, the soil of Saugy presented lots of rocks, which could contribute to a lower physical organization of the aggregates and thus poorer growing conditions, which are more prone to be improved by organic amendments [49].

4.6. Practical Implications

Declining soil fertility and the overuse of chemical fertilizers are increasing issues, especially in developed and developing countries. In addition to the cost, this excessive employment of chemicals has negative effects on the environment, i.e., the leaching and pollution of ground waters, eutrophication, greenhouse gas emissions, etc., and it ultimately will impact human health. It is thus important to find more sustainable alternatives to ensure sufficient food production. One of those alternatives that attracts attention is to apply biochar-based materials, such as activated carbon. AC is already used for the treatment of wastewater, but its application as a soil amendment remains underexplored. Our study demonstrated the potential of this material to reduce nitrate and OC leaching. However, it did not ameliorate crop yield. Activated carbon is thus a potential alternative that could stabilize nutrients in the soil. In addition, as it has high stability, it does not need to be applied every year. Therefore, AC can be used to reduce the negative impact of chemical fertilizers on the environment and protect ground waters, therefore having additional impacts on water bodies, such as improve biodiversity and ecosystem functions. However, to further increase its benefits, especially on crop yield, it needs to be improved by coating nutrients on its surface, using waste materials such as guano, and/or making it in a shape easily applicable on the field, giving an all-in-one solution. Such area needs more testing, as the price of AC is still high (around 1000 € per ton), and it would be higher with additional modifications; the most effective production process and the appropriate dose, a compromise between the cost and the benefits, need to be found. Finally, our study revealed that N use in the field could be slightly reduced without impacting yield, which would have both economic and environmental benefits: lower cost for the farmer and thus higher profits, less potential loss of N through leaching and atmospheric emissions, less acidification of the soil and eutrophication.

5. Conclusions

The potential of activated carbon to reduce nitrate leaching, enhance soil fertility and improve crop yield was evaluated at different scales, including leaching column tests, laboratory pot experiments, and field trials. The results showed that the reduction in nitrate leaching was dependent on the feedstock used and activation methods of AC, and it was inherent to the soil properties. The application of the AC to the soil reduced the loss of organic carbon through leaching and increased the total C content in the soil while slightly decreasing the activities of the enzyme involved in the C cycle, which shows potential for C storage and a reduction in CO₂ emissions. Activated carbon amendment also stabilized N in the soil. However, crop yields were not improved by AC addition.

We confirm our first hypothesis regarding the reduction in C and N loss through leaching at the microscale and partly at the meso-scale (only at the initial time). Conversely, we reject our second hypothesis that AC increases N cycle enzymes, which could indicate that N mineralization, and thus, emission, was not induced. Finally, we reject our last hypothesis on the amelioration of crop yield. More work needs to be conducted on such material before it can be deployed in the field to improve its benefits. For instance, coating essential nutrients such as nitrate will allow for stabilizing the nutrients in the soil and prevent their fast leaching, reducing the need to apply it regularly. Similarly, coating microorganisms involved in nutrient cycles, such as those solubilizing phosphate, will allow a higher soil microbial activity and thus will contribute to render nutrients more available for plants.