Agave-bagasse biochar and tequila vinasse: effect on soil CO 2 mineralization and carbon pool stability

1 Introduction

The global tequila demand has reached unprecedented heights, with Mexico producing a record-breaking 651.4 million liters in 2022 alone (1). However, this industrial success masks a mounting environmental crisis. For every liter of tequila bottled, the process leaves behind a staggering trail of waste: approximately 1.4 kg of dry bagasse and up to 12 liters of vinasse (2). At current production scales, this equates to millions of tons of organic by-products that, if poorly managed, threaten the very ecosystems they come from. While the principles of a circular bioeconomy offer a pathway to transform these residues into valuable resources, the reality remains hazardous (3). Untreated vinasse, characterized by acidic pH levels, temperatures reaching 90 °C, high electrical conductivity, and elevated levels of phenolic compounds and heavy metals; and decomposing bagasse pose severe risks to soil, water quality, and human respiratory health (4–6). Addressing this waste stream is no longer just an industrial preference; it is an ecological necessity.

Several strategies have been proposed to convert these residues into useful products, including composting, anaerobic digestion for biogas production, microalgae treatment, and fertigation (7–9). However, the widespread adoption of these alternatives has been limited, largely due to the high investment costs involved, particularly for small and medium-sized enterprises (10). Pyrolysis offers a cost-effective alternative for small-scale producers by converting agave bagasse into biochar, a stable, carbon-rich soil amendment. This method has gained considerable attention in recent years for its dual environmental and agronomic benefits (11). Biochar is a highly porous, carbon-rich material recognized for its capacity to improve soil quality by adsorbing essential nutrients and immobilizing heavy metals (12). Due to its high surface area and functional groups, biochar serves as an effective medium for the treatment of complex industrial effluents like vinasse. Specifically, literature indicates that biochar can adsorb phytotoxic phenolic compounds found in vinasse (chlorogenic, vanillic, syringic, and coumaric acids), thereby mitigating the inhibitory effects these substances have on plant growth (13, 14). By acting as a stabilizing agent for the high mineral and organic load present in vinasse (15), biochar offers a dual benefit: protecting the soil ecosystem from toxic runoff while facilitating the safe integration of liquid by-products into agricultural irrigation.

Beyond immediate waste remediation and soil improvement, the integration of biochar into tequila-producing regions represents a significant opportunity for long-term climate mitigation. Greenhouse gases (GHGs), the primary drivers of global warming, arise from both human activity and microbial processes in soil. Soils hold more carbon than the atmosphere and are central to the global carbon cycle, acting as both a source and a sink of GHGs (16). Carbon mineralization, the enzymatic decomposition of organic matter into inorganic compounds such as CO₂ or CH₄, is a key microbial process in this cycle. Microbial metabolism significantly influences both carbon loss and accumulation in soils (17). Since the advent of agriculture, an estimated 456,000 tons of carbon have been lost from soils globally (18), with annual losses of 0.5 to more than 2 Mg C ha^-¹ linked to the conversion of natural ecosystems to cropland (19). In this context, soil carbon sequestration (through strategies that increase storage or reduce loss) has emerged as a vital tool in climate mitigation efforts. Among these, biochar application has shown remarkable potential, with a global average of 13.0 Mg ha^-¹ of sequestered soil organic carbon reported (20). Furthermore, biochar can stabilize microbial necromass, enhance microbial activity, and reduce GHG emissions (21, 22). It also modulates microbial secondary metabolism: promoting flavonoid secretion (e.g., chrysosplenetin) by methanogens and denitrifiers (which increases methane production while reducing nitrous oxide emissions), and suppressing isoflavone production (e.g., puerarin) by Bacillus spp., which slows down organic matter mineralization and CO₂ release (23). These mechanisms underscore the multifunctional role of biochar in enhancing soil carbon stability and contributing to climate change mitigation.

While the potential of biochar for waste remediation and carbon sequestration is well-documented, the specific synergy between agave-bagasse biochar and tequila vinasse on soil carbon dynamics remains largely unexplored. In particular, it is unknown whether the combination of these two waste streams produces additive, synergistic, or antagonistic effects on soil carbon mineralization and pool stability. Therefore, the objective of this study was to evaluate the conversion of agave bagasse into biochar and the revalorization of tequila vinasse for their subsequent application as soil amendments. CO₂ evolution was monitored over the incubation period and the cumulative mineralization data were fitted to a double-exponential kinetic model, enabling the partitioning of the mineralizable carbon into a fast-turnover labile pool (C₁) and a slow-turnover pool (C_s), each governed by distinct first-order rate constants (k₁ and k_s). Both C₁ and C_s represent carbon that is ultimately mineralized to CO₂; C_s is therefore a proxy for the stabilized (slowly cycling) fraction rather than a direct measure of sequestered carbon, which corresponds to the applied carbon that is never respired over the observation window. This modeling framework allows the quantification of the proportion of added carbon that is rapidly mineralized versus that which is effectively stabilized in the soil matrix. By resolving these kinetic parameters across treatments, this study seeks to provide a mechanistic basis for evaluating agave-bagasse biochar and tequila vinasse as circular-economy amendments that address the environmental liabilities of the tequila industry while promoting long-term soil carbon storage.

2 Materials and methods

2.1 Sampling and material collection

Soil, agave bagasse, and raw vinasses were sourced in collaboration with a local tequila producer in Amatitán, Jalisco, Mexico, a major region to agave cultivation. Surface soil (5–10 cm depth) was collected from three distinct locations within the agave rhizosphere to capture spatial variability in soil properties. Following FAO guidelines to target the layer with the highest biological activity, the samples were pooled and homogenized (24). The composite soil was sieved through a <2 mm mesh to ensure uniform particle size and facilitate the removal of coarse organic debris, thereby minimizing experimental variance. This processed soil was then partitioned and transferred into 1 L glass flasks for the microcosm setup, according to the treatments detailed in Supplementary Table 1.

2.2 Preparation of biochar and vinasses

Biochar was produced from dried agave bagasse via slow pyrolysis at 300 °C using a laboratory oven (UF55, Memmert GmbH). The heating rate was controlled between 0.1 and 1 °C s^-¹, with a residence time of 30 minutes. The pyrolysis temperature was selected to achieve partial carbonization while retaining a chemically heterogeneous carbon structure and oxygen-containing surface functional groups relevant to the subsequent enrichment of the biochar with tequila vinasse. Low-temperature biochars produced at approximately 300–400 °C generally retain more aliphatic components, volatile matter, and polar surface functionalities than biochars produced at higher temperatures, whereas increasing pyrolysis temperature promotes aromatic condensation, fixed-carbon formation, and resistance to degradation (25, 26).

Biochar enrichment followed the protocol described by Schmidt et al. (27). Briefly, the biochar was immersed in raw vinasse at a volumetric ratio of 1:5 (w/v) and stirred continuously for 2–4 hours at 20–25 °C to facilitate the adsorption of minerals and organic compounds into the porous structure. Following the contact period, the suspension was decanted to separate the solids. The resulting impregnated biochar was then dried at 60 °C for 3 hours.

2.3 Microcosm assembly

Treatments were set in 1 L glass bottles, placing the amount of soil and additives (biochar, compost, fertilizer, water and/or vinasses), sieved through a 2 mm mesh. The flask lids were adapted for sampling through a valve system. Based on similar studies (28), the microcosms were kept in a climate-controlled room at 25 °C ± 2 °C in an incubator for 30 days, after a 7-day acclimatization period during which the soil reached its baseline state before the treatments were applied. Since final sampling was staggered among experimental runs, the actual elapsed incubation time ranged from 744 to 792 h. The experiment was carried out in a dark incubator to modulate the conditions of microorganisms underground.

2.4 CO₂-C headspace monitoring and cumulative respiration

Carbon mineralization was tracked as CO₂-C accumulating in the microcosm headspace over a 30-day incubation period, sampled through access port 1 on days 0, 1, 2, 3, 4, 5, 8, 11, 14, 17, 20, 23, 26, 29, and 30, maintaining a sampling progression until the total run time established at 30 days was completed. Headspace CO₂ concentrations recorded with the Biogas 5000 analyzer (Geotech, UK) were converted to a carbon-mass basis (g CO₂-C g^-¹ dry soil) using the ideal-gas law with the measured headspace volume of each flask, the incubation temperature (25 ± 2 °C) and the dry mass of soil per microcosm (350 g). Following each measurement, the microcosms were purged with atmospheric air through both access ports to reestablish aerobic conditions and prevent the accumulation of metabolites that could inhibit microbial activity. For every replicate, the per-interval values were summed to obtain a monotonic cumulative emission trajectory; the trajectory was anchored at the origin (t = 0 h, CO₂-C = 0) so that each curve represents total cumulative respiration from the start of the incubation. All treatments were run in triplicate (n = 3) except the vinasse-family control SC-2 (n = 2).

Cumulative respiration (AUC). To compare the total respiratory output of each treatment, the area under the cumulative CO₂-C trajectory (AUC) was integrated for every replicate with the composite trapezoidal rule over the full set of sampling times, AUC = Σ½(t^I+₁−t^I)(y^I+₁+y^I), where y is the cumulative CO₂-C concentration at time t. AUC therefore integrates both the rate and the persistence of mineralization into a single magnitude (units: g CO₂-C·h·g^-¹ soil), and was used as the response variable for inferential comparisons of cumulative respiration among treatments. AUC was selected as the primary integrated response because it uses information from all sampling intervals and avoids basing the principal treatment comparison on a single sampling occasion whereas using parameters such as the maximum observed respiration rate or the final cumulative CO₂-C may be sensitive to sampling frequency and short-term variability or do not capture differences in the timing of mineralization (29, 30). Treatments were grouped into the three experimental families defined in Section 2.3: biochar (SC-1, BC-5, BC-10, BT-5, BT-10), vinasse (SC-2, V-50, V-100, VT-50, VT-100) and additive (SC-3, FER, COM). Each family was analyzed against its own co-incubated soil control.

2.5 Physicochemical analysis

The physicochemical characterization of the raw agave bagasse, biochar, and soil samples was performed through thermal and spectroscopic analyses. Differential scanning calorimetry and thermogravimetric analysis (DSC–TGA) were carried out using a simultaneous thermal analyzer (Waters SDT 650, TA Instruments, New Castel, DE) under a nitrogen atmosphere to evaluate the thermal stability, mass loss behavior, and decomposition profiles of the materials. Approximately 10–20 mg of dried sample was placed in 90 µL-alumina crucibles and heated from room temperature to 700 °C at a constant heating rate of 10 °C min^-¹, followed by a 30 min isothermal hold at the final temperature. The obtained thermograms were used to identify moisture release, volatile matter degradation, and carbonaceous fraction stability. Post-incubation soil thermograms were analyzed following the thermal fractionation protocol of Plante et al. (31): mass loss between 200 and 375 °C was assigned to the labile fraction of mineral-associated organic matter (SOM Labil), and mass loss between 375 and 550 °C to the recalcitrant fraction (SOM Recalcitrant). The soil thermal stability index (STSI) was calculated as the ratio SOM Recalcitrant/SOM Labil, and T₅₀ was defined as the temperature at which 50% of the total SOM mass loss (200–550 °C) had occurred. Carbonate content was quantified from the isothermal mass loss at 700 °C.

Fourier transform infrared spectroscopy (FTIR) analysis was performed (IRTracer-100 FTIR Spectrophotometer, Shimadzu). Spectra were recorded within the range of 4000–400 cm^-¹ using the attenuated total reflectance (ATR) mode, with a resolution of 4 cm^-¹ and 32 accumulated scans per sample. The resulting spectra were analyzed and compared to determine the presence of functional groups, as well as possible changes within samples.

Vinasse pH was measured using a calibrated digital pH meter (HI2210, HANNA instruments, Woonsocket RI USA) and for electrical conductivity (EC) measurements, a conductivity meter (HI2020-01, edge, HANNA instruments) was used. The instruments were calibrated and measurements were performed at room temperature (25 ± 2 °C). In both cases, the electrode was rinsed with deionized water between measurements and gently dried with lint-free tissue. The sample was disposed directly into a 50 mL tube for measurement, and the reading was recorded once stabilization (≤0.01 units variation over 30 s) was achieved.

For the physicochemical parameter set comprising: BOD₅, COD, total Kjeldahl nitrogen, total phosphorus, fats and oils, chlorides, total phenols, the Mexican environmental protection regulations (NMX-AA) were followed.

Because biochar production was constrained by the finite quantity of agave bagasse available, so only a limited volume of biochar-treated vinasse (VT) could be generated. Because the full NMX-AA effluent panel (BOD₅, total phosphorus, fats and oils, total suspended solids, chlorides, and total phenols) requires more than 4 L of sample, analytical effort on VT was prioritized for the parameters that most directly capture the changes introduced by biochar contact (pH, electrical conductivity, chemical oxygen demand, and total Kjeldahl nitrogen); the remaining parameters are reported as not determined (n.d.). The complete raw-vinasse (V) characterization provides the reference composition of the starting effluent.

Proximate analysis for total solids (TS), moisture content, volatile solids (VS), and fixed carbon (FC) were carried out following standard methodologies (32). For TS and moisture content (%), a well-mixed volume sample was placed in a pre-weighed porcelain crucible and dried at 105 °C for 24 h or until constant weight. TS (%) was calculated as:

TS(%)=(Dry weightWet weight)×100

For VS and FC, the dried residue from the TS determination was combusted in a muffle furnace (TE-M12D, Terlab, Mexico) at 575 °C for 4 h. After cooling in a desiccator, the crucible was weighed again. VS (%) was calculated from the mass loss during ignition:

VS(%)=[(Dry weight−Ash weight)Wet weight]×100

Finally, FC (%) was calculated as the difference TS (%) − VS (%).

2.6 Statistical analysis

Analyses were performed in R v4.4.1 (33). Before hypothesis testing, the assumptions of each one-way model were verified within every family: normality of the residuals with the Shapiro–Wilk test (evaluated on the pooled linear-model residuals, which is robust when a group has n < 3) and homogeneity of variance with Levene’s test centered on the median. A pre-registered decision rule was then applied: when both assumptions held (α = 0.05), treatment effects on AUC were tested with one-way ANOVA followed by Tukey’s HSD test; otherwise, the non-parametric Kruskal–Wallis test followed by Dunn’s test with Bonferroni correction was used. In the present dataset all three families satisfied both assumptions (Shapiro–Wilk p ≥ 0.36; Levene p ≥ 0.63), so ANOVA + Tukey HSD was applied throughout, ensuring a uniform and fully parametric inferential framework. Pairwise outcomes were summarized as compact-letter displays, with the letter “a” assigned to the group of highest mean AUC. Cumulative emission curves are plotted as treatment means with standard-deviation ribbons; boxplots show the median, interquartile range and full range of the replicate-level AUC, with individual replicates overlaid. The complete, reproducible analysis script and the replicate-level AUC table are provided as Supplementary Material.

2.7 Kinetic modeling of carbon mineralization

The cumulative CO₂ emission kinetics were analyzed and modeled using MATLAB (The MathWorks, Inc., Natick, MA, USA). To evaluate the mineralization dynamics and distinguish between carbon pool stabilities, the experimental data were fitted to a double-exponential kinetic model, as described by (34), using non-linear least squares regression (lsqcurvefit function). The model is described by the following equation:

CO2−C=Cs(1−e−kst)+Cl(1−e−klt)

Where CO2−C represents the cumulative mineralized carbon (g CO₂-C g^-¹ soil) at time t (hours) emitted as CO₂; Cs and Cl correspond to the stable and labile organic carbon pools (g CO₂-C g^-¹ soil), respectively; and ks and kl denote the specific mineralization rate constants (h−1) for the stable and labile pools. The goodness-of-fit for the non-linear regression was assessed using the coefficient of determination (R2) and the Root Mean Square Error (RMSE).

It should be emphasized that within this framework both C₁ and C_s are mineralizable pools; C_s quantifies the carbon respired slowly rather than the carbon durably sequestered. The stabilized/sequestered fraction proper corresponds to applied carbon that is not respired during the incubation and is therefore not captured by C₁ or C_s. Throughout this work, C_s is accordingly interpreted as a kinetic proxy for stabilization potential, and the terms “stabilization” (protective mechanisms and slow turnover) and “sequestration” (net durable storage) are used in this restricted sense.

3 Results and discussion

3.1 Material properties

3.1.1 Physicochemical properties

The physicochemical properties of the field-collected soil from Amatitán (see Section 2.1) are summarized in Table 1. According to the INEGI municipal edaphological survey (35), the dominant soil unit at the sampling site corresponds to a Luvisol (36), consistent with the valley-floor position of the agave plot. The soil exhibited a neutral pH (7.0), low electrical conductivity (102 ± 13 µS cm^-¹), and a cation exchange capacity of 9.50 ± 0.33 cmol kg^-¹. The ash fraction dominated the soil matrix (81.63%), with a total organic carbon of 10.65 ± 0.02% and soil organic matter of 18.37 ± 0.03%.

Table 1. Physicochemical properties of the field-collected soil (5–10 cm depth).

Two biochar variants were used in this study: pristine agave-bagasse biochar (BC) and vinasse-impregnated biochar (BT). Their proximate and physicochemical properties are summarized in Table 2. Both materials exhibit similar volatile matter contents (BC: 69.46 ± 3.38%; BT: 69.58 ± 1.12%) and fixed carbon fractions (BC: 24.31 ± 3.41%; BT: 25.49 ± 1.16%), with low ash contents (BC: 6.42 ± 0.32%; BT: 4.93 ± 0.27%) and high organic matter (BC: 93.58 ± 0.32%; BT: 95.07 ± 0.27%). The elevated volatile matter (~70%) relative to biochars produced above 500 °C (typically < 30%; 25) is consistent with the low pyrolysis temperature (300 °C), which favors partial carbonization and retention of thermally labile functional groups. Total organic carbon was comparable between materials (BC: 54.28 ± 0.18%; BT: 55.14 ± 0.16%). Vinasse impregnation induced a decrease in pH from 7.07 ± 0.05 (BC) to 5.62 ± 0.03 (BT), attributable to the organic acids present in tequila vinasse (pH 3.55; Table 3), and a reduction in EC from 1153.0 ± 31.2 to 936.7 ± 106.7 µS cm^-¹, suggesting partial displacement of soluble mineral species during the impregnation step. Moisture content was higher in BT (5.56 ± 0.08%) than BC (3.21 ± 0.13%), reflecting the hygroscopic contribution of sorbed vinasse constituents within the biochar pore network.

Table 2. Proximate and physicochemical properties of pristine agave-bagasse biochar (BC) and vinasse-impregnated biochar (BT), both produced at 300 °C.

Table 3. Physicochemical characterization of raw tequila vinasse (V) and biochar-treated vinasse (VT).

n.d., not determined; not measured for biochar-treated vinasse (VT) owing to the limited volume of treated effluent available (see Section 2.5).

Raw tequila vinasse (V) and biochar-treated vinasse (VT) were also characterized (Table 3). Raw vinasse was strongly acidic (pH 3.55), with high organic load (BOD₅ 11,815 mg/L; COD 25,755 mg/L; BOD₅/COD = 0.46) and a total phenolic content of 192.52 mg/L, consistent with the chlorogenic, vanillic, syringic, and coumaric acid profile reported for tequila vinasse (37). Biochar treatment raised the pH to 5.04 and increased EC from 2,572 to 4,440 µS cm^-¹, reflecting the acid-neutralizing capacity of the biochar and leaching of its soluble mineral constituents. Phenolic content was not determined for VT; however, the pH increase is indirectly consistent with partial adsorption of phenolic acids by the biochar surface (38).

3.1.2 Surface functional groups

FTIR spectra of the field-collected soil, raw agave bagasse, pristine biochar (BC), vinasse-impregnated biochar (BT), and tequila vinasse are presented in Figure 1. Field-collected soil reflected its mineral-dominated character: the broad O–H/N–H absorption at 3367 cm^-¹ integrates contributions from clay-bound water, soil moisture, and amide-N of organic matter, while bands at 1674 cm^-¹ (amide I/conjugated C=O of humic substances) and 1517 cm^-¹ (aromatic C=C) confirm the presence of a humified, lignin-derived organic fraction. The dominant absorptions at 998 and 909 cm^-¹ fall within the Si–O stretching and Al–OH bending region of clay minerals (39), consistent with the high ash content (81.63%; Table 1). The vinasse spectrum exhibited a broad O–H/water absorption (3000–3500 cm^-¹), characteristic of the high moisture and alcohol content of distillery effluents, together with bands at 1490–1550 and 1550–1610 cm^-¹ associated with aromatic ring vibrations (40), consistent with the phenolic profile reported for tequila vinasse (chlorogenic, vanillic, syringic, and coumaric acids; Table 3). Aliphatic C–H stretching (2800–3200 cm^-¹) and C–O–C vibrations (950–1200 cm^-¹) in the vinasse spectrum can be associated with carboxylic acids, esters, and aliphatic hydrocarbons frequently present in distillery by-products (41).

The transformation sequence from raw bagasse through pyrolysis and vinasse impregnation is resolved in three spectral stages. First, pyrolysis attenuated the hemicellulose ester carbonyl at 1746 cm^-¹ and generated bands at 1674 and 1595 cm^-¹ (conjugated C=O and aromatic C=C, respectively), documenting the onset of aromatization; the retention of C–H stretching bands (2800–3200 cm^-¹) and high volatile matter (~70%; Table 2) confirms incomplete carbonization at 300 °C (42). Second, the C–O stretching bands characteristic of the cellulosic backbone (1256, 1154, 1078 cm^-¹) disappeared entirely in BC, confirming thermal decomposition of glycosidic and ether linkages. Third, these same C–O bands reappeared in BT, indicating that vinasse impregnation reintroduced polysaccharide-associated functionalities onto the biochar surface, while the pyrolysis-specific bands (1674, 1595, 1800 cm^-¹) were consumed through esterification or complexation with vinasse constituents. The aromatic C=C mode at 1517 cm^-¹ persisted across all materials, confirming that vinasse impregnation modifies surface chemistry without disrupting the pyrogenic carbon architecture.

3.2 CO₂-C headspace dynamics

Figure 2 resolves the carbon-mineralization response along two complementary axes: the upper row (A1–A3) shows the cumulative CO₂-C trajectories, while the lower row (B1–B3) shows the integrated cumulative respiration (AUC) per treatment with its statistical verdict. Across the three families a single pattern emerges: only the biochar family showed a statistically significant treatment effect on total cumulative respiration, whereas the vinasse and additive families did not. The mechanistic basis for this divergence is resolved by the kinetic decomposition in Section 3.3 (Figure 3).

Biochar family: a dose- and impregnation-dependent stimulation of respiration. By the end of the incubation, cumulative CO₂-C ranged from 1.72 × 10^-³ g g^-¹ soil in the unamended control SC-1 to 2.56 × 10^-³ g g^-¹ soil in BT-10, representing a 48.5 % increase over the control. Pristine biochar produced a graded response (BC-5: +8.8 %; BC-10: +35.1 % relative to SC-1), and vinasse-impregnation amplified it further at both doses (BT-5: +36.3 %; BT-10: +48.5 %). ANOVA detected significant differences only in this family (F₄,₁₀ = 10.84, p = 0.0012; Figure 2B1); Tukey’s HSD gave the ordered sequence SC-1 ≈ BC-5 < BC-10 ≤ BT-5 ≈ BT-10, mirroring both dose and surface modification.

Two features of this response deserve emphasis. First, pristine biochar at the low rate (BC-5) was statistically indistinguishable from the unamended soil, indicating that agave-bagasse biochar does not, by itself, provoke a large burst of CO₂, consistent with the predominantly aromatic, microbially poorly accessible carbon expected from slow pyrolysis and with the modest, near-zero effect of pristine biochar on soil CO₂ flux reported in global syntheses (20, 43, 44). Second, the increment that did appear scaled with dose and was systematically larger for the vinasse-impregnated material. Because impregnation loads the porous biochar matrix with the soluble, mineral- and phenolic-rich fraction of vinasse (27), the additional CO₂ from BT-5 and BT-10 is most parsimoniously attributed to mineralization of this sorbed labile fraction together with a possible substrate-driven priming of native soil organic matter (45, 46). The stable-pool expansion and its thermal confirmation are independent of the biological origin of the respired carbon; the priming component, however, remains an inference. Because neither the biochar nor the vinasse carbon was isotopically labelled, the respired CO₂ cannot be attributed to a specific source, and the additional respiration under BT cannot be unambiguously separated into mineralization of amendment-derived carbon, accelerated decomposition of native soil organic carbon (real priming), and accelerated microbial biomass turnover (apparent priming; 47). Distinguishing these components would require ¹³C or ¹⁴C isotopic tracing of the added materials (48, 49), which was beyond the scope of the present incubation. Importantly, a higher cumulative emission is not, on its own, evidence against carbon stabilization: whether the extra carbon respired by BT-10 reflects a genuine loss or the slow turnover of a large retained pool cannot be decided from cumulative totals and is resolved only by the pool-resolved kinetics in Figure 3. Figure 2 therefore deliberately frames, rather than answers, this question.

Vinasse family: organic-rich, yet respiratory silent. The vinasse family behaved in the opposite way to what its high organic load would suggest. The four vinasse treatments traced an exceptionally tight bundle of curves (Figure 2A2) and all finished at or below their co-incubated control SC-2, between 1.54 and 1.71 × 10^-³ g CO₂-C g^-¹ soil. AUC analysis returned no significant differences among treatments or against the control (F₄,₉ = 0.67, p = 0.629; all treatments share letter a; Figure 2B2).

This result is informative precisely because it is counterintuitive. Vinasse is a carbon- and nutrient-dense effluent, and direct land application is widely reported to stimulate soil respiration (37); the absence of any cumulative stimulation here points to constraints specific to the matrix and the incubation window. The acidic, saline and phenolic-rich character of tequila vinasse (the same chlorogenic, vanillic, syringic and coumaric acids highlighted in the Introduction) can transiently depress the activity of native decomposers (50), while the most readily oxidizable fraction of vinasse carbon may have been consumed within the first daily measurements, before the curves separate. The net outcome is a saturating trajectory that flattens early and never exceeds the unamended soil. From a circular-economy standpoint this is a favorable finding: under the doses tested, returning tequila vinasse to soil did not translate into a measurable cumulative CO₂ penalty over the incubation, which removes one of the commonly assumed environmental objections to its controlled agronomic reuse.

Additive family: highest emissions but driven by variability rather than treatment. Compost (COM) reached 2.82 × 10^-³ g CO₂-C g^-¹ soil (≈11% above SC-3), while inorganic fertilizer (FER) fell 12.7% below SC-3; AUC detected no significant differences (F₂,₆ = 1.12, p = 0.387; Figure 2B3) owing to large within-group dispersion. The additive treatments shifted the central tendency of cumulative respiration without shifting it reproducibly.

Synthesis across families. Read together, the three families occupy distinct respiratory regimes. Ranked by pooled family means, total cumulative emission decreased in the order additive (2.53 × 10^-³) > biochar (2.16 × 10^-³) > vinasse (1.66 × 10^-³ g CO₂-C g^-¹ soil), yet this ranking of magnitude is decoupled from the ranking of statistical structure: the family with the clearest, dose-ordered and significant treatment response was biochar, not the highest-emitting additive family. All families shared an early, substrate-driven respiration phase (roughly 36–51 % of the eventual cumulative total was reached within the first 120 h), after which their trajectories diverged. The functional message of Figure 2 is thus threefold: (i) agave-bagasse biochar, and especially vinasse-impregnated biochar, raises cumulative soil respiration in a dose-dependent and statistically significant manner; (ii) tequila vinasse applied directly to soil does not measurably increase cumulative CO₂ relative to unamended soil; and (iii) organic and inorganic additives move the central tendency of respiration but not reproducibly. Whether the carbon respired under the biochar treatments represents a net climate cost or the slow, sustained turnover of a large stabilized reservoir is the question that the kinetic decomposition (Figure 3) is built to answer; the cumulative and inferential evidence assembled here defines the magnitude of the effect and isolates biochar as the only amendment family with a robust functional signature.

3.3 Kinetic partitioning of soil carbon: double-exponential model fitting and stable pool dynamics

To resolve the mechanistic basis of the family-level divergence reported above (Section 3.2/Figure 2), cumulative CO₂–C emission curves were fitted to a double-exponential model under the biological constraint kl>ks. This decomposition partitions total mineralizable carbon into a fast-turnover labile pool (Cl, kl) and a slow-turnover pool (Cs, ks). Fits were excellent for all 13 treatments (R2≥0.996; CV(RMSE) ≤3.14%; Table 4), supporting reliable extraction of pool sizes and turnover estimated constants.

Table 4. Estimated kinetic parameters of the double-exponential decomposition model fitted to cumulative CO₂-C emissions across treatments, with biological constraint kl > ks. Cs and Cl: stable and labile organic carbon pools, respectively; ks and kl: first-order mineralization rate constants for the stable and labile pools; t½: half-life; Frac L and Frac S: proportion (%) of total mineralizable carbon allocated to the labile and stable pools, respectively; R²: coefficient of determination; CV(RMSE): coefficient of variation of the root-mean-square error.

Soil controls (SC-1, SC-2, SC-3) are shown in italics.

The biochar family produced a pronounced, dose-dependent expansion of the slow-turnover pool (Cs), with systematic amplification under vinasse impregnation at both doses (BT > BC; Figure 3A). BT-10 reached the largest Csof the entire study (0.0712 g g^-¹ soil), more than an order of magnitude above the unamended control SC-1 (0.0015 g g^-¹), followed by BC-10 (0.0279) and BT-5 (0.0230), a significant family-level effect (Kruskal-Wallis p= 0.026; Figure 3A). The slow-mineralizing fraction (FracS) rose correspondingly from 65.9% in SC-1 to 87.4–98.6% across the four biochar treatments (Table 4), indicating a fundamental shift in the partitioning of mineralizable carbon toward the slow-turnover compartment. The magnitude and direction of this response are consistent with the global meta-analytic synthesis of Gross et al. (20), which compiled 324 non-field (pot and incubation) observations and reported a mean relative SOC increase of 75% (95% CI 67–85%) following biochar application, with stronger effects in long-term incubations (>500 days, +163%) and at higher application rates (>20 Mg ha^-¹). Two specific contextual factors of the present experiment fall within ranges identified by that meta-analysis as favorable: the dose tested here (5–10% w/w, equivalent to 50–100 g kg^-¹) sits in the upper range of incubation studies surveyed, where Gross et al. (20) report significant SOC increases without saturation; and the agave-bagasse feedstock is plant-derived with an elevated C/N ratio (mean ~66 in their plant/wood-based subset), conditions that promote SOC accumulation through low microbial N availability and aromatic-C dominance. Their meta-analysis further documents that organic fertilizer co-applications significantly enhance SOC sequestration relative to biochar alone, providing an independent line of evidence for the BT > BC effect observed here, consistent with the vinasse impregnation acting as the organic co-amendment that amplifies the C-stabilizing action of agave-bagasse biochar.

Across the full dataset, Csand kswere inversely related on a log–log scale (Spearman ρ= –0.93, p<10−4; Figure 3C), with each family occupying a distinct region of the trade-off space. The biochar family populates the upper-left quadrant of Figure 3C (large Cs, slow ks), a kinetic fingerprint compatible with three concurrent stabilization mechanisms that operate over distinct timescales. First, the polyaromatic skeleton of pyrogenic carbon has intrinsically low decay rates, with mean residence times of 90–1600 years reported for biochars produced between 450 and 550 °C (20, 51). Second, biochar mineralization is suppressed in the later phase of incubation by sorptive protection of native SOC and dissolved organic carbon onto the biochar surface, an effect first quantified by Zimmerman et al. (46)and consistently reproduced in long-term incubations; in our 30-day experiment, this mechanism is most clearly expressed in BC-10 and BT-10, where FracS approaches 98%. Third, the BT > BC differential at both doses indicates that vinasse impregnation does not destabilize the pyrogenic matrix but instead loads the biochar with low-molecular-weight C and nutrients that are released progressively into the soil solution rather than mineralized in bulk (L. 52). For the vinasse family, the lower-right position in Figure 3C (small Cs, faster ks) is consistent with the documented biochemistry of distillery vinasse: organic acids (oxalate, lactate, acetate, malate), glycerol, ethanol and other low-molecular-weight compounds dominate the soluble C fraction (53), so a small persistent reservoir and rapid turnover of the remainder are exactly what the chemistry predicts. The additive family falls between these extremes, with a less steep family-level slope (Figure 3C) — a position consistent with compost and synthetic fertilizer modifying the kinetic structure of the soil to a smaller and more variable degree than the C-dense biochar amendments.

The rate-constant ratio kl/ksprovides a complementary index of kinetic partitioning between the two pools (Figure 3B). Biochar treatments dominated the upper end of this axis, with BT-10 reaching kl/ks≈ 308.7, where its labile and slow-turnover pools turning over at rates differing by more than two orders of magnitude; followed by BC-10 (126.7) and BT-5 (111.8); the family-level effect was significant (Kruskal-Wallis p= 0.023; Figure 3B). Vinasse treatments occupied the lower end of the spectrum (range 15.7–52.7; Figure 3B), and the compressed ratio confirms uniform mineralization of vinasse-derived C without a sharply separated persistent reservoir (consistent with the low-molecular-weight composition described above). The additive family fell between these extremes (range 61.2–159.8) and showed no significant within-family differences (p = 0.12; Figure 3B).

Within the additives, FER and COM occupied contrasting positions in kinetic space despite their statistically indistinguishable AUC (Figure 2B3), and their parameter signatures point to fundamentally different modes of action. FER reached Cs= 0.0173 g g^-¹ (~35% above SC-3 (0.0128)), paired with a sharply elevated kl/ks= 159.8 driven primarily by an accelerated kl(0.0177 h^-¹ vs. 0.0132 h^-¹ in SC-3, +34%) while Clremained close to control (0.000840 vs. 0.000926, −9%). The observed pattern, where mineral N addition consistently induces negative priming while organic C additions stimulate native SOC mineralization, indicates that inorganic fertilization does not reshape carbon kinetics by relieving microbial N limitation; instead, mineral N acts as an independent force that suppresses mineralization, while positive priming is governed by microbial energy acquisition which triggers the synthesis of SOM-degrading exoenzymes and accelerates microbial biomass turnover (54). COM, by contrast, increased both pools by closely matched proportions (Cs +8%, Cl+19%) while leaving kland kseffectively unchanged (kl 0.0110 vs. 0.0132 h^-¹ in SC-3; ks0.000180 vs. 0.000183 h^-¹), yielding a kl/ksratio (61.2) similar to control (72.2). This signature, proportional pool expansion at conserved turnover constants, is the kinetic fingerprint of mature, humified compost, which has already passed through the active phase of decomposition and contributes to soil C predominantly through stabilized humic substances that integrate into the existing soil C matrix without modifying its decay structure (55, 56). It also resolves the apparent paradox of the previous section, where COM produced the largest AUC of the study (+11% over SC-3): the extra respiration originates from the modestly enlarged Cl of compost itself (a +19% increase in absolute terms), not from accelerated mineralization of native SOC, confirming that compost adds readily mineralizable substrate without fundamentally reshaping the underlying kinetic matrix of the soil.

Taken together, the kinetic decomposition resolves each of the interpretive ambiguities left open by the cumulative respiration analysis (Section 3.2). The elevated CO₂ output of the biochar family does not represent a net carbon liability; rather, it is the surface expression of a vast slow-turnover pool whose turnover rate constant is one to two orders of magnitude slower than that of the labile fraction, yielding estimated half-lives of 83–979 days across the four biochar treatments versus 3–5 days for their corresponding labile pools (Table 4). The respiratory silence of the vinasse family is mirrored by its compressed kinetic structure: small C_s, fast k_s, and k₁/k_s ratios below 53, indicating that vinasse-derived carbon cycles through the soil without establishing a persistent reservoir, a profile that, while offering no carbon-sequestration benefit, confirms the absence of a cumulative CO₂ penalty under the tested doses. The double-exponential framework thus adds a mechanistic dimension that cumulative totals alone cannot provide and establishes vinasse-impregnated agave-bagasse biochar (BT) as the only amendment that simultaneously satisfies two conditions for effective soil carbon management: a large absolute increase in the slow-turnover pool and a rate-constant structure that ensures its long-term persistence. However, these conclusions rest on a mathematical partitioning of the mineralization signal; whether the modeled slow-turnover pool corresponds to a physically and thermally distinct carbon fraction in the post-incubation soil matrix is a question that requires independent, non-kinetic evidence. The following section addresses this question by examining the physicochemical properties and thermogravimetric signature of the incubated soils.

3.4 Effect on the physicochemical properties of the soil

To verify whether the theoretical projections obtained using the double exponential model correspond to actual structural changes in the soil matrix, the kinetic parameters were compared with the physicochemical properties analyzed at the end of the incubation period (Figure 4). This dual approach allows the rates of heterotrophic respiration to be linked to actual variations in the fractions of soil organic matter (SOM).

This confirms that tequila vinasse, under the doses tested, neither enriched nor degraded the organic, mineral, or thermal structure of the soil, reinforcing the absence of a cumulative CO₂ penalty and supporting its controlled agronomic reuse, in line with recent systematic reviews indicating that strategic vinasse application can improve soil properties without environmental deterioration when proper dosages are established (57). The additive family displayed internally divergent responses that mapped onto its distinct kinetic mechanisms: FER raised EC twenty-fold (6,573 vs. 327 µS m^-1 in SC-3) and depressed both STSI (1.10 vs. 1.93) and T50 (409.9 vs. 438.5 °C), consistent with salt-mediated negative priming and microbial necromass accumulation in the thermally labile window (31, 54), whereas COM shifted both thermal fractions proportionally with only a modest STSI decline (1.59), mirroring its kinetic signature of proportional pool loading at conserved rate constants.

The biochar family was the only group in which the soil showed a consistent, multi-variable departure from the unamended state (Figure 4, top row). pH rose from 6.95 ± 0.08 in SC-1 to 7.59 ± 0.01 in BT-10, consistent with the acid-neutralizing capacity of pyrogenic carbon (58), and EC increased from 122.7 to 228–295 µS m⁻¹ across all treatments, reflecting the release of soluble mineral constituents during incubation. More critically, the thermogravimetric indices reveal a dose-dependent enrichment of the soil organic matrix that directly corroborates the kinetically defined Cs expansion reported in Section 3.3: total SOM increased significantly in BC-10 (+65.8% over SC-1) and BT-10 (+23.0%), both of which showed significant departure from the control for labile SOM, STSI, and T50. The TGA-based detection of increased organic matter in biochar-amended soils is consistent with the thermogravimetric patterns reported by Yanardağ et al. (59), who documented that biochar specifically increased the labile SOC thermal fraction in a Luvisol, the same soil classification as the present study. However, a closer comparison between BC-10 and BT-10, the most carbon-enriched treatments of the study, uncovers a divergence that the kinetic model alone does not fully resolve.

BC-10 and BT-10 occupied overlapping statistical positions in the cumulative respiration analysis (both within the “a–ab” group; Figure 2B1) and shared comparably high stable fractions in the kinetic model (FracS 97.6 vs. 98.6%; Table 4), a partitioning consistent with the global meta-analysis of J. Wang et al. (60), who estimated a recalcitrant pool of ~97% with a mean residence time of 556 years across 128 isotope-traced biochar decomposition observations. This similarity might suggest equivalent modes of action. The physicochemical data reveal otherwise. BC-10 produced the largest thermogravimetric response of the entire study: SOM Total reached 10.31 ± 0.88% versus 7.65 ± 0.50% in BT-10, driven by a labile SOM of 5.00% (more than double the control) that depressed STSI to 1.08 and T50 to 404.6 °C, the lowest values across all thirteen treatments. BT-10, by contrast, showed more moderate thermogravimetric enrichment (SOM Labil 3.27%, STSI 1.36, T50 419.2 °C) yet harbored a Cs 2.6-fold larger than BC-10 (0.0712 vs. 0.0279 g CO₂–C g^-1 soil) and a half-life nearly three times longer (979 vs. 354 days). In other words, pristine biochar at 10% loaded more thermally detectable organic matter into the soil, but vinasse-impregnated biochar at the same dose retained a far larger and far more persistent carbon reservoir despite leaving a smaller thermal footprint.

This divergence is consistent with the surface-chemical transformation documented by FTIR (Section 3.1.2): vinasse impregnation consumed the pyrolysis-specific reactive groups (1674, 1595, 1800 cm⁻¹) through esterification or complexation with vinasse constituents while reintroducing polysaccharide-associated C–O functionalities onto the biochar surface. Hagemann et al. (61) identified an analogous process in co-composted biochar, demonstrating through high-resolution spectromicroscopy that an organic coating covering both outer and inner (pore) surfaces adds hydrophilicity, redox-active moieties, and mesoporosity, strengthening biochar–water interactions and nutrient retention, a mechanism the authors attributed to the coating itself, rather than to biochar surface oxidation. In the present study, vinasse impregnation functions as a comparable surface passivation process, creating an organo-mineral-coated pore network in which the volatile matter of the 300 °C biochar (~70%; Table 2) remains physically occluded and biologically inaccessible to soil microorganisms, even though it decomposes readily under the programmed oxidation conditions of the TGA (62, 63). The mechanistic importance of physical occlusion is further supported by Kalu et al. (64), who demonstrated that biochar pores accumulate dead microbial residues (necromass) and preserve added labile substrates through physical entrapment, simultaneously increasing microbial carbon-use efficiency and inducing negative priming of native SOC. Similarly, Burgeon et al. (65) showed that in century-old pyrogenic organic matter, 56.4% of total pyrogenic carbon was occluded in aggregates or sorbed onto mineral phases, concluding that long-term persistence is enhanced by occlusion and sorption processes beyond intrinsic chemical recalcitrance. In pristine BC-10, by contrast, the same volatile fraction retains exposed reactive surface sites that are partially accessible to microbial exoenzymes, consistent with (66), who reported that rice-straw biochar produced at 300 °C exhibited 520% higher hot-water-extractable organic carbon than biochar at 500 °C, yielding more thermally detectable SOM but a smaller and faster-turning kinetic pool. The BC-10/BT-10 comparison thus demonstrates that thermal lability and biological lability are governed by fundamentally different mechanisms, programmed oxidation versus enzymatic attack within a physically constrained pore environment, and that vinasse impregnation shifts the balance decisively toward biological persistence. This decoupling between thermal and biological recalcitrance is corroborated by Giannetta et al. (67), who showed that mineral-associated organic matter across multiple ecosystem types exhibits lower thermal stability than free particulate fractions yet is more persistent against decomposition, confirming that the association of labile organic molecules with soil minerals is a major stabilization mechanism independent of thermal resistance.

A cross-family comparison between the biochar and additive treatments further underscores the functional superiority of the waste-derived amendments. BT-10 and FER achieved nearly identical thermogravimetric enrichment (SOM Total 7.65 vs. 7.62%), yet their kinetic architectures are radically different: BT-10 harbored a stable pool 4.1-fold larger than FER (Cs = 0.0712 vs. 0.0173 g CO₂–C g^-1 soil) with a half-life 3.8 times longer (979 vs. 260 days), while FER imposed an EC of 6,573 µS m^-1, a salinity level twenty-five times higher than BT-10 (255 µS m^-1) that, sustained over repeated application cycles, would pose a considerable risk of soil sodification (68). BC-10 likewise outperformed COM in both organic matter loading (SOM Total +53%) and kinetic stability (Cs 2.0-fold larger, t½ 2.2-fold longer), despite compost being a widely endorsed soil conditioner. These contrasts are consistent with the meta-analytic finding of Joseph et al. (69) that biochars tailored through pre- or post-production treatments outperform conventional organic amendments in building soil organic carbon, particularly when site-specific constraints are addressed by appropriate biochar formulations. Critically, the biochar treatments in this study are derived entirely from tequila-industry waste streams (bagasse + vinasse), whereas compost and inorganic fertilizer represent external inputs with independent production costs and supply-chain dependencies, reinforcing the economic and logistic advantage of the biochar pathway within a circular bioeconomy framework.

Taken together, the post-incubation physicochemical analysis closes the evidentiary loop opened by the kinetic decomposition: the three-regime classification (biochar = structural modification, vinasse = null, additives = mechanism-dependent) is independently confirmed at the soil-matrix level. The BC-10/BT-10 contrast reveals that vinasse impregnation passivates the biochar surface to favor biological persistence over simple carbon loading, while the cross-family comparison with conventional additives demonstrates that the waste-derived biochar amendments achieve equal or superior organic matter enrichment with far greater kinetic stability and without the salinity penalty of inorganic fertilization.

4 Conclusions

This study integrated cumulative respirometry, double-exponential kinetic modeling, and post-incubation thermogravimetric analysis to evaluate the impact of agave-bagasse biochar (pristine and vinasse-impregnated), tequila vinasse, and conventional amendments on soil CO₂ mineralization and carbon pool stability over a 30-day laboratory microcosm incubation.

Directly measured, the biochar family was the only one to significantly raise cumulative CO₂-C (up to +48.5% for BT-10; ANOVA p = 0.0012), while post-incubation thermogravimetry independently confirmed a dose-dependent enrichment of soil organic matter and a shift in its thermal fractions (BC-10 SOM Total +65.8%; BT-10 + 23.0%). Tequila vinasse, whether applied raw or after biochar pre-treatment, changed no measured physicochemical or thermal property of the soil relative to the unamended control. Across the biochar treatments, pH rose (up to 7.59 in BT-10) and EC increased, whereas inorganic fertilizer imposed an EC of 6,573 µS cm^-¹, twenty-five times higher than BT-10.

Inferred from the double-exponential fits and model-based estimations, the additional biochar respiration reflects the slow turnover of a substantially enlarged slow-mineralizing pool (BT-10: C_s = 0.0712 g CO₂–C g^-¹ soil, FracS = 98.6%, k_s half-life ≈ 979 days) rather than a rapid carbon loss, a partitioning consistent with the 97%/3% ratio reported by J. Wang et al. (60). Comparison with pristine biochar at the same dose (BC-10) indicates that vinasse impregnation does not merely add carbon but modifies its interaction with the soil matrix: BC-10 enriched SOM more visibly by TGA, yet BT-10 retained a 2.6-fold larger and nearly three-times slower-cycling modeled pool.

These laboratory and model-based findings now would require longer timespan observations and field validation before agronomic recommendations can be made. Priorities are: (i) multi-season field trials under repeated application cycles to test whether the observed vinasse neutrality and biochar stabilization persist under cumulative loading; (ii) direct mass-balance combined with ¹³C or ¹⁴C isotopic tracing of the amendments, to quantify the genuinely sequestered fraction and to separate real from apparent priming; and (iii) life-cycle assessment of the integrated bagasse-pyrolysis + vinasse-impregnation chain at production scale.

From a circular-bioeconomy perspective, the results position vinasse-impregnated agave-bagasse biochar as a dual-purpose valorization strategy and a functionally competitive alternative to conventional soil amendments: BT-10 matched the thermogravimetric enrichment of inorganic fertilizer (SOM Total ≈ 7.6%) while delivering a slow-mineralizing pool 4.1-fold larger and a half-life 3.8-fold longer, without the extreme salinity loading of mineral fertilization. Because both biochar amendments are derived entirely from tequila-industry waste streams, a single process chain transforms the two largest by-products of tequila production into a soil amendment with strong long-term carbon-stabilization potential, providing a mechanistic foundation for the field-scale trials and life-cycle assessment identified above.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fsoil.2026.1911074/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Martin A. N. Anikwe, Enugu State University of Science and Technology, Nigeria

Reviewed by: Pin Han Chen, Providence University, Taiwan

Manasa P, Koneru Lakshmaiah Education Foundation, India