To produce the different textures of artificial soils, we have used quartz grains of varying particle sizes (EUROQUARZ, Laußnitz, Germany, and Quarzwerke, Frechen, Germany). We added goethite (<6 µm), illite (<7 µm), and bentonite (<63 µm) (ASPANGER, Aspang, Austria) to create soil-like reactive surfaces in the fine-silt and clay size fractions. The C content of the artificial soils was considered to be negligible due to the insignificant C concentration of the ingredients. The soil mixtures were prepared according to the proportions presented in Table 1. The texture classes were defined as (1) loamy sand, (2) sandy loam, (3) loam, (4) clay loam, and (5) silty clay.

The artificial soils were prepared by mixing the ingredients in a dry state, ranging from coarser- to finer-scale particles. As each component was added, 10 manual turns were executed to combine them. After all the ingredients were mixed, the containers were added to a horizontal shaker and shaken overnight at 140 rpm. After the overnight shaking period, each soil mix container was manually turned 30 times.

The experimental setup was a 5×2 factorial design testing five different textures of artificial soils with and without biochar application with three replicates. For the incubation experiment, we used 20 g of artificial soil per experimental unit in 120 mL glass flasks. We added ball-milled air-dried clover biomass (Trifolium sp.) (C : N ratio of 18) as an organic matter source to all samples at a rate of 27 mg C g⁻¹ soil to mimic the natural background organic carbon (OC) content of arable topsoils. Ball-milled organic matter was chosen to enhance the development of these soils since the inorganic components added were C-free. We have used dissolved organic matter extracted from a local crop field as an inoculum of soil microorganisms for the artificial soils, applied at a rate of 0.06 mL g⁻¹ soil, in accordance with Pronk et al. (2012). The artificial soils were incubated under 60 % of the maximum water-holding capacity to ensure microbial activity (Fig. S1 in the Supplement). The water used to add the inoculum was accounted for in the amount added. Biochar was produced from Norwegian spruce at 700 °C, with 7 min holding time, and was added to the soil at a rate of 50 mg biochar g⁻¹ soil. Biochar had a C content of 95.6 %. The added biochar had a size distribution between 0.063 and 2 mm, controlled by sieving before addition to the soil. The basic biochar properties are described in Table S1 in the Supplement.

We measured CO₂ production over 115 h using an automated incubation system, described in detail by Molstad et al. (2007), with modifications as described in Molstad et al. (2016). The system consists of an autosampler (CTC PAL) connected to an Agilent gas chromatograph (Model 7890A, Agilent, Santa Clara, CA, USA). The system allows for high-resolution analysis of headspace gas concentrations in airtight 120 mL serum bottles. Corrections were applied to adjust for sampling dilution, leakage, and CO₂ equilibrium state as a function of the material pH and soil solution volume (Appelo and Postma, 2005). We monitored the weight of the samples during the incubation period to check whether adjusting the water content was necessary. Since this was a relatively short incubation (5 d) conducted in a closed environment, the water loss was minimal, and we did not need to refill the flasks to maintain the initial moisture content.

Soil pH was measured in a 0.01 M CaCl₂ solution according to the ISO 10390:2021.

We performed the Shapiro–Wilk test to confirm the data's normal distribution. Once the data normality was confirmed, we conducted a two-way analysis of variance (ANOVA) and a post-hoc test using Fisher's least significant difference (LSD) test to compare biochar and control treatments within each soil textural group. We used linear regressions correlating the increase in clay and silt content with the variables of interest to compare the different textural classes.

We observed significant effects of both biochar application and texture on C mineralization and a significant interaction between these variables in the analysis of variance (ANOVA) at p<0.01. Biochar application significantly reduced C mineralization rates compared to the control in the coarser-textured soils, namely (1) loamy sand, (2) sandy loam, and (3) loam (Fig. 1), according to the LSD test at p<0.05. However, no differences were observed between biochar and control treatments for the finer-textured soils, namely (4) clay loam and (5) silty clay (Fig. 1).

The reduction in C mineralization promoted by biochar in the loamy sand (soil 1) was over 6-fold higher than in the silty-clay soil (soil 5). The influence of clay and silt content on C mineralization also differed depending on whether or not biochar was applied (Fig. 2). However, the impact of soil texture was minimal in magnitude compared to the changes promoted by biochar addition. The control samples' clay and silt contents generally decreased C mineralization (Fig. 2). Every milligram of silt- and clay-sized particles in the artificial soils reduced clover residue mineralization by 0.00007 mg g⁻¹ soil (Fig. 2) under a constant moisture level (60 % of the water-holding capacity), given the slope of the equation. After a 5 d incubation in the silty-clay soil (soil 5), the clover mineralization was 0.06 mg CO₂ g⁻¹ soil lower than in the loamy-sand soil (soil 1).

Soil pH (CaCl₂) was significantly influenced by increased soil texture and biochar application (Fig. 3). Overall, increases in pH due to biochar addition were higher with increasing clay and silt contents, i.e., by 0.68 in the silty clay and 0.24 in the loamy sand (Fig. 3a). The increase in silt and clay particles alone also promoted increases in soil pH (p<0.05) (Fig. 3b). Given the slope of the linear regression equation, every milligram of the clay and silt fraction promoted an increase of 0.0003 in soil pH (CaCl₂). This increase was 3 times higher (0.009) when biochar was applied (p<0.01) (Fig. 2b).

The significant interaction between biochar application and soil texture suggests that the extent of biochar's impact on the mineralization of plant residues depends on soil texture, as observed by different studies (Butnan et al., 2017; Gross et al., 2021; Wang et al., 2017). On the other hand, soil texture alone had minor effects on C mineralization. The differences in the magnitude of early plant mineralization depending on soil texture suggest that biochar had a higher impact on reducing early C mineralization from clover residues in sandier textures than in clay-rich soils, which has also been observed under natural soil conditions (Butnan et al., 2017). Biochar has known benefits in terms of enhancing soil properties associated with promoting C protection, such as aggregation (Wang et al., 2017; Juriga et al., 2021), sorption capacity (Siedt et al., 2021), water retention, and porosity (Obia et al., 2016) in natural soils. Such effects may have potentially influenced the decrease in the early mineralization of the added plant residues since the improvement of the soil structure in artificial soils, such as through aggregate formation, is already observed within the first days of their development (Pronk et al., 2012).

Improvements in soil structure and C stability upon biochar application are known to happen at a higher intensity in sandy soils compared to in more clayey ones (Buchkina et al., 2017; Butnan et al., 2017), which may justify biochar's better performance in reducing early plant mineralization in sandy soils in our study. Our observations corroborate general trends observed in non-field experiments (Gross et al., 2021), with higher biochar performance in sandy textures compared to in clay. On the other hand, other experiments in incubation setups, such as that of Wang et al. (2017) and the field experiment observations made by Gross et al. (2021), suggest an opposite trend. The reasons for these discrepancies seem to depend mainly on the initial condition of the evaluated soils. For example, in the incubation setup, comparing contrasting textures in Wang et al. (2017), the unamended samples showed no difference in terms of water-stable aggregates despite their contrasting textures, and the C content was higher in the sandy soil. This experiment resulted in higher impacts of biochar application to clay soil structure and C storage. On the other hand, in our experiment, the unamended samples with higher clay content had a significantly higher water-holding capacity (Fig. S1), indicating an overall better structure than the sandy soils, and the organic matter added to them was the same for all soils, which resulted in a better performance of biochar in more sandy textures. Therefore, the impacts of biochar on reducing C mineralization in our experiment suggest that the amendment can compensate for the lack of soil structural quality in promoting early C persistence in soils.

Likewise, Gross et al. (2021) pointed out the low initial soil C content of sandier textures as a reason for observing higher SOC sequestration potential in non-field experiments. The contrast with field experiments can also be due to factors that happen only in field trial setups, such as biochar leaching in sandy soils; these are factors which are not observed under pot or incubation setups. Therefore, our interpretations are more focused on the intrinsic relationships between soil texture and biochar and their impacts on C mineralization in a controlled environment. However, climatic factors and long-term effects need to be confirmed through long-term field trials. Nonetheless, since most of the works evaluating the performance of biochar in different soil textures vary with regard to other factors such as C content, mineral composition, and structural properties, our work offers important insights into understanding these interactions of biochar with soil particle size in C mineralization.

Also, the increase in fine particles in soils is assumed to decrease organic matter decomposability due to the increased opportunity for physicochemical protection (Hassink, 1992; Kravchenko and Guber, 2017), which justifies the minor reductions observed in the mineralization rates caused by increased silt and clay fractions (Fig. 2). However, despite the significant correlations, the magnitude of changes in mineralization caused by soil texture was negligible in this experiment, given the low slope values (Fig. 2). Therefore, our results suggest that soil particle size (i.e., texture) has played a milder overall role than the application of biochar in our short-term study.

In particular, in sand-rich soil, where the available mineral surface area (as well as the permanently charged surfaces of clay minerals) is low, biochar can deliver additional surface area for adsorption processes. The biochar effect is probably reduced by increasing clay content as the higher clay mineral surface area and the smaller pores may overrule the biochar effect on the physicochemical protection of organic matter (OM). Since the biochar added in our experiment was a standard size, between 0.063 and 2 mm, we can argue that the biochar composition was foremostly responsible for the observed effects on plant mineralization rates rather than a change in the proportion of particle sizes in each soil texture. Nevertheless, the results with regard to texture controlling soil organic matter mineralization are contrasting in the literature, with results being dependent on soil moisture and other experimental conditions (Li et al., 2020, 2022). Our experiment using artificial soils with precise compositions and an early formation stage helped shed light on the effects of textures and particle size on organic matter decomposition.

The pronounced biochar effect on pH is likely due to its high acid-buffering capacity. Biochar consists of both alkaline functional groups and mineral ash containing both base cations and secondary carbonates (Fidel et al., 2017), explaining why the biochar and clay and silt content have increased soil pH. The higher pH of the soil solution promotes the dissolution of CO₂ gas, thereby reducing the amount of gas released from soils. This can result in a measurement artifact of lower-than-actual mineralization rates in pH-enhanced soils (Ma et al., 2013). However, this effect was accounted for in our study as mineralization rates were corrected for amounts of CO₂ dissolved in solution as a function of solution volume and pH (Appelo and Postma, 2005).

The liming effect of biochar is recognized in a variety of soils (Bolan et al., 2023), with an overall higher effect in clay than in sandy soils (Ajayi and Rainer, 2017), which agrees with the findings of our experiment (Fig. 3a). Nonetheless, soil comparisons under natural conditions also vary in terms of mineral composition, making the isolated textural effect less clear (Ajayi and Rainer, 2017). In this sense, our artificial-soil setup helps to clarify this effect by demonstrating a significant role of clay and silt fractions in increasing the soil pH (Fig. 3b) in soils with the same mineral composition and organic matter content. Our results also showcase that this influence of clay and silt particles on soil pH is significantly boosted in biochar-amended soils as the slope of the linear correlation was 3 times higher in the biochar treatment than in the control (Fig. 3b).

Direct sorption of CO₂ into biochar has also been reported alongside N₂O adsorption. It can also be considered to be a mechanism through which biochar can reduce the presence of these gases in the atmosphere (Cornelissen et al., 2013). This effect was probably counterbalanced by a higher clay and silt content in the soils, which may have protected the biochar surfaces and thus prevented CO₂ sorption, justifying the relatively greater increases in C mineralization in relation to the function of clay and silt in biochar-amended soils (Fig. 2). A higher soil pH has been shown to increase the mineralization of organic residues in soils (Li et al., 2007; Khalil et al., 2005). This effect has been attributed to a higher microbial biomass and a more bacteria-dominated microbial community in high-pH soils (Li et al., 2007; Laura 1976). In addition, a high pH causes stress for the microbial community, resulting in lower carbon use efficiency, translating into lower assimilation for organic C and a higher CO₂ release (Li et al., 2007). Some authors propose that the higher mineralization of organic matter may foster the supply of nutrients like nitrogen and sulfur, further stimulating microbial activity (Barrow and Hartemink, 2023). Due to mineral surface availability, a high clay and silt content may have leveled out the biochar effect. Sorption of cationic nutrients on those mineral surfaces may have led to limited availability, reducing the CO₂ release in the mixtures with biochar and a high clay and silt content (Barrow and Hartemink, 2023). Therefore, we encourage further studies exploring the specific roles of microorganisms in artificial-soil setups to understand how they can influence these microcosms and how these results compare to natural conditions.

In conclusion, our study suggests that biochar has a greater capacity to promote reductions in the early mineralization of clover residues in sandy soils. This potential diminishes as the clay and silt content increases in the soil. Here, we show that the artificial-soil setup using a precise composition is a suitable tool for investigating the effects of factors that are difficult to control in natural soils due to their heterogeneity. Our mineralization rates are comparable to other studies using artificial soils (Vogel et al., 2014; Bucka et al., 2021a, b) but can differ from experiments done under natural conditions (Gross et al., 2021). Therefore, interpretations made here are mainly applied to understand the intrinsic relationships between soil particle size and the biochar application to soil C persistence but should not be accounted for for purposes of quantifying C emissions from determined soil types or land uses.