Biochars were analyzed according to the analytical guidelines of the European Biochar Certificate (EBC, 2024) by Eurofins Umwelt-Ost GmbH (Bobritzsch-Hilbersdorf, Germany). The organic carbon (C_org) content of the dry and ash free (daf) biochar content (C_{org_daf}) was calculated according to Equation 3.

Elemental analysis according to EBC standards was also performed on samples after density fractionation. A triplicate of 3 g RE-biochar (milled to <20 μm in a ball mill) was suspended in a sodium-polytungstate solution adjusted to a density of 2.5 g cm⁻³. The two fractions (<2.5 g cm⁻³ = biochar-dominated; >2.5 g cm⁻³ = rock dominated) were separated and washed with de-ionized water in a vacuum filtration system.

For biochars and RE-biochars of the 1st production batch, total concentrations of main and trace elements were calculated according to Equation 4:

Where c_E,i is the total concentration of element i in the RE-biochar, c_{E,i(biomass ash)} the concentration of element i in the biomass ash (Supplementary Table S10), c_ash(biochar) the content of biomass-derived ash in the given RE-biochar (section 2.6.1 and Table 2), c_E,i(rock) the concentration of element i in the rock powder (Supplementary Table S7), with c_rock being the rock content in the given RE-biochar (section 2.6.1 and Table 2). The c_E,i for main elements is calculated as wt% (oxide form), for trace elements in ppm. A comparison between calculated and measured C_E,i is presented in Supplementary Table S6. The liming potential, given as calcium oxide equivalents (CaOe), was calculated based on threefold determination of basic compounds following VDLUFA II.1, 6.3.2: 1995.

The bulk density was calculated from the weight to volume ratio of a 100–120 cm³ subsample, milled < 3 mm (standard procedure in comparative biochar analysis) and dried 24 h at 105°C. For the calculation of the water holding capacity dried subsamples were weighed into 25 mL glass filters in triplicate, subsequently submerged in water for 12 h until consecutive draining for 12 h on a sand bed. Based on recorded weight differences, the gravimetric water content and water-holding capacity were calculated. The calorific value was measured according to DIN 51900-3: 2005-1.

Specific surface area (SSA) and porosity were quantified by Brunauer Emmett Teller (BET) method, based on nitrogen isotherms at 77 K aquired on an AUTOSORB-IQ volumetric sorption device from 3P-Instruments GmbH & Co KG (formerly Quantachrome GmbH & Co KG, Odelzhausen, Germany). Between 0.1 g and 0.2 g of sample was weighed into the sample cell, then degassed at 150°C for 12 h, with a final vacuum <10⁻² Pa applied before measurement. The surface area was determined in accordance with DIN ISO 9277: 2014. The mesopore size distributions (PSD) were determined, based on DFT calculation (QSDFT-Kernel, carbon slit/cylindrical pores, adsorption branch) by using ASiQwin Software package (Anton Paar, Ostfildern-Scharnhausen, Germany; formerly Quantachrome Inc., Boynton Beach, US). The corresponding plots obtained are shown in Supplementary Figure S8.

Fourier transformation infrared spectroscopy (FTIR) was performed as attenuated total reflectance (ATR) measurement using an Invenio X FT-IR Spectrometer (Bruker Corporation, Billerica, United States of America). The calculated infra-red absorbance was obtained from 60 consecutive scans, each covering the wavenumber spectra of 400–4,000 cm⁻¹. The absorbance spectrum was smoothed, and the baseline corrected using the software Spectragryph v.1.2.26.1, which also assisted with peak labeling. Peak identification was based on de la Rosa et al. (2014), Hagemann et al. (2018), and Johnston (2017).

Hydropyrolysis (HyPy) was performed as described in Meredith et al. (2017). In brief, milled samples were mixed with 10 wt% ammoniummolybdate-tetrahydrate as a catalyst. The samples were heated in a reactor under 150 bar hydrogen pressure from ambient temperature to 250°C at 300°C min⁻¹, then from 250°C to 550°C at 8°C min⁻¹ and held there for 2 min. The sample residues following HyPy were weighed and analyzed for C_org, which is referred to as BC_HyPy and presented as wt% of the initial C_org content.

For scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX), samples were fixed with carbon pads (Plano GmbH, Wetzlar, Germany) to an aluminum sample holder and were sputter-coated with a 3–5 nm gold layer using a Cressington 108auto (TESCAN GmbH, Dortmund, Germany), then subjected to SEM analysis using a JSM-6610 LV (JEOL Ltd., Tokyo, Japan). A working distance of 11 mm and 15 kV acceleration voltage were used. Images were created as backscattered electron images (BES). A 20 mm² Oxford X-mas detector (Oxford Instruments, Abingdon, United Kingdom) was used for EDX.

Thermogravimetric analysis and differential scanning calorimetry (TG-DSC) were performed on a STD 650 TG-DSC system (Waters, New Castle, United States of America). The dry samples (5 mg) were heated at a rate of 10°C min⁻¹ to 1,000°C under N₂ atmosphere. The total weight loss recorded was corrected for the ash content of the sample, which includes the rock content, to represent the weight loss of dry and ash-free (daf) biochar. For diabase-containing RE-biochars, the total weight loss during TGA was further corrected for weight loss recorded for pure diabase due to the conversion of carbonate to CO₂ (Supplementary Figure S1; Supplementary Table S5). The daf biochar mass fraction resisting thermal degradation until 1,000°C was defined as BC_1000C in mass%.

The fraction of silicate rock (c_rock) in composite feedstock materials or RE-biochar was calculated according to Equation 5.

Where c_ash (%) is the ash content of the material, c_Corg (%) the content of organic carbon, and r the ratio of c_ash (%) to c_Corg (%) of the corresponding pure biomass or pure biochar. The fraction of biomass in the composite feedstock material or biogenic components in RE-biochar (c_bio) is calculated according to Equation 6.

The mass yield (y_ma) is calculated according to Equation 7.

Where m_feedstock is the mass of the feedstock (dry matter) and m_biochar is the mass (dry matter) of the resulting biochar.

The carbon yield (y_c) is calculated according to Equation 8.

Here, c_{Corg(biomass)} is the organic carbon content of the pristine biomass and c_{Corg(biochar)} the organic carbon content of the biochar.

The stoichiometric carbon sink (C-sink) potential of RE-biochar is calculated as the sum of the stoichiometric C-sink potential of its pyrogenic carbon content (PyC-Sink) according to Equation 9 and the stoichiometric C-sink potential of the inorganic carbon, generated from rock weathering (IC-Sink) according to Equation 10 (cf. Renforth, 2019) both given in tCO₂e t⁻¹.

Where m_{RE_biochar} is the mass of the RE-biochar, M_CO2 the molar mass of CO₂ and M_C the molar mass of carbon.

Here, the respective M_{CaO/MgO/K2O/Na2O/SO3/P2O4} refer to the molar mass of oxides, and CaO/MgO/ K₂O/Na₂O/SO₃/P₂O₄ (%) to the mass fraction of the given metals as oxides in the rock.

The factor of 2 accounts for the stoichiometry of the oxides and valence of the cations released. The factor of 0.85 represents the ratio between the charge of the released cations and the sequestered CO₂, which accounts for CO₂ losses, due to equilibration of the oceanic carbonate system once bicarbonate enters this final reservoir. The latter is based on current ocean temperature, salinity and pCO₂ (Renforth, 2019). For RE-biochars, the formula m_rock = m_{RE_biochar} * c_rock is used.

Production of RE-biochars was carried out in one continuous batch with the yield recorded in regular intervals and the coefficient of variation being calculated (section 2.1). Generally, analytics were performed on representative samples without repeated measures, thus n = 1 with no further statistical evaluation. Where repeated measures were performed (density fractionation; section 2.4.1), statistical evaluation of paired data was performed with GraphPad Prism version 10.4.1.

Carbon yield refers to the proportion of C_org in the feedstock that is converted to biochar-C_org. Pyrolysis of wood pellets (low ash, high lignin feedstock) resulted in a carbon yield of 37.3 ± 1.1%. The addition of rock powder tended to slightly reduce the carbon yield, with a mean carbon yield of 35.4 ± 2.4% for all W-P-based RE-biochars (Figure 2A). For S-P-based (high ash, low lignin feedstock) biochar, the carbon yield was 37.9 ± 1.3%, which was not affected by rock addition (37.4 ± 1.7% for all S-P-based RE-biochars, Figure 2B).

The highest carbon yield of 39.9 ± 0.6% was achieved for W-based biochar (non-pelleted) (Figure 2C), which was not affected by low rock powder additions of 10% but was reduced to 27–32% by higher rock powder additions (50BaW, 50DiaW). An inspection of the burning chamber after project completion revealed fine material residues—potentially both from biomass and rock powder, indicating mineral and carbon losses. Since the research pyrolysis unit lacked a gas filter, some fine particles were likely drawn into the pyrolysis gas combustion chamber. Consequently, the observed differences in mass balance may result from these losses. Therefore, caution should be exercised when drawing conclusions regarding the carbon yield and ash content of non-pelleted RE-biochars.

All (RE-)biochars showed H:C_org ratios ≤ 0.4. Pure biochars had a H:C_org of 0.2 (Figures 1D–F), which remained largely unchanged by 10% nominal rock powder addition, but increased to 0.33–0.40 for 50% nominal rock powder addition. Density fractionations of RE-biochar showed that the low-density fraction had a lower H:C_org ratio than the fraction with a density >2.5 g cm⁻³ (Figure 3).

The fraction of dry and ash free (daf) biochar resisting thermal degradation until 1,000°C under N₂ atmosphere, here referred to as BC_1000C, was determined using TGA. W-P showed a BC_1000C content of 68.9% (Figure 4A), which was increased by addition of rock powder at low dosages to 74.1% (+8%) and 75.1% (+9%) for 10BaW-P and 10DiaW-P, respectively, but decreased at high dosages to 47.9% (50BaW-P) and 32.2% (50DiaW-P), respectively. Most daf biochar was volatilized between 750°C and 1,000°C. The fraction of daf biochar lost in this temperature range increased from 9.7% in W-P to 29.2 and 27.8% in 50BaW-P and 50DiaW-P, respectively.

The highest rate of weight loss, referred to as d(T) (Supplementary Figure S7), occurred (closely aligned with the d(T) of W-P) at 673°C in 10BaW-P and 50BaW-P, shortly after the initial pyrolysis temperature (650°C) was exceeded. The d(T) peaks of 10DiaW-P and 50DiaW-P were shifted upwards to 714°C and 723°C, respectively. This temperature range coincided with weight loss from carbonates in pure diabase (Supplementary Figures S1, S9).

Straw biochar had a BC_1000C content of 76.5% (S-P, Figure 4B), which decreased to 63.5% (10DiaS-P) and 32.9% (50DiaS-P), respectively, when diabase was added. Remarkably, thermal stability was severely reduced in straw RE-biochars with basanite, resulting in a BC_1000C content of zero. For S-P, d(T) peaked already at 550°C (10BaS-P) and 593°C (50BaS-P), well below the initial pyrolysis temperature (Supplementary Figure S7). Like W-P-based RE-biochar, the BC_1000C content of W-based RE-biochars (Figure 4C) decreased at 50% nominal rock powder addition yet remained unaffected at 10% rock powder addition.

The BC_HyPy fraction corresponds to the fraction of C_org that consists of clusters >7 fused aromatic rings (Howell et al., 2022), a measure for the long-term stability of biochar. It accounted for 90.3–95.5% of initial C_org for all samples (Figures 4D–F), except for 10DiaW, which showed a considerably lower BC_HyPy content (81.8%). Increases in BC_HyPy by up to 2.5% in W-P RE-biochar with 10% nominal rock powder addition are within the typical error of ±2% in HyPy analysis, which is limited by the accuracy of the elemental analyzer (Meredith et al., 2012).

The SEC of RE-biochar behaved inverse to the rock powder content. Pure biochars showed a SEC in a range of 8 mS cm⁻¹ (W-P) and 25 mS cm⁻¹ (S-P), which decreased to 2.3–24.0 mS cm⁻¹ at low rock amendment rates, and to 0.3–3.6 mS cm⁻¹ at high rock amendment rates, i.e., by up to 99%. The correlation between SEC and H:C_org ratio is pronounced (Exponential one-pool; R² = 0.76; Supplementary Figure S5). Still, the SEC of RE-biochars was higher than the SEC of mass equivalent post-pyrolysis mixtures of biochar and rock powder (Supplementary Table S15).

An absorption spectrum typical of high-temperature biochar was measured for W-P (Figure 5). Distinct peaks in the band of 1,020–1,160 cm⁻¹ relating to vibrations of C-O bonds of polysaccharides and carbohydrates (Johnston, 2017) were absent, however, strong absorption signals obtained in the band of 1,580–1,610 cm⁻¹ indicate the presence of C-C bonds in aromatic moieties. Compared to W-P, the peak intensity for O=C=O (CO₂) at 2349 cm⁻¹ decreased slightly in 10BaW-P and 10DiaW-P, with a stronger signal reduction in 50BaW-P and 50DiaW-P. This relates to a reduced infra-red absorption by gaseous CO₂ within the sample. There was a strong positive correlation between peak intensity at 2349 cm⁻¹ and cumulative pore space quantified by gas-adsorption, particularly for W-P based RE-biochar (R² = 0.71, Supplementary Figure S15B). The RE-biochars 50BaW-P and 50DiaW-P showed lower peak intensity at 1580 cm⁻¹ compared to W-P, which relates to stretching vibrations of C-C bonds, an indication of aromatic carbon moieties (de la Rosa et al., 2014; Johnston, 2017). Note that a decreased peak intensity may also be caused by dilution effects, as 50BaW-P contains only 17.0% C_org, compared to 84.3% C_org in pure W-P. The correlation between peak intensity at 1580 cm⁻¹ and C_org content was strong (W-P based RE-biochar: R² = 0.88, Supplementary Figure S14B). It is worth noting that for wood-based RE-biochars, employing nominal additions of 10% rock powder, the peak intensity at 1580 cm⁻¹ did not decrease, despite a 25% decrease in C_org content (Supplementary Figure S14B).

Between 700–890 cm⁻¹, the bending vibration of out-of-plane C-H bonds located at the edge of aromatic structures, are observed. Here, different peaks could be distinguished, e.g., for aromatic rings with only one out-of-plane C-H bond (high wavenumber; low H:C_org of aromatic ring) and up to three or four out-of-plane C-H bonds (low wavenumber; high H:C_org of aromatic ring) (Supplementary Figure S4; Russo et al., 2014; Johnston, 2017; Chikamatsu et al., 2018). Peaks from aromatic rings with single out of plane C-H bonds overlapped with a peak at 880 cm⁻¹ that was also recorded for pure basanite and diabase. However, the signal is also present in every biochar, including W-P. Peaks from aromatic rings with three to four out of plane C-H bonds became visible only in 10DiaW-P and 50DiaW-P.

Generally, absorption by minerals were located at bands between 950–1,500 cm⁻¹ (Johnston, 2017), thus interference with the interpretation of the abovementioned aromatic and aliphatic carbon-based absorption bands was unlikely. The amended silicate rock material can be detected by peaks at 990 cm⁻¹ (Hagemann et al., 2018; Smidt et al., 2002), indicating vibrations of Si-O along with strong signals at 1420 cm⁻¹ for diabase amended RE-biochars, indicating the presence of carbonates including CaCO₃. Absorption bands for phosphate (1000–1,050 cm⁻¹) are situated in the same region but masked by Si-O vibrations.

The calculated content of essential (e.g., Mg, Ca) and beneficial (Si) macro- and micronutrients in RE-biochars increased proportionally with increasing rock powder content (Table 2). Taking the example of 50BaW-P, the SiO₂ content increased from 0.2% in W-P to 35.1%, the MgO content from 0.4 to 6.6% and the CaO content from 4.9 to 10.3% when compared to W-P. These differences were smaller for straw biochars, which originated from biomass rich in calcium and potassium. Phosphorus content in RE-biochars was lower than in pure biochar, since the amended basanite and diabase showed low concentrations of P₂O₅. The calculated concentration of trace metals (Cu, Zn, Ni, Cr) was proportionally increased in RE-biochar, where trace metal concentrations in the rock powder were higher than those in the pure biochar. Scanning electron microscopy coupled to EDX showed the spatial heterogeneity of RE-biochar elemental composition at the micron scale (Figures 6A–F).

The calculation of main- and trace metal contents in RE-biochar according to Equation 4 led to results generally well comparable to direct measurements following the EBC standard methodology (Supplementary Table S6; EBC, 2024) and showed absolute differences in the range of ±2 wt% over all elements observed. However, where concentrations of a given element were small, the difference between calculated and measured nutrient concentrations often translated into a larger relative error.

While the addition of 10% rock powder to biomass had negligible effects on the biochar pH (material pH measured in CaCl₂ suspension), RE-biochars produced from wood or straw with 50% rock powder addition showed a higher pH than pure biochars (Table 2). Straw biochar had a higher pH than any wood (RE-)biochar. Likewise, the liming potential, given as wt% of CaO-equivalents (sum of basic compounds) was higher for straw than for wood (RE-)biochars and increased with rock powder addition, which was most pronounced in diabase-based RE-biochars, increasing from 6.8 wt% in W-P to 21.6 wt% in 50DiaW-P and from 6.4 wt% in S-P to 21.2 wt% in 50DiaS-P.

All biochars released ions during the leaching experiment, which affected the EC and pH of the leachate water (Figures 7A–C; Supplementary Table S12). Within the time horizon of the leaching experiment, the cumulative ion release (normalized to 1 kg amendment) decreased with an increasing rock content in the RE-biochar. This was mainly mirrored in a lower K release, which was 97% lower in 50BaW-P when compared to W-P (Figure 7). However, the release of Na, Mg, and Ca tended to increase for RE-biochar compared to their pure biochar counterpart. Sodium release was enhanced strongest in basanite-based RE-biochar, being almost 5 times higher in 50BaW-P compared to W-P. Magnesium and Ca release were enhanced strongest in diabase-based RE-biochar, increasing by 165 and 68%, respectively, for 50DiaW-P (Figure 7A). Potassium release was highest in S-P and straw-based RE-biochars (Figure 7B). It is worth noting that K release from non-pelletized wood biochar was reduced by 30% compared to pelleted blends, where the biomass had experienced more damage on a cellular level during pelleting (Figure 7C). The leachate EC (Figures 7D–F) and pH (Figures 7G–I) generally peaked within the first week of the experiment. Pure biochar, followed by RE-biochars employing 10% nominal rock powder addition, induced the highest spikes in leachate EC, reaching close to 10⁴ mS cm⁻¹ for S-P (conductivity of the MilliQ water supplied to the columns <5 μS cm⁻¹). Similarly, the leachate pH peaked in the first week of the experiment. The strongest pH elevations were observed for S-P-based biochars. After six weeks, the leachate from W and W-P-based RE-biochar amended columns reached a pH of 8.1 ± 0.1 and the leachate from S-P-based RE-biochar amended columns a pH of 8.4 ± 0.2, while the pH in leachate from the non-amended sand matrix remained at 6.8.

The addition of 10 wt% rock powder adds 1.85 and 1.59 wt% AAEMs from basanite and diabase, respectively, while 50% rock powder adds 9.3 and 7.9%. Catalytic effects have been demonstrated with low AAEM additive levels, e.g., >0.73% ∑[Ca, Mg, K] added as wood ash (Grafmüller et al., 2022) or >1 wt% K from potassium acetate (Mašek et al., 2019). However, for catalytic effects to occur, the AAEMs must be reactive and come in contact with pyrolyzing organic compounds (Dalluge et al., 2017), e.g., as substances that melt or decompose during pyrolysis, like sodium acetate decomposing at >324°C (Dalluge et al., 2017; Mašek et al., 2019; GESTIS, 2025). In many silicate rocks, however, AAEMs are mainly covalently bound in crystalline aluminosilicates. When considering only the water-extractable AAEMs in the rock powder, even the addition of 50% basanite resulted in the addition of only 0.03% AAEMs, i.e., one order of magnitude below the effective dose used by Grafmüller et al. (2022) (Supplementary Table S8). Also, the melting point of such igneous silicate rocks, formed from molten rock, is higher than the pyrolysis temperature (650°C). In consequence, we expect most AAEMs to remain enclosed in rock particles during pyrolysis. Thus, AAEMs embedded in aluminosilicate structures show a low catalytic activity, resulting in limited impact of rock powder addition on carbon yield and speciation, even when pelleting provided a good biomass-to-rock contact area.

Rock additives have pronouncedly different thermal properties compared to wood or straw. The heat capacity of rock is about threefold lower than that of biomass (literature data summarized in Supplementary Table S13), meaning that less energy is required to heat up the rock material to a given temperature. Further, the thermal conductivity of rock is one order of magnitude higher than that of biomass (Supplementary Table S14), facilitating heat transfer. Combined, these effects of rock enrichment can create heat bridges that increase the heating rate of organic matter adjacent to the rock material, thus likely increasing pyrolysis intensity. At the same time, energy from exothermic biomass decomposition would be distributed faster. Even after initial pyrolysis, the heat capacity of wood and straw based biochar (Huang et al., 2023) and the thermal conductivity of biochar (Patwa et al., 2022), remain below those of basanite and diabase. An increased pyrolysis severity will result in a decreased carbon yield due to enhanced decomposition, while (locally) increasing aromaticity. To the best of our knowledge, the impact of heat capacity and conductivity of biomass on pyrolysis products have not yet been systematically investigated. Kray et al. in preparation, show that the SEC of biochar produced on a PYREKA at 650°C further increased when the residence times were prolonged beyond 15 min. This highlights the potential of an increased heating rate and, thus, higher pyrolysis intensity at constant temperature due to rock-enrichment. Higher pyrolysis intensity would shape RE-biochar properties and carbon speciation when produced at residence times ≤ 15 min, as done in the present study.

Rock particles may provide additional surfaces for the formation of secondary char. It was shown for soot formation that hydrocarbon radicals condense onto an initial particle, triggering chain reactions forming solid substances form gaseous precursors (Johansson et al., 2018). Rock particles likely provide these initial sites and become coated with secondary char. This was confirmed by the density fractionation of RE-biochar (Figure 3), which showed that carbonaceous compounds are physically associated with rock particles and have a higher H:C_org ratio than the lighter, non-rock-associated biochar fraction. Given the distinct differences in the H:C_org ratio, rock particles appeared to be rather coated with secondary char than cross-contaminated with bulk biochar. Typically, (primary-)biochar is less hydrogenated and shows a lower H:C_org ratio than secondary biochar (i.e., soot) (Feng et al., 2021). Carbon was also detected on rock particle surfaces via EDX analysis (Supplementary Figure S3). An increase in SEC of co-pyrolyzed RE-biochar, compared to the SEC of equivalent post-pyrolysis mixtures of pure biochar and rock powder (Supplementary Table S15), further supports the hypothesis of rock particles being coated with (conductive) secondary char during pyrolysis. This could not be observed with particles that were only cross contaminated with carbon from the primary biochar. To the authors’ best knowledge, coating of rock particles with pyrogenic carbon is a novel observation, not yet investigated. Further research is needed to determine whether secondary char coating (probably non-polar and hydrophobic) reduces rock weathering rates due to surface inertization. The surface coating of basaltic glass and dolerite with precipitated calcite (polar compound), as observed during a batch weathering experiment by Stockmann et al. (2011, 2013), had proven to be sufficiently porous (or discrete crystalline) and did not affect the dissolution kinetics of the materials studied by Stockmann et al. (2011, 2013).

Potential adverse effects of the pyrogenic coating, regarding rock weathering rates, may also be counteracted by the thermal treatment of the rock powder during co-pyrolysis and WHC of adjacent biochar. Thermal treatment showed the potential to increase weathering rates as observed in other studies using serpentine minerals, which also comprised 6% of the here employed diabase (Supplementary Table S7; Gerdemann et al., 2007; Kelemen et al., 2011). Also, the close association of rock to the wet bulk biochar can increase the availability of water to weathering processes, especially as the RE-biochars still showed WHC in the same range as biochars obtained from pure biomass. However, the leaching experiments conducted here were too short and dominated by the dissolution of the biogenic ash to answer these questions. Leaching experiments of 6–12-month duration may provide for a disentanglement of biogenic and geogenic elemental fluxes. Follow-up experiments should systematically investigate how secondary char coating affects the dissolution kinetics of silicate rock and thus CDR from enhanced rock weathering.

The formation of secondary char depends on the retention time of (intermediate) gaseous pyrolysis products in the particle. Increased retention time increases the chance of polymerization of these gases into secondary char (Anca-Couce et al., 2014). The addition of mineral additives can increase the retention time of the pyrolysis gases by enveloping or encapsulating biomass. However, for this purpose, the mineral is applied in very high doses (sometimes >50% of the total mass of the feedstock) and in dissolved form (Rosas et al., 2009) as suspended nanoparticles (Zhao et al., 2019), or a combination of dissolved and suspended matter (Rawal et al., 2016) and then dried, so that the formation of an (almost) gas-tight shell is plausible. Such an effect has not yet been demonstrated for rock powder and seems unlikely. Still, 10BaW-P showed a much lower SSA than expected (37 vs. 137 m² g⁻¹) and 60% less pore volume than W-P (0.04 vs. 0.1 cm³ g⁻¹), whereas in non-pelletized RE-biochars (Figure 8C), the SSA was at the expected level; however, pelleting barely affected biochar SSA (197 vs. 185 m² g⁻¹ for W-P and P, respectively). This does not necessarily mean that rock dust always has a direct influence due to pore clogging. Rock dust may also indirectly affect the pelleting and the properties of the pellets as shown for additives used in industrial pelleting, such as dolomite (Tarasov et al., 2013).

Generally, pelleting technology was found to have a strong effect on particle density. The feedstock for the 1st and 2nd RE-biochar batch were pelletized on different presses (section 2.1). In the 2nd batch, feedstock and RE-biochar bulk densities increased by approximately 100% compared to first batch made with a different press (selected treatments Figures 1A,B; Table 2), resulting in similar H:C_org ratios between W-P, 10BaW-P and 50BaW-P, i.e., not repeating the increase of H:C_org in 50BaW-P from the 1st batch (Figure 2). Also, rock-enrichment led to an increased particle density (Figures 1C–E), however, not beyond expected particle densities. These observations point out that, at the given application rates, rock powder did not affect the pelleting process in a manner that increased the density of the pellets’ biogenic fraction (e.g., due to increased friction/ pressure/ matrices temperature) but only increased the total particle density due to rock powder incorporation between and around biogenic particles. It is unlikely that this phenomenon leads to a complete and airtight closure of pores, and has, thus, a limited effect on vapor retention times.

Repolymerization of secondary char may lead to the closing of micropores (< 2 nm; c.f. Figures 7D–F). According to their stoichiometry (Giudicianni et al., 2021), secondary char and condensates would both contribute to an elevated H:C_org ratio of the RE-biochar. A strong reduction in pore volume was observed for most RE-biochars and includes a complete closure of the micropore volume in 50Ba- and 50Dia- RE-biochars. This corresponds to an increase in H:C_org for 50Ba- and 50Dia- RE-biochars. The total pore volume of the RE-biochar from feedstock with 10% diabase (10DiaW-P, 10DiaS-P) was less affected, and the H:C_org ratios remained at a similar level as in pure biochar. This suggests that a nominal addition of 10% fine diabase may not modify vapor retention times as basanite does; it could be explained by diabase’s high calcite content (22% CaCO₃), which partially (29%) disintegrates during co-pyrolysis at 650°C, as shown by the TGA analysis of pure rock powder (Supplementary Figure S2). Partial obstruction of pores and consecutive formation of secondary char are not exhaustive explanations for elevated H:C_org ratios. For example, in the non-pelletized mixtures (10BaW, 50BaW, 10DiaW, 50DiaW), no decrease in SSA and pore volume beyond expected dilution effects was measured, albeit a pronounced increase in H:C_org and a decrease in thermal stability occurred for 50BaW and 50DiaW, which cannot be fully explained. Further research should also focus on the effects that altered thermal properties (section 4.1.2) may have on pyrolysis intensity and thus on H:C_org, HyPy, SEC, and BC_1000C, also how changes in porosity affect nitrate capture and nitrous oxide emission reductions in soils (Borchard et al., 2019; Kammann et al., 2015).

Co-pyrolysis did not lead to synergistic effects altering the measured y_c and thus the exploited biogenic C-Sink potential. In the present study, only additive effects could be observed. The addition of rock powder (IC-Sink potential: 0.36–0.43 t CO₂e t rock⁻¹) to biochar (PyC-Sink potential: 2.67–3.09 t CO₂e t biochar⁻¹) led to a proportional decrease in the total C-sink potential when calculated per unit mass of RE-biochar produced (cf. Table 3). However, when the calculation is reframed from the perspective of available biomass (cf. Table 4), the additive contribution of the IC-sink potential increases the total C-sink potential.

Albeit, the limited influence of rock powder addition on y_C, as observed in the present study, could be increased to a relevant level. Wet impregnation, prolonged incubation, or ultrasonic immersion (Guo et al., 2012) of biomass with rock powder could mobilize AAEM beyond the water dissolvable fraction quantified in the present study. The use of phyllosilicates, which show lower contents of AAEMs but larger surface areas (Liu et al., 2020; Lu et al., 2020; Buss et al., 2024), may support the formation of more secondary char by particle coating. Moreover, the densification of blended feedstocks could increase the retention time of volatiles in the solid matrix to promote the formation of secondary char (Anca-Couce et al., 2014). Thus, the impact of feedstock pelleting and densification on the carbon yield and biochar properties needs to be further investigated, both in the presence and absence of rock powder.