Engineering low-carbon fiber cement with biochar: understanding its physicochemical properties and their impact on the composite performance and carbon footprint†

Biochar, which is inherently heterogeneous (particle size, chemical composition, etc.), was subjected to a mechano-chemical process. A planetary ball milling device (PM 400, Retsch, Germany) was used along with stainless-steel balls (bearing size: 6.35 mm in diameter, ball/powder ratio: 20:1) for 45 minutes. The ball milled (BM) biochar (labeled as BC_BM, see Table 1 for sample composition and Tables S5–S7† for biochar physio-chemical properties) was premixed with cement; the NBSK fiber was then added to the mixing bowl at a constant water/binder ratio of 0.5. As a benchtop small-scale mixer, a kitchen-grade Techwood 6QT (power output: 800 W) was used, operating at 600 rpm for 6 minutes. The mixing process was carried out at intermittent intervals (i.e., 2 minutes). It is an important step since scraping the cement pastes that adhere to the side and bottom of the mixing bowl, and these breaks ensure the mixture is as homogeneous as possible (see Fig. S1† for detailed FC fabrication steps).

After the mixing process, the resulting FC slurry was transferred into a metallic mold of dimension 30 × 20 × 0.8 cm and placed on a vibrating microplate shaker (from Thermo Fisher, Canada) for 15 min at 300 rpm for compaction. After compaction, the FC samples were air-cured for 28 days (demolded after 7 days and the samples were placed in a plastic bag). All the sample preparations were carried out at a temperature of 23 ± 2 °C and at a relative humidity of 50 ± 4 °C. For rheological characterization, slurry state samples were utilized, and for mechanical and microstructural characterization, hardened state samples were utilized. The sample composition is illustrated in Table 1.

From Table 1, it may be noted that the total binder mass (i.e. cement + biochar) was held constant for all samples prepared for this study. The individual proportions of cement and biochar were systematically varied for each sample, but the total binder mass remained constant (i.e., when the cement content was decreased, the biochar content was increased to match the fixed binder mass). The weight percent of all other FC constituents such as NBSK fibers, water, etc. was taken relative to the fixed binder mass. While we acknowledge that replacing cement in proportions with a lower density material such as biochar can affect the solid volume fraction, the use of fixed binder mass was done to maintain a fixed water/binder ratio. This isolates the specific effect of biochar on the rheo-mechanical performance of FC. Such experimental design is commonly adopted in cement/composite research and ensures consistency and direct comparison of results.^1,2,22

Compared to the conventional laser diffraction system employed for analysing the particle size, the QICPIC + RODOS configuration provides a unique advantage of capturing real-time 2D images of individual particles, allowing for simultaneous measurement of both particle shape and size. While laser diffraction based particle analysers rely on light scattering patterns assuming a spherical shape geometry to generate particle size distribution, image-based analysis does not rely on any assumptions but rather uses fractal dimensions (based on the captured 2D image of the particle) to compute particle shape and size effectively.²⁴ This makes QICPIC + RODOS particularly useful for analysing irregularly shaped particles such as biochar, where particle morphology significantly dictates the resulting composite material performance.²⁴ In addition to this, the high speed imaging using the QICPIC system enables the detection of agglomerates and shape outliers, which laser diffraction may not be able to distinguish, and this helps us to understand the particle size distribution more effectively.²⁴

The FC slurry (supplemented with various proportions of biochar) was fed into the sample holder and was subjected to two different analyses: (a) steady state viscometry (to understand the trends in viscosity with the shear rate) and (b) strain amplitude sweep (to compute static and dynamic yield stress). Steady-state viscometry was carried out within a high-to-low shear rate regime (1000–0.01 S⁻¹) at a temperature of 25 ± 1 °C.

To calculate the static yield stress, a strain amplitude sweep was performed, ranging from 0.01% to 1000%. For the dynamic yield stress, strain amplitude measurements were conducted in reverse order, from 1000% to 0.01%. Both tests were carried out on freshly prepared samples. The yield stress for both static and dynamic measurements was defined as the shear stress at the crossover point between the elastic (G′) and viscous (G′′) moduli.

The flexural stress σ_f and strain ε_f during a three-point bending test can be calculated using the following formulas:

	(1)

	(2)

At this juncture, it is important to emphasize that three point bending tests were conducted on reduced scale specimens in accordance with ASTM C1185. Due to this, the mechanical size effect – such as that described by Bazant's size effect law (which indicates how the nominal strength of a structure changes with size, particularly in quasi-brittle materials like concrete)^26,27 – and related scaling models²⁸ are known to influence the apparent strength of fiber reinforced cement composites; its impact may however be limited in thin FC samples (8 mm thickness), particularly employed for non-structural applications (such as in this case). Nonetheless, this factor needs to be considered when interpreting flexural test results, particularly when extrapolating to full scale panel behaviour or when translating our current use case scenario to more structural applications (such as for concrete applications).

For probing the phase ID of the biochar samples, the sample as received from the vendor (in powdered form) was packed into the sample holder of the XRD machine and analyzed. For probing the phase ID of cement hydration products with the addition of biochar, the 28 day cured FC sample containing biochar was subjected to mild mechanical pulverization using a 700 W mixer from Black + Decker, USA. The powdered sample was then sieved through a USA standard testing sieve (mesh size of MS 100, ASTM E-11); this was done to maximize the probability of probing the hydration products. The sieved samples were packed in the XRD sample holder and analyzed for their hydration products.

The cost to performance analysis was carried out to evaluate the economic viability and scaling up potential of FC substituted with biochar. The performance metric of interest chosen was the specific strength of 28-day cured FC composites, obtained by dividing the mean MOR by the density of the FC sample. Cost here refers to the total fabrication cost of the FC sample – which was estimated by summing the cost of all processing steps (including raw material extraction and conversion), based on the system boundary use in LCA. Raw material costs were obtained from prior technoeconomic analysis and published market research (see Fig. S6 & Tables S13, S14† for detailed cost breakdown). The cost to performance ratio (CPR) was calculated by dividing the total FC fabrication cost by the specific strength of each sample with varied biochar loadings.

	Fig. 1 Particle size and morphological characterization of biochar. Particle size distribution, DIA particle images and SEM micrographs of biochar (a–d) before (BC_RAW) and (e–h) after ball milling (BC_BM). In (g and h), the orange arrow indicates the collapsed pore structure.

Now, to gain a deeper understanding of the morphology of biochar particles, we transitioned from the laser diffraction technique to electron microscopy. Fig. 1(c, d) and (g, h) depict the SEM images of biochar particles before and after ball milling. As seen from Fig. 1(c and d), biochar inherently consists of a porous honeycomb-like structure – when biomass undergoes pyrolysis, the organic components (such as cellulose, lignin, etc.) present in it decompose, whereas the cell structure (cell walls) often remains intact, giving rise to such structures.³¹ Due to this, the resulting biochar particles tend to possess a high surface area and pore volume, which makes them an ideal candidate as an additive/admixture for various applications.

Now, in addition to the reduction in particle size, the process of ball milling can have implications on the morphology of biochar structure. As evidenced from Fig. 1(e and f), some of the larger pores present in the biochar structure seemed to have collapsed upon ball milling (marked by the orange arrows, depicted in Fig. 1(g and h)). At this juncture, it is critical to note that high energy milling can indeed jeopardise the structural integrity of biochar samples if the milling parameters are not optimised. Therefore, upon optimization we chose 45 minutes (milling time), 20:1(ball to powder ratio) and 400 RPM as optimum milling parameters for this study, which is aligned with previous reports.³²

	Fig. 2 Pore characterization and visualization of biochar. (a) N2 adsorption–desorption isotherms and (b) BJH pore size distribution of biochar before (BC_RAW) and after ball milling (BC_BM). (c–d) Representative SEM micrographs of biochar after ball milling (BC_BM).

Additionally, understanding the surface characteristics and carbon structure of the biochar is crucial to gauge the reactivity between alkaline cement matrix and biochar.³⁴

Fig. 3(a) depicts XPS analysis of biochar surface, which is primarily composed of C (92.4 at%) and O (6.3 at%). Also, trace amounts of Al, Ca, Si and P (0.3, 0.4, 0.4, and 0.2 at%, respectively) were detected.³⁵

	Fig. 3 Surface characterization of biochar. (a) Survey and (b–d) HR-XPS spectrum of raw biochar and (e–f) Raman spectra of raw and ball milled biochar under (e) 532 nm and (f) 785 nm laser excitation.

From the deconvoluted XPS spectrum of carbon (C1s) as shown in Fig. 3(b), it can be inferred that surface chemistry of biochar is composed of unsaturated CC bond, polar C–OH and C–O–C, O–CO, CO bonds, saturated C–C and C–H bonds^35,36 and π–π satellite^35–37 in relative proportions of 79.3%, 4.2%, 1.6%, 3.0%, 6.4% and 5.5%, respectively. The carbon-carbon bonds make up the amorphous and graphitic structure of the biochar, which is chemically stable under strong alkaline condition. However, the hydroxyl and carboxyl linkages could be reactive under alkaline cement matrix. Next, from the deconvoluted spectrum of O1s as shown in Fig. 3(c), it can be inferred that O1s₁ (532.07 eV) and O1s₂ (533.46 eV) with a relative proportion of 62.9% and 37.1%, respectively were prodominanlty part of carbonyl and ester groups, respectively.^35,38–40 Fig. 3(d) presents the high resolution spectrum of Si, and further in-detail deconvolution was not pursued as the proportions of silica present on the biochar surface is low.

Since biochar is a carbonaceous material, and ball milling can induce changes in its carbon structure, Raman spectroscopy was utilized to understand the effect of mechanochemistry.⁴¹ In Fig. 3(e and f), two characteristics bands appear at around 1354 cm⁻¹ (‘D’ band) and around 1596 cm⁻¹ ‘G’ band, which are attributed to disorder/defect and sp² hybridised carbon atom associated to the crystalline graphitic layer, respectively.^42,43 The ratio of the intensities of ‘D’ and ‘G’ bands can provide a measure of the graphitization degree of the biochar samples. If the ‘I_D/I_G’ ratio is smaller, it indicates better graphitization or ordering of the stacked graphitic chains in biochar, and a higher ratio is indicative of a less ordered graphitic carbon structure.⁴¹ Two different laser excitation wavelengths of 532 nm and 785 nm were used to probe the graphitization degree. The 532 nm laser is efficient at probing the surface level characteristics of the biochar sample, whereas employing a 785 nm laser will offer better penetration depth, thereby providing better insight into the bulk level characteristics of the biochar samples. Note that from Fig. 3(e and f), for two different excitation wavelengths, the ‘I_D/I_G’ ratio is observed to be slightly smaller in magnitude for ball milled samples compared with raw biochar samples (irrespective of laser excitation wavelength). This indicates possible mechanochemical effect by ball milling process on the carbon structure of biochar.

Though graphitization degree of biochar may not directly influence the rheo-mechanical performance of the fiber cement, higher graphitized biochar may potentially have a longer shelf life in the high alkaline environment of the cement matrix, which could prolong the overall performance (durability) of the composites.

Moreover, cement hydration in presence of water is an exothermic reaction, leading to the formation of strength imparting hydration products such as calcium silicate hydrate (CSH). This is a continuous, time sensitive process where networked structures evolves in the fiber cement slurry. As such, it often complicates rheological measurements, particularly for long duration tests such as stress ramp or creep tests.^46,47 Thus, these methods, especially in systems undergoing active reactions often suffer from poor reproducibility and may negate subtle effects introduced by these additives.⁴⁷

To tackle these limitations, we adopted oscillatory strain amplitude sweep tests, a relatively short-duration method that allows us to study the viscoelastic behaviour of the FC slurry, while ensuring minimal disruptions from ongoing hydration effects.⁴⁸ This method provides insight into early-age microstructural transitions and yields both static and dynamic yield stress and provides a direct measure of thixotropy of the FC slurry. The yield point is typically identified by the crossover of G′ and G′′ — the point at which the material transitions from elastic (solid-like) to a viscous (liquid-like) behavior.⁴⁶ The corresponding shear stress is then taken as the yield stress value.⁴⁶ This method provides a direct method for computing yield stress, unlike conventionally used viscometric models (such as e.g., Bingham and Herschel–Bulkley), which are fitted to the steady state viscometry data (an indirect method) to compute yield stress viscoelastic materials.⁴⁹

Another important factor that can influence the rheological measurement, especially while dealing with suspensions with high yield stress or fibrous content is the effect of wall slip and edge fracture.^50,51 To minimise these effects, we utilised vane in cup geometry – which is particularly useful for measuring the rheological properties of highly viscous samples compared to commonly employed parallel plate or concentric cylinder geometry. However, it may be noted that the complex stress distribution in the vane geometry is different from other geometries, therefore, absolute comparison of yield stress values is only possible if the test parameters and test geometry employed are similar. Given these critical constraints in terms of model assumptions, test methods, geometric effects, thixotropy and continuous cement hydration – our focus is to understand how biochar substitution affects the rheological properties of the FC slurry, within our well defined experimental framework.⁵¹ Importantly, the results obtained by employing these well-defined protocols turned out to be consistent and reproducible, with varying biochar contents, offering valuable insights into the rheological characteristics of biochar substituted FC slurry.

As observed from Fig. 4(a), ‘τ_static’ of the FC slurry is highly dependent on the concentration of biochar. At a low content (0–2) wt%, ‘τ_static’ did not change significantly, indicating that the occurrence of yield stress is mostly concerned with the colloidal force and electrostatic interaction between the cement clinker phases as well as due to the structural build of cement hydration products (CSH network formation, see Fig. 4(a and b)),⁵² note that for computing the ‘τ_static’ of the cement paste, the sample was kept at rest after the mixing process. However, additional factors also contribute to ‘τ_static’ development in this domain (see Fig. 3(a and b)), such as the presence of coarser biochar (>50 μm) may behave as rigid inclusions within the FC slurry or may act as interlocking structures causing resistance to flow.⁵³ Additionally, the presence of NBSK fibers in the FC slurry (8 wt% in this case) would result in complex flocculation and physical entanglement (high aspect ratio fibers tend to entangle under shear) leading to agglomeration and increased resistance to flow (see Fig. 4(b)).^54,55 Therefore, a combination of these factors contributes to yield stress development (however, we believe that the role of biochar is considered minimal in this region). Now, as the biochar concentration rises from 2 to 4 wt%, the ‘τ_static’ increases rapidly (see Fig. 4(b)). We refer this concentration range as the critical biochar concentration ‘C_BC’, where an abrupt increase in ‘τ_static’ is observed (concentration range at which biochar begins to significantly influence the rheology of FC paste). This rapid increase in ‘τ_static’ could be attributed to the network formation of biochar within the FC slurry, causing a resistance to flow,⁵⁶ during the mixing process of FC slurry, some of the fine biochar particles can potentially buckle upon each other (due to possible hydrogen bonding) and form bridges connecting cement particles (electrostatic interactions), Fig. 4(b)). This could be attributed to the presence of surface functional groups in fine biochar (such as hydroxyl, carboxyl, and carbonyl groups) as evidenced through XPS (see Fig. 3(a and b)) and may promote the formation of hydrogen bonds with each other or with NBSK fibers promoting agglomeration and increasing resistance to the flow of the FC slurry.

	Fig. 4 Yield stress analysis of fiber cement slurry as a function of biochar content. (a) Variation of ‘τstatic’ & ‘τdynamic’ and ‘τstatic/τdynamic’. (b) Proposed mechanism of ‘τstatic’ evolution with biochar. Note that Ca2+, Na+, K+, and OH− depicted in the figure represents the ions present in the cement paste pore solution.

Now, at high biochar concentrations (4–10 wt%), the ‘τ_static’ manifested a steady increase, reaching a maximum yield stress of 52.07 Pa at 10 wt% (98% increase vs. control, see Fig. 4(a)). Such an observation could be attributed to the following: (a) biochar particles forming networks that bridge cement particles, thereby increasing resistance to the flow of the cement slurry (vide supra) and (b) biochar particles adhering to cement particles during mixing, potentially acting as nucleation sites for OPC hydration products to develop¹⁴ (referred to as secondary CSH, see Fig. 4(b)). Based on the data and characterization results, our hypothesis is that secondary CSH contribution is possibly minimal at low biochar dosages. Fine biochar particles (<50 μm) are more likely to contribute to nucleation effects, whereas coarser biochar (>50 μm) particles may act as rigid inclusions with limited reactivity. Therefore, the relative proportions of fine biochar in low biochar loading regime dictates the extent of secondary CSH formation. Additionally, the water absorption of NBSK fibers and biochar can tune the W/B ratio of the FC slurry, reducing the free water available in the FC and increasing the yield stress.¹⁴ Also, biochar releases free water, maintaining moisture over time, which can promote hydration.¹⁴ Thus, the aforementioned effects, possibly in-tandem contributes to the evolution of yield stress in the FC slurry marked by the steady rise at high biochar dosages. Additionally, it was observed from the Fig. 4(a) that ‘τ_dynamic’ remained largely independent of the biochar content. Perhaps, the formation of network structures (as discussed earlier) was delayed due to the continuous straining of the FC samples during oscillatory testing. As such, even at higher concentrations, the addition of biochar did not significantly affect the τ_dynamic of the cement slurry.

In this juncture, the ‘τ_static/τ_dynamic’ ratio can be used as a direct measure of time dependent rheological properties such as thixotropy/rheopexy.^44,57 A ratio greater than unity suggests that the material is thixotropic in nature, meaning the slurry becomes less viscous under stress or strain and gradually recovers its viscosity once the stress is released. It is important to note that various factors can influence this behavior, and cement pastes often transition from thixotropic to rheopexic behavior over time. However, mixing and pumping of fiber cement slurry occur during the dormant phase of cement hydration, where the slurry/paste predominantly exhibits thixotropic behavior.⁵⁸ As shown in Fig. 4(a), the ‘τ_static/τ_dynamic’ ratio for all biochar concentrations is well above 1, indicating that all FC samples exhibit thixotropic behaviour. Since ‘τ_dynamic’ remained unaffected by the biochar content, the ‘τ_static/τ_dynamic’ ratio mirrored the trend observed in the ‘τ_static/τ_dynamic’ evolution with varying biochar concentrations. Specifically, the ratio remained relatively constant at 0–2 wt% biochar and increased significantly from 2 to 10 wt%. Unique physico-chemical properties of biochar facilitate the formation of a network structure within the slurry, leading to enhanced structural build-up. Consequently, as biochar content increases, the FC samples become increasingly thixotropic, which could be indicative of the improved restructuring and breakdown/rapid deflocculation⁵⁹ (upon shear/recovery at rest) of the fiber cement slurry constituents with time. Overall, thixotropic nature and its tunability could be advantageous for applications such as pumping and consolidation, where a material that maintains its structure under stress yet flows easily when needed.

Fig. 5(a) depicts the mechanical characteristics of FC with biochar pertaining to its bending behavior, a crucial criterion for non-structural building materials, like 8 mm thick building facades. The stress–strain curves in Fig. 5(a), reveals a relative rise in peak strength with increasing biochar content, accompanied by a reduction in post-peak deflection and post-crack toughness, suggesting loss/decline in ductile behaviour and energy absorption after the failure of the FC sample (vs. control). Such behaviours are common in FC composites and a similar trend was observed in our prior studies with the inclusion of cellulosic nano/micro-particles in FC.^1,2 Such a transition could be ascribed to the high packing density of biochar (ball milled) particles, as they render a void filling effect. Densification of the cement matrix makes the FC to become stiffer (brittle) compared with the control, albeit enhancing the ability of the composite to withstand higher mechanical load and delay the initiation of microcracks.^60,61

	Fig. 5 Mechanical characterization and performance evaluation of fiber cement with biochar as a function of biochar content. (a) Stress–strain curves and the calculated modulus of rupture (MOR), (b) MOR comparison between biochar and traditional carbon materials as SCMs. In Fig. 4d, G, GO, CB, GN, and MWCNT stand for graphene, graphene oxide carbon black, graphene nanoparticles and multiwalled carbon nanotubes, respectively.62–66 More details pertaining to the composition and the type of dispersant/superplasticizer employed for these materials are mentioned in the ESI in Table S9.† All tested composites were cured for 28 d under ambient conditions unless stated otherwise.

From Fig. 5(a), it may also be noted that the control fiber cement exhibited improved strain hardening behaviour under flexural loading, characteristics of effective fiber bridging and crack resistance. It is important to note that despite the improvement of peak strength, the inclusion of even a low dosage of biochar (BC_BM-2) resulted in a significant reduction in post-peak deflection and post-crack toughness, which indicates a decrease in strain hardening behaviour. Such a drastic change in the failure response could be attributed to several reasons, such as matrix densification,⁶⁷ interfacial stress concentration from rigid/agglomerates of biochar,²² potential disruption of fiber-matrix bonding⁶¹ or to reduced fiber dispersion. Such effects are typical of fiber cement composites and has been reported with the inclusion of fine fillers, like, nanosilica⁶⁷ and nano/micro-cellulosic additives,^1,2 and matrix modifiers such as fly ash. Fly ash, in particualr, alters fracture behavior by modifying the microstructure and interfacial properties⁶¹ due to poor/heterogeneous dispersion of fine fillers/additives at extremely low or high loadings.⁶⁸ Therefore, the observed shift reflects a loss of ductility and toughness due to microstructural and interfacial interactions, rather than a direct transition of failure mode (brittle fracture).

Further analysis of the stress–strain curves indicates that there is less room for the fibers to flex under load, limiting the load transfer, and thereby compromising the ductility of the composite. Note that fibers are tightly embedded (mechanical interlocking) in between the cement–biochar matrix. So, as the biochar and cement contents are varied proportionally while keeping the NBSK fibers content constant, we believe that further optimization—particularly through adjusting fiber dosage (increasing the content of NBSK fibers) and by optimising biochar loadings—could improve the strain hardening behaviour of the composite. In this study fiber dosages were intentionally kept constant at 8 wt%, following conventional loading practices adopted in FC research.¹ To this end, our prior work has shown that it is possible to counteract matrix densification and induce flexibility by carefully optimising fiber/additive loading in FC.² Although we have demonstrated this potential, further investigation with biochar is beyond the scope of this study.

Moreover, the particle size and morphology of biochar particles significantly influence the mechanical behaviour (i.e., flexural response) and failure modes of fiber cement. Ball milling of biochar homogenized particle size and shape as shown in Fig. 1(a–h) and Fig. S2†, which possibly contributed to improvement in packing density and matrix densification. As a result, peak strength was increased upto 22.3 (vol%) of biochar addition. However, the same ball milling process may also compromise the ductility (Fig. 5(a)), as finer and more angular particles (see Fig. 1(a–h) and Fig. S2†) can interfere with fiber dispersion and reduce fiber matrix bonding. Thus, it can limit the effectiveness of the NBSK fiber in bridging cracks. So, the failure response of these materials observed a shift from ductile to a more matrix dominated brittle fracture. Note that these interactions can differ significantly from the mechanisms discussed in prior studies, which focused solely on matrix modifications, where the ductility seems to have been enhanced with the addition of biochar.⁶⁹

Fig. 5a depicts the variation of MOR with the biochar content (0–10 wt%), which is the replacment of the OPC content in the matrix. Interestingly, at a low biochar content (i.e., 4 wt%), MOR increased by 17% and the trend continued until 8 wt%, reaching a maximum of 9.8 MPa (40% increase). Interestingly, irrespective of the replacement of cement with biochar, the calculated MOR has been higher than the pristine cement matrix for all biochar concentrations (with an exception at 2 wt%) where the MOR was maintained with that of control. Because the incorporation of biochar can have an impact on the packing density¹³ while its polar surface chemistry and porous morphology can host nucleation sites favouring the ongoing hydration reactions.⁷⁰ In addition to this, the presence of NBSK fibers and biochar can retain water, which is available during the mixing of FC slurry (affecting the overall W/B ratio of FC slurry) and release it over time as the FC cures. Such a phenomenon affords unhydrated cement to hydrate over time, facilitating a mechanism termed internal curing, resulting in strength development. All these are crucial parts of the strengthening mechanism, i.e., an increase in peak strength as evidenced by the high MOR values. On the other hand, upon increasing the content of biochar to 10 wt%, the mean MOR reduced to 6.1 MPa from 9.8 MPa (54% reduction vs. BC_BM-8). Such a trend has been reported in prior research where excessive biochar in the matrix acts as a potential stress concentrator owing to the particle agglomeration,¹⁶ especially when no surfactant or dispersing agents are added in the formulation. Another possible reason for such failures, especially at high biochar dosages could be attributed to the inherent porous nature of biochar that can lead to the formation of potential air voids in the tensile plane of the cement composites compromising structural integrity as reported by Khushnood et al.¹⁸ Moreover, the evolution of ‘τ_static/τ_dynamic’ trends observed in this study was in accordance with the MOR trends with increasing biochar content (vide supra) – indicating a direct correlation with these important rheo-mechanical properties (see Fig. S4(a and b)†).

Finally, from Fig. 5(b), MOR of FC with 8 wt% of biochar (BC_BM-8) is compared against traditional carbon materials, such as carbon black (CB),⁶⁴ graphene (G),⁶² graphene oxide (GO),⁶³ graphene nanoparticles (GN)⁶⁵ and MWCNT.⁶⁶ As exhibited in Fig. 5(b), BC_BM-8 MOR is significantly higher (34%, 36% and 17%, respectively) than GO,⁶³ GN⁶⁵ and MWNT,⁶⁶ respectively. However, MOR values of cement composites with G⁶² and CB⁶⁴ were comparable with BC_BM-8. A crucial point to note is that the reported carbon materials require dispersing agents/surfactants/superplasticiser and fine/coarse aggregates to improve particle distribution, which in turn contribute to flexural strength development (see Table S9† for more details).

Fig. 6(b) depicts the PXRD patterns of reference OPC (unhydrated) and fiber cement (hydrated, 28-day cured) containing biochar (BC_BM-8). The composition (phase ID) of the cement clinker phases present in OPC is depicted in the PXRD curves represented in Fig. 6(a) and is adopted from our past study.¹ PXRD of (BC_BM-8) revealed distinct peaks at 2θ = 9° and 35° for ettringite and 2θ = 17° for portlandite, respectively, which were absent in the PXRD of OPC, confirming cement hydration.¹ Peaks at 2θ = 16°, 29°, 32°, and 49° were ascribe to the CSH, which is considered as the primary hydration product in cured fiber cement (BC_BM-8).⁷² Peak overlaps between ettringite and portlandite at 2θ = 17° and between CSH and calcite peaks at 2θ = 29° were observed.¹ Additional crystalline phases beyond conventional OPC hydration products were not detected.

	Fig. 6 Microstructural and hydration product characterization of fiber cement. (a–c) SEM micrographs of fiber cement (BC_BM-8). In (a–b), the orange arrows indicate the growth of OPC hydration products on the biochar surface, and in (c), the orage circle indicates the formation of possible hydration products at fiber-cement interface. (d–e) PXRD patterns of fiber cement (BC_BM-8) and OPC as reference. (f) Raw 29Si MAS NMR spectrum of control and fiber cement (BC_BM-8). (g) Deconvoluted 29Si MAS NMR spectrum of fiber cement (BC_BM-8). (h) Summary of spectral deconvolution results. Note, for NMR analysis, NBSK fibers were not employed to negate its effect on hydration products and 7 days curing was sufficient. The PXRD diffractogram of raw biochar is shown in Fig. S5† and the OPC reference diffractogram has been collected from our past study.1 The powder diffraction file (PDF) numbers associated with the OPC clinker phases (anhydrous state) and hydrated state have been obtained from the ICDD database, which are as follows: Alite (00-055-0738), Belite (00-033-0302), Gypsum (01-074-1905), Tricalcium aluminate (01-074-7039), Brownmillerite (01074-3674), Silica (01-071-0261), calcite (01-086-2334), portlandite (00-001-1079), and Ettringite (01-0757554).75

Visual evidence from SEM suggests the possibility of OPC hydration product (e.g., CSH) formation, localised near biochar particles – both within the cement–biochar matrix Fig. 6(a and b) and within the vicinity of biochar particles (Fig. 6(c)). However, SEM analysis alone is inadequate to characterize the complex chemical composition of OPC hydration products. The semicrystalline CSH, is also challenging via conventional PXRD technique.⁷³ Additionally, the occurrence of overlapping peaks between CSH and other hydration products creates uncertainty in deciphering the appropriate peaks with confidence, making phase quantification difficult.⁷⁴ Therefore, the effect (if any) of biochar on the structural conformation of silicates in CSH deemed to be difficult via SEM and PXRD techniques. Note that further exploration with these techniques were not explored. Hence, in such a case, solid-state NMR could be useful.

²⁹Si MAS NMR has been widely used to study the silicate species (degree of hydration and polymerization) around CSH, which is the primary OPC hydration product, as discussed in the above sections.⁷⁶ The environment around silicate species (conformation of silicates) in ²⁹Si NMR is given by Qⁿ, where ‘n’ denotes the number of Si–O–Si linkages. Q⁰ represents the isolated silicate species (silicon oxygen tetrahedra) belonging to the unhydrated cement clinker phases (C₂S, C₃S) and has a chemical shift observed in the range (−68 to −76 ppm).^77,78 It's important to note that Q⁰ comprises multiple resonances caused due to the presence of alite and belite phases in OPC, with resonance due to C₃S, occurring at a higher chemical shift compared to C₂S.⁷⁸ This combined resonance results in the broadening of the Q⁰ peaks, therefore multiple peaks (at δ = −69.1, −71.3, and −73.8 ppm) are assigned to deconvolute and quantify the Q⁰ species (see Fig. 6(g)).

Now, as OPC undergoes hydration, Q¹ (−76 to −82 ppm) and Q² (−82 to −88 ppm) species are formed at the expense of Q⁰ species.⁷⁹ Therefore, there are tentative peaks at δ = −79 assigned to Q¹, at δ = −80.8 ppm assigned to Q² (1Al) and a peak at δ = −84.7 ppm, assigned to Q² sites respectively.⁸⁰ Additionally, prior research from Anderson et al., has confirmed the resonance of Q² (1Al) sites in coreesponding to the aluminate phase in C₃A and C₄AF. In such a case, one of the two bridging oxygen atoms is bonded to Al, and thereby replacing Si in the silicate tetrahedra.⁸¹ Q³ and Q⁴ (higher polymerised silicate networks) are typically absent in OPC hydration product phases and their presence is often observed with the addition of silicate based admixtures in the cement.⁸²

Fig. 5(f) depicts the raw ²⁹Si MAS NMR spectrum of 7 day cured control and fiber cement (BC_BM-8) sample, respectively. Fig. 6(g) describes the deconvoluted spectrum of the BC_BM-8 sample. From characteristic peaks occurring at chemical shifts (vide supra) are assigned to deconvolute and quantify the different silicate species. Fig. 6(h) summarises the results from the spectral deconvolution and by quantifying the amount of silicate species, degree of silicate hydration (α) is possible to quantitate using the formula α (%) = 100 − Q⁰ (%).⁸³ It may be noted that cement curing and subsequent strength development, occur progressively at an interval of 7 days (for e.g., 7, 14, 21, and 28 days). 7 day cured samples were used for NMR analysis, as hydration at this stage is advanced enough to identify distinct silicate species, without the complexity introduced by later stage silicate polymerisation.

From Fig. 6(h), a reduction in Q⁰ species (vs. control) indicates possible increase in the degree of hydration. This means that isolated silicate species (Q⁰ species) in clinker phases converted into polymerised silicate species (Q¹, Q²) in presence of biochar in fiber cement (excluding the effects of NBSK fibers). Additionally, as shown in Fig. 5(h), amount of Q¹ species increased (64%) while amount of Q² species did not increase. This means that the incorporation of biochar possibly have altered the environment around silicate species, favouring the formation of fragmented polymerised silicate species (Q¹) compared to the formation of polymerised silicate networks (such as Q² species), especially during the early age curing (i.e.,7 days in this case). This in a way validates the visual evidence that biochar may have acted as nucleation points for additional precipitation of hydration products (referred to as secondary CSH in this case) to develop.

We ascribe this ability of biochar to its physicochemical characteristics, especially surface chemical functionality. The inference derived from the high resolution deconvoluted XPS spectrum of C and O – indicates the presence of surface functional groups (in relatively high proportions) that are capable of being reactive with the cement clinker phases in the high alkaline environment present in the cement matrix. Additionally, evidence from past studies has shown that the presence of hydroxyl and carboxyl linkages in the cement admixtures/additive can form active nucleation sites which could promote the precipitation of early CSH. This happens because the hydroxyl and carboxyl functional groups present in biochar can potentially interact with calcium ions in cement pore solution (wet state while mixing) – which prompts the calcium ions to localise near hydroxyl groups. Simultaneously, silicate ions (from dissolved clinker phases) are also drawn to the hydroxyl groups, thus facilitating a close interaction between calcium and silicate species. Overall, based on the SEM, PXRD and NMR, the key message we can infer is that presence of biochar possibly promotes the formation of poorly crystalline CSH.⁸⁴ This early precipitation of CSH then acts as anchoring points for this process to proceed further, and the corresponding binding of these CSH phases accumulates to form connected CSH networks (at the end of curing, typically after 28 days in an air cured system). Note that similar nucleation effects have been reported with other carbon based admixtures such as graphene/graphene oxide in cement systems.⁸⁵