Sonochemical Biosilica Derived from Rice Husk Ash for Cementitious Composites in 3D Concrete Printing

Abstract

The study presents an approach to the synthesis of micro- and nano-sized biosilica from rice husk ash (RHA) and describes its effective incorporation into cementitious composites for 3D concrete printing (3DCP). It is demonstrated that the calcination of rice husk at 700 °C, followed by sonochemical treatment, leads to the formation of a nanoscale silica phase with high pozzolanic reactivity. X-ray powder diffraction (XRD), infrared spectroscopy (IR), differential thermogravimetric analysis (DTG), and scanning electron microscopy (SEM) show that the incorporation of nano-biosilica (NBS) into the cementitious composites accelerates the hydration process through a nucleation effect and pozzolanic reaction. This, in turn, densifies the hardened cement microstructure and improves compressive strength significantly. Laboratory 3D concrete printing tests demonstrate that adding 1.72 wt.% NBS improves shape retention, decreases layer slump, and improves interlayer bond strength. The results indicate the viability of rice husk ash-derived biosilica as a supplementary cementitious material (SCM) in 3DCP due to its positive influence on the concrete mortar properties and parameters.

1. Introduction

Over the past 15 years, 3D concrete printing (3DCP)—a construction technology based on the layer-by-layer extrusion of specially designed cement-based mixtures using automated robotic systems—has attracted significant worldwide attention [1,2,3,4,5]. Despite the well-documented advantages of 3DCP [4,5] in terms of design freedom, labor reduction, and construction speed, durability-related concerns remain a primary challenge. In particular, the interlayer bond zone is a primary concern due to elevated porosity and, consequently, increased water absorption compared to conventionally cast concrete [6,7,8]. Additionally, the contraction and autogenous shrinkage of 3DCP mixtures, driven by their characteristically high cement content, critically affect the dimensional stability and long-term performance of printed elements [9]. In both cases, the durability and integrity of the material used must be addressed through rational microstructural engineering of the hardened paste. Therefore, the use of finely dispersed mineral modifiers such as silica fume, metakaolin, and analogous pozzolanic admixtures could upgrade the material’s properties and structure. This approach positively influences the hydration products of hardened cement phases, especially portlandite (Ca(OH)₂). Therefore, to develop durable and reliable 3DCP mixes, it is essential to upgrade the paste microstructure through the incorporation of micro- or nano-sized mineral fillers and supplementary cementitious materials (SCM), including calcined clays, fly ashes, and ground granulated blast-furnace slag [10]. Recent publications have reported the use of industrial waste that can simultaneously reduce the clinker content in extrusion-based mixes through partial cement substitution, impart favorable rheological properties to the fresh paste, and improve the mechanical performance of the hardened matrix [11,12,13].

In recent years, industrial mineral wastes have attracted considerable research interest worldwide, partly due to their potential for construction materials production. Agricultural residues such as bamboo leaf ash (more than 10 wt.% of cement replacement) [14], corn straw (5–10 wt.% cement replacement) [15], sugarcane bagasse ash (used in combination with eggshell powder as a mineral admixture in concrete) [16], sunflower stem ash (as a pozzolanic cement additive) [14], and rice husk (up to 25 wt.% cement replacement) could be effectively incorporated into various cementitious binder systems [17,18,19].

Among these wastes, rice husk (RH) merits special attention due to its global availability and the scale of its production. RH is the protective outer shell of rice grain and has several distinctive properties. From a chemical point of view, the RH composition is similar to wood and mainly contains cellulose (28–48 wt.%), lignin (12–16 wt.%), and hemicellulose (23–28 wt.%) [20]. Compared with wood, RH contains a significant amount of silicon dioxide (SiO₂), which accounts for its high thermal resistance.

Global rice production exceeds 750 million tons per year. As a result, approximately 150 million tons (20 wt.%) is generated during milling and husk removal operations [21]. The Krasnodar region is the main rice-producing region in Russia. In 2024, rice production in the region reached 862,000 metric tons, which was 10.5% higher than in 2023 [22]. According to current projections, output may increase to 1.3 million tons by 2030 due to the expansion of cultivated areas and improvements in irrigation infrastructure. This trend highlights the need for efficient utilization of rice-processing residues.

Currently, rice husk is either landfilled or openly burned. This fact leads to increased environmental pollution and hinders the effective recycling of RH. However, rice husk can be converted into rice husk ash (RHA) through controlled thermal treatment (calcining). RHA is well known to contain more than 85 wt.% amorphous SiO₂ with high pozzolanic reactivity toward the phases of hardened Portland cement. RHA is often considered a low-carbon or potentially carbon-neutral additive in the literature because the biogenic CO₂ released during combustion is partly offset by CO₂ uptake during rice growth [23,24,25].

A bibliometric analysis of over 1000 publications conducted by Amin et al. [26] demonstrated that 10–20 wt.% cement replacement by RHA improves compressive strength, durability, and microstructural densification. That review noted [26] that the pozzolanic reactivity of RHA could be compared with condensed silica fume (CSF). Both promote the formation of a denser material structure.

Amran et al. [27] and Nayak et al. [28] reported that approximately 10 wt.% of cement substitution with RHA reduces total porosity by 25–34% and retards chloride ion penetration depth by 35–37%. A reduction in drying shrinkage of up to 50% relative to control mixtures was also noted.

Hu et al. [29] established a direct dependence between RHA fineness and pozzolanic activity: ultrafine grinding to d_RHA < 5 µm leads to the greatest compressive strength gain. It was reported that RHA with a specific surface area exceeding 20,000 m²/kg exhibits a pozzolanic reaction comparable with condensed silica fume. This fact is accompanied by rheological improvements in the fresh cement paste. The reduction in portlandite content in the hardened cement paste after 28 days of curing reached 40%.

Considering the above, it is reasonable to reduce the particle size of RHA to increase its efficiency in cementitious systems. Finer RHA enhances pozzolanic reactivity, refines the phase composition and microstructure of the hardened paste, and improves the rheological properties of the fresh mortar. All these factors indicate the appropriateness of using ultrafine RHA as a component of extrusion-based mixes for 3D concrete printing applications. This research assumed that sonochemical treatment of biosilica derived from RHA could simultaneously enhance the hydration of Portland cement, refine the pore structure, and improve interlayer bonding. The scientific novelty of this work lies in a fundamentally different processing route for rice husk ash-derived nano-biosilica. In the present study, nano-biosilica was formed directly in an aqueous medium from RHA by prolonged sonochemical treatment. The resulting modifier, obtained as a stable aqueous colloidal suspension, was introduced into the cementitious system. Intermediate drying and dry grinding steps were intentionally avoided in order to minimize the agglomeration of nanosized particles and to ensure their uniform distribution within the cement matrix.

2. Materials and Methods

2.1. Materials

This study focuses on plant-based waste generated during the dehusking and milling of rice grains at agricultural enterprises in the Krasnodar region (Krasnodarzernoprodukt-Expo LLC, Krasnodar, Russia). The chemical composition of the inorganic fraction of rice husk (RH) and of the rice husk ash (RHA) produced by the calcination of RH in a muffle furnace is given in

Table 1. Chemical composition of rice husk and rice husk ash.

Composition	CaO	SiO2	Al2O3	Fe2O3	MgO	SO3	Na2O	K2O	LOI
RH	0.61	15.64	0.24	0.12	0.45	0.18	0.48	0.28	82.0
RHA	3.36	86.48	1.33	0.64	1.93	0.45	2.09	1.57	1.68

.

Ordinary additive-free cement of strength class 42.5 (CEM I 42.5 N), produced in accordance with the Russian national standard GOST 31108-2020 “Common Cements. Specifications” [30], which is harmonized with the European standard EN 197-1:2011 “Cement—Part 1: Composition, Specifications and Conformity Criteria for Common Cements” [31], was used as a binder component.

2.2. Phase Composition, Microstructural Characterization and Other Methods

The phase qualitative composition of the additives and modified hardened cement paste was examined by X-ray diffraction (XRD) using a DRON-3M diffractometer equipped with a copper anode (Cu Kα radiation, λ = 1.5406 Å). Scans were recorded in the 2θ range of 10–80° with a step size of 0.02° and a scanning rate of 4°/min, with an exposure time of 0.3 s per step. The diffraction patterns were subsequently processed and analyzed using the Origin graphical software package.

The quantitative analysis of crystalline hydrates in hardened cement was carried out by differential scanning calorimetry (DSC/DTA) and differential thermogravimetric analysis (DTG) using an STA2000 synchronous thermal analyzer (Xiang Yi Instruments, Wuhan, China). In addition, changes in sample mass as a function of temperature were recorded. Thermal analysis was performed at a heating rate of 10 °C/min in the temperature range from 30 to 1000 °C under a flowing nitrogen atmosphere (50 mL/min). The test materials were placed in corundum crucibles.

The microstructure of the materials was investigated by scanning electron microscopy (SEM) using a JEOL 1610LV microscope (JEOL, Tokyo, Japan) with a resolution of 15 nm at 1 kV, at magnifications from ×1000 to ×50,000 and an accelerating voltage of up to 20 kV. The elemental composition was determined by energy-dispersive X-ray spectroscopy (EDS) for electron probe microanalysis using an SSD X-Max Inca Energy system coupled with the SEM. EDS measurements were performed at an accelerating voltage of up to 20 kV to identify Si, O, and trace elements (K, Ca, Mg, Al, Fe, Cl) in the samples.

The qualitative identification was carried out using a Fourier transform infrared (FTIR) spectrometer equipped with an attenuated total reflectance (ATR) accessory, Nicolet 380 (Thermo Fisher Scientific Inc., Waltham, MA, USA).

The equipment was provided by the Mendeleev University of Chemical Technology Collective Use Center (Moscow, Russia).

The compressive strength of hardened cement samples was determined using an MICIS-300/10-1M (Armavir, Krasnodar Krai, Russia) testing press. The loading rate was 1.5 ± 0.125 MPa/s. After casting, the specimens were cured at a temperature of 20 ± 2 °C and a relative humidity of at least 95% until the testing age.

The true density of the hardened cement paste was determined by precision helium pycnometry using an AccuPyc 1340 Automatic Helium Pycnometer (Micromeritics Instrument Corp., Norcross, GA, USA), with an accuracy of ±0.0001 g/cm³.

2.3. Determination of the Additives’ Pozzolanic Activity by Calcium Hydroxide Uptake

The pozzolanic activity of the RHA-based additives was evaluated from the amount of calcium oxide (CaO) consumed from a saturated calcium hydroxide (Ca(OH)₂) water solution. Two procedures were used to determine the pozzolanic activity: an accelerated method according to GOST R 56593-2015 “Mineral additives for concretes and mortars. Test methods” [32] and the classical Butt’s method [33]. Both methods quantify the amount of CaO removed from a saturated Ca(OH)₂ solution by the active tested mineral additives. In Butt’s method, the experiment is carried out for 30 days, whereas in the accelerated method, the test duration is 2 days.

Before the tests are carried out, 5 L of saturated Ca(OH)₂ water solution was prepared; the solution may be used only if the CaO concentration is greater than 1.05 g per liter. In Butt’s procedure, 2 g of the test material are placed into a graduated cylinder and 100 mL of saturated Ca(OH)₂ solution is added. Then, the cylinder is tightly sealed, and the suspension is kept in the dark for 2 days. After that, 50 mL of the solution is titrated with 0.05 N hydrochloric acid (HCl), and the results are recorded. The cylinder containing the rice husk ash (RHA) or rice husk (RH) is replenished with 50 mL of Ca(OH)₂ solution, shaken, and allowed to equilibrate again. The same procedure is repeated every 2 days over a period of 30 days, and then all partial results are summed and the pozzolanic activity of the mineral additive is calculated.

The accelerated method reduces the test duration by preheating the saturated Ca(OH)₂ water solution containing the additive at 85–90 °C for 8 h, followed by cooling and titration with aqueous HCl.

The amount of CaO absorbed per gram of additive was calculated according to the equation:

$$\mathrm{CaO} = 1.4·2·{\displaystyle \frac{V_0·(V_2-V_1)}{V_2}},$$

(1)

where 1.4 is the titer of a 0.05 mol/L HCl solution with respect to CaO (differential correction coefficient), mg·mL⁻¹·g⁻¹;

V₀—the volume of HCl consumed for the titration of 50 mL of saturated Ca(OH)₂ solution without the mineral additive that was tested, mL. This sample was not heated in the drying oven (“cold” solution);

V₁—the volume of HCl consumed for the titration of 50 mL of the test solution containing the tested mineral additive, mL;

V₂—the volume of HCl consumed for the titration of 50 mL of saturated Ca(OH)₂ solution without the mineral additive, mL. This sample was heated in the drying oven, as were the samples of additives tested (“hot” solution).

Titration procedures were carried out in the presence of methyl orange indicator until the solution color changed to a pale pink.

2.4. Rice Husk Ash Preparation and Sonochemical Treatment

Rice husk ash (RHA) was produced by the calcination of raw rice husk in a laboratory muffle furnace at temperatures of 600–800 °C, followed by rapid cooling in air. The heating rate was 5 °C/min, and the holding time at the maximum temperature was 1 h.

To obtain micro- or nanosized RHA particles, a sonochemical treatment in an aqueous medium was implemented using an ultrasonic magnetostrictive disperser. The titanium horn (waveguide) of the disperser operated in a reciprocating mode at a frequency of 30 kHz with an oscillation amplitude of 0.1 µm. At this amplitude, the power transmitted to the processed dispersion was 616 W, while the device consumed 1500 W from the mains. The device was powered by alternating current at 220 V, with a nominal frequency of 50 Hz, and the ultrasonic power density was approximately 0.7 W/cm³.

The mass fraction of RHA in the final aqueous suspension was approximately 10 wt.% (Figure 1a). The total sonochemical treatment time for an 800 mL dispersion of 77 g RHA in distilled water was 50 min (10 cycles of 5 min), in order to avoid overheating and prevent particle agglomeration caused by elevated processing temperatures (not higher than 60 °C). The RHA grinding products consisted of a colloidal suspension (Figure 1a) and a sediment consisting of larger settled particles (Figure 1c).

The resulting colloidal suspension was used directly in the same aqueous medium in which the sonochemical treatment had been carried out (without subsequent drying). This method of preparing and using nano-biosilica, which eliminates intermediate drying steps, prevents the agglomeration of even the finest nanoparticles and thereby promotes the most efficient utilization of the nanomaterial’s reactive potential.

The biosilica fraction in the colloidal suspension was determined using sedimentation analysis with a torsion balance, according to the Stokes law-based estimation of the sedimentation cut-off size. According to this equation, which is applicable for estimating the settling velocity of micron- and submicron-sized particles, the settling time of SiO₂ particles in water is about 12 h for particles with a diameter of 0.5 μm and about 3 days for particles with a diameter of 0.2 μm. Given the idealized nature of this calculation, which does not account for particle shape, agglomeration, or other factors affecting sedimentation behavior, the settling period for particles of about 0.2 μm was extended to 7 days. After this period, the particles remaining suspended in the dispersion may be operationally regarded as a fine fraction expected to consist predominantly of particles smaller than approximately 200 nm and were therefore used in the subsequent stage of the study.

The average particle size of the dry solids (micro-biosilica) was 1.7 µm. The ratio of the nanoparticle fraction remaining suspended in the colloidal solution to the micro-biosilica sediment was 30/70 on a dry matter basis.

The sedimentation method was used to estimate the operational sedimentation cut-off and to distinguish the coarse settling fraction from the sedimentation-stable fine fraction in the dispersion medium, whereas SEM analysis was employed to examine particle morphology and to confirm the presence of nanosized particles in the fine fraction.

2.5. Laboratory 3DCP. Printing Parameters and Experimental Methods

The final stage of the research was carried out using a 3D concrete printer under laboratory conditions, with a throughput of up to 25 kg/min of dry mix. The time interval between successive layers was 10 min through the printing process. Printing was performed by layer-by-layer extrusion at a speed of 50 mm/s using a circular nozzle with a diameter of 45 mm, with a positioning accuracy of ±1 mm.

The slump resistance of extruded cement-based materials was evaluated using a method based on GOST R 59096-2020 “Materials for additive manufacturing in construction. Test methods” (Russian Standard), adapted to the laboratory conditions of the present study. No independent rheometer-based measurements of yield stress, plastic viscosity, or thixotropy were performed. The extruded material was deposited onto a steel sheet wetted with a damp cloth, which was positioned under the print nozzle of the 3D printer, and layer upon layer was printed. After sixty seconds, the layer width was measured with a ruler at three points spaced evenly along the printed layer. The arithmetic mean of these measurements was then calculated and compared with the nominal layer width; the deviation was required not to exceed 5%.

Interlayer bond strength was determined using an ONIKS-1.AP adhesion tester (Russia). In this study, the ONIKS-1.AP.020 modification was used, with a load range of 3–20 kN corresponding to a bond strength range of 0.1–40 MPa and a relative load-measurement error not exceeding ±2%. Steel loading discs were bonded to the surface of the printed specimens with a two-component epoxy adhesive; after the adhesive had cured, the discs were loaded in normal tension until failure, and the maximum pull-off force was recorded. The bond strength was calculated as the ratio of the pull-off force to the area of the steel disc.

3. Results and Discussion

3.1. Phase Composition of Rice Husk and the Products of Its Calcination

As previously mentioned, rice husk consists of cellulose (C₆H₁₀O₅)_n, lignin (C₃₁H₃₄O₁₁)_n, hemicellulose (pentosans (C₅H₈O₄)_n and hexosans (C₆H₁₀O₅)_n), and silicon dioxide (SiO₂). Inorganic silica in rice husk (approximately 20 wt.%) is presumed to be present predominantly in an amorphous, chemically reactive form, which distinguishes this residue from many conventional mineral additives, in which silica occurs mainly as crystalline quartz or cristobalite. RH exhibits a textured, micro-relief surface, which is visible at 100× and 250× magnification (Figure 2a and Figure 2b, respectively).

Another similar type of rice crop residue is rice straw, i.e., the dry stalks and leaves of the rice plants remaining in the field after harvesting and threshing. The SiO₂ content in rice straw is slightly lower; however, rice straw differs fundamentally from rice husk in its chemical composition due to its elevated chlorine content (associated with plant physiology), which may constrain the use of this residue in concrete mixtures (Figure 2c) from a chemical point of view.

Peng et al. [34] performed a detailed investigation of the fate of chlorine during the pyrolysis of rice straw. It was shown that, already at 200–400 °C, the fraction of water-soluble Cl in the char decreases from 60.3% to 40.1% due to the release of CH₃Cl and HCl, whereas at 800 °C and above, chlorine is almost completely volatilized. These findings are corroborated by Igwebike-Ossi [35], who concluded that the use of rice husk is preferable because traces of chlorine in the husk are either absent or do not exceed 0.026%, i.e., several times lower than in straw ash. Similar values were obtained in the present study. The Cl content in rice straw was significantly higher (9.46 ± 0.52 wt.%) than in rice husk (≈0 wt.%). Thus, RH was selected as the target material.

It is known that the organic constituents of RH, such as cellulose and hemicellulose, are completely decomposed at temperatures up to 500 °C, whereas lignin degradation proceeds up to 700 °C and higher, leaving a substantial carbonaceous residue [36].

It is also widely accepted that any residual organic impurities in RHA are highly undesirable because of the strongly alkaline environment of the pore solution in hydrating cement (pH > 13). For example, lignin, a high-molecular-weight aromatic polymer with a three-dimensional structure, dissolves in alkaline solutions as a result of the ionization of phenolic hydroxyl groups [36,37] by Equation (2):

$$\mathrm{Ar}–\mathrm{OH} + {\mathrm{OH}}^{-} \stackrel{\mathrm{p}\mathrm{H}>10}{\to } \mathrm{Ar}–{\mathrm{O}}^{-} + {\mathrm{H}}_2\mathrm{O}$$

(2)

where Ar represents an aromatic hydrocarbon moiety, most commonly a benzene ring (C₆H₆) or a substituted benzene derivative. Cellulose, which is also present in RH, is a polysaccharide that does not contain phenolic groups available for ionization; it is essentially insoluble in water and alkaline solutions, but it swells under highly alkaline conditions and loses its initial properties. Hemicellulose behaves in a similar way under these conditions, although its more complex molecular structure makes it more soluble in a strongly alkaline environment. Therefore, a preliminary treatment of RH to remove its organic components is required [38]. Otherwise, the performance of cement-based materials and products could be significantly impaired. Consequently, RH was calcined at 700 °C.

XRD analysis of RH before calcination confirmed the presence of amorphous silica (Figure 3a). Amorphous SiO₂ is characterized by a broad diffuse halo in the 2θ range of approximately 20–24° [38]. Similarly, RHA samples obtained from RH calcination in a muffle furnace at 700 °C were examined (Figure 3b). The same halo of amorphous silica is also present in the latter diffractogram. Moreover, the halo increases with area. This observation appears to suggest that RHA has the highest chemical reactivity and is essentially free of organic impurities, in contrast to RH.

To verify that the organic fraction had been completely removed from the RH after calcination, both compositions (RH and RHA) were characterized by IR spectroscopy (Figure 4). In the RH spectrum (Figure 4a), characteristic absorption bands of methyl (CH₃₋) and methylene (CH₂) groups are observed. The absorption bands at 2920.15 cm⁻¹ and 1421.42 cm⁻¹ are assigned to CH₂ vibrations, whereas the band at 2850.99 cm⁻¹ corresponds to CH₃ groups, in agreement with the assignments reported by Yang et al. [39]. A band at 1649.20 cm⁻¹ is attributed to a carbonyl (C=O) group of organic origin, associated with lignin and cellulose. Vibrations characteristic of silica appear at 1082.49, 798.55, and 470.19 cm⁻¹. A broad band at 3410.56 cm⁻¹ arises from hydrogen-bonded OH groups, whereas the bands at 1515.31 and 1376.37 cm⁻¹ are associated with non-hydrogen-bonded (free) OH groups or the aromatic skeletal vibration of lignin [40,41].

After calcination (Figure 4b), all organic absorption bands disappear in RHA. The spectrum retains only an O–H stretching band at 3434.07 cm⁻¹ and three intense Si–O bands at 1094.69, 800.12, and 468.06 cm⁻¹. This change in the IR response indicates that the obtained RHA is essentially free of organic impurities and can be used as a highly reactive modifier in cementitious systems.

3.2. Structure of Rice Husk Ash as a Function of the Treatment Conditions

To facilitate a more effective introduction of RHA into the cement binder, it was subjected to sonochemical treatment in water. As a result of this treatment, two types of biogenic silica-based products were obtained from the initial RHA fractions (in this paper, the prefix “bio-“ is used to indicate that the silica originates from plant-derived ash):

• A fine precipitate (micro-biosilica) with an average particle size between 1 and 10 µm (MBS) and a specific surface area (SSA) of 20,568 cm²/g (Figure 5a,b).
• A relatively stable colloidal suspension consisting of nano-biosilica (NBS), as observed in SEM images of the suspension filtrate (dry solids content of approximately 2.9–3 wt.% (Figure 5c)).

The SEM image of the fine precipitate obtained from RHA at 250× magnification reveals the presence of relatively large particles (Figure 5a). Some of them are agglomerated into clusters up to 35 µm in size (Figure 5b). The particles are primarily spherical within a 5–15 µm size range. At 5000× magnification, small agglomerated domains of nano-biosilica can be found on the surface of the larger particles (Figure 5b).

After drying the colloidal suspension, a solid precipitate is formed, which was examined by SEM (Figure 5c,d). The ultrafine particles produced by sonochemical treatment are found to be predominantly rounded (spherical or sub-spherical), with apparent sizes of approximately >200 nm, confirming the nanoscale nature of the material. In this case, the product can be referred to as nano-biosilica (NBS). However, the drying process leads to the agglomeration of the ultrafine particles, resulting in aggregates larger than 2 µm and an increase in the apparent particle size (Figure 5d). Even so, the nanoscale primary particle size is supported by the observed sedimentation behavior of the suspension, both under quiescent conditions and under sonochemical treatment.

3.3. Pozzolanic Activity of Rice Husk Ash Depends on Its Particle-Size Fraction and Treatment Method

The pozzolanic activity of RH and RHA arises from the acid–base reaction between silicic acid H₄SiO₄, released by reactive amorphous silica, and calcium hydroxide Ca(OH)₂. This action leads to the formation of calcium silicate (C–S–H) hydrates by Equation (3). The rate of this reaction depends on the specific surface area of the SiO₂ particles, the degree of its amorphization, and the pH level of the pore solution (particularly, on the OH⁻ concentration, which catalyzes the breaking of Si–O–Si bonds).

At the early stage of cement hydration, after water addition, the pore solution becomes saturated with Ca(OH)₂ within about 2 h, and a maximum supersaturation of around 1.7 g CaO/L is reached after several hours [42,43]. In this highly alkaline environment (pH 12.5–13.5, typical of cement paste), the solubility of amorphous silica is much higher than under neutral conditions.

At high pH levels, silicic acid dissociates and the aqueous speciation is dominated by H₃SiO₄⁻ ions (with a smaller fraction of H₂SiO₄²⁻), which rapidly react with Ca²⁺, as shown in Equation (3), to form low-basicity calcium hydrosilicates of the C–S–H (I) type. The chemical processes occurring in the cement paste at this stage can be divided into several steps by the equations:

$${\mathrm{SiO}}_{2\left(\mathrm{amorphous}\right)} + 2{\mathrm{H}}_2\mathrm{O} \stackrel{\mathrm{p}\mathrm{H}>13.5}{\to } {\mathrm{H}}_4{\mathrm{SiO}}_4 \stackrel{\mathrm{rapid}}{\leftrightarrow } {\mathrm{H}}_3{{\mathrm{SiO}}_4}^{-} +{\mathrm{H}}^{+} \stackrel{\mathrm{p}\mathrm{H}>12.5}{\leftrightarrow }{\mathrm{H}}_2{{\mathrm{SiO}}_4}^{2-} +{\mathrm{H}}^{+}$$

(3)

$${\mathrm{H}}_2{{\mathrm{SiO}}_4}^{2-}+\mathrm{Ca}{2}^{+}+{\mathrm{H}}_2\mathrm{O} \to \mathrm{C}–\mathrm{S}–\mathrm{H} \left(\mathrm{I}\right)$$

(4)

It is highly likely that the remaining amorphous silica interacts with Ca(OH)₂ during the later stages of hardening.

Rice husk waste exhibits significant pozzolanic activity (Figure 6). RH is characterized by an activity of 36.5 mg/(g of additive) according to the accelerated determination method and 395.2 mg/(g of additive) according to Butt’s method. The activity of RHA after calcination reaches 45.2 mg/(g of additive) and 420.6 mg/(g of additive) using similar methods, respectively, which is approximately 25% higher than the corresponding values for the initial RH. The difference in the activity of nano-biosilica (NBS) and micro-biosilica (MBS) is not significant, likely due to the limited stability of the sol containing NBS and its gradual agglomeration after the pre-drying process. As a result, particle growth is observed, leading to a reduction in the amount of CaO absorbed from the saturated solution during the testing process. So, MBS/NBS pozzolanic activity was 52.1 mg/(g of additive) according to the accelerated method and 480.6 mg/(g of additive) according to Butt’s method.

These results are consistent with the IR data, which show that the intensity of the absorption bands in RHA (after calcination) is higher relative to the initial RH. Further sonochemical treatment of RHA increases the pozzolanic activity of both micro-biosilica (MBS) and nano-biosilica (NBS). Considering the pozzolanic activity, particle size factor and the influence on hydration via the nucleation effect, NBS was selected for further investigation in the cement–additive system.

3.4. Properties of Cement Containing Nano-Biosilica Additives

To investigate the effect of varying NBS content on the properties and structure of cement paste, cubic specimens with a side length of 2 cm were prepared at a W/C ratio of 0.4 (

Table 2. Test compounds and their properties.

Composition	NBS, wt.%	W/C	Compressive Strength 1 Day, MPa *	Compressive Strength 28 Day, MPa *	Real Density, g/cm3
Control mix (CM)	0	0.4	12.9 ± 0.6	37.8 ± 1.5	2.2244 ± 0.0001
Mix 1 (M1)	0.44		13.6 ± 0.7	45.1 ± 1.9	2.3230 ± 0.0001
Mix 2 (M2)	0.88		16.5 ± 0.6	53.8 ± 2.5	2.3860 ± 0.0001
Mix 3 (M3)	1.3		18.5 ± 1.5	64.8 ± 2.6	2.4464 ± 0.0001
Mix 4 (M4)	1.72		23.6 ± 2.1	78.0 ± 2.8	2.4928 ± 0.0001

* The compressive strength values are presented together with the standard deviation (S) calculated from the test results of four specimens for each mix at the corresponding curing ages. The coefficient of variation of the test results for all investigated mixes did not exceed 8.7% at 1 day and 4.9% at 28 days of hardening, confirming acceptable experimental variability.

).

It was observed that the strength of the cement paste increases monotonically with increasing NBS content, both at the early stage and after 28 days of hardening. At an NBS content of 0.44 wt.%, a slight increase in strength is observed after 1 day of hardening (13.6 MPa for M1 versus 12.9 MPa for the control mix). The effect becomes more pronounced after 28 days (45.1 MPa versus 37.8 MPa, respectively), indicating accelerated hardening and a higher degree of hydration.

A further increase in the NBS content from 0.88 to 1.3% enhances this effect, raising the 1-day strength to 16.5 and 18.5 MPa and the 28-day strength to 53.8 and 64.8 MPa, respectively. This corresponds to a 40–70% increase in strength relative to the control mix. The maximum effect is observed at 1.72% NBS, where the compressive strength at both early and later stages of hardening is nearly twice that of the control mixture. The observed increase in strength is attributed to the densification of the hardened cement paste, which was approximately 10% denser than the control mix (2.4928 g/cm³ versus 2.2244 g/m³). The densifying effect of NBS is probably related primarily to its nucleation, filler, and pozzolanic action, leading to additional C–S–H formation, which has higher real density. These changes are significant and most likely result from quantitative and qualitative modifications at the microstructural level, which were investigated in the subsequent stage of the study.

Despite the superior strength performance of the mixture containing 1.72% NBS, further increasing the NBS content beyond this level appears impractical for several reasons. These include a marked decrease in the pH of the cement paste due to the high pozzolanic activity of NBS, as well as severe particle agglomeration and aggregation after sonochemical treatment is stopped, followed by the sedimentation of agglomerates. This hinders the uniform distribution of NBS particles within the cement matrix. The colloidal suspension obtained by sonochemical treatment contains suspended NBS particles with sizes ranging up to 200 nm. Therefore, it is reasonable to use this dispersion after sonochemical treatment as the mixing water for the binder in order to avoid nanoparticle agglomeration during drying. The water content of the suspension during sonochemical treatment was adjusted to ensure subsequent specimen preparation at a water-to-cement ratio of 0.4.

The substantial increase in the strength of the modified compositions is primarily attributed to the nucleation effect. Owing to their high specific surface area, nano-SiO₂ particles provide additional nucleation sites for the hydration products of the original cement phases (particularly C–S–H gel). They also fill micro- and nanoscale voids, thereby densifying the structure of the cement matrix. Together with the pozzolanic reaction, this increases the number of nucleation sites for hydrate phases, shortens the induction period of hydration, and accelerates C–S–H growth.

When two cement pastes with opposite strength characteristics were examined by the SEM method—the control mix (Figure 7a,c,e) and the paste containing 1.72% NBS (Figure 7b,d,f)—noticeably different microstructures were observed.

At 1000× magnification, the matrix of the control mix paste appears relatively loose (Figure 7a): individual large hydrate crystals and unhydrated cement phases with voids between them are clearly visible. The microstructure is dominated by crystalline phases, whereas the C–S–H gel is distributed unevenly and forms isolated clusters (Figure 7c,e), resulting in a well-developed system of connected pores and microcracks. Morphologically, the calcium silicate hydrate in the control paste is most often observed as loose flocculent accumulations and individual fine fibrous formations, locally needle-like in shape (Figure 7e), which do not form a continuous interconnected framework throughout the volume. In some areas, the structure also contains clusters of prismatic ettringite crystals up to 2 µm in length.

The XRD patterns of the compositions CM, M2 and M4 show the phase assemblage typical of hydrated Portland cement, including portlandite (CH), anhydrous alite and belite (C₃S/C₂S) phases, hydroaluminate phases (AF_t, C₄AH₁₃, C₃AH₆, and AF_m), and related hydration products (Figure 8). In all samples, intense CH reflections are clearly observed at 2θ ≈ 18° and 34°, whereas the main reflections of clinker and hydrate phases appear in the 2θ range of about 29–30° and 32–34°.

The CH peak at 17.9° is the most informative for portlandite because it is the principal reflection with the highest intensity. Comparative analysis of the three diffractograms shows that, for the composition containing 1.72% NBS (Figure 8c), the intensity of the portlandite reflections decreases by 22.6% relative to the control mix (Figure 8a). At the same time, the intensity of the C₃S/C₂S peaks at 32.1°, 32.5°, 41.2°, and 51.6° decreases by 26–35% (depending on NBS dosage). This indicates a substantially higher degree of cement hydration in the M2 and M4 mixtures. NBS acts as a nucleation center for X-ray amorphous C–S–H gel formation, thereby accelerating the hydration of clinker phases. At an NBS content of 0.88% (Figure 8b), the decrease in the intensity of the anhydrous clinker phases is only 5–8%, while the portlandite peaks at 17.9° and 50.9° increase by about 10%. Both effects fall within the uncertainty associated with comparing absolute peak intensities across different diffractograms, where a typical error of about ± 15% may be expected.

The data obtained by XRD do not allow the C–S–H content in the investigated compositions to be quantified directly, either from the background or from the amorphous halo. The difference in background intensity between the compositions (0–7%) is within the experimental error. Because of the poorly crystalline, effectively X-ray amorphous nature of C–S–H, this phase is barely discernible in the diffractograms, and any increase in its content in the presence of the modifier can only be inferred indirectly from the decrease in the intensity of the reflections of the original anhydrous crystalline phases. Therefore, at the next stage, DTA (DSC) and DTG were performed to obtain a more accurate quantitative assessment of the phase composition of the mixes (Figure 9).

The release of water from C–S–H (II), with a CaO/SiO₂ ratio of about 1.5–2, proceeds stepwise and is most intense in the temperature range from 105 to 400 °C. Thermal effects and mass losses in this range provide direct evidence of the amount of high-basicity calcium silicate hydrates formed in the system. In turn, the low-basicity C–S–H (I) modifications, with a CaO/SiO₂ ratio below 1.5, are less stable and lose chemically bound water more gradually over the range from 20 to 500 °C.

Both NBS compositions contain approximately 33% more hydrate-bound water than the control mix in the 105–400 °C range, taking into account that all samples were dried before testing. This confirms a substantially higher degree of hydration due to the nucleation effect, as well as the additional formation of C–S–H through the pozzolanic reaction. In the control cement mix, the first DTG peak in this temperature range is moderate in amplitude and corresponds to the removal of capillary water, the dehydration of AF_t (ettringite), and part of the C–S–H. For the composition containing 0.88% NBS, the amplitude of this peak increases markedly, indicating a higher fraction of hydrate phases, primarily C–S–H (I), containing weakly bound water. For the composition with 1.72% NBS, the 0–400 °C peak is even more obvious, and the total mass loss in this range reaches a maximum of 15.95%, compared with 9.35% for the control mix.

As previously shown by the authors, the addition of 1.72% NBS resulted in the strength of the hardened cement paste being nearly twice that of the control mix. The observed effect is attributed not only to the pozzolanic activity of the additive, but also to the structure-forming influence of NBS on the morphology of the hydrate phases, particularly to the increased content of the high-strength fine-crystalline C–S–H phase. This conclusion is supported by the XRD, DTA, and SEM results.

Unlike most published studies, where the addition of nano-SiO₂ to cementitious materials at 28 days results in reduced or comparable portlandite contents, in the present case DTG shows a moderate increase in CH at NBS dosages of 0.88–1.72%, together with a substantial increase in C–S–H. These findings are consistent with the XRD results, which clearly show a decrease in the intensity of the alite and belite peaks by more than 30%, indicating an increase in their hydration products, namely C–S–H and portlandite.

This effect may be related to the fact that, at relatively low NBS dosages, the nucleation effect and the acceleration of clinker-phase hydration prevail over the pozzolanic reaction, while the stoichiometric capacity of nano-biosilica to consume Ca(OH)₂ remains limited because of its low content in the cement. To assess the potential of the applied NBS dosages to participate in the pozzolanic reaction with Ca(OH)₂, a simplified stoichiometric calculation was performed using the following equation:

$$\mathrm{Ca}(\mathrm{OH}{)}_2 + {\mathrm{SiO}}_2 + 2{\mathrm{H}}_2\mathrm{O} \to \mathrm{Ca}{\mathrm{H}}_2{\mathrm{SiO}}_4·2{\mathrm{H}}_2\mathrm{O}$$

(5)

According to Equation (5), 1 mol of SiO₂ consumes 1 mol of Ca(OH)₂. Based on the molar masses of these phases, conversion to a mass ratio shows that 1 g of SiO₂ can theoretically bind up to about 1.23 g of Ca(OH)₂. Accordingly, per 100 g of binder, an NBS dosage of 0.88% theoretically reacts with 1.08 g of CH, whereas a dosage of 1.72% NBS theoretically reacts with 2.12 g of CH.

According to the DTG data, the CH content in the control cement mix (CM) was 17.7 wt.%, while in the compositions containing 0.88% and 1.72% NBS, it was 19.3% and 19.6%, respectively (that is, about 1.6–1.9% higher than in the control mix). This is observed even under the assumption of complete binding and 100% participation of silica in the reaction, which is not attainable under real conditions.

Thus, the simplified calculation shows that the nucleation effect prevails over the pozzolanic effect when NBS is used at dosages below 1.72%, which is consistent with the experimental data showing an increased portlandite content in the system according to DTG. The accelerated formation of low-basic C–S–H around NBS reduces the fraction of large capillary pores and refines the pore size distribution, thereby increasing the density of the hardened material and improving the interlayer bond.

3.5. Properties of Cement–Sand Mortar Based on a Binder Modified with Nano-Biosilica

To verify the laboratory findings, a trial 3D printing was carried out using cement–sand mortars prepared with the control mix binder and with the binder modified by 1.72% NBS. The mortars were designed as a fine-grained cementitious mix and consisted of Portland cement CEM I 42.5 at a dosage of 505 kg and quartz sand with a fineness modulus of 2.1 at 1515 kg. The rheological properties were adjusted using the polycarboxylate-based superplasticizer Melflux 1641F at 2.97 kg, while mix stability and water retention were improved by the cellulose ether Mecellose FMC 22501 at 1.91 kg. In addition, 148.5 kg of a colloidal nano-biosilica solution was incorporated as a nanomodifying component (for NBS mortar only), and 120 kg of tap water was used as the mixing water. The values are given with reference to a volume of 1 m³ of cement mortar.

Measurement of the geometric parameters of the extruded layers showed that the layer width of the control composition was 58 mm (Figure 10b,d) after 60 s of layer-upon-layer printing, whereas the composition containing 1.72% NBS produced a layer width of 53.5 mm (Figure 10a,c), i.e., 8.5% lower than that of the control mix. The nominal layer width of the NBS-containing mortar increased only from 53.0 to 53.5 mm after 60 s (

Table 3. Test mortars and their properties.

Composition	Nominal Layer Width After Extrusion, mm	Layer Width After 60 s, mm (Two Layers)	Interlayer Bond Strength After 28 Days of Curing, MPa
Control mortar	55.5	58.0	1.8 ± 0.04
NBS mortar	53.0	53.5	2.1 ± 0.04

), whereas the control mortar increased from 55.5 to 58.0 mm. This result indicates a lower tendency to slump and better shape retention of the deposited filaments under the adopted test conditions.

In addition, the interlayer bond strength increased from 1.8 ± 0.04 to 2.1 ± 0.04 MPa (Table 3), corresponding to an improvement of approximately 17%, which is correlated with the microstructural observations discussed in Section 3.4.

Since no dedicated rheological measurements were performed, these results should be interpreted as indirect evidence of enhanced structural build-up rather than as a direct quantification of yield stress or thixotropy. As noted above, nano-silica intensifies particle flocculation and accelerates C–S–H formation, leading to a faster recovery of the material structure after extrusion and, consequently, to better retention of the geometry of the deposited layer.

Visual assessment of the print quality revealed substantial differences between the two compositions. The layers printed with NBS exhibited a more coherent and uniform surface, clearer boundaries between adjacent layers, and no pronounced extrusion defects (Figure 10c). In contrast, the control composition showed greater slumping in some areas and less stable cross-sectional geometry during printing (Figure 10d). These observations are consistent with the results reported by Liu et al. [44], who found that nanostructured silica improves shape retention in 3D printing and also improves interlayer bonding. Therefore, the mixture can be considered operationally printable under the selected process parameters, although pumpability was not quantified in a separate standardized test. Similar findings were reported by Slavcheva et al. [45], who stated that cement systems modified with nanosized SiO₂ particles increased the plastic strength by 4–5 times.

4. Conclusions

• This study demonstrates an integrated approach that combines sonochemical processing of rice husk ash with multiscale characterization and laboratory-scale 3D concrete printing tests, and provides direct evidence that nano-biosilica improves both hydration kinetics and the structure of hardened material for 3DCP.
• Thermal activation of rice husk by calcination at 700 °C, followed by sonochemical treatment, makes it possible to obtain both a fine micro-biosilica precipitate (1–10 µm) and a colloidal suspension containing nano-biosilica (<200 nm), as confirmed by sedimentation analysis with a torsion balance and SEM analysis.
• The pozzolanic activity of RHA after calcination increases by approximately 25% compared with RH (36.5 mg/g of additive by the accelerated method and 395.2 mg/g of additive by Butt’s method for RH, whereas 45.2 mg/g of additive and 420.2 mg/g of additive, respectively, for RHA). At the same time, NBS and MBS activities after sonochemical treatment were similar but higher by approximately 15% compared with RHA (52.1 mg/(g of additive) and 480.6 mg/(g of additive) using similar methods).
• The incorporation of nano-biosilica in cement at dosages of 0.44–1.72 wt.% provides a monotonic increase in the strength of the hardened cement paste at both early and later curing ages. At a dosage of 1.72%, the 28-day strength nearly doubles relative to the control mix, reaching 78.0 MPa versus 37.8 MPa.
• Combined XRD and DTG/DTA analyses showed that NBS both accelerates the hydration of clinker phases through a nucleation effect and participates in the pozzolanic reaction, resulting in an increased amount of low-basicity C–S–H(I). SEM additionally confirmed substantial densification of the cement-stone structure, while only a moderate change in portlandite content was observed after modification with NBS.
• Trial 3D printing of cement–sand mortar containing 1.72% nano-biosilica demonstrated reduced layer slumping, clearer geometry, and better shape retention of the printed elements. Interlayer bond strength was also improved by 17.0% (from 1.8 to 2.1 MPa).

Overall, the findings highlight nano-biosilica derived from rice husk ash as a promising, locally available SCM for engineering the microstructure and performance of 3DCP mixtures.