Introduction

Nanomaterials have significantly received the attention of engineers, chemists, biologists and multidisciplinary researchers in recent years due to their high specific surface area and unique size-dependent physicochemical properties, which differ substantially from those of their bulk forms1,2. These characteristics can enhance surface-mediated phenomena such as adsorption, catalysis, charge transfer and interfacial reactivity, thereby expanding their potential applications across various scientific fields3,4. High specific surface area zinc oxide nanoparticles, in particular, have attracted remarkable research interest due to their versatile physicochemical, optical and biological properties and wide applicability in emerging technologies5,6. Zinc oxide offers a broadly tunable morphology, high optical transparency, good electron mobility, a high exciton binding energy of 60 meV and a wide band gap of 3.37 eV. Its stability under thermal and mechanical stress, broad radiation absorption spectrum and excellent photostability further reinforce its status as a multifunctional material7,8 and enable its use in numerous advanced technologies such as transparent electronic devices, UV-emitting components, optoelectronic and piezoelectric systems, solar cells, chemical sensors and spintronic devices9,10,11,12,13,14. The specific surface area and pore architecture of ZnO play a decisive role in determining functional performance for applications governed by surface-dependent processes (e.g. adsorption, photocatalytic degradation, antibacterial activity and heterogenous catalysis). Increasing the SSA of ZnO enhances the density of accessible active sites, facilitates mass transport and improves interfacial interactions with target species. Consequently, ZnO nanoparticles have been widely employed in the degradation of organic pollutants, wastewater treatment and gas purification15,16,17,18. High specific surface area of active zinc oxide has also been reported to improve adsorption capacity, catalytic efficiency and antimicrobial performance, which turns it a promising candidate for environmentally oriented technologies19,20,21. Zinc oxide nanoparticles are considered biocompatible and exhibit low toxicity, as Zn2+ ions are required for normal growth in animals, humans and plants due to their involvement in many enzyme systems, metabolic activities and oxidation–reduction processes. Moreover, high SSA ZnO nanoparticles play an increasingly important role in construction materials, biomedical technologies and functional coatings. In cement, concrete, ceramics and composite matrices, ZnO nanostructures can improve mechanical strength, thermal stability and UV resistance. In biomedical and cosmetic applications, they are employed in sunscreens, wound dressings, drug-delivery carriers and tissue-engineering scaffolds due to their biocompatibility, antimicrobial properties and controlled Zn2+ release22,23,24,25,26.

A variety of physical, chemical and bio-assisted synthesis approaches have been developed to produce ZnO nanostructures with a broad range of surface properties and architectures27. Physical synthesis of ZnO nanoparticles typically involves top-down techniques such as laser ablation, ball milling and physical vapor deposition (PVD). They generally offer high material purity and precise structural control, which are advantageous for electronic applications; however, these techniques typically require specialized equipment and high energy input28,29. Chemical methods (e.g. co-precipitation, hydro/solvothermal, sol-gel, emulsion/micro-emulsion, spray pyrolysis, chemical vapor deposition (CVD), etc.) are widely adopted due to their relative simplicity and flexibility in controlling particle size and morphology30. Nevertheless, many conventional chemical synthesis strategies face trade-offs between process complexity, energy demand, morphology control and reproducibility, particularly when targeting porous or high SSA ZnO31,32. For example, sol-gel and precipitation methods often suffer from particle agglomeration and limited control over particle size distribution33, whereas CVD and hydro/solvothermal typically require high temperatures, sophisticated equipment and complex procedures34. Recently, green synthesis approaches employing plant extracts or microorganisms have been explored as environmentally benign alternatives. However, their broader application may be constrained by biological variability, scalability challenges, incomplete mechanistic understanding and regulatory constraints35,36.

The hydrolysis of zinc-based precursors (e.g. zinc hydroxides, zinc carbonates, basic zinc carbonates, zinc acetates and zinc acetate derivatives) represents a common chemical route for synthesizing ZnO nanoparticles. In particular, the thermal decomposition of basic zinc carbonates, including Zn3CO3(OH)4·2 H2O, Zn4CO3(OH)6·H2O and Zn5(CO3)2(OH)6, at appropriate temperatures leads to the formation of active zinc oxide37. The conversion of balk zinc oxide powder into hydrozincite through CO2 carbonation under mild alkaline conditions, followed by its thermal decomposition, has been reported as an effective route for producing porous ZnO with enhanced specific surface area. Studies have shown that the formation of basic zinc hydroxycarbonate (hydrozincite) is favored in moderately alkaline media (typically pH ≈ 9–11), where carbonate precipitation is promoted while avoiding incomplete carbonation at lower pH or excessive Zn(OH)2 redissolution at higher alkalinity, leading to improved precursor stability and reproducibility. During the carbonation, ZnO suspended in water reacts with dissolved CO2 to form hydrozincite, which subsequently decomposes into ZnO upon moderate temperature calcination while retaining a porous, high SSA structure. Notably, Kowali et al. reported up to a 30-fold increase in specific surface area (from 1.8 to 61.8 m2/g) and a 20-fold increase in pore volume—primarily macro and mesopores—through transformation of low SSA ZnO by carbonation and thermal treatment. Hydrozincite formation in aqueous media proceeds according to the reactions described via Eqs. (1) and (2), which can be summarized as Eq. (3)38,39:

$$\:5ZnO+10{H}^{+}\leftrightarrow\:5{Zn}^{2+}+5{H}_{2}O$$
(1)
$$\:5{Zn}^{2+}+2C{O}_{2}+8{H}_{2}O\leftrightarrow\:{Zn}_{5}{\left(C{O}_{3}\right)}_{2}{\left(OH\right)}_{6}+10{H}^{+}$$
(2)
$$\:5ZnO+2C{O}_{2}+3{H}_{2}O\leftrightarrow\:{Zn}_{5}{\left(C{O}_{3}\right)}_{2}{\left(OH\right)}_{6}$$
(3)

In Table 1, several chemical synthesis methods of high SSA ZnO are compared, all based on carbonation of a starting zinc source and calcination of a zinc-containing precursor (e.g. basic zinc hydroxycarbonate or hydrozincite).

Table 1 Comparison of chemical synthesis methods of high SSA ZnO through carbonation and calcination.
Full size table

The functional performance of ZnO nanomaterials is strongly governed by their size, morphology and pore size, which can be precisely controlled through adjustments in synthesis routes, process conditions, precursor composition, pH or reactant concentrations27. Despite extensive prior work on carbonation-decomposition routes, systematic evaluation of the combined effects of these parameters and their interactions remains limited. Statistically guided optimization strategies have recently been applied to other high SSA porous materials in environmental research, demonstrating their general utility in multivariable process design43. In this study, the response surface methodology (RSM) is employed as a statistically robust design-of-experiments approach to optimize and evaluate the key synthesis parameters governing hydrozincite formation and thermal decomposition. This work aims to provide reproducible and statistically guided optimization of high SSA zinc oxide by quantitatively modeling the relationships between process variables and ZnO textural properties.

Materials and methods

Reagents

In this study, bulk zinc oxide powder (Sepid Oxide Shokuhieh Co., Qom, Iran) with the specific surface area of 4.5 m2/g was employed as the primary reagent for the synthesis of experimental samples. Aqueous ammonia solution (2.5 wt%) was used to adjust initial pH of Zinc oxide and distilled water suspension. Carbon dioxide gas was also used for the carbonation of low SSA zinc oxide and production of basic zinc hydroxycarbonate (hydrozincite).

Preparation of samples

In each experiment, a suspension containing 10 g of bulk zinc oxide powder and an appropriate volume of distilled water (100, 150 or 200 mL) was prepared in a 250 mL beaker. A magnetic heater-stirrer was used to mix the suspension at a constant speed of 600 rpm. After a few minutes of mixing at room temperature, the initial pH of the suspension was adjusted by the gradual addition of 2.5 wt% aqueous ammonia solution and monitored via a digital pH meter. Once the desired initial pH (ranging from 9 to 11, depending on the synthesis conditions) was achieved, the suspension was heated to the target reaction temperature. When the desired temperature was reached, the carbonation reaction was started by immersing a CO2 sparger in the beaker; so that CO2 gas was bubbled into the suspension at a constant flow rate of 100 mL/min. The carbonation time for each synthesis was measured from the moment the sparger was inserted. During carbonation, the beaker was covered to prevent water evaporation and maintain a constant liquid-to-solid ratio (L/S) throughout the process. When the carbonation time was over, the suspension was allowed to cool to room temperature and the final pH was measured, typically ranging between 6 and 7. The solid product was then separated by vacuum filtration and washed with 500 mL of distilled water. All samples were dried in an oven at 110 °C for 20 h. Subsequently, 2 g of each dried hydrozincite sample was calcined in a furnace at 400 °C for 3 h with a heating rate of 3 °C/min and decomposed to the final high SSA ZnO. Figure 1 schematically illustrates the overall synthesis procedure including the CO2 carbonation and subsequent thermal decomposition of hydrozincite.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Synthesis of high SSA ZnO through CO2 carbonation and thermal decomposition.

Design of experiments (DOE)

The statistical experimental design was applied to investigate and optimize the effects of controlling parameters using the response surface methodology (RSM) based on the Box-Behnken design (BBD). This statistical approach was selected due to its high efficiency in evaluating both individual and interactive effects of multiple process parameters through a limited number of experimental runs. Four key independent parameters of synthesis temperature (factor A), carbonation time (factor B), initial pH of suspension (factor C) and liquid-to-solid ratio (L/S) (factor D) were selected and examined and a total of 27 experimental runs were designed. The response (i.e. dependent variable) was the specific surface area (m2/g) of the final ZnO samples, determined through BET surface area analysis. The factors and their levels are summarized in Table 2.

Table 2 Factors and levels for Box–Behnken design (BBD).
Full size table

When all 27 runs were performed, the response variable (SSA) was analyzed and then recorded in the Design Expert 13. In the response surface methodology (RSM), the mathematical relationships between input variables and the output (response) are approximated using polynomial equations. Derived from experimental data and the designed set of runs, these equations model how variations in process parameters influence the response44. As shown in Eq. (4), the general form of the response surface equation typically includes linear, quadratic and interaction terms among the variables:

$$\:Y={\beta\:}_{0}+\sum\:{\beta\:}_{i}{X}_{i}+\sum\:{\beta\:}_{ii}{X}_{i}^{2}+\sum\:{\beta\:}_{ij}{{X}_{i}X}_{j}$$
(4)

Where the \(\:Y\) represents the predicted response; \(\:{X}_{i}\), \(\:{X}_{i}^{2}\) and \(\:{X}_{j}\) are the independent variables; \(\:{\beta\:}_{0}\) is the constant coefficient; \(\:{\beta\:}_{i}\) represents the coefficient of linear effect; \(\:{\beta\:}_{ii}\) represents the coefficient of quadratic (second-order) effect and \(\:{\beta\:}_{ij}\) is the coefficient of interaction effect45. Such models not only enable the prediction of response under any desired set of conditions, but also serve as powerful tools for analyzing the effect of each factor and optimizing the synthesis parameters.

Characterization

All synthesized high SSA ZnO samples were characterized using Brunauer–Emmett–Teller (BET, BELSORP MINI X) analysis to determine their specific surface area and porosity. Furthermore, the optimum sample was examined by additional analyses. It was tested by the Fourier transform infrared spectroscopy (FTIR, Perkin Elmer) to investigate the functional groups and interaction and connection of chemical constituents. It was also analyzed using the X-ray diffraction (XRD, Bourevestnik, DRON-8) to assess the crystalline structure. The scanning electron microscopy (SEM, TESCAN, Vega II XMU) was applied to examine the morphological features of optimum sample. These complementary techniques provided a comprehensive understanding of the physicochemical characteristics of the optimized sample.

Results and discussion

Development of RSM model using experimental data

As explained previously, the Response Surface Methodology (RSM) was employed to determine the optimal conditions for the synthesis of high SSA zinc oxide, considering four influential factors: synthesis temperature (factor A), carbonation time (factor B), initial pH of suspension (factor C) and liquid-to-solid ratio (factor D), with BET specific surface area as the response parameter. 27 experiments were carried out under the specified conditions proposed by a three-level Box Behnken design (BBD) with three central points and the specific surface area of final product was measured via BET analysis. The central point experiments were performed as independent synthesis runs and the corresponding SSA values were measured independently; these replicate runs were used to estimate the pure experimental error and assess the reproducibility of synthesis process. Table 3 summarizes the synthesis conditions of each run and corresponding response values. The measured specific surface area (SSA) values ranged from 23.2 to 32.2 m2/g, while the pore volume (Vp) varied from 0.220 to 0.570 cm3/g and the average pore radius ranged from 17.68 to 34.46 nm, indicating notable variations induced by different synthesis parameters. These findings highlight the necessity of statistical modeling to understand the influence of each factor and to identify the optimum synthesis conditions for maximizing the SSA.

Table 3 Experimental data and response values (SSA) for synthesized high SSA ZnO samples.
Full size table

Statistical assessment and model adequacy

In the framework of Response Surface Methodology (RSM), the experimental data were fitted using polynomial regression models to develop a statistically reliable model which could describe the correlation between the process variables and the response variable. Sequential model sum of squares, lack-of-fit tests and model summary statistics were used to evaluate and determine the most appropriate regression model for describing variations in SSA under different synthesis conditions. According to the results of sequential model comparison (Table 4), the 2FI model was selected as the most appropriate model. This choice is supported by a significant improvement in model fitting from the linear to the two-factor interaction model (p-value < 0.0001), while the addition of quadratic terms did not contribute substantially towards explaining the variations of SSA (p-value = 0.9623) and even reduced the predicted R2, indicating a decrease in the model’s predictive capability. In statistical design, the preferred model is the simplest one that provides adequate accuracy; therefore, terms that do not significantly improve predictive performance are eliminated. The cubic model was also discarded due to aliasing, as the experimental design did not provide sufficient information to independently estimate all cubic terms.

Table 4 Summary of model-fitting results for predicting the response variable (SSA).
Full size table

According to Table 5, the analysis of variance (ANOVA) results confirmed that the 2FI model was statistically significant (F-value = 33.37, p-value < 0.0001). Among the main factors, synthesis temperature (A), carbonation time (B) and liquid-to-solid ratio (D) significantly influenced the SSA (p-value < 0.05), while the initial pH (C) was not significant (p-value = 0.1115). Additionally, four two-factor interactions (i.e. AB, AC, BD and CD) were significant, indicating complex interdependencies among the synthesis parameters. The lack-of-fit (LOF) measures the model’s adequacy to reflect data excluded from the regression for effective predictions45. The 2FI model showed a non-significant lack of fit (p = 0.5617), indicating good agreement with the experimental data without overfitting and a lower predicted residual sum of squares (PRESS = 14.00) compared to linear or quadratic models, confirming its superior predictive ability.

Table 5 Summary of model-fitting results for predicting the response variable (SSA).
Full size table

Equation 5 represents the 2FI regression model describing the effect of synthesis parameters on the specific surface area (SSA) in terms of coded variables:

$$\begin{aligned} SSA\:\left( {m^{2} /g} \right) = & 27.45 + 1.70 \times A + 0.6728 \times B + 0.2585 \times C \\ & - 1.17 \times D - 1.49 \times AB - 1.30 \times AC - 0.3515 \times AD \\ & + 0.2823 \times BC - 2.16 \times BD + 0.6215 \times CD \\ \end{aligned}$$
(5)

The 2FI model produced an R2 of 0.9542, an adjusted R2 of 0.9257 and a predicted R2 of 0.8583. The reasonable agreement between the adjusted and predicted R2 values (Δ < 0.2) signified the model’s high predictive capabilities. The residual standard deviation of 0.5316 and a low coefficient of variation (CV = 1.94%) further validated the model’s accuracy and reliability. Additionally, the adequate precision of 23.2159 (≫ 4) indicated that the signal-to-noise ratio was desirable and the model offered sufficient resolution for navigating the design space (Table 6).

Table 6 Summary of fit statistics.
Full size table
Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Diagnostic plots: (a) Normal plot of residuals; (b) Residuals vs. predicted; (c) Predicted vs. actual; (d) Cook’s distance.

The adequacy of regression model is further analyzed using the diagnostic plots. The normal plot of residuals demonstrated an almost normal distribution (Fig. 2a). In the residuals versus predicted plot, the experimental points were randomly scattered around the zero axis, which confirmed homogeneity of variances with no systematic error (Fig. 2b). In the predicted versus actual plot, the points were clustered along the diagonal line, implying the model’s robust predictive performance (Fig. 2c). All Cook’s distance values were below the crucial threshold, confirming the statistical reliability of the data set (Fig. 2d).

Among the four examined parameters, synthesis temperature (A), carbonation time (B) and liquid-to-solid ratio (D) demonstrated statistically significant impacts on the specific surface area (p < 0.05), while the influence of initial pH (C), though positive, was not significant (p ≈ 0.11). The factor A exhibited the most significant positive impact (+ 1.70), presumably due to regulated particle development at higher temperatures that mitigates early-stage agglomeration. The factor B had a notable positive effect (+ 0.6728), correlating with a higher conversion to zinc hydroxycarbonate and formation of a more porous ZnO structure. The liquid-to-solid ratio (D) had a substantial negative impact (–1.17), indicating that excessively dilute suspensions hinder nucleation and hence decrease the surface area. Despite the factor C was not significant individually (+ 0.2585), it played a secondary role in interactions with other variables.

The ANOVA test also indicated that four two-factor interactions (i.e. AB, AC, BD and CD) were statistically significant in the 2FI model. The AB (temperature × time) and AC (temperature × pH) interactions, both with negative coefficients, demonstrating that increasing temperature in combination with prolonged carbonation or higher pH reduces the specific surface area due to accelerated crystal growth and formation of denser precipitates (Figs. 3 and 4). The BD interaction (time × L/S ratio) also exhibited a strong negative effect, suggesting that extended carbonation at higher concentrations promotes secondary particle growth and lowers the surface area (Fig. 5). Conversely, the CD interaction (pH × L/S ratio) showed a positive contribution, implying that under alkaline conditions, higher liquid-to-solid ratios favor the development of more porous structures and thus enhance the surface area (Fig. 6). The findings revealed that while each parameter independently affects the response, simultaneous variations can produce unexpected or even opposing effects. This underscored the importance of using interaction models for effective optimization and experimental design in such systems.

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

Contour and 3D surface plots illustrating the interaction effect of synthesis temperature and carbonation time on specific surface area (SSA) of final ZnO product.

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

Contour and 3D surface plots illustrating the interaction effect of synthesis temperature and pH on specific surface area (SSA) of final ZnO product.

Fig. 5
Fig. 5The alternative text for this image may have been generated using AI.
Full size image

Contour and 3D surface plots illustrating the interaction effect of carbonation time and liquid-to-solid ratio on specific surface area (SSA) of final ZnO product.

Fig. 6
Fig. 6The alternative text for this image may have been generated using AI.
Full size image

Contour and 3D surface plots illustrating the interaction effect of pH and liquid-to-solid ratio on specific surface area (SSA) of final ZnO product.

Optimization using desirability function

Aimed at determining the optimal synthesis conditions for maximizing the specific surface area (SSA), the optimization procedure was performed using the desirability function in Design Expert 13. The procedure identified an optimum set of conditions with an overall desirability of 0.9478 and an SSA desirability of 0.8992, confirming the robustness of the optimized model (Fig. 7a). The optimum values included a synthesis temperature of 70 °C, a carbonation time of 1 h, an initial pH of 9 and a liquid-to-solid ratio of 20. These conditions yielded a predicted SSA of 31.318 m2/g, closely matching the experimentally observed maximum. The desirability profiles further showed that all factors exhibited values near 1, indicating that the target ranges were readily attainable (Fig. 7b). Therefore, the developed model demonstrated strong predictive accuracy while reducing experimental effort and laboratory costs.

Fig. 7
Fig. 7The alternative text for this image may have been generated using AI.
Full size image

(a) Desirability function aimed to maximize the response variable (SSA); (b) Desirability ramp plots for optimum synthesis conditions.

Validation of RSM model

To validate the reliability and predictive capability of the statistical model developed in this study, a zinc oxide sample was synthesized under the optimum conditions identified by response surface methodology. The specific surface area was measured through BET analysis (29.1 m2/g), which closely matched the model prediction (31.3 m2/g) with a relative error of 7% (Table 7). This strong consistency confirmed that the 2FI model provided an effective and accurate representation of the correlation between synthesis parameters and resulting surface properties of ZnO.

Table 7 Summary of model-fitting results for predicting the response variable (SSA).
Full size table

The optimum sample was further analyzed using characterization techniques. A Type IV isotherm with an H3 hysteresis loop was observed in the nitrogen adsorption-desorption isotherm of optimum ZnO sample, which is characteristic of mesoporous solids formed by non-rigid aggregates of plate-like or layered particles (Fig. 8a). At low relative pressures (p/p0 < 0.3), the gradual increase in the adsorption curve reflects a monolayer-multilayer adsorption mechanism occurring on the ZnO surface, indicating the presence of accessible external surface area. A sharp rise is observed in the adsorbed volume as the relative pressure increases (p/p0 = 0.8-1.0), reflecting the substantial capillary condensation within an interconnected mesoporous network. The H3 hysteresis loop between the adsorption (ADS) and desorption (DES) curves provides strong evidence for the mesoporous structure with slit-like pores, which is typical of agglomerated plate-shaped nanoparticles or layered particle arrangements. This pore geometry is consistent with the formation and subsequent thermal decomposition of hydrozincite platelets during the carbonation-calcination process. The mesoporosity generated through this mechanism contributes significantly to the observed enhancement in specific surface area and pore volume, both of which are critical for surface-driven applications such as adsorption and heterogenous catalysis46.

Fig. 8
Fig. 8The alternative text for this image may have been generated using AI.
Full size image

(a) N2 adsorption (ADS)-desorption (DES) isotherm of BET analysis for optimum ZnO sample; (b) BJH desorption pore size distribution (PSD) of optimum sample.

The BJH desorption pore size distribution (PSD) curve further supports the mesoporous nature of the synthesized ZnO, with significant pore volume contribution in the range of 20–40 nm (Fig. 8b). The presence of multiple peaks indicates a hierarchical pore system, combining smaller mesopores (< 10 nm) and larger mesopores/macropores (> 50 nm). Such bimodal or broad pore distribution is advantageous for facilitating mass transfer, improving surface accessibility and reactivity and reducing diffusion limitations, thereby enhancing the functional performance of porous ZnO in catalytic, adsorption and biomedical applications47.

As shown in Fig. 9, the SEM images of optimum ZnO sample display a nanostructured morphology characterized by plate-like and interconnected particles that form a porous framework. This structural arrangement is consistent with the BET–BJH analysis, which confirmed a high specific surface area accompanied by a hierarchical mesoporous system. The interconnection of plate-shaped primary particles generates slit-like voids and interparticle spaces, consistent with the H3 hysteresis loop observed in nitrogen adsorption-desorption analysis. The porous and interconnected network facilitates the exposure of numerous active sites, while the mesoporous channels enhance molecular diffusion and mass transfer. These features are particularly beneficial for surface-dependent processes, including adsorption and catalytic reactions. The observed morphology also reflects the underlying formation mechanism of ZnO during the carbonation-calcination route. During the carbonation step, partial dissolution of ZnO in the aqueous medium and its reaction with CO2 lead to the precipitation of plate-like hydrozincite, which gradually forms a layer around the original ZnO core. As the carbonation progresses, the growth and aggregation of hydrozincite platelets causes the particles to transform from irregular rod- or hexagon-like shapes into more uniform spherical agglomerates, developing a porous precursor structure. During the calcination at 400 °C, the hydrozincite decomposes and releases CO2 and H2O while preserving the platelet-like morphology38. The secondary porous ZnO layer covering the primary particles aligns with the microstructural features seen in the SEM images and supports the formation of a hierarchical ZnO structure with increased surface area. The retention of plate-like features after calcination explains the persistence of slit-shaped mesopores and supports the structure-property relationship inferred from BET analysis. Generally, the SEM observations confirm that the optimized synthesis conditions effectively control particle growth and suppress dense agglomeration, leading to the formation of an active, high SSA ZnO architecture suitable for surface-driven applications.

Fig. 9
Fig. 9The alternative text for this image may have been generated using AI.
Full size image

SEM images of optimum sample.

Figure 10 presents the FTIR spectra of optimum sample before and after calcination. The broad absorption band in the range of 3200–3600 cm−1 corresponds to the O–H stretching vibrations of structural hydroxyl groups together with a small contribution from physisorbed moisture. In this region, hydrozincite typically exhibits a broad envelope with a more pronounced component near 3300 cm−1. The weak band at approximately 1640 cm−1 originates from the H–O–H bending mode of physically adsorbed water, which is commonly observed in materials synthesized via wet-chemical routes. The pair of strong bands at 1500–1510 and 1380–1415 cm−1 are attributed to the asymmetric stretching mode (ν3) of CO32− groups, serving as characteristic diagnostic carbonate vibrations of hydrozincite (basic zinc carbonate). A weak band observed in the 1030–1060 cm−1 range corresponds to the symmetric stretching mode (ν₁) of carbonate ions, typically arising from weak crystal-field interactions. The medium intensity band near 950 cm−1 is assigned to the in-plane bending vibration of O–H groups, a distinct feature of hydrozincite. Additionally, the bands at 870–880 cm−1 correspond to out-of-plane bending of CO32− ions, while those in the 680–730 cm−1 region represent the in-plane bending mode (ν4) of carbonate species. The bands appearing between 450 and 600 cm−1 arise from Zn–O lattice vibrations associated with hydrozincite; these modes disappear after calcination due to the complete decomposition of precursor and formation of ZnO. The absorption band at 2330–2350 cm−1 is attributed to gaseous CO2 in the optical path and should not be considered in phase identification. The FTIR spectrum of the calcined sample displays a dominant and broad band in the range of 430–520 cm−1, corresponding to intrinsic Zn–O lattice vibrations, which are indicative of the wurtzite structure of ZnO. A relatively weak and broad feature near 3400 cm−1 is associated with the stretching of O–H groups or surface-adsorbed water, while the weak band at around 1630 cm−1 corresponds to the H–O–H bending vibration of physically adsorbed water. The complete disappearance of carbonate related bands confirms the full decomposition of hydrozincite (basic zinc hydroxycarbonate) during calcination48,49.

Fig. 10
Fig. 10The alternative text for this image may have been generated using AI.
Full size image

FTIR spectra of optimum sample (grey: hydrozincite; blue: active zinc oxide).

The X-ray diffraction (XRD) analysis was conducted over a 2θ range of 0–80°, enabling crystalline phase identification of the optimum sample before and after calcination. The diffraction pattern of the uncalcined sample exhibited the characteristic signatures of hydrozincite (Zn5(CO3)2(OH)6), including a dominant peak at approximately 13° (2θ) alongside several well-defined peaks within the 28–36° interval. These features confirm the presence of basic zinc carbonate precursor prior to thermal decomposition (Fig. 11a). In the calcined sample, intense diffraction peaks emerged at 31.7°, 34.4°, 36.2°, 47.5°, 56.6° and 62.8° (2θ). These peaks correspond to the reference pattern of hexagonal wurtzite ZnO as documented in JCPDS card No. 36-1451, indicating the complete decomposition of hydrozincite and formation of crystalline ZnO (Fig. 11b).

Fig. 11
Fig. 11The alternative text for this image may have been generated using AI.
Full size image

XRD patterns of optimum sample: (a) hydrozincite (before calcination); (b) active zinc oxide (after calcination).

The sharpness and high intensity of these reflections signify good crystallinity of the calcined product, despite the relatively mild calcination temperature employed. Calculated using the Scherrer equation, the average crystallite size of calcined sample was approximately 12.6 nm50,51. This nanoscale crystalline size contributes to an increased external surface area by limiting crystal growth during the thermal decomposition. However, it is important to note that the high specific surface area observed in this study cannot be solely attributed to small crystallite size. Instead, the dominant contribution to SSA arises from the mesoporous architecture generated during the carbonation-decomposition process, as evidenced by the type IV isotherm with an H3 hysteresis loop and the hierarchical pore size distribution revealed by BET-BJH analysis. In this context, crystallite size primarily governs the size of the primary ZnO domains, while the overall specific surface area is largely controlled by interparticle porosity and the platelet-derived pore network inherited from the hydrozincite precursor. This is in accordance with similar observations reported for porous metal oxides, where the mesoporous architecture and interparticle void network dominate the specific surface area, while crystallite size primarily influences the external surface contribution rather than the bulk porosity52. The preservation of crystallinity alongside high SSA highlights the effectiveness of the optimized synthesis conditions in decoupling crystal growth from pore collapse, enabling the formation of porous ZnO with both structural order and enhanced surface accessibility.

An RSM model is generally considered statistically adequate when the predicted R2 value is relatively high (typically exceeding 0.80) and its difference from the adjusted R2 remains below approximately 0.20, indicating its reliable predictive performance within the experimental domain. Therefore, the predicted responses are expected to agree with the experimental data within a reasonable margin of error. Moreover, based on established statistical principles of experimental design, a relative prediction error within the range of 5–10% is regarded acceptable for nanomaterial synthesis and other complex engineering systems, where random experimental fluctuations, minor uncontrolled process variations, batch-to-batch inconsistencies and instrumental limitations are unavoidable53. Given the observed relative error of 7%, which lies well within the acceptable limits, the developed model demonstrates strong predictive accuracy for optimizing the synthesis of high surface area ZnO.

Future directions and outlook

While the present study focuses on statistically guided optimization of synthesis parameters and elucidation of structure-property relationships governing the specific surface area of ZnO, further investigations are warranted to expand the functional scope of the optimized material. Future work may explore elemental doping or composite formation (e.g. ZnO-carbon or ZnO-metal oxide hybrids) to tailor surface chemistry, defect density and electronic properties for enhanced performance in photocatalytic, adsorption or antibacterial applications. In addition, evaluating the behavior of synthesized high SSA ZnO in realistic water matrices or complex environments would provide valuable insight into its stability, surface reactivity and practical applicability54. Recent studies on high surface area porous materials have demonstrated that the optimization of textural properties, combined with surface modification, can significantly influence adsorption efficiency and toxicity mitigation in environmental systems, highlighting promising directions for extending the present synthetic strategy55. Finally, coupling the optimized carbonation-decomposition route with application-driven performance testing will enable more comprehensive assessment of the material’s functionality and guide future process development.

Conclusion

In this study, high SSA ZnO nanoparticles were synthesized through controlled CO2 carbonation and thermal decomposition of a hydrozincite precursor. The response surface methodology (RSM) enabled the systematic assessment of effect of key parameters governing specific surface area and the optimization of synthesis conditions within the experimental design space. Validation experiments confirmed the reliability and predictive accuracy of developed model and characterization techniques and analyses verified the formation of a porous ZnO nanostructure with enhanced specific surface area, highlighting the effectiveness of carbonation-decomposition route combined with statistical optimization for preparing ZnO nanoparticles suitable for surface-driven applications.