Zinc nitrate hexahydrate (Zn (NO₃) ₂·6H₂O, 99.99% purity) and ammonium hydroxide (NH₄OH, 99% purity) were purchased from Sigma-Aldrich and used as received without additional purification. Fresh saffron (Crocus sativus L) flowers were collected from the Pampore region in Jammu and Kashmir, India. The anthers and petals were separated by hand, thoroughly washed with distilled water, and used for extract preparation.

Fresh saffron (Crocus sativus L.) flowers were obtained from the Pampore region of Jammu & Kashmir, India, during the harvesting season. The flowers were washed thoroughly with distilled water to remove dust and other surface impurities. The anthers and petals were manually separated and air-dried at room temperature in the shade for 24 hours to reduce surface moisture.

To prepare the aqueous extracts, 5 grams each of dried saffron flower anthers and petals were separately washed thoroughly and sonicated to remove any adhering dirt and soil particles. Each sample was then mixed with 150 mL of distilled water and heated at 70°C for one hour with constant stirring to facilitate the extraction of bioactive compounds. After heating, the mixtures were allowed to cool naturally to room temperature. The solutions were then filtered through muslin cloth to remove solid residues, yielding clear aqueous extracts. These filtrates were used immediately in the green synthesis of ZnO nanoparticles, serving as natural reducing, capping, and chelating agents in the co-precipitation process.

2.3. Green Synthesis of ZnO Nanoparticles Using Saffron Flower Anther (Z2) and Petal (Z4) Extract: Co-Precipitation Method

Zinc nitrate hexahydrate (Zn (NO₃) ₂·6H₂O, 99.99% purity) and ammonium hydroxide (NH₄OH, 99% purity) were both purchased from Sigma-Aldrich. First, zinc nitrate hexahydrate was weighed according to the required stoichiometric calculations and dissolved in 40 mL of double-distilled water. This solution was then stirred continuously at room temperature for approximately 30 minutes. Subsequently, 100 mL of aqueous extract obtained from saffron flower anthers was added dropwise to the solution, acting as a natural chelating and stabilizing agent. The pH was then monitored, and ammonium hydroxide (NH₄OH) was added dropwise until a pH of 10 or higher was achieved. The reaction mixture was heated to 70°C and maintained under stirring for 2 hours. The precipitate formed was centrifuged at 9,000 rpm for 5 minutes and washed five times—twice with double-distilled water, once with ethanol, and again twice with double-distilled water. The product was then dried in an oven at 90°C for 12 hours, followed by calcination in a muffle furnace at 600°C for 2 hours. The obtained ZnO nanoparticles were finally ground using a mortar and pestle to obtain a fine homogeneous powder, as shown in Fig. 1.

A similar synthesis procedure was followed for the petal extract, with 100 mL of aqueous extract obtained from saffron flower petals replacing the anther extract in the initial step as shown in Fig. 2. The rest of the reaction conditions, including pH adjustment, heating, centrifugation, washing, drying, and calcination, remained the same, ensuring that any observed variations in nanoparticle properties could be attributed primarily to the different phytochemical compositions of the petals and anthers.

The XRD patterns of ZnO nanoparticles synthesized via the green co-precipitation route using (a) anther extract and (b) petal extract are essentially identical and match the hexagonal wurtzite phase of ZnO (space group P6₃mc), in agreement with the standard ICDD/JCPDS card no. 36-1451 (Ref 20) shown in Fig. 3. The characteristic reflections appear at 2θ ≈ 31.7° (100), 34.4° (002), 36.2° (101; most intense), 47.5° (102), 56.6° (110), 62.8° (103), 66.4° (200), 67.9° (112), and 69.1°–72.6° (201/202/204, weak), consistent with previously reported wurtzite ZnO diffraction data (Ref 21). No detectable impurity peaks from Zn (OH)₂ or other zinc phases were observed, confirming the formation of phase-pure ZnO in both cases (Ref 22). The nearly identical relative intensities suggest no strong preferred orientation, and the comparable peak widths indicate similar crystallite sizes and micro strain for anther- and petal-derived samples; any small broadening differences can be explained by phytochemical-mediated nucleation and growth effects rather than changes in crystal structure (Ref 23). Overall, the plant part used (anther vs. petals) does not alter the crystal phase or lattice symmetry but may subtly influence crystallite size distribution and defect density, consistent with earlier findings for plant-extract-mediated ZnO synthesis (Ref 22).

The Raman spectra of the green-synthesized ZnO made with anther extract (Z2) and petal extract (Z4) both show the fingerprints of wurtzite ZnO in Fig. 4. In (Z2), the main peaks occur at 205, 275, 331, 436.9, 577, and 642 cm⁻¹; in Z4, they are at 201, 272, 325, 428, 570, and 642 cm⁻¹. The intense band near 437 cm⁻¹ (Z2: 436.9; Z4: 428) is the E₂(high) mode of wurtzite ZnO; its downshift by ~ 9 cm⁻¹ in Z4 is consistent with tensile micro strain and/or phonon-confinement effects typical for smaller/defective ZnO nanocrystals (canonical bulk E₂(high) ≈ 437 cm⁻¹) (Ref 24). The features at ~ 331 cm⁻¹ (Z2) and ~ 325 cm⁻¹ (Z4) correspond to the E₂(high)–E₂(low) combination, while ~ 205 cm⁻¹ (Z2) and ~ 201 cm⁻¹ (Z4) are the 2E₂(low) overtone; both are standard second-order ZnO bands. The peaks at ~ 275–272 cm⁻¹ arise from TA(M) + LA(M) combinations, again expected for wurtzite ZnO (Ref 25). The bands in the LO region differ most: Z2 shows A₁(LO) at ~ 577 cm⁻¹, whereas Z4 is at ~ 570 cm⁻¹ with lower relative intensity; LO-mode position/intensity are highly sensitive to point defects (oxygen vacancies V_O, zinc interstitials Zn_i) and resonance effects stronger, slightly higher-wavenumber LO features generally indicate greater defect concentration and different carrier–phonon coupling, so Z2 likely contains more V_O/Zn_i than Z4 (Ref 26). The weak shoulder near ~ 642 cm⁻¹ in both samples is attributable to A₁(LO) + TA(M) multiphonon scattering (Ref 25). Overall, both extracts yield phase-pure wurtzite ZnO, but Z4 is more downshifted in E₂(high) and LO modes (428/570 cm⁻¹) pointing to higher tensile strain and/or smaller phonon-coherence length with fewer LO-active defects than (Z2); by contrast, (Z2)’s stronger ~ 577 cm⁻¹ LO band suggests more native defects (V_O/Zn_i) introduced by its phytochemical environment

during growth (Ref 27).

The XPS analysis of ZnO nanoparticles synthesized by the two green routes (anther extract, Z2, and petal extract, Z4) reveals the characteristic Zn and O signals of wurtzite ZnO with no significant impurity peaks, confirming the purity of both samples in Figs. 5 and 6. In both spectra, the Zn 2p region displays two strong peaks at ~ 1021 eV (Zn 2p₃/₂) and ~ 1044 eV (Zn 2p₁/₂) with a spin–orbit splitting of ~ 23 eV, consistent with the Zn²⁺ oxidation state (Ref 28,29).

The main differences are observed in the O 1s region: Z2 shows a dominant lattice oxygen peak at ~ 530 eV accompanied by a pronounced shoulder at ~ 531–532 eV, which is attributed to oxygen vacancies, hydroxyl groups, and chemisorbed oxygen species (Ref 30,31) In contrast, Z4 also exhibits the lattice oxygen feature at ~ 530 eV but with a weaker high-binding-energy component, suggesting a lower concentration of surface defects. This variation can be traced back to differences in the phytochemical composition of the two extracts, which affect nucleation and defect passivation during ZnO formation (Ref 32). The stronger defect-related O 1s intensity in Z2 implies a higher density of oxygen vacancies and Zn interstitials, known to enhance photocatalytic activity through improved charge trapping and reactive oxygen species generation, whereas Z4 may exhibit relatively lower photocatalytic reactivity but improved structural and chemical stability (Ref 30,31). Overall, the comparative XPS results highlight how the choice of biological precursor in green synthesis not only dictates particle formation but also tailors defect chemistry at the nanoscale, thereby tuning functional performance.

The HRTEM analysis of the ZnO nanoparticles synthesized via the green route reveals their nanocrystalline nature, with particles exhibiting well-resolved lattice fringes. The measured interplanar spacing of ~ 0.26 nm corresponds to the (002) plane of the hexagonal wurtzite ZnO structure (JCPDS card No. 36-1451), in agreement with earlier reports (Ref 33,34). The nanoparticles display quasi-spherical morphology in (Fig. 7a-b) with an average particle size of ~ 26 nm, in (Fig. 7c), suggesting that biomolecules present in the plant extract regulate the nucleation and growth process, thereby limiting excessive agglomeration (Ref 35). The SAED pattern further supports the high crystallinity of the ZnO nanoparticles, showing sharp and concentric diffraction rings that can be indexed to the (100), (002), (101), and (110) planes of wurtzite ZnO in (Figure (7d)). The absence of diffuse halos in the diffraction pattern confirms the negligible presence of amorphous phases. Together, the HRTEM and SAED results provide direct evidence that the green synthesis method yields highly crystalline ZnO nanoparticles with well-defined lattice ordering, making them suitable for applications where crystallinity plays a key role, such as photocatalysis and optoelectronics (Ref 36).

The photocatalytic activity of the green-synthesized ZnO nanoparticles (Z2) was evaluated using three organic dyes—methylene blue (MB), rhodamine B (RhB), and crystal violet (CV)—under visible light irradiation. The UV–Vis absorption spectra (Figure (8a–c)) revealed a gradual decrease in the main absorption bands of all three dyes with increasing irradiation time, confirming the photocatalytic degradation of the pollutants. The normalized concentration plots (C/C₀ vs. time, (Figure (8d–f)) indicated that ZnO effectively decomposes all dyes, with CV exhibiting the fastest decline, followed by MB, while RhB showed the slowest degradation (Figure (8g–i)) Quantitative analysis using the pseudo-first-order Langmuir–Hinshelwood model (Figure (8j–l)) yielded apparent rate constants of 0.00957 min⁻¹ for MB, 0.00631 min⁻¹ for RhB, and 0.0122 min⁻¹ for CV, corresponding to half-lives of ~ 72.4 min, ~ 109.8 min, and ~ 56.8 min, respectively. The linearity of the ln(C₀/C) versus time plots validates the first-order kinetic model, confirming that the process is surface-reaction limited rather than diffusion-controlled. The higher rate constant of CV compared to MB and RhB suggests stronger adsorption and interaction at oxygen-vacancy sites and surface defect states of ZnO, which facilitate charge transfer and enhance reactive oxygen species (ROS) generation (Ref 37–39). In the case of RhB, the slower degradation rate is consistent with previous studies and can be attributed to a more complex degradation pathway involving stepwise N-deethylation, which produces intermediate species and leads to a characteristic blue shift in the absorption spectrum during irradiation (Ref 40). These findings highlight the importance of surface defect-mediated processes in ZnO photocatalysis, where oxygen vacancies play a crucial role in prolonging charge carrier lifetime and enhancing ROS formation (Ref 41). Overall, the green-synthesized ZnO nanoparticles demonstrate efficient photocatalytic performance against structurally diverse dyes, supporting their potential application in wastewater treatment, while the obtained rate constants are comparable to or higher than those reported for other plant-extract-assisted ZnO systems (37, 38).

The photocatalytic performance of the green-synthesized ZnO nanoparticles (Z4) was investigated against three model dyes—methylene blue (MB), rhodamine B (RhB), and crystal violet (CV)—under visible light irradiation. The UV–Vis absorption spectra (Figure (9a–c)) showed a progressive decrease in the characteristic absorption peaks of the dyes with increasing irradiation time, confirming efficient photocatalytic degradation. The concentration decay curves (C/C₀ vs. time, (Figure (9d–f)) and degradation efficiency plots (Figure (9g–i)) further demonstrated the activity of ZnO, with MB undergoing the fastest degradation, followed by CV and RhB. Kinetic fitting of the data with the pseudo-first-order Langmuir–Hinshelwood model (Figure (9j–l)) yielded apparent rate constants of 0.01526 min⁻¹ for MB, 0.00674 min⁻¹ for RhB, and 0.00771 min⁻¹ for CV, corresponding to half-lives of ~ 45.4 min, ~ 102.8 min, and ~ 89.9 min, respectively. The linearity of the ln(C₀/C) versus time plots validates the first-order model, indicating that the degradation is surface-reaction limited rather than diffusion-controlled. The comparatively higher rate of MB degradation may be attributed to its stronger adsorption and charge transfer interactions at ZnO defect sites. In contrast, the slower degradation of RhB is consistent with its multi-step N-deethylation pathway, which produces intermediate species before full mineralization. These findings highlight the role of oxygen vacancies and defect-mediated charge separation in ZnO photocatalysis (Ref 37–41). Importantly, the obtained rate constants are comparable to or higher than those reported for other plant-extract-assisted ZnO nanoparticles, such as hibiscus-extract ZnO (k ≈ 0.004–0.008 min⁻¹ for MB) (Ref 43), and hydrothermal green-synthesized ZnO (k ≈ 0.005–0.010 min⁻¹ for MB and RhB) (Ref 42), underlining that the present bio-synthesized ZnO nanoparticles exhibit competitive or superior photocatalytic efficiency. Taken together, these results demonstrate that green ZnO systems not only provide a sustainable synthetic route but also deliver high photocatalytic activity, supporting their practical potential in wastewater treatment applications.

A comparison of the photocatalytic performance of ZnO nanoparticles synthesized with saffron anther extract (Z2) and saffron petal extract (Z4) highlights the strong influence of the plant precursor on degradation kinetics. For MB, Z4 exhibited a significantly higher rate constant (0.01526 min⁻¹, t½ ≈ 45.4 min) compared to Z2 (0.00957 min⁻¹, t½ ≈ 72.4 min), suggesting that the petal-derived ZnO possesses more favorable surface states for MB adsorption and charge transfer.

In contrast, Z2 outperformed Z4 in CV degradation, with k = 0.0122 min⁻¹ (t½ ≈ 56.8 min) versus 0.00771 min⁻¹ (t½ ≈ 89.9 min), indicating that the anther-derived ZnO provides more active defect sites for interaction with CV molecules. For RhB, both samples showed comparable activity (Z2: 0.00631 min⁻¹; Z4: 0.00674 min⁻¹), consistent with its multi-step N-deethylation mechanism that depends less on surface chemistry. These results demonstrate that even under identical co-precipitation conditions, the phytochemical composition of the chosen extract (anther vs. petal) can tailor ZnO nanoparticle properties, leading to dye-specific variations in photocatalytic efficiency. Such tunability underscores the importance of extract selection in green synthesis strategies for optimizing ZnO-based photocatalysts in wastewater treatment.