One-Step Plasma–Solution Synthesis of Prussian Blue and Copper Hexacyanoferrate Composites for Selective Photocatalytic Dye Degradation

Abstract

This work presents a novel one-step plasma–solution synthesis of Prussian Blue (PB) and copper hexacyanoferrate (Cu-PBA) nanoparticles via underwater pulsed DC discharge. For the first time, the direct plasma-assisted formation of these coordination polymers is reported. The obtained materials were examined by X-ray diffraction, Fourier-transform infrared spectroscopy, Raman spectroscopy, and scanning electron microscopy (SEM). These analyses confirmed that the desired phases had formed, along with small amounts of oxide byproducts (α-Fe₂O₃, CuO) arising from the erosion of the electrodes. Photocatalytic activity was evaluated through the degradation of organic dyes (Reactive Red 6C, Rhodamine B, and Methylene Blue) under UV-light irradiation. Both catalysts achieved complete dye degradation within 90 min of UV irradiation (after an initial 30 min dark adsorption step, total experiment time 120 min). Notably, selective performance was observed: PB exhibited higher activity toward the cationic dye Methylene Blue, while Cu-PBA was more effective for the anionic dye Reactive Red 6C. This selectivity is attributed to the specific oxide impurities forming heterojunctions that facilitate charge separation and generate distinct reactive oxygen species. The plasma–liquid method offers a rapid and environmentally benign route to functional PBA-based composites, with potentially scalable characteristics pending further engineering optimization. These findings highlight the potential of utilizing synthesis-induced impurities to tailor photocatalytic selectivity for water purification applications.

1. Introduction

First prepared in the early 1700s, Prussian blue (PB), with the general formula Fe₄[Fe(CN)₆]₃·nH₂O, is the oldest synthetic coordination compound known to date [1,2]. Its unique mixed-valence structure, where Fe³⁺ and Fe²⁺ ions are bridged by cyanide ligands (−C≡N–C≡N−), forms a rigid open framework with a face-centered cubic lattice [3,4]. This structure exhibits remarkable tolerance to isomorphous substitution of iron ions by other transition metals (e.g., Co, Ni, Cu, Mn, Zn). This type of substitution gives rise to an extensive class of compounds called Prussian blue analogues (PBAs). Their general formula is A_x[Fe(CN)₆]₂·nH₂O, where A represents an alkali metal and M stands for a transition metal [5,6]. The ability to tune the composition by varying the metal centers allows for precise control over the electronic, optical, and magnetic properties of these materials [7,8]. Depending on the choice of transition metal and synthesis conditions, PBAs can crystallize in different symmetries, including cubic, monoclinic, rhombohedral, and tetragonal phases, which further diversify their functionality [9,10].

The synthesis of PB and PBAs has been extensively studied, with co-precipitation being the most conventional and widely used route [11,12]. This method typically involves the slow addition of a solution containing transition metal ions (e.g., Fe³⁺, Cu²⁺, Co²⁺) to a hexacyanoferrate solution, resulting in a precipitate [13,14]. The reaction is usually carried out at room temperature and the product is collected by centrifugation or filtration, followed by washing and drying [15]. Although straightforward and cost effective, the co-precipitation method tends to produce particles with a broad size distribution and requires meticulous control of parameters like concentration, temperature, pH, and mixing rate to obtain the desired morphology and crystallinity [16]. For instance, Arbenin et al. demonstrated that by simply changing the divalent metal ion (Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺) in the synthesis, they could obtain PBAs with different crystal structures (cubic for most metals, but trigonal for Zn) and particle sizes ranging from nanometers to micrometers [17]. This highlights the sensitivity of the co-precipitation process to the nature of the precursors. To address the limitations of conventional co-precipitation, a variety of alternative strategies have been developed. Hydrothermal and various involve heating the reaction mixture in a sealed vessel at elevated temperatures and pressures, allowing for better control over crystallinity and yielding uniform nanocubes with well-defined morphologies [18,19]. Electrochemical deposition is another powerful technique, particularly for fabricating thin films of PBAs on conductive substrates for sensor or battery applications [20,21]. This method offers precise control over film thickness and morphology by adjusting the applied potential or current density. Microemulsion techniques, which use surfactants to create nanoscale reaction droplets, have been successfully employed to produce ultra-small PBA nanoparticles [22]. Additionally, template-assisted synthesis using mesoporous silica or carbon matrices can generate materials with hierarchical porosity and enhanced surface area [23,24]. More recently, mechanochemical synthesis (ball milling) has emerged as a solvent-free, environmentally friendly alternative for rapid PBA production [25]. Mechanochemical synthesis has been further advanced by Wang et al. [26], who demonstrated scalable production of high-quality PBAs for potassium-ion batteries with high yield and crystallinity. Beyond direct PBA synthesis, recent studies have focused on producing functional derivatives from PBA precursors via thermal treatment. For example, Yang et al. [27] successfully synthesized MO/MFe₂O₄ nanoparticles derived from Prussian blue analogues, demonstrating enhanced lithium storage performance. Despite the diversity of available methods, many of them are multi-step, time-consuming, or require the use of organic solvents, surfactants, or complex instrumentation, which complicates the purification process and limits their scalability and environmental friendliness [28]. The search for a simple, rapid, and sustainable synthetic route that allows good control over particle properties remains an ongoing challenge.

PB and PBAs have drawn considerable attention due to their adaptability and diverse uses. These properties stem from their porous structure, large surface area, and ability to undergo reversible redox reactions [16]. Recent advances in the field have been comprehensively reviewed by Zheng et al. [29], who summarized synthesis strategies and emerging applications of PBA derivatives, including photocatalysis, energy storage, and sensing. One of the most prominent areas of research is electrochemical energy storage, where PBAs are extensively studied as cathode materials for rechargeable batteries. Their rigid open framework, featuring interstitial sites, enables reversible intercalation and deintercalation of not only ions such as Na⁺, Li⁺, but also ions such as Mg²⁺, Ca²⁺, and Al³⁺ [30,31,32,33]. This makes them promising candidates for beyond-lithium-ion battery technologies.

Another major area of application is environmental remediation, particularly the removal of hazardous radionuclides from aqueous solutions. The ion-exchange properties of PBAs make them exceptionally efficient sorbents for Cs⁺, Sr²⁺, and Co²⁺, which are common components of radioactive nuclear waste [34,35,36]. The selectivity for Cs⁺ is especially high due to the nearly perfect size matching between the Cs⁺ ion (3.25 Å) and the channel windows of the PBA framework (approx. 3.2 Å) [37]. Le et al. compared the sorption performance of Cu₂[Fe(CN)₆], Co₂[Fe(CN)₆], and Ni₂[Fe(CN)₆] and found that the copper analogue exhibited the highest adsorption capacity for all three ions, which was attributed to the weaker binding of Cu²⁺ to the cyanide ligands, making it more amenable to ion exchange [38]. PBAs have also been explored for the removal of organic dyes and heavy metals from wastewater [39,40].

The photocatalytic potential of PB and PBAs has gained increasing attention [41,42]. While pristine PBAs can absorb visible light due to intervalence charge transfer transitions, their photocatalytic activity is often limited by rapid electron–hole recombination [43]. A highly effective strategy to overcome this limitation is to couple PBAs with a suitable semiconductor, such as TiO₂, ZnO, or WO₃, to form a heterojunction [44,45]. In such hybrid systems, the PBA can act as an efficient co-catalyst. Upon photoexcitation of the semiconductor, photogenerated electrons or holes can be transferred to the redox-active centers of the PBA. This spatial separation of charge carriers significantly suppresses recombination, while the PBA provides active sites for surface redox reactions [46]. For instance, Phul et al. demonstrated that coating ZnO origami nanostructures with a CoFe-PBA shell creates a p-n heterojunction, which dramatically improves the oxygen evolution reaction (OER) under solar irradiation, achieving a 2.4-fold increase in activity compared to bare ZnO [47]. Similarly, Vaijayanthi et al. reported that a Ce-PBA/N-doped carbon composite exhibited excellent photocatalytic degradation of multiple pollutants (ciprofloxacin, Cr(VI), and crystal violet) under visible light, with the PBA framework providing the redox-active Ce³⁺/Ce⁴⁺ and Fe²⁺/Fe³⁺ centers [48]. These studies underscore the synergistic effects achievable by combining PBAs with other functional materials, opening up new avenues for the design of advanced photocatalysts.

Despite these advances, the synthesis of PBA-based functional materials, especially hybrids for photocatalysis, typically involves multiple stages: first, the synthesis of the support or host material, followed by a separate step for PBA deposition or growth. This multi-step nature can be time-consuming and may lead to issues with interfacial contact or non-uniform coating.

This study presents a novel one step procedure for producing Prussian blue and its copper-based counterpart (Cu₂[Fe(CN)₆]) via plasma solution synthesis. The process takes advantage of the highly reactive conditions created by atmospheric pressure discharge plasma in contact with a liquid medium to facilitate the formation of these complex materials. To our knowledge, no previous study has reported the direct single-step plasma-assisted synthesis of PB and PBA nanoparticles. This method removes the necessity for multiple processing stages, aggressive reagents, or extended reaction durations, providing a fast and environmentally friendly route to these valuable materials with potential for scalability. Detailed characterization of the resulting particles is carried out using XRD, FTIR, Raman, and SEM, followed by a discussion of the proposed formation pathway. This work demonstrates that plasma solution synthesis can serve as a powerful and flexible technique for preparing functional coordination polymers intended for photocatalytic and other advanced uses, possibly allowing the direct in situ fabrication of PBA-based hybrid systems.

2. Materials and Methods

The synthesis of Prussian blue and its copper analogue was carried out in a plasma chemical cell employing a pulsed underwater DC discharge. The discharge medium consisted of an aqueous solution of potassium hexacyanoferrate at a concentration of 0.5 g/L. Two series of experiments were conducted. In the first series, a discharge was generated between iron electrodes, while in the second series, copper electrodes were used. Both electrode types were made from 1 mm diameter wires with 99.99% purity, supplied by Tangda (Shenzhen Tangda Technology Co., Ltd., Shenzhen, China). A detailed description of the experimental apparatus can be found in reference [49]. To initiate the discharge, a BP 0.25 2 DC power source (TD ARS THERM, Novosibirsk, Russia) capable of delivering up to 10 kV was employed together with a 1000 Ω ballast resistor. Emission spectra were acquired using an AvaSpec 3648 spectrometer (Avantes B.V., Apeldoorn, The Netherlands) fitted with a 1200 lines/mm diffraction grating and a 25 μm slit. The optical detector (Avantes B.V., Apeldoorn, The Netherlands), which was equipped with a collimating lens, was placed 5 cm away from the plasma region. Voltage and current profiles were captured with an ADS 2072 digital multichannel oscilloscope (AO ANP ELIKS, Moscow, Russia). The discharge was operated at a current of 0.25 A for a duration of 30 min. After the discharge was switched off, the solution was stirred for one hour to allow the reaction to reach completion. The resulting precipitate was then collected by centrifugation, washed twice, and centrifuged again at 6000 rpm for 15 min using a UC 6000E centrifuge (ULAB, Nanjing, China). Finally, the precipitates were dried at 80 °C for one hour.

The product yield was assessed using a gravimetric approach. Following drying, the powders of Fe₄[Fe(CN)₆]₃ and Cu₂[Fe(CN)₆] were weighed on an analytical balance with a precision of ±0.1 mg. The yield was computed as the actual product mass divided by the theoretical mass derived from the reaction stoichiometry, assuming full consumption of the limiting reagent to form the corresponding hexacyanoferrate. The measured yield for Prussian blue was 75%, whereas the copper analogue gave a yield of 79%.

X ray diffraction (XRD) was employed to determine the phase composition of the synthesized materials. A D2 Advance diffractometer (Bruker, Billerica, MA, USA) with CuKα radiation was used, and data were collected across a 2θ range from 10° to 70° with a step size of 0.02°.

Fourier transform infrared (FTIR) spectra were recorded on a VERTEX 80v spectrometer (Bruker Optics, Billerica, MA, USA) covering the 4000-to-400 cm⁻¹ region. The powder samples were mixed with KBr (99.99% IR grade, Acros Organics, Medley, FL, USA) at a 1:100 ratio by weight and then compressed into pellets for measurement.

Raman measurements were conducted using a Confotec NR500 system (Sol Instruments, Minsk, Belarus) equipped with a 532 nm solid-state diode laser operating at 25 mW. The spectrometer included a specialized attachment that allowed spectra acquisition at controlled sample temperatures.

Surface morphology was examined by scanning electron microscopy (SEM) using a Quattro S instrument (Thermo Fisher Scientific, Brno s.r.o., Czech Republic). The specific surface area was measured with a Nova Series 1200E analyzer (Quantachrome, Boynton Beach, FL, USA) based on nitrogen adsorption–desorption at 77 K.

XPS measurements were carried out on a Multiprobe RM ultra-high vacuum system (Omicron Electronics Deutschland GmbH, Erlangen, Germany). The analysis chamber maintained a residual gas pressure below 10⁻¹⁰ mbar to avoid surface contamination and oxidation. Photoelectrons were excited using MgKα X-rays (1253.6 eV) and collected with a hemispherical energy analyzer (125 mm radius, 1.5 mm analysis spot, Omicron Electronics Deutschland GmbH, Erlangen, Germany). The analyzer was operated in constant transmission mode (50 eV) with an absolute resolution of ≤0.25 eV, using a 6 mm circular input aperture and five 3 × 10 mm rectangular output slits.

Aqueous suspension absorption spectra were recorded on an SF-56 spectrophotometer (OKB Spectr, Saint Petersburg, Russia) in the 200–1000 nm range.

Zeta potential was measured using a Zetasizer Nano ZS (Malvern Panalytical, Malvern, WR14 1XZ, UK). The synthesized powders (0.5 mg/mL) were dispersed in deionized water by ultrasonication for 10 min. Each measurement was repeated five times, and averages are reported with standard deviations.

Fluorescence spectra were acquired on a FluoTime 300 fluorometer (PicoQuant GmbH, Berlin, Germany). For comparison, a reference Prussian blue sample free of iron oxide impurities was prepared via conventional co-precipitation. A 0.1 M FeCl₃·6H₂O solution (50 mL) was added dropwise to a 0.1 M K₄[Fe(CN)₆]·3H₂O solution (50 mL) under vigorous stirring at room temperature. The resulting dark blue precipitate was collected by centrifugation, washed three times with deionized water, and dried at 80 °C for 12 h. This reference sample is denoted as “PB-ref”.

The photocatalytic performance of the prepared samples was assessed by monitoring dye degradation in an aqueous solution under dark conditions and UV irradiation. A dye mixture was used, containing Reactive Red 6C (RR6C, anionic, λ_max = 533 nm), Rhodamine B (RhB, zwitterionic xanthene, λ_max = 554 nm), and Methylene Blue (MB, cationic thiazine, λ_max = 667 nm), each at 1.2 mg L⁻¹. The reaction took place in an 800 mL cylindrical reactor equipped with a water-cooled quartz jacket. UV light was supplied by a 250 W high pressure mercury lamp (Shenzhen Chengxing Guangjin Precision Technology, Shenzhen, Guangdong, China, peak emission at 365 nm). For each test, 0.03 g of the synthesized powder was added to 800 mL of the dye solution. Dye concentration changes were measured using an SF-56 spectrophotometer (OKB Spektr, Saint Petersburg, Russia) across 350–750 nm. The degradation efficiency for each dye was calculated using Equation (1):

$$Dye removal\left(\%\right)={\displaystyle \frac{C_0-C_t}{C_0}}·100\%,$$

(1)

where C₀ is the initial dye concentration and C_t is the concentration after a given irradiation time.

3. Results and Discussions

3.1. Optical Emission Spectroscopy of Discharge

Optical emission spectroscopy (OES) was used to detect the species produced during a pulsed underwater discharge and to understand the electrode erosion mechanism. Figure 1 shows the emission spectra of the discharge in a K₃[Fe(CN)₆] solution with copper and iron electrodes, recorded across the 200–900 nm range. The emission spectra for both electrode types display a similar set of lines across most of the range, indicating a dominant contribution from the solution components to the emission spectrum. The most intense lines in the visible (350–450 nm) regions correspond to iron atoms [50]. Their presence is expected and results from the dissociation and excitation of [Fe(CN)₆]³⁻ hexacyanoferrate ions within the plasma. All recorded spectra feature atomic hydrogen lines H_α (656 nm) and H_β (486 nm), along with atomic oxygen lines at 777 nm and 844 nm [50], which points to efficient water molecule dissociation within the discharge region. The hydroxyl radical band at 306 nm is also observed, confirming the formation of reactive oxygen species [51]. Near-infrared potassium lines (766–770 nm) are present in the discharge emission spectra, consistent with the precursor composition [50]. The most notable difference between the spectra appears in the 320–330 nm range. In the discharge spectrum with copper electrodes, two intense lines at 324.7 and 327.4 nm correspond to resonance transitions of atomic copper [50]. These lines are completely absent in the spectrum with iron electrodes. This observation directly confirms the erosion of copper electrodes under the pulsed discharge, resulting in copper atoms entering the plasma phase, becoming excited, and emitting characteristic radiation. Conversely, in the discharge with iron electrodes, the intensity of the iron lines is somewhat higher than in the copper electrode discharge. Additionally, the main iron lines (e.g., 372 nm) may overlap with intense iron ions lines originating from the solution. Thus, OES analysis confirms that the plasma generates a broad spectrum of active species (OH, H, O) capable of initiating chemical reactions, while electrode erosion releases metal ions into the solution, which participate in the formation of the target hexacyanoferrate phases.

3.2. Optical Properties of Samples

The optical properties of the synthesized materials were examined by UV-visible spectroscopy. Figure 2a shows the absorption spectra of Fe₄[Fe(CN)₆]₃ and Cu₂[Fe(CN)₆] over the 200–1000 nm range. Prussian blue exhibits a broad absorption band from 350 to 550 nm, peaking near 400 nm (Figure 2a). This band arises from intervalence charge transfer between Fe²⁺ and Fe³⁺ ions connected via cyanide bridges [52]. Absorption in the red region is negligible. In the UV region (200–350 nm), an intense band appears, which is assigned to ligand to metal charge transfer transitions within the hexacyanoferrate anion [53]. For the copper analogue Cu₂[Fe(CN)₆], a significant change in optical properties is evident. The absorption edge shifts to approximately 370 nm, and the intense intervalence charge transfer (IVCT) band characteristic of Prussian blue is absent. This is explained by the fact that in the Cu₂[Fe(CN)₆] structure, all iron ions are in the +2 oxidation state, while copper ions are also in the +2 state, resulting in the absence of the mixed valence necessary for the low-energy IVCT transition. The primary contribution to absorption in the visible region arises from ligand-to-metal charge transfer transitions and possibly d–d transitions in Cu²⁺ ions. Additionally, the absorption spectrum of Cu₂[Fe(CN)₆] exhibits a small peak around 380–400 nm. According to the literature, CuO nanoparticles display an absorption edge in the 350–450 nm range [54]. Therefore, the peak at 380–400 nm confirms the formation of a CuO impurity phase during plasma–liquid synthesis with copper electrodes.

The band gap (E_g), determined by the Tauc method for direct allowed transitions, was found to be 3.4 eV for Fe₄[Fe(CN)₆]₃ and 2.8 eV for Cu₂[Fe(CN)₆] (Figure 2b). The value measured for the copper analogue matches well with previously reported results for copper hexacyanoferrate (2.5–2.8 eV) [53], thereby confirming that the desired phase has been obtained. However, for Prussian blue, the E_g of 3.4 eV significantly exceeds the values reported for bulk samples (1.7–2.3 eV) [52,55]. To resolve this discrepancy, we performed additional analyses. First, an indirect transition Tauc plot for Fe₄[Fe(CN)₆]₃ was constructed (Figure 2c), which revealed a linear region corresponding to E_g = 2.16 eV, consistent with literature values [56]. Second, XPS analysis (Section 3.5) showed a Fe²⁺:Fe³⁺ ratio of 0.75:1 in the plasma-synthesized sample, deviating from the ideal 1:1 stoichiometry of Prussian blue. This deviation suppresses the low-energy intervalence charge transfer responsible for the ~1.9 eV absorption, while higher-energy ligand-to-metal charge transfer (LMCT) transitions dominate the spectrum, resulting in an apparent E_g of 3.4 eV in the direct transition plot. We therefore conclude that the 3.4 eV value does not represent a true band gap but rather reflects the suppression of IVCT due to non-stoichiometry induced by the non-equilibrium plasma synthesis conditions. The indirect transition model (E_g = 2.16 eV) offers a more precise characterization of the electronic structure of Prussian blue produced via plasma synthesis.

Quantum confinement effects are ruled out, as the crystallite size (24 nm) and particle size (50–200 nm) exceed the typical regime for quantum confinement in Prussian blue.

3.3. Structural Features of Synthesized Composites

The plasma–solution synthesis method demonstrated high efficiency in producing both materials. Gravimetric analysis of the final products after washing and drying revealed a yield of 75% for Fe₄[Fe(CN)₆]₃ and 79% for Cu₂[Fe(CN)₆]. The slightly higher yield of the copper analogue may be attributed to differences in the kinetics of the plasma-induced erosion of the copper electrode or to the higher rate of complexation of Cu²⁺ ions with the hexacyanoferrate anions.

X-ray diffraction was employed to examine the phase composition and crystal structure of the products from plasma–solution synthesis. Figure 3a displays the XRD patterns for the materials prepared using iron (Fe-PB) and copper (Cu-PBA) electrodes. A detailed examination of the peak positions and intensities confirms the successful formation of the target Prussian Blue analogues, although minor impurities resulting from electrodes erosion are also present.

The XRD pattern of the material synthesized using iron electrodes displays a set of diffraction peaks characteristic of classical Prussian Blue, Fe₄[Fe(CN)₆]₃ [3]. The positions of the main peaks at 17.5°, 24.7°, 35.2°, 39.5°, 43.8°, 50.8°, and 54.5° are nearly identical to those of the copper analogue and confirm the formation of the isostructural cubic framework [14]. The peak at ~24.7° is the most intense, corresponding to the (220) reflection of the PB structure.

Similar to the copper-based synthesis, the diffraction peak of the unreacted K₃[Fe(CN)₆] precursor at ~13.1° is completely absent, confirming the effectiveness of the washing.

A more detailed inspection of the pattern reveals notable asymmetry and broadening of the peak at approximately 34°. The position of the shoulder at ~34.1° is close to the reflection of several iron oxides, including hematite (α-Fe₂O₃) [57]. Given the reductive/oxidative nature of the plasma environment and the erosion of iron electrodes, the formation of an iron oxide, such as hematite, is highly plausible [57]. Peaks at 35.4°, 49.5°, 56.5° can also be attributed to the hematite phase.

The XRD pattern of the product obtained using copper electrodes exhibits a series of well-defined diffraction peaks that can be indexed to a face-centered cubic (fcc) structure with the space group Fm-3m, characteristic of Prussian Blue analogues [58]. The most intense peaks appear at 2θ angles of approximately 17.5°, 24.8°, 35.2°, 39.7°, 43.8°, 50.8°, and 54.5°. These reflections are indexed to the (200), (220), (400), (420), (422), (440), and (600) planes, respectively, and are consistent with previously published patterns for copper hexacyanoferrate [14]. The presence of these sharp peaks confirms that the target compound is the primary crystalline phase in the sample. Notably, the characteristic diffraction peak of the precursor salt, potassium hexacyanoferrate, typically located at 13.1° [4], is absent from the pattern. This unequivocally demonstrates the high efficiency of the post-synthesis washing procedure in removing unreacted starting materials.

However, a careful examination of the pattern reveals a small but distinct peak at approximately 35.6°, accompanied by a weak peak near 38.8°. These reflections are not typical of the cubic PBA structure. The peak at around 38.8° corresponds to the second most intense line of tenorite, the monoclinic phase of copper(II) oxide (CuO) [58]. Its presence, along with the peak at approximately 35.6°, indicates the formation of a copper oxide impurity. This is a direct consequence of plasma-induced erosion of the copper electrodes during synthesis, leading to the incorporation of Cu²⁺ ions that partially precipitate as an oxide phase [59].

For a cubic system, the interplanar spacing is related to the lattice parameter a and the Miller indices (hkl) by the Wulff–Bragg equation:

$${\displaystyle \frac{1}{d_{hkl}^2}}={\displaystyle \frac{h^2+k^2+l^2}{a^2}}$$

(2)

The lattice parameter can be calculated as:

$$a=d_{hkl}·\sqrt{h^2+k^2+l^2}$$

(3)

The average lattice parameter for Fe₄[Fe(CN)₆]₃ is a = 10.17 ± 0.02 Å, and for Cu₂[Fe(CN)₆], it is a = 10.12 ± 0.02 Å. These values are in good agreement with the literature data for the corresponding compounds [3,4]. The decrease in the lattice parameter upon substituting Cu²⁺ for Fe³⁺ is attributed to the smaller ionic radius of Cu²⁺ (0.73 Å) compared to that of high-spin Fe³⁺ (0.78 Å).

The average crystallite size (i.e., the coherent scattering regions) was estimated using the Scherrer formula:

$$D= {\displaystyle \frac{K\lambda }{\beta cos\theta }}.$$

(4)

In this equation, K represents the dimensionless shape factor (taken as 0.9 for spherical particles), λ denotes the CuK_α radiation wavelength (1.5406 Å), β is the peak’s full width at half maximum expressed in radians, and θ stands for the Bragg angle. The most intense reflection, (220), was used for the calculations, as it is least affected by instrumental factors. For Fe₄[Fe(CN)₆]₃, the average crystallite size is 24 nm, whereas for Cu₂[Fe(CN)₆], it is 26 nm. These values confirm the nanocrystalline nature of the synthesized materials, which is characteristic of plasma–liquid synthesis.

The XRD analysis confirms that the one-step plasma–solution synthesis is a viable method for producing both copper hexacyanoferrate and classical Prussian Blue. A common and noteworthy feature for both syntheses is the unavoidable formation of a metal oxide phase, which originates from the erosion of the metal electrodes by the plasma discharge. This phenomenon is well-documented in plasma–liquid systems, where electrode material is sputtered and subsequently oxidized in the reactive environment [59].

The chemical composition and structural characteristics of the synthesized compounds were analyzed using IR spectroscopy. Figure 3b presents the IR absorption spectra of classical Prussian blue, synthesized via underwater discharge with iron electrodes, and its copper analogue, obtained using copper electrodes in a plasma–solution system. In the high-frequency region of both spectra, a broad absorption band appears between 3200 and 3600 cm⁻¹, corresponding to the stretching vibrations of O–H groups (ν(O–H)), along with a band near 1600–1615 cm⁻¹, attributed to the bending vibrations of water molecules (δ(H–O–H)). The presence of these bands indicates the hydrated nature of the synthesized coordination polymers, characteristic of Prussian blue-type compounds [15]. The most informative region of the IR spectrum for hexacyanoferrates lies between 2000 and 2200 cm⁻¹, where the stretching vibrations of the cyano groups (ν(C≡N)) are observed [60]. The spectrum of Fe₄[Fe(CN)₆]₃ exhibits an intense band with a maximum absorption at approximately 2086 cm⁻¹, corresponding to vibrations of the bridging cyanide ligands that connect Fe²⁺ and Fe³⁺ ions [61]. In the spectrum of the copper analogue Cu₂[Fe(CN)₆], this band undergoes broadening and a change in profile within the 2000–2200 cm⁻¹ range, reflecting the coordination of Cu²⁺ ions to the nitrogen atoms of the cyano groups and indicating the formation of the target phase [61]. The low-wavenumber region of the spectra shows bands associated with stretching and bending modes of metal ligand bonds [62]. In both samples, characteristic bands appear at approximately 599 cm⁻¹ and 471 cm⁻¹. These are assigned to Fe–C stretching vibrations (ν(Fe–C)) and δ(Fe–CN) bending vibrations, respectively [63], confirming that the hexacyanoferrate anion is present in both structures. Differences appear in the 800–900 cm⁻¹ range: the Cu₂[Fe(CN)₆] sample shows a band at approximately 855 cm⁻¹, associated with vibrations of the Cu–N≡C bond [64], whereas the Fe₄[Fe(CN)₆]₃ spectrum contains a band near 808–810 cm⁻¹, attributed to vibrations of the Fe–N bonds within the bridging structures of Prussian blue [65]. A band at approximately 1385 cm⁻¹, observed in both spectra with comparable intensity, is assigned to the stretching vibrations of carbonate ions [66]. This effect likely arises from the absorption of atmospheric carbon dioxide by the solution under conditions of a local pH increase induced by the plasma discharge. In summary, the IR spectroscopy data confirm that the plasma–solution method enables the synthesis of both Prussian blue and a copper hexacyanoferrate(II) phase with a structure analogous to Prussian blue, wherein Fe³⁺ ions at lattice sites are replaced by Cu²⁺ ions.

Additional information about the structure and bonding in the synthesized compounds was obtained using Raman spectroscopy. Figure 3c presents the Raman spectra of Fe₄[Fe(CN)₆]₃ and Cu₂[Fe(CN)₆] within the 240–2500 cm⁻¹ range.

In the high wavenumber range, the spectra of both samples are primarily characterized by strong bands arising from cyano group stretching vibrations (ν(C≡N)) [67]. A notable feature is the appearance of two separate bands in this region for each material, suggesting the presence of crystallographically distinct cyanide bonds in the structure.

For the original Prussian blue, Fe₄[Fe(CN)₆]₃, bands are observed at approximately 2084 cm⁻¹ and 2129 cm⁻¹. The lower-frequency component (~2084 cm⁻¹) is characteristic of C≡N vibrations in Fe²⁺–CN–Fe³⁺ bridging fragments, while the higher-frequency band may be attributed to the presence of [Fe(CN)₆]⁴⁻ groups in a different local environment or to Fe³⁺–CN–Fe²⁺ inverse bridges [68].

The spectrum of the copper analogue Cu₂[Fe(CN)₆] also exhibits two bands; however, their positions are significantly shifted to approximately 2125 cm⁻¹ and 2170 cm⁻¹. The most notable feature is the emergence of the high-frequency component at around 2170 cm⁻¹. This shift toward higher frequencies clearly indicates the coordination of Cu²⁺ ions to the nitrogen atoms of the cyano groups, resulting in the formation of Cu²⁺–N≡C–Fe²⁺ fragments [69]. The higher electronegativity of Cu²⁺ compared to Fe²⁺/Fe³⁺ causes electron density withdrawal from the cyanide ion, which increases the force constant of the C≡N bond and, consequently, raises the stretching vibration frequency.

The low-frequency region of the spectra also contains important information. Both samples exhibit a band at approximately 475 cm⁻¹, attributed to deformation vibrations of the Fe–C≡N bonds (δ(Fe–CN)) [70]. A key difference is observed in the 500–550 cm⁻¹ region: the Cu₂[Fe(CN)₆] spectrum features a characteristic band at around 535 cm⁻¹, which is absent in the Fe₄[Fe(CN)₆]₃ spectrum. This band can be confidently assigned to the stretching vibrations of the Cu–N bond (ν(Cu–N)), formed between copper ions and cyanide bridges [70]. The band at approximately 270 cm⁻¹, present in both spectra, corresponds to deformation vibrations of the crystal lattice [4].

Although the diffraction data indicate the presence of a small amount of oxide phases (CuO in the sample with copper electrodes and, presumably, α-Fe₂O₃ in the sample with iron electrodes), the corresponding bands are not observed in the Raman spectra. This can be explained, first, by the resonance enhancement of the signal from the vibrations of the cyanide groups in hexacyanoferrates, which dominates the Raman spectra [70]. Second, oxide phases formed due to electrode erosion may exist as nanosized or poorly crystallized inclusions, whose Raman signals are significantly weaker and broader compared to those from the highly ordered PBA structure. Thus, the combination of XRD and Raman spectroscopy provides a comprehensive understanding of the phase composition: XRD analysis reveals the presence of crystalline impurities, while Raman spectroscopy confirms that the main and dominant phase in the samples is the target hexacyanoferrate.

Thus, the Raman spectroscopy data not only correlate with the IR spectroscopy results but also provide additional evidence for the incorporation of Cu²⁺ ions into the hexacyanoferrate structure. Furthermore, they indicate the presence of two types of cyanide bonds in both compounds studied.

3.4. Morphological Features of Synthesized Samples

The porous structure of the synthesized materials was analyzed using low-temperature nitrogen adsorption–desorption. The calculated textural properties are summarized in

Table 1. Texture parameters of samples.

Parameter	Sample
	Fe4[Fe(CN)6]3	Cu2[Fe(CN)6]
SBET (m2/g)	8.74	3.82
SBJH (m2/g)	6.57	2.87
Dpore (nm)	4.05	4.82
Vpore (cm3/g)	0.011	0.005

. The specific surface area (S_BET) was found to be 8.74 m²/g for Fe₄[Fe(CN)₆]₃ and 3.82 m²/g for Cu₂[Fe(CN)₆].

These values are significantly lower than the typical range for classical hexacyanoferrates synthesized by conventional methods (200–600 m²/g) [71]. This substantial difference can be attributed to the unique characteristics of plasma–liquid synthesis. First, the high nucleation and particle growth rates under nonequilibrium plasma conditions lead to the formation of small nanoparticles that tend to strongly agglomerate. Agglomeration into dense aggregates drastically reduces the surface area available for nitrogen adsorption and decreases the pore volume, as confirmed experimentally. Second, structural imperfections inherent to nonequilibrium synthesis can disrupt the regular pore network typical of ideal hexacyanoferrate crystals. Despite the low absolute values, a correlation exists between the textural properties and the phase composition of the samples. The higher specific surface area of Fe₄[Fe(CN)₆]₃ compared to Cu₂[Fe(CN)₆] aligns with X-ray diffraction data indicating the presence of a CuO impurity phase in the copper sample, which likely has a denser packing and further reduces porosity. The average pore diameter for both samples is approximately 4–5 nm, corresponding to the mesopore range and characteristic of nanoparticle aggregates [72]. The low pore volumes (0.011 and 0.005 cm³/g) further support the hypothesis of dense particle agglomeration. In summary, plasma–liquid synthesis enables the production of nanosized hexacyanoferrates; however, their textural characteristics are influenced more by particle morphology and agglomeration degree under nonequilibrium plasma discharge conditions than by the crystalline structure itself.

Scanning electron microscopy was used to investigate the morphology of the prepared samples. Figure 4 shows micrographs of the Fe₄[Fe(CN)₆]₃ and Cu₂[Fe(CN)₆] powders synthesized by the plasma–liquid approach. The microstructure of the PB sample is characterized by irregularly shaped agglomerates measuring 1–2 μm, composed of primary nanoparticles sized between 50 and 200 nm. The agglomerate surfaces are relatively smooth, with interparticle pores forming the material’s porous structure. This morphology is typical of nanostructured hexacyanoferrates synthesized under nonequilibrium conditions [72]. In contrast, the Cu₂[Fe(CN)₆] sample exhibits a different morphology: the primary particles have a more regular cuboid shape, ranging from 100 to 300 nm, and form agglomerates up to 2–3 μm in size. The particle surfaces display increased roughness, likely due to the presence of a CuO impurity phase on the surface, as identified by X-ray diffraction analysis. The more pronounced crystallite faceting aligns with the somewhat larger crystallite size calculated using the Scherrer formula, indicating a more ordered crystalline growth in the copper analog. These morphological differences may result from distinct crystallization mechanisms during the erosion of iron and copper electrodes in the plasma–liquid discharge, as well as the influence of impurity oxide phases on particle growth.

3.5. X-Ray Photoelectron Spectroscopy Studies of Samples

X-ray photoelectron spectroscopy was employed to determine the elemental composition and chemical states on the surfaces of the synthesized materials. Figure 5 displays the survey spectra of the PB and Cu-PBA samples recorded using a MgKα X-ray source. The spectrum for Fe₄[Fe(CN)₆]₃ shows peaks associated with C 1s, N 1s, O 1s, and Fe 2p. The presence of two components in the Fe 2p₃₂ spectrum at approximately 708.5 eV and 710.5 eV confirms the mixed-valence structure of classical Prussian blue, which contains Fe²⁺ ions as part of the [Fe(CN)₆]⁴⁻ anion and Fe³⁺ ions occupying cationic lattice sites [15]. In the Cu₂[Fe(CN)₆] spectrum, in addition to the C 1s, N 1s, O 1s, and Fe 2p peaks (with only the Fe²⁺ component at ~708.5 eV), intense peaks corresponding to Cu 2p_3/2 (~933 eV) and Cu 2p_1/2 (~953 eV) are observed, along with characteristic Cu 2p_3/2 satellite features in the 940–945 eV region [73]. These satellites are diagnostic of the Cu²⁺ oxidation state. The absence of the Fe³⁺ component in the Fe 2p spectrum and the presence of Cu 2p satellites clearly indicate that copper in the structure exists as Cu²⁺, substituting Fe³⁺ ions at lattice sites and forming the Cu₂[Fe(CN)₆] phase. The low oxygen content (~5–8 at.%) is attributed to the formation of oxide phases, as confirmed by X-ray diffraction.

For a detailed analysis of the chemical states of the elements, high-resolution spectra of Fe 2p, Cu 2p, and N 1s were recorded. Figure 6 presents the spectral fitting results obtained with a mixed Gaussian Lorentzian function following Shirley background correction. The Fe 2p_3/2 spectrum of Prussian blue was resolved into two main components (Figure 6a). The peak at 708.5 eV corresponds to low-spin Fe²⁺ ions, which are part of the hexacyanoferrate anion [Fe(CN)₆]⁴⁻. The peak at 710.5 eV is attributed to high-spin Fe³⁺ ions occupying cationic sites in the Prussian blue crystal lattice [15]. An additional broad component in the 715–720 eV region is a characteristic shake-up satellite for high-spin Fe³⁺. The Fe²⁺:Fe³⁺ integrated intensity ratio was approximately 0.75:1, which is in good agreement with the Fe₄[Fe(CN)₆]₃ stoichiometry. The N 1s spectrum (Figure 6b) shows a dominant peak at 397.8 eV, which is assigned to nitrogen atoms in cyanide groups bonded to iron(III) ions (Fe³⁺–N≡C) [15]. Additionally, a small peak is observed at 401.5 eV, which can be attributed either to plasmonic energy loss from the main N 1s peak or to nitrogen-containing contaminants (e.g., ammonium ions) adsorbed on the surface [74]. The intensity of this peak does not exceed 5% of the main peak’s intensity, indicating a minor amount of surface impurities. For the copper analogue Cu₂[Fe(CN)₆], the Fe 2p_3/2 spectrum contains only one peak at 708.5 eV, corresponding to Fe²⁺ in [Fe(CN)₆]⁴⁻ (Figure 6c). The absence of the Fe³⁺ component confirms that, in the copper analogue, iron exists exclusively in the +2 oxidation state. The Cu 2p_3/2 spectrum (Figure 6d) exhibits a main peak at 933.0 eV and a characteristic broad satellite (shake-up) in the 940–945 eV region, which is a diagnostic feature of Cu²⁺ ions [73]. The splitting between the main peak and the satellite is approximately 7.5 eV, typical for copper(II) compounds with nitrogen-containing ligands. The N 1s spectrum (Figure 6e) is asymmetric and can be deconvoluted into two components at 397.8 and 398.8 eV. The first component corresponds to the nitrogen of the cyano groups bound to iron (Fe–C≡N), and the second to the nitrogen coordinated to copper ions (Cu–N≡C). This confirms the formation of a heterometallic bridge structure Fe–C≡N–Cu. Thus, the high-resolution XPS data fully correlate with the results of XRD, IR, and Raman spectroscopy and unambiguously confirm the successful substitution of Cu²⁺ for Fe³⁺ ions in the hexacyanoferrate structure.

3.6. Proposed Mechanism of Composite Formation During Plasma–Liquid Synthesis

Based on the analysis of product phase composition, molecular structure, and plasma diagnostics using optical emission spectroscopy, a mechanism for the formation of target compounds was proposed. A key role in this process is played by a pulsed discharge initiated in a K₃[Fe(CN)₆] solution. OES analysis shows that plasma exposure causes intense dissociation of water molecules, as evidenced by the atomic hydrogen lines H_α and H_β, atomic oxygen, and the hydroxyl radical band OH. The resulting active species (H•, OH•, O) act as oxidizing and reducing agents at various stages of the synthesis. Simultaneously, the plasma causes fragmentation of hexacyanoferrate ions, as indicated by the presence of iron lines for both types of electrodes. The released Fe²⁺/Fe³⁺ ions and cyanide-containing fragments participate in further reactions. Direct spectroscopic confirmation of this was obtained by analyzing the discharge with copper electrodes: intense resonance lines of atomic copper were recorded in the spectrum at 324.7 and 327.4 nm. These lines are completely absent from the spectrum with iron electrodes, clearly indicating the entry of copper into the plasma phase. Based on the data obtained, the synthesis mechanism can be described as follows. When iron electrodes are used, erosion leads to the release of Fe²⁺ ions into the solution. Under the action of oxidizing agents, they are converted to the Fe³⁺ state. At the same time, hydrated electrons cause partial reduction of the starting [Fe(CN)₆]^3- ions to [Fe(CN)₆]⁴⁻. The reaction between Fe³⁺ ions and [Fe(CN)₆]⁴⁻ hexacyanoferrate species results in the precipitation of insoluble Prussian blue. A competing process is the hydrolysis of some Fe³⁺ ions, forming an iron oxide impurity, as confirmed by X-ray diffraction data. In the case of copper electrodes, eroded copper enters the solution predominantly as Cu²⁺ ions (via the Cu+ stage followed by oxidation). These ions directly interact with hexacyanoferrate anions, forming a copper analogue. A competing process is the hydrolysis of some Cu²⁺ ions in the near-cathode region, leading to the formation of a CuO impurity phase, consistent with X-ray diffraction data. Thus, plasma–solution synthesis is a complex, multi-stage process in which the generation of active species from water and electrode erosion under the action of the discharge are key factors determining the formation of the target hexacyanoferrate phases. It should be noted that the formation of the hexacyanoferrate framework in our plasma–solution system critically depends on the continuous supply of transition metal cations (Fe²⁺/Fe³⁺ from iron electrodes or Cu²⁺ from copper electrodes) released via electrode erosion. Inert electrodes (e.g., Pt, graphite) would not erode under the same discharge conditions and therefore cannot provide the necessary cations to precipitate the PBA phase. Consequently, a true “impurity-free PBA” reference cannot be obtained under identical plasma conditions. This inherent limitation is acknowledged; however, the comparison of the two composites (with different intrinsic oxide impurities) still allows us to correlate the nature of the oxide phase with the observed selectivity.

3.7. Photocatalytic Activity of Synthesized Materials

Throughout this work, we refer to the synthesized materials as ‘composites’ because the oxide impurities (α-Fe₂O₃ or CuO) are not merely passive by-products; they form heterojunctions with the PBA matrix, actively influence charge separation, and impart functional synergy. This usage aligns with the common definition of a composite as a material comprising two or more distinct phases that together exhibit enhanced or emergent properties. The photocatalytic performance of the prepared materials was evaluated by examining the decomposition of three organic dyes belonging to different classes: Reactive Red 6C, Rhodamine B, and Methylene Blue. The experiment included a dark stage (30 min) to establish adsorption equilibrium, after which the system was irradiated with UV light for 90 min. Thus, the total experiment time was 120 min, with 90 min of active irradiation. All kinetic curves show the dark stage (0–30 min) and the irradiation stage (30–120 min) separated by a vertical dashed line. Figure 7 shows the kinetic curves for dye degradation. Control photolysis experiments were performed under identical UV irradiation conditions in the absence of any catalyst. The degradation of all three dyes after 120 min did not exceed 3% for Reactive Red 6C, 5% for Rhodamine B, and 6% for Methylene Blue (Figure 7a, curves “no catalyst”). This confirms that the dye degradation observed in the presence of the synthesized composites is primarily due to photocatalytic activity rather than direct photolysis. The samples exhibit different adsorption capacities, determined by the nature of the dye and the surface charge of the catalyst. Prussian blue adsorbs the cationic dye MB most effectively (51.3%), while the adsorption of the anionic azo dye RR6C and the zwitterionic RhB does not exceed 8–9%. This indicates the negative surface charge of Fe₄[Fe(CN)₆]₃, which favors the electrostatic attraction of cationic molecules. The copper analogue adsorbs all dyes significantly better, especially the anionic RR6C (27.5%) and RhB (21.0%). The increased adsorption of RR6C may indicate a less negative surface charge of Cu₂[Fe(CN)₆], which facilitates the attraction of anionic dyes. Differences in textural characteristics may also contribute to adsorption capacity. When irradiated with UV light, both catalysts completely degrade all three dyes within 90 min.

To elucidate the differences in adsorption behavior, zeta potential measurements were performed for both samples in aqueous suspension. The zeta potential of Fe₄[Fe(CN)₆]₃ was measured to be −24 ± 2 mV, while Cu₂[Fe(CN)₆] exhibited a significantly less negative value of −7 ± 1 mV. These results confirm that Prussian blue possesses a more negative surface charge under the experimental conditions, which favors electrostatic attraction of cationic dyes such as Methylene Blue (51.3% adsorption). In contrast, the near-neutral surface charge of Cu₂[Fe(CN)₆] reduces electrostatic repulsion of anionic dyes, explaining its higher adsorption of Reactive Red 6C. These zeta potential data directly support the proposed role of surface charge in the observed adsorption selectivity.

The reusability of the synthesized photocatalysts was evaluated over three consecutive degradation cycles (Figure 8). After each cycle, the catalyst was recovered by centrifugation, washed with deionized water, and dried at 80 °C for 1 h. Both samples maintained >85% of their initial activity after three cycles, indicating good stability. The modest drop in activity could result from some surface oxidation or from nanoparticle loss during the recovery process.

Kinetic analysis using a pseudo-first-order model shows that the rate constants (k) are in the range of 0.032–0.048 min⁻¹ (

Table 2. Kinetic parameters of photocatalysis for samples.

Dye	k1, min−1	R12	k2, L mg−1 min−1	R22
Fe4[Fe(CN)6]3
RR6C	0.041 ± 0.003	0.95	0.0042 ± 0.0005	0.89
RhB	0.044 ± 0.007	0.94	0.0048 ± 0.0006	0.87
MB	0.038 ± 0.002	0.93	0.0039 ± 0.0004	0.88
Cu2[Fe(CN)6]
RR6C	0.048 ± 0.005	0.96	0.0050 ± 0.0005	0.91
RhB	0.045 ± 0.008	0.95	0.0046 ± 0.0007	0.88
MB	0.032 ± 0.003	0.92	0.0021 ± 0.0003	0.84

). A correlation is observed between adsorption and photocatalytic activity for RR6C and MB: for the anionic Reactive Red 6C, the copper analog Cu₂[Fe(CN)₆] (adsorption 27.5%) exhibits a higher rate constant (k = 0.048 min⁻¹) compared to Prussian blue (k = 0.041 min⁻¹, adsorption 8.4%). For cationic Methylene Blue, Fe₄[Fe(CN)₆]₃ (51.3% adsorption) outperforms its copper counterpart (k = 0.038 min⁻¹ versus 0.032 min⁻¹). For Rhodamine B, despite different adsorption (7.6% for Fe₄[Fe(CN)₆]₃ and 21% for Cu₂[Fe(CN)₆]), the photocatalytic activity of both samples is virtually identical (k ≈ 0.045 min⁻¹). This may indicate the dominance of other degradation mechanisms for RhB, such as direct oxidation by photogenerated holes or the involvement of reactive oxygen species generated in the bulk solution rather than on the catalyst surface.

To verify the appropriateness of the pseudo-first-order model, the kinetic data were also analyzed using the pseudo-second-order model:

$${\displaystyle \frac{t}{C_t}}={\displaystyle \frac{1}{k_2{C}_0^2}}+{\displaystyle \frac{t}{C_0}},$$

(5)

where k₂ is the pseudo-second-order rate constant (L mg⁻¹ min⁻¹) and C₀ is the initial dye concentration. The fitting results are summarized in Table 2.

Comparison of the correlation coefficients shows that the pseudo-first-order model generally provides a better description of the degradation kinetics than the pseudo-second-order model. This is consistent with the low initial dye concentration used in this study, under which pseudo-first-order kinetics are typically observed. However, the MB/Cu-PBA system exhibits a relatively lower for the pseudo-first-order fit, suggesting the presence of competing processes (e.g., simultaneous adsorption and degradation) or a more complex degradation pathway for this particular dye–catalyst pair. Nevertheless, the pseudo-first-order model remains widely adopted in the literature for comparative photocatalytic studies and was therefore retained for consistency.

The difference in rate constants for RR6C degradation between Prussian blue (0.041 ± 0.003 min⁻¹) and the copper analogue (0.048 ± 0.005 min⁻¹) was evaluated using a two-tailed paired t-test. The p-value was <0.05, indicating that the difference is statistically significant despite the modest absolute difference (~17%). All reported rate constants are averages of three independent experiments with standard deviations as shown.

The observed selectivity under UV light irradiation can be explained by considering the band alignment of the heterojunctions formed between the hexacyanoferrate phases and the oxide impurities (α-Fe₂O₃ in Fe₄[Fe(CN)₆]₃ and CuO in Cu₂[Fe(CN)₆]).

Under UV excitation (365 nm, 3.4 eV), electrons are photoexcited in both components of each heterojunction. In the Fe₄[Fe(CN)₆]₃/α-Fe₂O₃ heterojunction, the conduction band of α-Fe₂O₃ (ca. +0.5 eV vs. NHE) is more positive than that of Fe₄[Fe(CN)₆]₃ (ca. −0.2 eV vs. NHE), favoring electron transfer from PB to α-Fe₂O₃. Accumulation of valence band holes occurs on PB, enabling the oxidation of water or hydroxide ions and subsequent generation of hydroxyl radicals (•OH, +2.7 eV vs. NHE). These radicals preferentially attack the aromatic rings of cationic dyes such as Methylene Blue, which is efficiently adsorbed on the negatively charged PB surface (51.3% adsorption).

In the Cu₂[Fe(CN)₆]/CuO heterojunction, CuO has a narrower band gap (~1.7 eV) and a more negative conduction band than Cu₂[Fe(CN)₆]. Under UV irradiation, photogenerated electrons migrate to the CuO surface, where they reduce dissolved oxygen to superoxide radicals (O₂⁻•, −0.33 eV vs. NHE). Superoxide radicals are known to cleave azo bonds (–N = N–) in anionic dyes such as Reactive Red 6C. The higher adsorption of RR6C on Cu₂[Fe(CN)₆] (27.5%) compared to PB (8.4%) further enhances the degradation rate. Thus, the selectivity is due to the band alignments and reactive oxygen species generated under UV excitation. The role of the oxide impurities is to facilitate charge separation and to direct the photogenerated carriers toward specific redox pathways.

To verify that the oxide impurities indeed form heterojunctions that facilitate charge separation, steady-state photoluminescence spectra were recorded for the plasma-synthesized composites and for a conventionally prepared impurity-free Prussian blue reference (Figure 9). The PL intensity of the Fe₄[Fe(CN)₆]₃/α-Fe₂O₃ composite was reduced by approximately 35% compared to PB-ref, and the Cu₂[Fe(CN)₆]/CuO composite showed a ~40% reduction. Such PL quenching is a well-established indicator of efficient spatial separation of photogenerated electron–hole pairs and is characteristic of heterojunction formation [44]. The observed quenching directly demonstrates that the incidental oxide impurities (α-Fe₂O₃ in the iron-based composite and CuO in the copper-based composite) act as charge-separating components, thereby contributing to the enhanced and selective photocatalytic activity discussed above. For the copper analogue, an impurity-free reference could not be obtained by the same co-precipitation route. Nevertheless, the PL intensity of the plasma-synthesized Cu₂[Fe(CN)₆]/CuO composite was still ~40% lower than that of the conventionally prepared pure PB-ref, indicating a strong charge-separating effect of the CuO impurity.

Both synthesized materials exhibit high photocatalytic activity under UV light, with clear selectivity associated with the surface charge and the nature of the dye. Prussian blue is most effective for the degradation of cationic dyes (MB), while its copper analog is preferred for anionic azo dyes (RR6C). The obtained results demonstrate the potential of using hexacyanoferrates synthesized by the plasma–liquid method for photocatalytic water purification from various classes of organic pollutants and open up possibilities for the creation of selective photocatalysts by varying the metal composition in the structure.

Therefore, a comparison of the proposed plasma–liquid method with conventional approaches is summarized in

Table 3. Comparison of plasma–liquid synthesis with conventional methods for Prussian blue analogues.

Parameter	Co-Precipitation [11,12]	Hydrothermal [18,19]	Electrochemical [20,21]	Mechanochemical [24]	This Work (Plasma–Liquid)
Synthesis time	Minutes–hours	Hours–days	Minutes–hours	10–60 min	30 min
Number of steps	≥2	≥2	1–2	1	1
Organic solvents	Often	Often	Rarely	No	No
Additional reagents (stabilizers, reducers)	Yes	Yes	No	No	No
Product yield, %	60–90	70–95	50–80	70–90	75–79
SBET, m2/g	200–600	150–500	100–400	50–200	3.8–8.7
Scalability	High	Moderate	Moderate	High	Potentially scalable

. As follows from the table, the plasma–liquid synthesis offers several advantages: it is a truly one-step process (unlike co-precipitation or hydrothermal methods), requires no organic solvents or additional reagents, and is completed within 30 min with a high yield (75–79%). However, a notable drawback is the relatively low specific surface area (3.8–8.7 m²/g) due to nanoparticle agglomeration under the non-equilibrium plasma conditions. Despite this limitation, the obtained materials demonstrate complete photocatalytic degradation of three organic dyes within 120 min, confirming their high activity. The scalability of the method is currently under investigation; preliminary results indicate that the discharge cell can be scaled up by increasing the electrode length and solution volume, which we plan to report in a follow-up study.

4. Conclusions

In this study, a novel one-step plasma–solution synthesis method was successfully developed for the preparation of Prussian blue (Fe₄[Fe(CN)₆]₃) and its copper analogue (Cu₂[Fe(CN)₆]) using an underwater pulsed DC discharge. For the first time, the direct plasma-assisted formation of Prussian blue and copper hexacyanoferrate nanoparticles was demonstrated.

The formation of the desired face centered cubic (Fm 3m) phases was confirmed by XRD, FTIR, Raman spectroscopy, and XPS, with crystallite dimensions falling between 24 and 26 nm. Minor oxide impurities—α-Fe₂O₃ in the iron-based synthesis and CuO in the copper-based synthesis—were detected, originating from electrode erosion during the plasma discharge. Optical properties and band gap measurements using UV–visible spectroscopy revealed that the copper analogue exhibits a band gap of 2.8 eV, consistent with literature values. For Prussian blue, an apparent band gap of 3.4 eV was observed, which is attributed to the predominance of indirect optical transitions.

Both materials completely degraded three organic dyes—Reactive Red 6C, Rhodamine B, and Methylene Blue—within 90 min of UV irradiation (after an initial 30 min dark adsorption step, total experiment time 120 min). The degradation kinetics followed a pseudo-first-order model, with rate constants ranging from 0.032 to 0.048 min⁻¹.

It should be noted that Prussian blue exhibited greater activity toward the cationic dye Methylene Blue, whereas the copper analogue was more effective against the anionic azo dye Reactive Red 6C. This selectivity is attributed to the formation of heterojunctions with impurity oxide phases, which facilitate charge separation and generate distinct reactive oxygen species—hydroxyl radicals for Prussian blue and superoxide radicals for Cu-PBA.

The adsorption capacity of the samples correlated with their photocatalytic activity for the charged dyes (MB and RR6C), confirming the role of electrostatic interactions between the catalyst surface and the dye molecules.

In summary, the plasma–solution synthesis method provides a rapid, single-step approach to producing functional Prussian blue analogues with tunable selectivity for photocatalytic applications. The inherent formation of oxide impurities, rather than being a disadvantage, can be leveraged to tailor the photocatalytic properties of the materials. This work opens new avenues for the direct, in situ synthesis of PBA-based heterojunctions for water purification and other advanced applications.