Hybrid graphenic and iron oxide photocatalysts for the decomposition of synthetic chemicals

A comparative analysis was conducted to assess the effectiveness of the developed Fe/g-C hybrid in relation to other photocatalysts described in the literature (Table S1). The assessment was based on the degradation of PFAS achieved, considering the degradation time, UV fluence, and the photocatalyst’s dosage normalized to the initial PFAS concentration (Fig. 1) and was limited to the studies were all parameters (particularly UV fluence rate) are reported. Please refer to Eq. 1 below:$$ {NDE} \\ =\frac{{Degradation} \, {Efficiency}\left( \% \right)}{{time}\left[h\right]\times {UV}{fluence\; rate}\left[{mW}/c{m}^{2}\right]\times {catalyst}{dosage}\left[g/L\right]/{PFOA}{concentration}\left[{mg}/L\right]}$$
(1)
where NDE stands for normalized degradation efficiency presented in Fig. 1. The results showed that the Fe/g-C hybrid outperformed other photocatalysts described in the literature, with the use of readily available materials and a straightforward fabrication procedure, which enhances its accessibility and cost-effectiveness for potential commercialization. The Fe/g-C hybrid’s favorable attributes suggest it as a potential option for widespread PFAS degradation.Fig. 1: Normalized decomposition efficiency (NDE) grouped in 3 different initial PFOA (perfluorooctanoic acid) concentrations.Yellow boxes correspond to decomposition efficiency of different heterogeneous photocatalysts reported in the literature (Table S1). The data for Fe/g-C (iron doped carbon) featured in this study are indicated in orange. The error bars (whiskers) represent the range from the upper quartile (Q3) to the maximum value for the upper whisker and from the lower quartile (Q1) to the minimum value for the lower whisker.Iron oxide/graphenic carbon (Fe/g-C) hybrid photocatalysts were obtained by pyrolysis of cellulose impregnated with different contents of iron chloride (FeCl3). Details on preparation procedures and characterizations are described in the synthesis section in the Methods. Scanning electron microscopy (SEM) images in Fig. 2A–C showcase the morphological characteristics of graphenic carbon (g-C) and Fe/g-C samples. The presence of tube-like structures suggests that certain features of the original cellulose materials are retained post-pyrolysis, see Fig. 2A. These structures appear to be somewhat fragmented and overlapping, creating a dense and irregular texture. In Fig. 2B, 32 wt% Fe/g-C, provides a more zoomed-out view, revealing a heterogeneous mixture of particles where again, tube-like morphologies are noticeable. This scale gives a sense of the overall distribution of particles within this sample. In Fig. 3C magnifies an area of the iron-doped carbon, showing a dense agglomeration of particles with iron oxide crystallites. Unlike the more fibrous and elongated structures seen in Fig. 2A and B, this image shows a rough, granular surface texture. The iron oxide nanoparticles appear to be well-dispersed across the carbon substrate, which is a desirable trait for enhancing photocatalytic activity due to improved charge separation.Fig. 2: Morphological and structural characterization of pure graphenic carbon (g-C) and iron doped carbon (Fe/g-C).A–C SEM images of pure g-C, 32 wt% Fe/g-C, and a magnified region showing iron oxide crystallite grown on g-C surface, respectively (scale bars: A = 50 μm, B = 50 μm, C = 10 μm) (D) pXRD pattern. E FTIR spectra. F Raman spectra showing the D (1373 cm−1) and G (1589 cm−1) bands.Fig. 3: Photocatalytic removal of PFOA employing graphenic carbon based photocatalysts and the influence of iron doping and environmental conditions on degradation efficiency.A Comparison of PFOA removal efficiency ac, under UV light, indicating an increase in removal efficiency with higher Fe content with [PFOA]0 = 1 mg L−1. B Impact of initial PFOA concentration on its removal using a 32 wt% Fe-doped g-C hybrid photocatalyst, demonstrating decreased removal efficiency with higher concentrations of PFOA. C Removal efficiency of PFOA under dark conditions, UV radiation (254 nm), and simulated solar light (AM 1.5 G), showing the enhanced removal capability under simulated solar light with [PFOA]0 = 1 mg L−1. All experiments were conducted with a photocatalyst dosage = 1 g L−1, temperature = 22 ± 2 °C, and UV fluence rate = 1.42 ± 0.05 mW cm−2. Trend lines are provided to aid visual interpretation of the data.The chemical structure of Fe/g-C hybrids was further confirmed by the powder X-ray diffraction (pXRD) pattern, Fourier transform infrared (FTIR) and Raman spectra, respectively as shown in Fig. 2D–F. X-ray Photoelectron Spectroscopy (XPS) was also applied to investigate the surface electronic structure of the iron oxide (Figs. S5 to S7). These results are highly similar and consistent with previous reports26,27,28.Powder samples were subjected to XRD to evaluate their crystalline structure in a medium and long-range order. According to Fig. 2D, the XRD pattern of the pure carbon (g-C) did not exhibit the presence of a crystalline phase. However, as the iron concentration was increased, implying the increased crystallinity and the prevalence of hematite as an iron oxide phase. The diffraction peaks were consistent with the presence of α-Fe2O3 (JCPDS#33-0664), g-Fe2O3 (JCPDS#39-1246), Fe3O4 (JCPDS#19-0629), α-FeOOH (JCPDS#12-0412), graphite and amorphous carbon network29 as indicated in Fig. 2D.FTIR spectra of g-C and iron doped samples contain intensive bands for aromatic C=C stretching vibrations at about 1584 cm−1 30. The intensity of this peak remains relatively consistent, suggesting that the iron oxide doping does not substantially alter the sp²-hybridized carbon network. Carbonyl (C=O) stretching vibrations occur at 1705 cm−1 31, which decrease in intensity as a function of iron concentration. This suggests that the introduction of iron might influence the formation or exposure of carbonyl functionalities in the graphenic structure. The presence of band centered at 1192 cm−1 indicate the presence of C–O groups26. This peak’s intensity and shape could vary based on the interaction of these oxygen-containing groups with the iron oxide particles. Also, a broad band in the range from 3500 to 3200 cm−1 is apparent and attributed to O–H stretching vibrations. Additionally, an absorption band at 542 cm−1 was observed in the samples which was assigned to Fe–O stretching. The trend that can be observed is that as the iron content decreases, the peaks attributed to iron oxide (if the 542 cm–1 peak indeed represents Fe-O) become less pronounced, while the organic functionalities of the graphenic carbon remain relatively constant. This aligns with expectations since the fundamental carbon lattice structure should remain largely intact, while the dopant level varies.Raman spectroscopy clarifies structural changes in the graphenic carbon structure as a function of iron content (Fig. 2E). In contrast to pXRD, Raman measurements monitor structural properties at short-range orders. The Raman spectra of pure g-C and Fe/g-C show a characteristic G band centered at 1589 cm−1 and a D band centered at 1373 cm−1 32. The identified bands correspond to the vibrational modes of carbon sp2 atoms. The ID/IG ratio, representing the intensity ratio of the D and G bands, is frequently used as a structural order indicator (see Table S2). Graphitic carbon materials with long-range order in graphitic layers have a higher ID/IG ratio than pure graphite when defects such as aromatic cluster borders appear. A carbon material composed of a few aromatic clusters of small size will have a higher D mode intensity depending on how many sixfold rings are in the cluster. An elevated ID/IG ratio indicates the order in this case. As iron content in hybrid photocatalyst increased, ID/IG ratios linearly increased (Figure S2), indicating a higher degree of order. The enhanced structural order as a function of iron content can lead to improved charge transfer and reduce recombination of electron-hole pairs resulting in a faster photocatalytic degradation rate of PFAS33.The textural properties of carbonized materials derived from renewable resources (e.g., cellulose) are key parameters affecting their functionality34. Specific surface area, pore volume and size were assessed by N2 physisorption isotherms and the results for pure g-C and 32 wt% Fe/g-C were listed in Table 1. The BET surface area of g-C and 32 wt% Fe/g-C were ~56 and 427 m2 g−1, respectively. The average absorption pore width was obtained by the automatic DFT system calculation, which increases as a function of iron content. The observed isotherms (Figure S3) fall under Type IV, typical for mesoporous materials, with hysteresis loops indicative of capillary condensation within these pores35. The evolution of hysteresis loops with increased iron doping suggests alterations in pore connectivity and shape, hinting at the complexity introduced by the iron oxide. The specific surface area, as revealed by the BET analysis and summarised in Table S3, increases with iron oxide doping. The increase in surface area suggests that the addition of iron oxide to the graphenic carbon matrix contributes positively to creating additional surface area. This could be due to the creation of additional pores or to the iron oxide nanoparticles preventing the collapse of existing pores during the synthesis process, thus maintaining or enhancing the overall porosity of the material36.Table 1 Textural analysis values (specific surface area and pores volumes) and optical properties of pure graphenic carbon (g-C) and 32 wt% iron doped graphenic carbon (Fe/g-C)From the UV-Vis diffuse reflectance spectroscopy and to evaluate the optical bandgap from the absorption spectra, the Kubelka-Munk theory was applied37. Optical band gaps in a range of 1.28 to 1.19 eV could be derived for the composites (Fig. S4 and Table S4). These values are smaller than those typically reported for graphene with oxygen-containing functional groups (2.5 eV)38 and Fe2O3 (2.67 eV)39 but are comparable to those reported for other amorphous carbon with diverse functionalization40. Furthermore, the high bandgap could result in reduced photocatalytic activity. However, the large surface area and porous size may facilitate its photoactivity.In our custom-designed photoreactor (Fig. S8), the photocatalytic decomposition of PFOA (C0 = 1 mg L−1) was conducted with photocatalyst dosage of 1 g L−1 doped with different iron content ranging from 1.6 to 32 wt% Fe. As shown in Fig. 3A, owing to the low molar absorption coefficient of PFOA at 254 nm, εPFOA, 254 nm of 3.5±0.9 M−1cm−1, the concentration of PFOA remained unchanged in UV alone experiments. When employing Fe/g-C with iron concentrations of 1.6, 3.2, 4.8, and 6.4 wt% Fe, the reductions in PFOA concentration were modest. at 3.7%, 5.8%, 7.2%, and 13.4% respectively, over a 6-hour irradiation period (UV fluence of 30.7 J cm−2), Conversely, Fe/g-C hybrids 16 wt% and 32 wt% Fe content markedly improved PFOA decomposition, achieving removal efficiencies of 66.4% and 89.7% within the same duration, 6 hours. In view of the marginal role played by direct photolysis (without photocatalyst) on PFOA decomposition, it suggests that the process entails two sequential mechanisms. The first step is the adsorption of PFOA on Fe/g-C hybrid photocatalyst surface, which is followed by a subsequent step of hole transfer from the surface of the hybrid photocatalyst to the adsorbed PFOA molecules facilitating the breakdown of PFOA to its shorter chain components (see Fig. S13).Due to the increased decomposition of PFOA in the presence of 32 wt% Fe/g-C, compared to the other catalysts within the current work, we selected this photocatalyst composition to investigate the photodegradation of PFOA and effects of key parameters (e.g., initial concentration of PFOA) on the degradation process. As shown in Fig. 3B, increasing the dosage of catalyst considerably enhanced the removal of PFOA which was mainly due to the further availability of complexation sites for PFOA molecules. It is also clear from the results that the catalyst shows effective PFOA removal even in low dosages of 0.1 g L−1, where approximately 79% of the PFOA was removed within 6 hours.The initial concentration of PFOA may affect the ratio of available complexation sites on catalyst surface to the amount of PFOA and consequently the extent of its removal; thus, the initial concentration should be selected wisely to be illustrative of real case scenarios. The reported values of initial PFAS concentrations in real aquatic environments vary in a wide range of ng L−1 in natural waters (e.g., surface- and groundwater) to mg L−1 in industrial wastewater or concentrated waste stream of physical separation techniques41. As shown in Figure S12, increasing the initial concentration of PFOA from 0.1 to 5 ppm decreased the removal efficiency from nearly complete removal of PFOA to 86% within 6 hours. Such observation implies that the catalyst could result in fast removal of PFOA in low concentrations commonly observed in natural waters, while achieving substantial removal ( ~ 85%) at PFOA dosage representative of concentrated waste streams. Moreover, the results confirm that the Fe/g-C hybrid is an effective photocatalyst for PFAS adsorption and decomposition offering a more efficient approach than other methods currently being used for this process42.Owing to the ability of iron to form a complex with PFOA43, Fe/g-C hybrids could play a dual-role as both adsorbent and photo-catalyst to degrade the captured PFOA. Figure S12A and B compares the PFOA removal using 32 wt% Fe/g-C under UV and a control experiment in the presence of the same dosage of catalyst but in dark (without UV irradiation) condition. Interestingly, the catalyst containing 64 wt% Fe/g-C did not demonstrate enhanced PFOA degradation (Fig. S12C). Within 5 hours, 63.7% and 89.7% of PFOA were removed within 6 h in dark and UV conditions, respectively. Such observation confirms the dual role of employed catalyst where PFOA molecules were firstly adsorbed on the surface of photocatalyst and then degrade by UV irradiation and in turn faster removal of PFOA was observed in UV process. Additionally, thermogravimetric analysis (TGA), as shown in Fig. S1B, affirms the catalyst’s stability within the experimental temperature range, providing vital insights into its thermal behavior ensuring the reliability of our experimental results.Scaling up a heterogeneous photocatalyst for PFAS decomposition requires not only the ability to manufacture the catalyst in large quantities but also to identify if the photocatalyst is stable over a longer period of time. This was investigated by performing five consecutive batch cycles ( ~ 30 hours) using 32 wt% Fe/g-C photocatalyst (Fig. 4A). After 6 h in each cycle, upon taking 2 mL sample for analysis, an aliquot (0.1 mL) of PFOA stock solution (500 mg L−1) was added to the solution to increase the concentration of PFOA back to roughly 1 ppm and the solution was mixed for five minutes to ensure sufficient contact with the photocatalyst and then take a sample for initial concentration of next cycle. As shown, >90% of PFOA was removed with the first 4 cycles and nearly 88.5% of PFOA was removed in the last cycle. Furthermore, the photocatalyst morphology did not exhibit any notable change after 30 hours of experiment, which indicates the Fe/g-C maintains its structure during extended irradiation. The results imply that the photodegradation of PFOA also stability and recyclability of the photocatalyst, allowing for repeated use without even use of chemical regeneration.Fig. 4: Recyclability and temporal stability of the iron doped graphenic carbon (Fe/g-C) with >90% PFOA removal and fluoride (F-) adsorption.A PFOA removal in 5 consecutive cycles using 32 wt% Fe/g-C hybrids. Experimental conditions: [PFOA]0 in each cycle = 1 mg L−1, photocatalyst dosage = 1 g L−1, temperature = 22 ± 2 °C, and UV fluence rate = 1.42 ± 0.05 mW cm−2. B SEM images of photocatalyst after the 5th. C XPS O 1 s and F 1 s before and after PFOA decomposition.Fluoride is in the indicative byproduct of defluorination of PFOA. Figure S14 shows the total recovery (mass balance) of fluoride, i.e., the molar ratio of total fluoride in the form of PFOA, its generated intermediates (shorter chain PFCAs), and fluoride (F-), versus time. The total fluoride recovery using 32 wt% Fe/g-C hybrid decreased to 42.2% within 1.5 h and remained relatively constant afterward. Compared to our earlier observation using a molecular photo-mediator (UV/VUV/sulfite system) where fluorine recovery was >90% in all irradiation time44, such fluoride recovery is notably lower which could be mainly due to the adsorption of generated fluoride to the surface of catalyst. The affinity of iron oxides for fluoride ions has been confirmed in previous reports45. In a control experiment using sodium fluoride (NaF) with the same concentration of fluoride in 5 ppm PFOA (3.45 ppm F-), we observed that 47.0% of initial fluoride was adsorbed on to catalyst surface (32 wt% Fe/g-C). It was mainly due to the presence of iron oxide on the catalyst surface since no adsorption of fluoride was observed using pure g-C catalyst (Fig. S14). Additionally, the iron concentration resulting from ion leaching post-photodegradation was determined by ICPMS. The leached iron fell within the range of 69.948 to 70.210 ppm.To provide insights into the adsorption process and photodegradation of PFOA, we conducted high-resolution XPS analysis on the 32 wt% Fe/g-C hybrid before and after the reaction (Fig. 4C). After quenching the reaction, an additional peak assigned to F 1 s occurred at 648.5 eV, confirming the fluoride adsorption on the photocatalyst surface after cycle 1. The O 1 s XPS spectra before and after the reaction (Fig. 4C) were deconvoluted into three peaks, which were assigned to hydroxyl (OH) groups at 531.4 eV, lattice oxygen in α-Fe2O3 at 530.2 eV, and at 533.2 eV oxygen of carbonyl groups46. The binding energy and relative content of oxygen groups are shown in Table S8. These data reveal that the oxygen-related functional groups on the photocatalyst hybrids underwent a reduction following the photocatalytic reaction. This implies that, during this process, the photo-excited electrons within α-Fe2O3 spontaneously migrate towards the g-C network, resulting in the reduction of the oxygen-related functional groups present on the g-C support. It appears that the oxygen-containing functional groups on the g-C support play a role as an electron acceptor, which serves to decelerate the recombination of charges47,48,49. Consequently, this facilitates the direct transfer of holes, ultimately contributing to the efficient photocatalytic oxidation occurring on the Fe/g-C surface.These findings provide strong evidence for an effective approach to PFAS degradation. We have demonstrated that through a simple, economical synthesis, an abundant heterogeneous hybrid photocatalyst can rapidly decompose PFOA, achieving rates of ≥85% in just 3 hours with a UV fluence of 30.7 W cm−2. In contrast to previous assumptions, our findings demonstrate that high rates of PFOA degradation can be achieved without resorting to expensive and complex methods Furthermore, the observed consistency in maintaining decomposition rates of ≥85% for 30 hours under a UV fluence of 1.42 ± 0.05 mW cm−2 underscores the promise of this approach. With these heterogeneous photocatalysts, achieving approximately 90% PFOA decomposition in 6-hour batch-type experiment, further exploration of other sustainable sources of graphenic carbon for PFAS degradation is warranted. This work establishes that graphenic carbon-doped composites can effectively address persistent organic pollutants in water, with the added benefit of reducing the required UV fluence rates, thus making PFAS degradation more efficient and practical.