Converting Light Energy to Chemical Energy: A New Catalytic Approach for Sustainable Environmental Remediation.

We report a synthetic approach to form cubic Cu2O/Pd composite structures and demonstrate their use as photocatalytic materials for tandem catalysis. Pd nanoparticles were deposited onto Cu2O cubes, and their tandem catalytic reactivity was studied via the reductive dehalogenation of polychlorinated biphenyls. The Pd content of the materials was gradually increased to examine its influence on particle morphology and catalytic performance. Materials were prepared at different Pd amounts and demonstrated a range of tandem catalytic reactivity. H2 was generated via photocatalytic proton reduction initiated by Cu2O, followed by Pd-catalyzed dehalogenation using in situ generated H2. The results indicate that material morphology and composition and substrate steric effects play important roles in controlling the overall reaction rate. Additionally, analysis of the postreacted materials revealed that a small number of the cubes had become hollow during the photodechlorination reaction. Such findings offer important insights regarding photocatalytic active sites and mechanisms, providing a pathway toward converting light-based energy to chemical energy for sustainable catalytic reactions not typically driven via light.

Introduction

Given the current global energy state, a push for rapid access to sustainable energy, such as solar cells and renewable fuels, has emerged.[1,2] Accordingly, an expansion in research in the area of photocatalysis has occurred over the past decade.[3] Although significant effort has been exerted to design and fabricate highly active photocatalysts to replace conventional catalytic materials,[4,5] little progress has been made toward reducing the overall energy, materials, and resources necessary to create and maintain these light-harvesting systems. As such, existing catalytic technologies must evolve toward adopting sustainable synthetic practices that minimize environmental and economic impacts. To this end, the ability to easily synthesize and characterize functional nanomaterials with controlled size, shape, composition, and overall structural morphology will allow us to achieve catalytic technologies with optimized reactivity.[6−8]

While photocatalysis is appealing, it is traditionally limited to redox chemistries. Thus, the application of photobased approaches to non-photo-responsive catalytic reactions remains exceedingly rare.[9] One approach to address this limitation is to fabricate multicomponent photocatalytic architectures with multiple catalytic domains. For this approach, semiconductor materials are of particular interest because of their photoinduced charge transfer properties, where fast transfer of charges across the semiconductor interface is critical for high energy conversion efficiency.[10,11] These charge transfer processes can be greatly influenced by the presence of a noble metal cocatalyst[12] coupled to the semiconductor, which can readily accept photoinduced electrons to efficiently transfer them to surface adsorbed acceptor molecules such as to H+ to produce H2.[13−15] Additionally, these metal nanoparticle cocatalysts frequently act as an electron reservoir and consequently promote charge separation within the semiconductor-metal assembly.[16] Solar H2 production from photocatalytic water splitting is one of the most notable applications of photoinduced electron transfer, along with being a promising route to achieving renewable energy.[17] Advancing toward sustainable chemical processes, photocatalytically generated H2 could be employed in other catalytic reactions, such as hydrogenation or hydrodehalogenation, which could also occur at the metal nanoparticle surface. Use of such a light-promoted tandem catalytic system has been demonstrated recently.[9,18,19] For instance, Hirai and co-workers have reported the N-monoalkylation of primary amines in alcohol solvents by tandem photocatalytic and catalytic reactions on TiO2 loaded with Pd particles (Pd/TiO2).[19] In separate work, we reported the use of Cu2O cubes with galvanically deposited Pd nanoparticles on the oxide surface (Cu2O/Pd) as catalysts for the reductive dechlorination of polychlorinated biphenyls (PCBs), a well-known environmental persistent organic toxicant present at many contaminated sites around the world.[9] For this, photoactivation of the oxide component results in H2 production, facilitated by Pd, via proton reduction. This H2 was subsequently activated on the Pd metal surface from which PCB dechlorination was processed in a tandem-like fashion. From this, it is clear that semiconductor-metal multicomponent materials can be tailored to facilitate light-promoted tandem catalysis, catering to reactions not typically driven via light.

In this contribution, we demonstrate that the Cu2O/Pd tandem photocatalytic activity is highly sensitive to both the overall composite material morphology and the substrate structure for nontraditional photocatalytic reactions (Scheme ). To examine the structural effect, Cu2O cubes were generated where varying amounts of Pd nanoparticles were galvanically deposited on the oxide surface (0–15 wt %). The materials were extensively characterized to confirm their morphology, size, composition, and component arrangement using scanning electron microscopy (SEM), transmission electron microscopy (TEM), energy-dispersive X-ray spectroscopy (EDS), inductively coupled plasma optical emission spectrometry (ICP-OES), and powder X-ray diffraction (XRD). Light-promoted tandem catalytic activities of the materials were investigated by monitoring the reductive dechlorination of PCBs. Under light irradiation, H2 was produced via proton reduction by the Cu2O material facilitated by the Pd nanodomains. Subsequently, the in situ generated H2 was used for Pd-catalyzed dehalogenation of PCBs. For this, PCB structures that positioned the Cl substituent at the ortho, meta, and para positions were reductively dechlorinated to generate the final biphenyl product, and the corresponding reaction rate constants were quantified. Comparison of the rate constants demonstrated a general dechlorination reactivity order of para > meta ≥ ortho, with maximal catalytic rates obtained using Cu2O/Pd materials with 9 wt % Pd deposited at a composite material catalyst loading of 2 mg/mL. These findings provide intriguing information concerning the active sites and composite interface of photocatalytic materials, as well as how such structure-defined properties affect the overall reaction kinetics for photoinitiated tandem systems. They also increase the understanding of material structural effects over catalytic reactivity, opening pathways toward new structures with enhanced reactivity or translation of such photobased approaches to new catalytic processes, all of which are highly important for environmental remediation of halogenated organic compounds such as PCBs.

Scheme 1

Reactivity of Cu2O/Pd Composite Structures for Photodriven Tandem Catalysis

Materials and Methods

Chemicals

CuSO4 and Na2CO3 were obtained from BDH Chemicals; Pd(CH3COO)2 was acquired from Strem Chemicals. HR-GC hexanes and 200 proof ethanol were attained from EMD Millipore and Pharmco-AAPER, respectively. Finally, polyvinylpyrrolidone (PVP; MW ∼ 29 000 g/mol), sodium citrate, glucose, 2-chlorobiphenyl (PCB 1), 3-chlorobiphenyl (PCB 2), 4-chlorobiphenyl (PCB 3), and biphenyl were purchased from Sigma-Aldrich. All chemicals were used as received without further purification. Milli-Q water (18 MΩ·cm) was used for all experiments.

Synthesis of Cu2O/Pd Cubes

Cu2O cubes were prepared according to synthetic methods established by Sui et al.[20] In detail, 3.0 g of PVP was dissolved in 180 mL of an aqueous 0.038 M CuSO4 solution with vigorous stirring in a round-bottom flask. Upon complete dissolution of the PVP, the solution was stirred for an additional 10 min. Next, 40 mL of an aqueous 0.37 M sodium citrate and 0.61 M sodium carbonate solution mixture was slowly added dropwise with continuous stirring. This was followed by the addition of 50 mL of an aqueous 1.4 M glucose solution, which was again slowly added dropwise with continuous stirring. After complete glucose incorporation, the reaction was stirred for 10 min. The flask was then placed in a 70 °C water bath for 2 h without stirring. Once complete, the dark orange precipitate was filtered through a 0.2 μm polycarbonate membrane, thoroughly washed with water and ethanol, and dried under vacuum at 60 °C for at least 12 h.

The as-synthesized Cu2O cubes were subsequently coated with Pd nanoparticles via a galvanic exchange reaction.[9] The production of the Cu2O/Pd cubes is described using the deposited Pd mass % to differentiate the samples. The fabrication of 1 wt % Pd Cu2O/Pd materials is described below; however, changes to the mass of Pd added were employed to reach the appropriate metal loading. In a round-bottom flask, 11.3 mg of Pd(CH3COO)2 was dissolved by slowly adding 200 proof ethanol until a total volume of 150 mL was reached with stirring and sonication. Upon complete dissolution of the Pd(CH3COO)2, 450 mg of the Cu2O material was added to the solution and sonicated until the Cu2O powder was well-dispersed. The solution was stirred for 24 h in the dark as the galvanic exchange reaction proceeded. The Cu2O/Pd 1 wt % cubes were then filtered, washed with ethanol, dried under vacuum at 60 °C, and then stored in a vacuum desiccator.

Photocatalytic Reductive Dechlorination of PCBs

For each photocatalytic experiment, 100 mg of the Cu2O/Pd materials was added to a 60 mL borosilicate glass vial capped with a Teflon-coated silicone septum closure. To this, 50 mL of a 50/50 ethanol/water (v/v) solution of the specific PCB congener at a concentration of 25 μM was added after being bubbled with N2 for at least 30 min. This resulted in a catalyst loading of 2 mg/mL. Next, the headspace of the vial was purged with N2 and the Cu2O/Pd materials were suspended by sonication. Once the materials were fully dispersed, the reaction was vigorously stirred while being irradiated with a 450 W medium pressure Hg-vapor lamp operating at ∼60 mW/cm2, based on the manufacturer’s specifications, where the sample to light source distance was ∼10 cm. To analyze the reaction progression, 500 μL aliquots of the reaction mixtures were taken at predetermined time intervals in at least triplicate trials. Each aliquot was combined with 500 μL of hexanes in order to extract the unreacted PCB and the biphenyl product for gas chromatography (GC) analysis. Separate reactions were also conducted using the exact same procedures; however, the amount of Cu2O/Pd particles was adjusted to 50 and 150 mg to give 1 and 3 mg/mL catalyst loadings, respectively. All experiments were conducted under ambient conditions and at room temperature.

Characterization

SEM was performed using a FEI/Philips XL-30 field emission SEM equipped with an Advanced Detector Technologies EDS detector; TEM studies were completed using a JEOL JEM-1400 TEM. ICP-OES analysis was performed using a Varian Vista-Pro CCD simultaneous ICP-OES spectrophotometer. Powder XRD analysis was completed on a Philips MRD X’Pert diffractometer using Cu Kα radiation. Finally, reaction progression analysis was conducted on an Agilent 7820A GC equipped with a flame ionization detector and on an Agilent 5975C GC-MS. All chromatograms, unless otherwise noted, showed only the presence of PCB, biphenyl, and hexane.

Results and Discussion

Materials Synthesis and Characterization

As illustrated in Scheme , to achieve light-driven PCB degradation, the photocatalytic capability of Cu2O was combined with the hydrodehalogenation activity of Pd to produce a composite structure for photodriven tandem catalysis. For this, H2 generation is processed via proton reduction by Cu2O photoexcitation facilitated by the Pd, followed by reductive dehalogenation using the in situ generated H2 at the Pd component. As previously reported, the Cu2O/Pd materials have shown significant tandem catalytic activity for the reductive dehalogenation of PCBs.[9] To understand both the fabrication of the composite structure and how this structure affects the reaction kinetics, the Pd content of the composite material was varied between 0 and 15 wt % Pd. In this regard, the Cu2O/Pd materials were synthesized to have a final deposited Pd mass of 1, 3, 6, 9, or 15%, where the percent value listed represents the actual Pd mass % (relative to the Cu2O mass) deposited onto the oxide surface. Prior to Pd deposition, Cu2O cubes were generated with an edge length of 665 ± 55 nm (Figure a). Figure also presents SEM images of the Cu2O/Pd materials after the galvanic deposition process. In general, high yields of well-defined cubic composite structures were generated for all of the samples. Figure b specifically presents the Cu2O/Pd 1% materials. Here, distinct cubes were generated with an average edge length of 663 ± 69 nm. Such dimensions indicate that the addition of 1% Pd by mass does not substantially alter the size of the cubes as compared to that of the initial Cu2O structures. As the Pd mass increased to 3, 6, 9, and 15% (Figure c–f), a shift in particle size was observed; although it was evident that cubic materials were again prepared, the average particle size increased to 689 ± 32, 738 ± 67, 752 ± 63, and 814 ± 91 nm, respectively. This increase in particle size likely arises from the deposition of the Pd materials, as confirmed by TEM analysis (discussed below). For all particle sizing, at least 100 cubes were measured over multiple SEM images (Supporting Information, Figure S1). Interestingly, an ombré color effect from light to dark was observed for the color of the materials as the Pd content increased. To this end, a dark amber color was noted for the Cu2O/Pd 1% sample, whereas the Cu2O/Pd 3% particles were light brown in color. As the Pd mass increased to 6, 9, and 15%, dark brown to black colors were observed for these materials, respectively, indicative of the significantly higher amounts of Pd deposited.

Figure 1

SEM images of the Cu2O/Pd composite structures with a Pd mass percent deposited of (a) 0% (bare Cu2O cubes), (b) 1%, (c) 3%, (d) 6%, (e) 9%, and (f) 15%.

To thoroughly examine the structural effects of the various Pd amounts on the composite materials, TEM analysis was conducted (Figure ). Figure a presents the bare Cu2O cubes prior to Pd deposition, where the left image shows the overall cube structure and the right image displays a high-magnification analysis of the oxide edge. Unfortunately, due to the material thickness, imaging of the cube facet is not possible. In this sample, it is clear that the surface of the metal oxide cube is smooth, as anticipated. Figure b presents the analysis for the Cu2O/Pd 1% structures, where it is evident that the Pd nanoparticles were directly deposited on the cube surface. To this end, the metallic nanomaterials were imaged on the Cu2O surface as the light gray rough region at the oxide edge as compared to the Pd-free materials. As Figure b depicts, the size and spatial distribution of the noble metal components were disperse, likely arising from the galvanic deposition process, and prior elemental mapping studies confirmed that the Pd remains on the oxide surface.[9] As the Pd content increased in the Cu2O/Pd 3, 6, 9, and 15% (Figure c–f) samples, the deposited Pd layer topology of the surface became progressively rougher and flake-like, suggesting that multiple layers of Pd materials were incorporated.

Figure 2

TEM images of (a) bare Cu2O cubes and (b–f) Cu2O/Pd materials with a deposited Pd mass of (b) 1%, (c) 3%, (d) 6%, (e) 9%, and (f) 15%. The left panel displays a single cube, and the right panel shows the zoomed-in area indicated in the left image.

Additional characterization to confirm the composite material morphology was conducted (Figure ). To quantify the Pd content in the composite structures, EDS was employed. From the EDS analysis of the Cu2O/Pd 9% materials (Figure a), Pd was successfully deposited, as evidenced by a strong peak at ∼2.8 keV corresponding to the Pd Lα X-ray line. As anticipated, Cu was also present in the sample arising from the Cu2O core component. Similar analyses were conducted for the Cu2O/Pd 1, 3, 6, and 15% materials. The EDS assessment indicated that the deposited Pd mass percent depended on the added Pd mass percent (relative to the Cu2O sample mass) in the galvanic exchange reaction. The plot presented in Figure b shows the linear relationship between the Pd mass percent added into the galvanic exchange reaction and the Pd mass percent deposited on the Cu2O surface. For instance, when 153 mg of Pd(CH3COO)2 (17% Pd mass) was added in the reaction with 450 mg of Cu2O, 9.1 ± 0.1% Pd mass was deposited, as measured by EDS. In general, diminished deposition amounts were observed as compared to the reaction stoichiometry, most likely due to incomplete Pd2+ reduction at the Cu2O interface as the filtrate was pale yellow in color for the higher Pd loadings. Pd wt % amounts for the Cu2O/Pd materials were also confirmed with ICP-OES and are presented in Table S1 of the Supporting Information. Finally, the crystallinity of the materials was studied via powder XRD (Figure c). Diffraction patterns of the as-synthesized structures were compared with the pattern for bulk Cu2O. The diffraction patterns for all of the Cu2O/Pd materials are consistent with the cubic phase of Cu2O, displaying reflections at 29.4, 36.3, 42.2, 61.4, 73.6, and 77.4° 2θ, corresponding to the (110), (111), (200), (220), (311), and (222) lattice planes of Cu2O, respectively.[21] Furthermore, no peaks arising from Cu0 metal, CuO, or Pd were observed. Such results suggest that Cu2O is the dominant species present in the composite materials, consistent with the SEM, TEM, and EDS results. Additionally, the Pd layer on the Cu2O surface is quite thin, especially when considering the thickness of the oxide material, and beyond the detection limit of the technique.[22−24]

Figure 3

Cu2O/Pd characterization. (a) EDS analysis of the Cu2O/Pd 9% materials. For this example, when sufficient Pd(CH3COO)2 was added to the bare Cu2O cubes to reach 17% Pd mass added (relative to the Cu2O mass), 9% Pd mass was deposited, as measured by EDS. (b) Linear relationship between the Pd mass percent added in the reaction and the Pd mass percent deposited. (c) XRD patterns of the Cu2O/Pd materials compared to that of Pd-free Cu2O cubes (intensities are offset for clarity).

Taken together, significant structural differences can be noted in these materials based upon the effect of the Pd content. For this study, the major difference in the preparation process for the different materials was the amount of Pd introduced during the galvanic exchange reaction, where the amount of Pd deposited on the Cu2O cubes was directly proportional to the amount of Pd2+ added into the galvanic exchange reaction. This generally resulted in an increasingly rough and flakey material surface as greater amounts of Pd were deposited; however, all of the different final composite structures maintained a cubic morphology, where the size increased proportional to the amount of Pd deposited on the oxide surface. While the XRD measurements indicate that only Cu2O is present in the samples, EDS and ICP-OES confirmed the presence of surface Pd for all of the samples, which was visually observed by TEM. This effect is attributed to the thinness of the Pd layer on the surface of the oxide cube.[24]

Photocatalytic Reactivity

The photocatalytic reductive dechlorination of PCBs was examined using the Cu2O/Pd 1, 3, 6, 9, and 15% particles irradiated with a Hg-vapor lamp in a photochemical cabinet. The reactivity of the materials was evaluated using three monochlorinated PCB congeners to ascertain steric effects on the material reactivity: PCB 1, PCB 2, and PCB 3. These PCB substrates were specifically selected due to their positioning of the Cl group at the ortho, meta, and para positions, respectively. Such structural features are known to affect the catalytic reactivity based on steric constraints.[25−27] For these reactions, 100 mg of particles was mixed into 50 mL of the 25 μM PCB solution, resulting in a 2 mg/mL catalyst loading, followed by light irradiation for 250 h. Figure a presents the dechlorination analysis of PCBs 1, 2, and 3 using the Cu2O/Pd 1% materials. By monitoring the dechlorination of the PCBs over time, the kinetics of the reactions were determined and compared. For this system, PCB 3 (chlorine in the para position) was dechlorinated the most efficiently. In this regard, 89% of PCB 3 was dechlorinated after 250 h, resulting in a pseudo-first-order rate constant kPCB3 of (7.7 ± 0.7) × 10–3 h–1.[28,29] Such a result was anticipated because PCB 3 is the least sterically hindered congener, as compared to the other substrates. When PCBs 2 and 1 were employed in the reaction system using the Cu2O/Pd 1% tandem catalysts, kPCB2 and kPCB1 values of (2.3 ± 0.5) × 10–3 and (2.4 ± 0.7) × 10–3 h–1 were noted, respectively (Figure and Supporting Information, Table S2).

Figure 4

Overall reaction analysis for the photocatalytic reductive dechlorination of PCBs 1, 2, and 3 using the Cu2O/Pd materials at a catalyst loading of 2 mg/mL in the reaction: (a) Cu2O/Pd 1%, (b) Cu2O/Pd 3%, (c) Cu2O/Pd 6%, (d) Cu2O/Pd 9%, and (e) Cu2O/Pd 15%. Note that lines are added to guide the eye.

Figure 5

Comparison of the pseudo-first-order rate constants for PCB photodechlorination for each of the Cu2O/Pd materials at a catalyst loading of 2 mg/mL.

Identical analyses were conducted for all of the composite structures prepared at the selected Pd loadings (Figure ). For all of the particles prepared, the rate constants for PCB photodechlorination generally maintained the para (kPCB3) > meta (kPCB2) ≥ ortho (kPCB1) trend (Figure ). In this regard, when the Cu2O/Pd 3% sample catalyzed the reaction, rate constants of (3.4 ± 0.5) × 10–3, (2.7 ± 0.1) × 10–3, and (6.3 ± 0.3) × 10–3 h–1 were observed for the PCB 1, 2, and 3 substrates, respectively. Note that this is the only sample where kPCB2 < kPCB1. The k values determined for the Cu2O/Pd 6% catalyzed reaction were (2.5 ± 0.2) × 10–3 (PCB 1), (4.0 ± 0.2) × 10–3 (PCB 2), and (7.7 ± 0.3) × 10–3 h–1 (PCB 3), which were slightly higher than those noted for the Cu2O/Pd 3% system. Interestingly, the material with 9 wt % Pd content (Cu2O/Pd 9%; Figure d) was the most efficient dechlorination system. Complete dechlorination of PCB 3 was practically achieved in 75 h. At this time point, 98% of the substrate was dechlorinated, giving rise to a kPCB3 value of (50.8 ± 2.9) × 10–3 h–1. Almost complete dechlorination (95%) of PCB 2 was reached after 250 h of irradiation, with a corresponding kPCB2 value of (11.0 ± 0.7) × 10–3 h–1. For the PCB 1 substrate, diminished reactivity was noted, as anticipated, where 39% of this reagent was dechlorinated in 250 h, with a kPCB1 value of (1.9 ± 0.1) × 10–3 h–1. When using the material with the highest Pd loading studied (Cu2O/Pd 15%; Figure e), reduced reactivity was observed as compared to the Cu2O/Pd 9% sample. In this regard, when the Cu2O/Pd 15% sample catalyzed the reaction, rate constants of (2.4 ± 0.2) × 10–3, (6.3 ± 0.3) × 10–3, and (36.5 ± 1.3) × 10–3 h–1 were observed for the PCB 1, 2, and 3 congeners, respectively. These lower k values, as compared to Cu2O/Pd 9%, may be due to the extensive Pd coverage of the oxide core, therefore obstructing light absorption. Nevertheless, these results indicate that the reactivity of the Cu2O/Pd materials for PCB photodechlorination is maximized for 9 wt % Pd on the Cu2O core.

When considering all of the Cu2O/Pd materials studied, a trend was evident where the rate of reductive dehalogenation was correlated to the location of the Cl substituent in the PCB molecule; catalytic removal of chlorine is typically favored in the order of para > meta ≥ ortho positions.[25−27] In general, while all of the synthesized structures were photocatalytically reactive for the reductive dechlorination of all of the PCB congeners studied, the most significant dechlorination was observed for PCB 3, regardless of which Cu2O/Pd sample was used as a catalyst. From this, it is evident that substrate steric effects play an extremely important role in the material reactivity. In particular, the rate constants for PCB 1 were the lowest among all of the Cu2O/Pd materials, ranging over a small window of (1.9–3.4) × 10–3 h–1. This suggests that the amount of Pd deposited on the Cu2O has little effect on the dechlorination of PCB 1, most likely due to its steric hindrance. Furthermore, dechlorination of PCBs 2 and 3 was observed to be enhanced as the mass of Pd increased in the catalytic materials. This likely arises from the optimized inorganic morphology that displays increased Pd materials for reductive dechlorination, as well as the accessibility of the Cl substituent in the substrate structure. Taken together, these catalytic results suggest that the Pd content, composite structures, and substrate steric effects work synergistically to influence the overall reactivity.

On the basis of the changes in particle structure, such information can be correlated to the observed reactivity to determine structure/function relationships. In general, the greatest reactivity for the photodechlorination reaction was observed from the Cu2O/Pd 9% sample. Such results were quite interesting as this material did not possess the greatest amount of surface-deposited Pd. To elucidate the structure/function relationships of the composite material, a set of control experiments was carried out. To probe the photocatalytic effect of the Cu2O materials, a solution of PCB 3 was bubbled with H2 for 5 h in the presence of Cu2O/Pd 9% particles at a catalyst loading of 2 mg/mL while in the dark (Supporting Information, Figure S2a). After 3 h, complete dechlorination of PCB 3 was observed with only a modest amount of biphenyl being detected. It is likely that biphenyl was the dominant product generated, as indicated by GC analysis; however, due to the catalytic setup, aerosolization of the biphenyl occurred, resulting in diminished amounts detected. To confirm this aerosolization effect, a solution of biphenyl was bubbled with H2 for 5 h in the absence of a photocatalyst while in the dark (Supporting Information, Figure S2b). Within 2.5 h, complete aerosolization of biphenyl was observed. Taken together, these results suggest that the production of H2 is the rate-determining step for the tandem catalytic process. In this regard, the high degree of Pd coverage for this sample is anticipated to facilitate H2 generation to greater degrees than those materials with lesser amounts of Pd, as observed herein, due to diminished charge recombination effects; however, when more Pd is deposited on the surface, as in the Cu2O/Pd 15% sample, lower reactivity is observed due to diminished light absorptivity by the Cu2O core. In addition to activating the reductive dehalogenation step of the tandem catalytic process, the Pd metal surface likely increases the overall surface area of the composite structure, providing additional adsorption sites for H+ ions. Pd could then reduce adsorbed H+ ions through the electrons transferred from the Cu2O conduction band to form H2.[30] Previous BET surface analysis has confirmed that surface-deposited Pd increases the overall surface area of the composite material.[9] Therefore, it is possible for H2 generation to occur at both the Cu2O and Pd sites. Altogether, these results provide important insights into the catalytic functions of the individual components of the composite structure and their overall roles in the tandem catalytic process.

To confirm that the observed reactivity arose from the particles and not light-based degradation of PCBs, a series of additional control studies was conducted. For this, solutions of PCBs 1, 2, and 3 were irradiated for 250 h in the absence of the photocatalyst (Supporting Information, Figure S3). In general, while all of the PCB congeners demonstrated varying degrees of photobased degradation, biphenyl formation was negligible. For instance, PCB 1 was quickly converted into other oxidative, and nonreductive, dechlorination products, such as hydroxybiphenyl, as detected via GC-MS, with minor generation of biphenyl. Furthermore, in agreement with previous studies,[28,29] it was observed that PCB photodegradation was favored in the order of Cl substitution at the ortho > meta > para positions under light irradiation in the absence of a photocatalyst. Such a trend is in direct opposition to the current results, where photocatalytic reductive dehalogenation of PCBs favored a Cl substitution order of para > meta ≥ ortho positions, with biphenyl as the only product formed. Taken together, this suggests that other photodriven chemical processes that do not follow reductive dechlorination pathways are occurring in the absence of the catalyst to generate such products as hydroxylated PCBs. It should be noted that hydroxylated PCBs have enhanced toxicity compared to that of the parent substrate;[31−33] thus, reductive pathways of degradation, as provided by the Cu2O/Pd material, are preferred.

While the above control reactions confirmed the Cu2O/Pd reductive dechlorination reactivity, it is essential to determine at what concentration the catalyst loading affects the tandem photocatalytic performance. This is especially important for transitioning energy- and material-intensive processes to photocatalytic routes in order to maximize reactivity while minimizing the consumption of resources. In this regard, modified photocatalytic reactions using the Cu2O/Pd 9% materials were carried out where the catalyst mass employed was varied to provide loadings of 1 and 3 mg/mL. Note that all of the reactions discussed above have a catalyst loading of 2 mg/mL. From this analysis, shown in Figure a, different reactivities were observed for the various catalyst loadings for the dechlorination of PCB 3. Interestingly, under these conditions, 2 mg/mL was the optimal catalyst loading for the dechlorination process. As mentioned above, under these conditions 98% of PCB 3 was dechlorinated in 75 h, with a kPCB3 value of (50.8 ± 2.9) × 10–3 h–1 (Figure b and Supporting Information, Table S3). The 1 mg/mL catalyst loading gave a decreased kPCB3 value of (17.6 ± 0.7) × 10–3 h–1. A further diminished kPCB3 value of (9.9 ± 0.9) × 10–3 h–1 was observed for the 3 mg/mL catalyst loading. These results indicate that 2 mg/mL is the most efficient catalyst loading for the dechlorination process. At a lower catalyst loading, the decreased amount of catalytic materials presented led to lower reactivity, whereas for the 3 mg/mL catalyst loading, the high material concentration resulted in inefficient light absorption by the materials, causing diminished reactivity.

Figure 6

Catalyst loading analysis using the Cu2O/Pd 9% materials for the photocatalytic reductive dechlorination of PCB 3: (a) overall reaction analysis using the indicated catalyst loadings and (b) comparison of the pseudo-first-order rate constants for each of the indicated catalyst loadings.

From the above studies, it is clear that the Cu2O/Pd materials control the overall photodriven reductive dechlorination process and that the reaction rates are highly sensitive to the composite structure and catalyst loading. To probe whether the photocatalytic reductive dechlorination reaction affects the material structure, SEM and TEM analyses were conducted on the Cu2O/Pd 9% materials after being used for the dechlorination of PCB 3 at a catalyst loading of 2 mg/mL, as shown in Figure . Additional SEM and TEM images of the materials presented for the postreaction analysis can be found in the Supporting Information, Figure S4. Imaging of the Cu2O/Pd 9% cubes postdechlorination showed that while the majority of the structures remained intact, some of the particles became hollow (∼29%). In this regard, Figure a shows a large area SEM image of the materials where fully intact materials are present, although hollow structures are also evident. Figure b shows a zoomed-in SEM image of the cubic materials in the red box of Figure a. In this image, it is apparent that significant internal structural changes have occurred for this set of materials. TEM analysis of these materials further exposes the dramatic structural changes to the particles that become hollow during the reaction (Figure c). In this image, it is clear that the Cu2O/Pd materials have been internally changed, whereby the oxide component was potentially partially etched while the Pd layer remained intact. To examine the material composition of the postdechlorinated particles, EDS analysis was performed. These assessments indicated that the Pd content (relative to the Cu2O mass) of the composite structures increased from 9.1 ± 0.1% Pd mass before the reaction to 9.6 ± 0.6% Pd mass after the reaction occurred, suggesting that the morphological changes arose predominantly from partial etching of the oxide component.

Figure 7

SEM and TEM images of the Cu2O/Pd 9% structures after the photocatalytic reductive dechlorination of PCB 3 using a catalyst loading of 2 mg/mL. (a, b) SEM and (c) TEM images. Panel (b) shows a zoomed-in image of the red box in panel (a).

To identify and understand the structural modifications to the particle morphology due to photocatalysis as opposed to photocorrosion, a set of control experiments was conducted. In this study, at a catalyst loading of 2 mg/mL, the Cu2O/Pd 9% materials were photoirradiated in the reaction solvent for 250 h in the absence of PCBs. SEM analysis showed that the majority of the cubes were still whole (71% of the sample); however, 29% of the materials appeared to be pitted, suggesting that photocorrosion had occurred (Supporting Information, Figure S5a).[8,34,35] This pitted morphology was significantly different than that observed after the photodechlorination reaction, which generated hollow structures, suggesting that the hollowing effect was not due to photocorrosion. To reaffirm this hypothesis, a second control study was completed where bare Cu2O cubes at a catalyst loading of 2 mg/mL were photoirradiated in the solvent for 250 h. In this analysis, pronounced surface reconstruction was observed (Supporting Information, Figure S5b), displaying jagged overgrowths on the oxide. Although no pitting was evident in the bare Cu2O sample, hollow cubes were also not observed, supporting the hypothesis that photocorrosion does not drive Cu2O/Pd material etching.

These postreaction analyses provide important insights concerning the photocatalytic mechanism of the Cu2O/Pd composite structures. In the reaction system with PCB 3, hollowing of the cubes is evident. This control study suggests that the hollow cubes are a result of photodechlorination, which may affect the recyclability of the materials vide infra; pitting is evident from the Cu2O/Pd 9% materials and particle surface reconstruction is observed in the bare Cu2O cubes after being photoirradiated in ethanol/water. Additionally, the postdechlorination wide-area EDS analysis shows that the Pd mass percent is higher after the reaction. Taken together, these postreaction analyses suggest that photodechlorination is responsible for the hollowing of the cubes. Although speculative, it is possible that the shell of the hollow cubes is composed mainly of Pd; however, additional studies are required to determine this fine level of detail.

Because of the observed structural changes of the Cu2O/Pd after the first catalytic cycle, a reduction in the photocatalytic activity for subsequent catalytic cycles could be observed. As such, recyclability studies of the composite materials as catalysts for PCB photodechlorination were conducted using Cu2O/Pd 7.5% materials with PCB 3 at a catalyst loading of 2 mg/mL. Figure S6 in the Supporting Information presents the dechlorination analysis and the resulting rate constants for the recyclability studies. For this process, when the 75 h reaction period of the first cycle was completed, the Cu2O/Pd 7.5% material was filtered, washed, and dried. Once the material was completely dry, it was weighed, and the appropriate volume of PCB 3 was added in order to maintain a catalyst loading of 2 mg/mL. For the first cycle, 88% of PCB 3 was dechlorinated after 75 h, resulting in a kPCB3 value of (29.8 ± 2.0) × 10–3 h–1. For the second cycle, 60% of PCB 3 was dechlorinated after 75 h, giving a kPCB3 value of (10.6 ± 1.2) × 10–3 h–1. As is evident, diminished reactivity was noted for the second reaction cycle, likely arising from the catalytic material degradation. It is evident that the redox potential of Cu2O intrinsically exists within its band gap energy, which might lead to self-oxidation or self-reduction and depletion of copper species into the solution.[36] Additionally, Cu2O oxidation to CuO at the composite material surface could be occurring resulting in diminished reactivity; however, partial reduction back to Cu2O should restore the reactivity. Additional studies are required to determine the actual basis of the reactivity changes.

Optimization of this tandem catalytic system is still in progress; however, in comparison to other catalysts used for the dechlorination of PCBs, the Cu2O/Pd composite materials are highly unique in that light is used as the energy source to drive the reaction. For this, the most established class of materials for PCB dechlorination is based on Pd/Fe nanostructures. For example, Wang et al. reported the synthesis of nanoscale Pd/Fe particles for the dechlorination of PCBs.[37] Within 17 h, complete dechlorination of the PCB congeners of Aroclor 1254 by Pd/Fe nanoparticles was observed. The initial PCB solution mixture concentration was 5 mg/L, and the catalyst loading was 50 mg/mL. Although complete dechlorination was achieved within 17 h, a less concentrated PCB solution and a significantly higher catalyst loading than those employed herein were used. In an additional approach, Zahran et al. synthesized Pd/Fe bimetallic nanotubes that demonstrated high reactivity.[38] Dechlorination of 25 μM 3,3′4,4′-tetrachlorobiphenyl using Pd/Fe bimetallic nanotubes with 0.9 wt % Pd was achieved in 25 h at a catalyst loading of 0.25 mg/mL. This high dechlorination efficiency was attributed to the high surface area to volume ratio of the hollow nanotubes structure. For these systems, the production of H2 by Fe0 in aqueous solution leads to iron corrosion. As such, H2 production and iron reactivity decrease over time; therefore, Pd/Fe materials are not sustainable as a result of surface oxidation and precipitation of iron oxides/hydroxides on the surface of the iron.[39,40] While these materials are the most reactive for PCB dechlorination, they are not directly comparable to the Cu2O/Pd systems due to the significant differences in H2 generation (Fe oxidation vs photocatalysis). While the Cu2O/Pd materials demonstrated lower reactivity, they represent a new avenue to materials with potential long-term reactivity for in-field photodechlorination using sunlight as the energy source. Further studies are underway to enhance the reactivity of the tandem catalysts, including pathways to enhance their recyclability.

Conclusions

In summary, we have generated cubic Cu2O/Pd composite structures with tandem photocatalytic reactivity. It was demonstrated that the Pd content in the material affects the particle surface morphology, where higher amounts of Pd result in structures with a flakey topology. Additionally, the cubic shape of the particles was maintained regardless of the material composition. Although simple, this architecture is strategic for transitioning energy- and material-intensive reactions that require H2 as a reagent to more sustainable, photocatalytic methods. The as-synthesized Cu2O/Pd particles demonstrated high tandem catalytic performance for the reductive dehalogenation of PCBs 3 and 2, but they showed slow dechlorination rates for PCB 1. The data indicate that the overall dechlorination reaction rate results from a synergistic effect of several key factors, including the surface morphology and composition of the materials and the steric effects of the substrate. Postreaction analysis showed that some of the Cu2O/Pd cubes had become hollow following dechlorination. The light-driven tandem catalytic system demonstrated herein exemplifies how current photocatalyst materials for H2 production can be applied to reactions not typically fueled by light. The fundamental understanding of material structural effects over catalytic function and reactivity is important in the design of photocatalytic systems for sustainable reactivity.