Improved Photocatalytic Performance of TiO₂–Nitrogen-Doped Graphene Quantum Dot Composites Mediated by Heterogeneous Interactions

Abstract

Photocatalysis is typically monitored via analysis of phases in isolation and focuses on the removal of a target analyte from the solution phase. Here we analyze the photocatalytic action of a TiO₂–nitrogen-doped graphene quantum dot (NGQD) composite on a target analyte, phenol, using comprehensive multiphase NMR (CMP-NMR) which observes signals in solid, solution, and gel phases in situ. Phenol preferentially interacts with the composite photocatalyst compared to pure TiO₂, increasing its effective concentration near the catalyst surface and its degradation rate. The presence of NGQDs in the composite reduced the fouling of the catalyst surface and caused a reduction of photogenerated intermediates. Increased heterogeneous interactions, likely mediated by π–π interactions, are hypothesized to cause each of these improvements in the observed photocatalytic performance by TiO₂–NGQDs. CMP-NMR allows the elucidation of how the photocatalytic mechanism is enhanced via material design and provides a foundation for the development of efficient photocatalysts.

Photocatalysis has long been studied as an effective method of sustainably degrading organic contaminants in water.¹⁻³ A significant body of literature has been assembled on techniques to improve the photocatalytic efficiency of TiO₂, the preeminent material in the field, via doping,¹ morphology control,⁴ and the formation of composites.⁵ One subclass of composites, TiO₂ coupled with graphene quantum dots (GQD), has received heightened interest in recent years due to their improved photocatalytic performance, often attributed to improved lifetimes of charge carriers in the composite compared with the pure material.^6,7

Photocatalytic wastewater treatment is a heterogeneous process involving colloidal catalysts and dissolved small molecules. Several studies have demonstrated how modulating heterogeneous interactions via matrix modification can affect catalytic performance.⁸⁻¹⁰ Despite this, few reports in photocatalysis investigate the role these heterogeneous interactions have, aside from measuring adsorption of the small molecule to the catalyst surface. Our group recently published a report monitoring photocatalysis using Comprehensive MultiPhase Nuclear Magnetic Resonance (CMP-NMR).¹¹ CMP-NMR creates a series of NMR spectra of individual phases defined by their diffusion (solutions, gels) and dipolar interactions (solids). These include the solution phase (bulk solution molecules far from colloidal surface), the solid phase (molecules tightly bound or adsorbed to the colloidal surface), and a restricted diffusion phase (molecules at the interface between the previous phases which are interacting with the colloid but not tightly bound).¹²⁻¹⁴ A detailed description of the origin and interpretation of these spectra in photocatalytic systems has been provided in previous studies.¹¹ This previous work identified that molecules in a diffusion restricted phase, which were distinct from those bound to the colloidal surface, were preferentially targeted by reactive oxygen species (ROS) generated at the catalyst surface.

This study demonstrates that the heterogeneous interactions between analyte and catalyst can be modulated by changing the material composition of the photocatalyst, as was demonstrated with a novel titanium dioxide–nitrogen-doped graphene quantum dot (TiO₂-NGQD) composite photocatalyst. Through these improved analyte–catalyst interactions the TiO₂–NGQD photocatalyst was shown to have increased photocatalytic efficiency, while also being able to retain and subsequently degrade photogenerated intermediates.

Phenol was chosen as a model pollutant in this study due to its industrial relevance and relatively simple NMR spectrum (Figure S1). It was degraded by the TiO₂–NGQD catalyst under Xe arc irradiation (broadband UV and visible light). Diffusion Ordered SpectroscopY (DOSY) is a 2D NMR technique that measures the average diffusion constant of molecules in a sample. When performed with a CMP-NMR probe at magic angle spinning (MAS), it averages the signals arising from species in both the freely diffusing solution phase and the diffusion-restricted gel phase. When an NMR-active small molecule is present with a catalyst, its measured diffusion constant is an average of the measured populations and can appear “slower” depending on the population present in the diffusion restricted gel phase and therefore can be utilized to infer the degree of interaction between the catalyst and analyte. The measured diffusion constants for freely diffusing phenol and phenol in the presence of TiO₂–NGQD and P25 TiO₂ are shown in Figure 1.

Figure 1.

DOSY spectra of phenol in the presence of P25 TiO₂ (green), TiO₂–NGQD (blue), and in the absence of any catalyst (black).

In the absence of a catalyst (phenol control), the diffusion of solution-phase phenol is as shown in Figure 1 (black). Adding TiO₂ (P25) resulted in a slower average diffusion (Figure 1, green), likely because a fraction of the population interacted with the P25 surface and thus experienced restricted motion. In the presence of TiO₂–NGQD, the measured diffusion signals of phenol slowed down by an even larger amount, suggesting that it has a much larger degree of interaction between the colloid and phenol (Figure 1, blue).

CMP-NMR can be used to monitor different phases present within a heterogeneous system and, when applied to a photocatalytic reaction, how those populations change during the course of the reaction.

At various time points during the reaction, the solution was sampled and a CMP-NMR analysis was performed (Figure 2). Through the course of the degradation, a change in each of the aforementioned phases is observed. At time 0 h, taken after adsorption equilibrium is given time to establish but prior to illumination, a significant portion of the phenol population is present in the gel phase (Figure 2B) with a moderate portion present in the solution phase (Figure 2A) and no detectable signals present in the solid phase. Over the course of illumination, phenol in the solution phase is completely removed, which is consistent with photocatalysis’ well-documented success at mineralizing organic contaminants.¹⁵ Solid-phase signals are observed to grow over initial time points, indicating that phenol becomes tightly bound to the surface of the colloidal catalyst, suggesting early stages of surface fouling. With increased irradiation the solid signals are reduced, suggesting that the catalyst is capable of regenerating the surface with sufficient time and light. This is an improvement over P25 TiO₂ whose surface was found to foul under similar conditions.¹¹

Figure 2.

CMP-NMR spectra throughout the photodegradation of phenol by TiO₂–NGQD (2% by mass) under Xe irradiation. Solution state signals (A) arise from molecules far from the surface; gel-phase signals (B) arise from molecules experiencing restricted diffusion; solid signals (C) arise from rigidly bound molecules collected by CP-MAS.¹¹ The cartoon (not to scale) depicts select interactions occurring in various phases of interest.

The restricted diffusion phase is perhaps the most interesting, as it is not observable by other standard analytical techniques. The signals in this phase arise from molecules whose microscopic diffusion is limited relative to purely dissolved molecules but retain the relaxation properties of solvated molecules from the perspective of the NMR. In the phenol/TiO₂–NGQD system, we interpret these molecules as undergoing reversible adsorption (or physisorption) with the colloidal surface.¹² At time 0 h, we observe that a significant portion of phenol signals are present in this phase, corroborating the large degree of interaction with the catalyst observed via DOSY. After illumination, the population of the gel phase decreases drastically and remains at a low level for the remainder of the degradation. This potentially suggests that these molecules are preferentially targeted by the photocatalytic process.

Average diffusion coefficients measured for these later time points show a similar “speeding up” of the phenol peaks consistent with a decrease in the gel-phase population (Figure S2). The influence of light on this sharp decline is further evidenced by noting no significant change when comparing with an identical sample which is kept for 72 h in the absence of light (Figure S3). The phenol undergoes no noticeable change in phase distribution unless catalyzed by light. The interpretation that molecules physiosorbed to the surface are those undergoing degradation suggests a photocatalytic degradation mechanism mediated by ROS generation. In this system, phenol’s limited diffusion ensures it remains in the vicinity of the catalyst’s surface where short-lived ROS can be generated and subsequently reacted with.

The results of the CMP-NMR and DOSY suggest a notable degree of interaction between phenol and TiO₂–NGQD creating the “gel phase” when allowed to come to equilibrium. The phenol present in this gel phase is responsible for the observed “slowing” of phenol via the DOSY. We attribute this interaction to attractive π–π interactions between the graphitic lattice of the NGQDs¹⁶ and the aromatic ring of phenol. Based on the CMP-NMR spectra, no phenol signals are measured in the solid phase (Figure 2C), suggesting that the phenol does not tightly bind, consistent with relatively weak degrees of interaction such as van der Waals π–π interactions. Further, no signals arising from the NGQDs appear in any phase of spectra taken in this study, likely due to their colloidal nature limiting solution signals and natural abundance of ¹³C limiting solid-state detection limits (phenol used was 99.9% ¹³C).

TiO₂–NGQD has shown improved performance when removing common industrial pollutants from the solution phase.^6,17 An emerging area of concern in pollutant removal is the fate of degradation byproducts or intermediates, especially when intermediates are more toxic or have increasing environmental lifetimes compared to their parent analytes. Previous reports have shown that photocatalytically generated intermediates escape the surface of the catalyst¹¹ and must be reabsorbed for subsequent degradation, eventually leading to full mineralization as CO₂ and H₂O. The photocatalytic efficiency of P25 and two composites with varying NGQD loadings (2%, 8% NGQD by mass of TiO₂) were studied via quantitative NMR so that reaction intermediates could be accurately identified. For all catalysts, hydroquinone was detected as the primary photogenerated intermediate, and its production and subsequent degradation as a function of illumination time by various catalysts are shown (Figure 3).

Figure 3.

Evolution of photogenerated hydroquinone throughout the degradation of phenol by P25 TiO₂ and TiO₂–NGQD composites (% w/w). Inset shows the concurrent degradation of phenol in the same trial.

P25 TiO₂ (red line) was found to produce the highest concentration of hydroquinone, which was an order of magnitude larger than that of the composite photocatalysts. The maximum observed hydroquinone concentration does not directly correlate to the degradation rate of phenol as the TiO₂–NGQD composites (both 2%, blue line, and 8%, purple line) were found to have higher phenol degradation efficiencies compared to P25 TiO₂. The maximal hydroquinone concentration measured in solution decreases with increasing amounts of NGQDs, suggesting that the NGQDs can significantly improve the ability of the photocatalyst to retain intermediates and promote the complete mineralization of contaminants.

Photocatalysis is a fundamentally surface-driven process, as ROS are generated at the interface of the catalyst and solution, and their short lifetimes do not allow for significant diffusion into solution. For an organic contaminant, such as phenol, to undergo photocatalytic degradation, it must first diffuse to the surface of the catalyst and remain close enough to interact with the generated ROS. As intermediates are produced during the degradation, they may either diffuse far from the catalyst surface or be retained for subsequent degradation. The presence of NGQDs in a composite material was found to improve the degree of interaction between the catalyst and the dissolved pollutants. This enabled faster rates of analyte degradation and lower maximal concentrations of intermediates escaping the catalyst surface, both significant improvements in the photocatalytic performance.

Through CMP-NMR we observed enhanced heterogeneous interactions between colloidal catalysts and pollutants and hypothesized their impact in improving photocatalytic performance. By leveraging these interactions through material design, we enable a new blueprint for the development of efficient catalysts.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jpclett.4c00335.

• Experimental material and methods along with additional figures (PDF)

Author Contributions

^† M.P.C., R.T.L., and R.G.B. contributed equally.

The authors declare no competing financial interest.