CFD modelling of an immobilised photocatalytic reactor for phenol degradation

Water is considered as one of the most vital natural resources for the sustainability of life. Growing population and industrial development have led to a decline in clean water sources and an increase in wastewater generation (Riffat & Husnain 2022). The National Institution for Transforming India (NITI Aayog) report on ‘Composite Water Management Index (2018)’, mentioned a total of 600 million Indian population encountering high water crisis. The country with 70% of contaminated water is placed 120th among 122 countries in the water quality index (WQI) (NITI Aayog 2018). Contaminated water is a complex mixture of organic and inorganic components, microbes and chemicals, which impart adverse effects on both aquatic and terrestrial environments (Riffat & Husnain 2022). Recalcitrant compounds are resistant to degradation and form toxic intermediate products under conventional treatment methods (Mohamad Said et al. 2021). Phenolic compounds are recalcitrant organics categorised as pollutants of priority concern by the European Union and the United States Environmental Protection Agency (USEPA) (Anku et al. 2017). Industrial sources of phenol include petrochemicals, pulp and paper industries, pharmaceuticals, textiles and many more (Saputera et al. 2021). Advanced oxidation processes (AOPs) use highly reactive oxygen species, which helps in the complete mineralisation of pollutants to CO₂, H₂O and mineral acids (in presence of halogens) (Andreozzi et al. 1999; Joseph et al. 2009). Hydroxyl radical with an oxidation potential of 2.8 V is the most reactive oxidising species used in water and wastewater treatment (Ameta & Ameta 2018). AOPs were found to be efficient in the removal of emerging and recalcitrant pollutants with low biodegradability. In comparison, AOPs were found to be comparable in cost with that of conventional treatment methods (Andreozzi et al. 1999; Bracamontes-Ruelas et al. 2022). Photocatalysis, an AOP involves semiconductor activation by means of irradiation to completely mineralise several organic pollutants. TiO₂ is the most preferred semiconductor due to its high chemical stability, non-toxicity and low cost (Ameta & Ameta 2016). The coupling of photocatalysis with biological treatment methods was reported to be efficient in the reduction of total organic carbon (TOC) in a study conducted by Lin et al. (2020).

Photocatalytic reactor designs should consider factors such as working regime, photocatalyst, irradiation, as well as the volume of liquid treated. A well-designed reactor speeds up the treatment process with a reduced power consumption hence making the treatment technology more economical. Scaling up of photocatalytic reactors is one of the major challenges in this research area (Enesca 2021). Computational fluid dynamics (CFD), a computer simulation technique is widely used in fluid flow, heat transfer and chemical reaction analysis. CFD programmes due to their adaptability are effective tools that can be used for industrial and non-industrial purposes. The technique offers several advantages over experimental methods including reduction in time and cost and improved accuracy of results (Versteeg & Malalasekera 2007). The choice of a CFD solver depends on factors such as domain type, mesh type, flow type, flow compressibility, type of reference frame, customisation capabilities and so more (Liu & Zhang 2019). ANSYS Fluent is one such user-friendly commercial CFD package used to simulate fluid flow, heat transfer and fluid–catalyst interactions with accuracy, using the finite volume method approach (Jeong & Seong 2013). Modern user interfaces in commercial CFD software help streamline different problem configurations and results analysis. User-defined functions in ANSYS Fluent can be coded externally as a supplementary function and can used to attain accurate solutions (Barraza-Jiménez et al. 2019). A study by Vezzoli et al. (2011) used ANSYS Fluent solver to design a flat plate photocatalytic reactor for phenol degradation over titanium dioxide photocatalyst. The investigation was focused on the impact of different operational parameters on the reaction kinetics. These standard CFD programs are equipped with data visualisation tools such as 2D and 3D plots, coloured graphical representations and animations. The availability of features such as different perspective views, vector plots, and grid displays has simplified the analysis processes (Barraza-Jiménez et al. 2019; Mann 2008; Ruiz-Soto et al. 2020). Duran et al. (2010) conducted a performance prediction study of annular photocatalytic reactors under different hydrodynamic models over a wide range of flow rates and intrinsic reaction rates. In a study by Devia-Orjuela et al. (2019) a baffled flat plate photocatalytic reactor under UV light was used for the degradation study to understand the effect of baffles on degradation efficiency. In this particular research, it was observed that the baffles helped in fluid mixing, which in turn enhanced the degradation efficiency. Ahmed et al. (2022) modelled a flat plate photocatalytic reactor using ANSYS Fluent solver for phenol degradation. The analysis reported that the degradation increased with an increase in radiation and a decrease in phenol concentration and flow rate. Ruiz-Soto et al. (2020) demonstrated hydrodynamic analysis on a photocatalytic reactor using ANSYS Fluent. A study by Gao et al. (2023) was focused on the optimisation of geometric structure and reaction conditions to enhance photocatalytic performance leading to the development of a honeycomb reactor as the optimum design. The current study focuses on the design and comparison of immobilised photocatalytic reactors in the modelling of hydrodynamics, species transport and chemical reaction kinetics using ANSYS Fluent. Phenol degradation was the simulated reaction in the study, reaction conditions of the same were validated from previous works of Ahmed et al. (2022) and Vezzoli et al. (2011).

ANSYS Fluent is a commercial CFD package that utilises the finite volume method to solve the governing equations of flow, which include conservation of mass, momentum, energy and species transport equations as described below.

The workflow followed in a CFD project consists mainly of four steps. Problem definition is the initial phase of study wherein suitable assumptions and simplifications are fed into the model in regard to problem domain, initial and boundary conditions. The pre-processing stage inputs the problem onto CFD through user-friendly interface. Steps included in this stage are domain definition, grid generation, selection of physical and chemical phenomenon to be modelled and definition of material properties and boundary conditions (Versteeg & Malalasekera 2007). Grid generation is an important step in pre-processing, as the accuracy of the simulation results depends on the quality of meshin (Ferziger et al. 2020). The solving stage deals with solving fluid flow (conservation of mass, momentum, energy, species transport) equations using numerical solution techniques. Most of the CFD codes utilised the finite volume method as the numerical technique (Versteeg & Malalasekera 2007). The post-processing stage analyses and visualises the results in the form of vector plots, contour plots, X–Y plots, streamlines, animations, and particle tracking and can be exported as reports (Liu & Zhang 2019).

ANSYS Fluent 2022 R1 software package is the CFD solver used in the present study. Fluent operates in a single window workflow known as ANSYS workbench, which is a modern and user-friendly interface for streamlining CFD processes.

Flow and reaction conditions used in the study were validated from the previous study by Ahmed et al. (2022) titled ‘Phenol degradation of waste and stormwater on a flat plate photocatalytic reactor with TiO₂ on a glass slide: An experimental and modelling investigation’. 2D geometry simulated the flow conditions and 3D geometry simulated the reaction conditions used in the study.

The reactor design considered overall outer dimensions as 810 mm × 300 mm × 30 mm, with a constant flow rate of 0.3 L/s. Flow and reaction conditions used in the study are given in Table 1, along with the dimensions of the reactor.

Table 1

Geometry and boundary conditions

Geometry .
Overall outer dimensions	810 mm (length)
	300 mm (breadth)
	30 mm (depth)
Inlet area dimensions	600 mm2
Flow Conditions .
Inlet velocity	0.5 m/s
Flow type	Turbulent
Turbulence model	Realisable k-epsilon
Outlet pressure	101,325 Pa
Reaction Conditions .
Reaction	Pseudo first-order kinetics
Reactive species	O2
Reactive surface	Bottom wall
Inlet phenol concentration	7.78 × 10−7 kmol/m3
Temperature	300 K
Walls	No-slip condition

Geometry .
Overall outer dimensions	810 mm (length)
	300 mm (breadth)
	30 mm (depth)
Inlet area dimensions	600 mm2
Flow Conditions .
Inlet velocity	0.5 m/s
Flow type	Turbulent
Turbulence model	Realisable k-epsilon
Outlet pressure	101,325 Pa
Reaction Conditions .
Reaction	Pseudo first-order kinetics
Reactive species	O2
Reactive surface	Bottom wall
Inlet phenol concentration	7.78 × 10−7 kmol/m3
Temperature	300 K
Walls	No-slip condition

View Large

From Figure 12(b), it can be observed that in the velocity field, minor recirculation zones were present only near the inlet region. However, the recirculation zones in the reactive zones were eliminated along with the removal of flow channelisation present in the serpentine flow path reactor. Thus the addition of baffles has improved the flow homogeneity along with enhanced mixing of fluid.

From Figure 14, it can be inferred that the production of carbon dioxide occurs when the phenol gets oxidised. CO₂ production increases from 0 at inlet to 2.32 × 10⁻⁶ at outlet as shown in Figure 14(a), at the plane located at 0.002 m from the bottom wall. From Figures 13(a) and 14(a), it can clearly be observed that as the reaction progresses, phenol gets converted to carbon dioxide across the reactor. Figure 14(b) shows the distribution of carbon dioxide across the entire reactor.

Based on the results from these models, it can be inferred that the introduction of baffles within the serpentine flow field had a very significant effect on the hydrodynamics as well as the performance of the reactor, mainly because of the mixing of flow at each bend. The serpentine type of geometry has exhibited significant channelisation, but these effects are negated to some extent by the use of transverse baffles across the flow path, this can be due to the disturbance created by the baffles, which are placed across the flow direction. The transverse baffles with horizontal orientation were found to provide better performance compared to vertical and slanting ones. The efficiency also tends to increase with the baffle number and length. At a flow velocity of 0.5 m/s and inlet molar concentration of 7.78 × 10⁻⁷, the horizontal baffled reactor exhibited an efficiency of 43.1% for a residence time of 18.44 s in phenol degradation.

This study was focussed on CFD modelling of photocatalytic reactors using ANSYS Fluent, which would ease the current complexity involved in scaling up and commercialisation of photcatalytic reactors. CFD techniques are effective in the analysis of fluid flows and reactor design, which would reduce time and cost in laboratory experiments. The investigation compared different reactor configurations for the degradation of phenol in an immobilised reactor. A simple cuboidal reactor exhibited a removal efficiency of 27%. To improve the residence time and contact time between the catalyst and pollutant, a serpentine flow path was introduced in the cuboidal reactor, which improved the phenol degradation efficiency to 42%. However, the serpentine flow path presented flow channelisation, which resulted in a non-uniform flow pattern, which may affect pollutant and radiation distribution within the reactor. To cater to this issue, baffles were introduced in the serpentine flow path, which improved the flow homogeneity. The effect of baffle type, number and dimensions on phenol degradation efficiency were also compared. Of which horizontal baffles of curved shape, with 20 mm diameter and 60 mm length were found to have the highest removal efficiency. The optimum number of baffles was found to be three in a stretch. The final reactor model considered was the serpentine flow path horizontal baffled reactor with the optimum baffle conditions, which exhibited an efficiency of 43.1% under a flow rate of 0.3 L/s.

The authors thank the National Institute of Technology Karnataka, Surathkal for the research facilities provided. DPB acknowledges the Ministry of Education, India for the PG fellowship.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.