Controlled Delivery of H2O2: A Three-Enzyme Cascade Flow Reactor for Peroxidase-Catalyzed Reactions

Peroxidases are promising catalysts for oxidation reactions, yet their practical utility has been hindered by the fact that they require hydrogen peroxide (H2O2), which at high concentrations can cause deactivation of enzymes. Practical processes involving the use of peroxidases require the frequent addition of low concentrations of H2O2. In situ generation of H2O2 can be achieved using oxidase-type enzymes. In this study, a three-enzyme cascade system comprised of a H2O2 generator (glucose oxidase (GOx)), H2O2-dependent enzymes (chloroperoxidase (CPO) or horseradish peroxidase (HRP)), and a H2O2 scavenger (catalase (CAT)) was deployed in a flow reactor. Immobilization of the enzymes on a graphite rod was achieved through electrochemically driven physical adsorption, followed by cross-linking with glutaraldehyde. Modeling studies indicated that the flow in the reactor was laminar (Reynolds number, Re < 2000) and was nearly fully developed at the midplane of the annular reactor. Immobilized CAT and GOx displayed good stability, retaining 79% and 84% of their initial activity, respectively, after three cycles of operation. Conversely, immobilized CPO exhibited a considerable reduction in activity after one use, retaining only 30% of its initial activity. The GOx-CAT-GRE system enabled controlled delivery of H2O2 in a more stable manner with a 4-fold enhancement in the oxidation of indole compared to the direct addition of H2O2. Using CPO in solution coupled with GOx-CAT-GRE yields of 90% for the oxidation of indole to 2-oxyindole and of 93% and 91% for the chlorination of thymol and carvacrol, respectively.


Table of Contents
Where  ⃗ is the Reynolds-averaged velocity vector (m/s),  is the density (kg/m 3 ),  is the static pressure (N/m 2) , t is time (s), ̿ is the effective stress tensor (N/m 2 ).
For the residence time distribution (RTD) following species transport equation was used to simulate multispecies flow where,   is the mass fraction of the tracer of species  and   is the mass diffusion coefficient of the species  in the mixture.The viscosity and the mass diffusivity of the tracer in the mixture were considered constant.For the determining the RTD of the reactor, the tracer was numerically injected from the inlet surface.It was ensured that more than 99.8% per cent of the injected tracer has passed through the outlet.The RTD was also determined separately by giving specifying mass fraction a on the inner wall of the mid annular space.

b. Boundary conditions and post processing
The flow and residence time distribution (RTD), simulations were performed in the geometry shown in Figure 1 for a flow rate of 0.08mL/min.Velocity was calculated based on the inlet annular area and the volumetric flow rate.The calculated velocity magnitude was given as flat velocity inlet boundary condition to the inlet surface of the reactor.The direction of the velocity was normal to the inlet surface.The outlet surface with the same dimensions as the inlet surface was defined as a pressure outlet (constant pressure boundary condition).
Water was used as a working fluid with a constant density (ρ) of 1000 kg/m 3 , and the dynamic S6 viscosity (µ) of 0.001 Pa.s.For the RTD simulation of the secondary species, the tracer was assumed to have the same properties as water.The mass diffusivity of the tracer was assumed constant with a magnitude of 2 x 10 -9 m 2 /s.
For the RTD study, simulated tracer response after a step input was analyzed using the classic tracer theory presented by Levenspiel 2, 3 .The mean average residence time  ̅ (s), the variance of the curve  2 , E -θ curve and the Dispersion number (/ ) (where  is the dispersion coefficient), were calculated as: Where m i is the mass fraction of the tracer (-), and. =   ̅ ⁄ (-) is the non-dimensional mean residence time, and E is the mean residence time distribution function (s -1 ).

c. Solution to model equation
The model equations were numerically solved by the finite volume method 4-7 using the S7 commercial CFD code Ansys Fluent (Ansys Inc, Version 2021 R2).The governing equations were spatially discretized using the Third-order MUSCL scheme.The pressure and velocity coupling were considered using the SIMPLE algorithm.The pressure was discretized using a PRESTO!scheme.The under relaxation factors were 0.3 and 0.7 for pressure and momentum, respectively.For flow simulation steady state simulation was performed.All the parameter were converged to 10 -6 .
For residence time distribution simulation, converged flow field from the flow simulation was used and only species transport equations were solved.The convergence criteria of 10 -7 were met for the species.Simulation for flow time of 5τ was sufficient to reach a flat mass-weighted averaged concentration of tracer at the outlet of the reactor.The time step was calculated by generally accepted thumb rule of  1000 The simulated outlet concentration data were collected and processed for the RTD curve (E) and the Dispersion number ( / ) according to Equations 4-8.

d. Mesh
The 3-dimensioanl mesh was used for the numerical simulation was generated in Ansys meshing platform.This was strategically split into sweepable blocks in Ansys SpaceClaim to accommodate only hexahedral and pave elements in the entire fluid volume.The curvature and proximities were captured and were also scoped critically, to ensure high quality elements in the critical regions.Biasing was given to critical edges and faces to ensure appropriate growth of cells and transition of cell size.Figure 3-2 shows the mesh images at critical locations.

Figure S3 .Flow modeling of annular reactor 1 .
Figure S3.Plots of selective oxidation of indole to 2-oxindole; at three enzymatic cascade

Figure S5 :
Figure S5: Mesh at different locations