Graphitic carbon nitride photocatalysis: the hydroperoxyl radical role revealed by kinetic modelling

Metal-free graphitic carbon nitride (GCN) is an optical semiconductor with the advantage of in situ H2O2 generation parallel to pollutant removal. The photocatalytic degradation mechanism using GCN is attributed to...


Text S1. Theoretical calculation of photocatalysis rate
The photocatalysis rate (kphoto) of GCN with a 417 nm LED radiation was calculated as Equations S1-S3.
Where r0 is the initial rate of the photocatalytic reaction (M s -1 ), I0 is the incident radiation , kobs is the photocatalytic degradation rate of PhOH, Ci is the initial concentration of PhOH, and Cf is the theoretical concentration at one second of reaction calculated as Equation S4.The final reaction time was one second because, in this condition, it is assumed that phenol molecules are only degraded via the photocatalytic process over the total number of photons absorbed [33].
The absorbed photons are difficult to evaluate in a photocatalytic system due to the absorption, transmission, and scattering of the semiconductor particles.As reported in the literature [34][35][36][37], it was calculated considering the reactor geometry (irradiation area) and the molar absorptivity as the number of reactant molecules transformed divided by the number of photons of the light incident inside the rector.The average apparent quantum efficiency (j) is the ratio of 90% converted reactant molecules over the photons entering the reactor (Eq.S5) [36,38,39].
Where N90 are 90% transformed moles at its matching time (t90, 90 seconds), I is the radiation intensity (Einstein m -2 s -1 ), A is the number of catalytically active sites estimated by multiplying the specific surface area (SBET, in m 2 g -1 ) of the catalyst particles and its mass (g), n is the frequency (c/l), and h is the Planck constant.In this calculation, the SBET was assumed as 87 m 2 g -1 according to our previous work [40].
It must be noted that the BET surface areas suggest the number of adsorption sites, but it does not necessarily represent the number of catalytically active sites.However, this assumption is commonly accepted as a conservative estimation for the active sites when the photocatalytic process depends on the surface characteristics [39,41].

Text S2. Sensitivity analysis
The Normalised Residuals Sum of Squares (NRSS) value represents the average deviation of the resolved model from the experimental data.It is an indicator of the capacity of the model to reproduce the experimental data.The average RSS is defined as Equation S.6: Where Mi is the model response at a given system condition and time, Di is the experimental data at the same given condition and time, and n is the total number of measured data points over all conditions and time.To analyse the effect of changes on specific rate constant values [42], the NRSS was calculated by running the model while varying one rate constant with all others held fixed at their optimised values.In this case, the Di is the altered model response at the same condition and time as the control model.

Figure S4 .
Figure S4.Kinetic model of the recombination e -/h + reaction in the LED-417/GCN

Table S1 . Kintecus model reactions.
Reactions and rate constants included in the kinetic model

Table S2 .
Photocatalytic results of PhOH degradation and H2O2 generation in the

Table S3 .
Percentages of PhOH removal depending on the reactive species and the percentage of dissolved oxygen.Data obtained from Kintecus modelling.

Table S4 .
Percentages of H2O2 formation depending on the reaction pathway and the amount (%) of dissolved oxygen (DO).Data obtained from Kintecus modelling.

Table S5 .
Percentages of H2O2 removal depending on the reaction and the amount (%) of dissolved oxygen (DO).Data obtained from Kintecus modelling.The rest of the reactions (R43, R45) that could remove H2O2 had no contribution to the model.