Bifunctional Europium for Operando Catalyst Thermometry in an Exothermic Chemical Reaction

Abstract Often the reactor or the reaction medium temperature is reported in the field of heterogeneous catalysis, even though it could vary significantly from the reactive catalyst temperature. The influence of the catalyst temperature on the catalytic performance and vice versa is therefore not always accurately known. We here apply EuOCl as both solid catalyst and thermometer, allowing for operando temperature determination. The interplay between reaction conditions and the catalyst temperature dynamics is studied. A maximum temperature difference between the catalyst and oven of +16 °C was observed due to the exothermicity of the methane oxychlorination reaction. Heat dissipation by radiation appears dominating compared to convection in this set‐up, explaining the observed uniform catalyst bed temperature. Application of operando catalyst thermometry could provide a deeper mechanistic understanding of catalyst performances and allow for safer process operation in chemical industries.


Thermometric Performance and Temperature Determination
The Boltzmann equation is the intensity ratio of the two thermally coupled states, here the 5 D1 and 5D0 of Eu 3+ , and A is a constant, is applied to calculate the energy gap between the thermally coupled states for calibration measurements and for the determination of the catalyst temperature. When Boltzmann thermometers are applied, the relative sensitivity of the thermometer can be expressed as [2] ( where ∆ is energy gap between the two thermally coupled states, is the Boltzmann constant and T is the temperature in K. The temperature uncertainty can be expressed by [3] ( . 3) where A and B are the integrated count rates in the regions of interest for the 5 D1 and 5 D0 emission.

Data Analysis Procedure
The raw spectral data was converted before it was suited for thermometric applications. The first step was the dark subtraction to correct for stray light. Subsequently, any faulty measurements, i.e. measurements approaching the maximum count rate and low signal measurements (below 10% of the maximum count rate) that result in a large , were removed from the dataset. Next, the wavelength was converted to energy scale according to ( . 4) = 1 * 10 −7 ( −1 ).
As the x-axis went from evenly spaced intervals to non-evenly spaced intervals, the spectral intensity has to be corrected accordingly. [4] The Jacobian transformation was applied where I and E are the intensity and energy at a specific wavelength. Next, the spectra were normalized to 100 and the regions of interest were fitted with multiple Lorentzians and a baseline, given by the basic function where , , and are the peak amplitude, peak width, peak center and a constant ( Figure S1). The area of the sum of the Lorentzian (without z) was applied as input for the Boltzmann equation (Eq.S1).

Heat Transfer Calculations By Convection and Radiation
Heat generated by the reaction (Qr) was calculated according to with a methane conversion ≈ 0.3, inlet CH4 concentration 4 = 0.1 ≈ 1.5 mol m -3 , gas flow = 3 ⋅ 10 −7 m 3 s -1 and reaction enthalpy Δ = 158 ⋅ 10 3 J mol -1 . If all heat would be withdrawn by reaction gas mixture, the temperature increase of the gas (DTg) can be calculated with with a gas density ≈ 0.5 kg m -3 and a heat capacity of the gas ≈ 1150 J kg -1 K -1 . As the inlet gas temperature is assumed to be equal to the oven temperature ( Figure S5), convection by reaction mixture is presumably not the predominant cooling mechanism due to the large calculated DTg. Therefore, needs to be withdrawn radially outwards by either conduction or radiation. The heat generated by reaction per sieved catalyst particle (Qr,p) is calculated with ( . 9) , = .
Assuming spherical catalyst particles, the number of particles, , is approximated by where s is the Stefan-Boltzmann constant (s = 5.6703*10 -8 W m -2 K -4 ). Even at extremely low values of material emissivity, particle emission Qe,p is at least 2 orders of magnitude larger than the heat generated per particle , . Hence, any increase in particle temperature due to reaction would immediately be lost by radiation to the environment. Therefore, the Figure S1. Regions of interest for (A) I2 and (B) I1. Three Lorentzians were used as input for I2 and five Lorentzians were used to fit the I1 region, of which only the 5 D0 → 7 F2 were used as input for I1.
( . 12) , = 4 = 6.7 with the outer reactor surface = 3.3 ⋅ 10 −4 m 2 and = 773 . Again, Qe,reactor is more than 2 orders of magnitude larger then . Hence, any temperature increase due to reaction will quickly be emitted by radiation, until equilibrium at oven temperature is reached. It can therefore be concluded that the bed temperature most likely is constant throughout its volume, at oven temperature. Still, heat is generated by the reaction and the catalyst bed temperature must be higher than the surrounding with ,0 the total mission without reaction. However, the material emissivity of EuOCl is unknown and determination is outside the scope of this work. Nevertheless, assuming a particle emissivity of 0.1 would result in a ΔT of 5K, which is in the right order of magnitude. The calculations enabled us to explain the trends qualitatively but a more detailed model and more carefully controlled experiments are needed for the quantitative description of the temperature increase.   Figure S4. (A) Excitation (emission recorded at 619 nm) and emission (excited at 375 nm) spectrum of EuOCl at room temperature. [5] From the excitation spectrum, it becomes apparent that Eu 3+ can be excited in the intra 4-F transition at 375nm. The emission spectrum clearly yields the Eu 3+ emission spectrum. (B) Emission spectrum of EuOCl at 500 °C excited at 375nm where the Eu 3+ emission are labelled according to the energy diagram of Eu 3+ ( Figure 1A). [6]   . From 0 -12 min TOS, the Tcat followed the Toven closely. After 12 min TOS, the apparent Tcat dropped below the Toven due to the larger uncertainty in the temperature measurement and the changing emission spectrum. The fitted (B) 5 D1 and (C) 5 D0 revealed that at 500 °C, a maximum in the counts was obtained, after which it rapidly decreased at lower temperatures. Chlorination of the catalyst at a temperature below 500 °C quenched the luminescence signal, reducing the counts. The decrease in counts was also reflected in the R 2 of the fits, which decrease rapidly, especially at temperatures lower than 490 °C.