The Catalytic Effect of Iron and Alkali and Alkaline Earth Metal Sulfates Loading Series on the Conversion of Cellulose-Derived Hydrochars and Chars

The catalytic effect of minerals on biomass conversion was studied focusing on Fe as well as alkali and alkaline earth metals as the metallic inorganic elements typically present in minerals found in biomass. A mineral-free reference hydrochar and an analogous char material based on cellulose were systematically doped with sulfates of the different metallic inorganic elements in various amounts via impregnation, thereby excluding differences originating from the counterion and the carbon matrix. Thermogravimetric reactivity measurements were performed in diluted O2 and CO2, and the derivative thermogravimetry curves were fitted using the random pore model. This procedure enabled a quantification of the apparent activation energy decrease due to doping as well as the influence of doping on the carbon structural parameter. Fe sulfate was always among the most active minerals, and alkali metal sulfates were typically more active than alkaline earth metal sulfates. The only exception was the high activity of very small Ca sulfate loadings during gasification. A saturation behavior of the kinetic parameter upon increasing the mineral loading was observed. The Langmuir-type modeling of this dependence further revealed that catalytically influenced devolatilization results in a char with higher oxidation reactivity, whereas for gasification, thermal annealing dominates. The systematically derived parameters provide a comprehensive description of catalytic effects, taking into account the type of mineral, the applied loading, the used atmosphere, and the fuel morphology. The derived activation energies can be used to include catalytic effects into combustion models.

. DTG profiles of MH800 doped with increasing loadings of individual metal sulfates measured in 20% O2/He or 50% CO2/He.

S4 Development and Validation of the MH fitting
First, isothermal mole fraction variations of O2 or CO2 in He for the undoped MH model fuel were performed at constant temperatures to derive the 2 and 2 for oxidation and gasification, respectively. Fixing the derived apparent reaction orders, isothermal temperature variations at 20% O2/He and 50% CO2/He were used to derive kinetic parameters such as ℎ and . Sets of three isothermal measurements obtained in both atmospheres applying varied mole fractions of reactive gas and temperatures were fitted using the RPM (equation (2)). Figure S4 shows the isothermal measurements performed for oxidation of the undoped MH fuel. Figure S4. RPM fitted isothermal measurements of MH with an O2 mole fraction variation at 400 °C (left) and a temperature variation in 20% O2/He (right). RPM fit as dashed lines.
Clearly, the degree of char conversion at one temperature strongly depended on the applied concentration of the reactive gas. However, the full agreement between experimental data and fit was impaired by the simultaneous devolatilization during the first half of conversion.
Nevertheless, in accordance with the 53 wt% of volatiles in the sample, the agreement of fit and experiment was much better for conversion degrees above 50 %. Hence, this range of the measurements was used for the quantification of the apparent reaction order for oxidation. The calculated 2 of 0.39 is slightly smaller compared with values reported for different carbon materials in literature between 0.5 and 1.0 and also smaller than the value reported for MH800 ( 2 = 1.14) 1,2 . Here, the larger saturation of the carbon matrix with abundant oxygen S8 functional groups in the hydrochar may explain the difference in the apparent reaction orders between the hydrochar and the char. Fixing 2 , isothermal temperature variations at a constant molar ratio of O2 were used to derive the pre-exponential factor ℎ and the apparent activation energy , as well as the structural parameter .
The isothermal measurements in CO2 atmosphere were analyzed applying the same procedure as shown in Figure S5.  Table S4. In accordance with literature 1 , 2 is smaller than 2 and only marginally smaller than the typically reported values for 2 between 0.4 and 0.7 for various biomass chars. While the , 2 of MH for oxidation amounting to 132 kJ mol -1 is 14 kJ mol -1 smaller than the derived , 2 of the MH800 char, thereby describing a more reactive hydrochar compared with the char, the , 2 of MH for gasification is 39 kJ mol -1 larger than the corresponding , 2 of MH800.
Nevertheless, the hydrochar is still also more reactive during gasification compared with the char, as the ℎ, 2 of the hydrochar is two orders of magnitude larger than that of the char.
Considering that CO2 gasification takes mainly place in macro-und mesopores 3,4 , the high contribution (almost 50%) of these large pores to the surface area of the hydrochar may explain the difference in ℎ, 2 for both fuels. In comparison, in MH800 macro-und mesopores only contribute by around 7% to the overall surface area.
Concerning the structural parameter describing the porosity of the fuel, a higher is expected for gasification. A value close to zero describes a sample with high porosity in which conversion proceeds overall in the sample volume including smaller pores 5,6 . In contrast, if conversion only takes place in larger pores or if pores are blocked by minerals, an increase S10 of is expected, explaining the seven orders of magnitude difference between 2 and 2 observed for the undoped MH fuel.
In order to limit the influence of overlapping devolatilization and char conversion in the isothermal measurements, especially during oxidation experiments, the RPM fitting of the falling branch of the DTG curves was tested. Another advantage of this method was the use of only one temperature-programmed measurement instead of the need to measure three isothermal measurements, saving a lot of time considering the number of doped samples.
Fixing and ℎ derived from the isothermal measurements, the fitting results for the DTG curves of undoped MH in 20% O2/He and 50% CO2/He are summarized for comparison with the results of the isothermal measurements in Table S4. A good agreement of fit and experimental data was achieved as shown in Figure S6, deriving almost identical values for and in both atmospheres compared with the isothermal measurements.  Figure S7 and Figure S8. For a better comparison, the calculated parameters derived from isothermal measurements as well as from DTG fitting are also summarized in Table S4.  While not as perfect a match as for the undoped sample, both fitting approaches resulted in similar and for 1.2% K-MH in oxidation and gasification, respectively. Hence, the DTG fitting approach by RPM was successfully validated also for doped samples. Referring to the DTG fitting results in the following, the apparent activation energy decreased due to the catalytic effect of the 1.2 wt% K, resulting in a decrease by 16 kJ mol -1 (12%) during oxidation and a decrease by 36 kJ mol -1 (13%) during CO2 gasification. As predicted earlier, the structural parameter of the doped sample slightly increased in comparison with the undoped sample S12 indicating that mineral deposits may inhibit mass transport. Nevertheless, it can be assumed that the reactions still take place in the whole volume of the sample, as the value is still close to zero.  Table S6. Langmuir-type parameters of the maximum effect ΔEA,max in kJ mol −1 (and its corresponding percentage compared with the undoped MH in brackets) and the strength of loading dependence in wt% −1 as well as the strength of loading-dependent deactivation in wt% −1 and the power derived for the loading dependence of the activation energy difference during oxidation and gasification.