Overall Reduction Kinetics of Low-grade Pyrolusite Using a Mixture of Hemicellulose and Lignin as Reductant

Manganese is widely used in many fields. Many efforts have been made to recover manganese from low-grade pyrolusite due to the depletion of high-grade manganese ore. Thus, it is of practical significance to develop a clean, energy-saving and environmentally friendly technical route to reduce the low-grade pyrolusite. The reported results show that biomass wastes from crops, crop waste, wood and wood waste are environmentally friendly, energy-saving, and low-cost reducing agents for roasting reduction of low-grade pyrolusite. Kinetics of the reduction reactions is necessary for an efficient design of biomass reduction of pyrolusite. Therefore, it is important to look for a general kinetics equation to describe the reduction of pyrolusite by different kinds of biomass, because there is a wide variety of biomass wastes, meaning that it is impossible to investigate the kinetics for each biomass waste. In this paper, thermal gravimetric analysis and differential thermal analysis were applied to study the overall reduction kinetics of pyrolusite using a mixture of hemicellulose and lignin, two major components of biomass. Overall reduction process is the overlap of the respective reduction processes. A new empirical equation based on the Johnson–Mehl–Avrami equation can be used to describe the respective reduction kinetics using hemicellulose and lignin as reductants, and the corresponding apparent activation energy is 30.14 kJ mol−1 and 38.91 kJ mol−1, respectively. The overall kinetic model for the reduction of pyrolusite by the mixture of hemicellulose and lignin can be simulated by the summation of the respective kinetics by considering their mass-loss fractions, while a unit step function was used to avoid the invalid conversion data. The obtained results in this work are necessary to understand the biomass reduction of pyrolusite and provide valuable assistance in the development of a general kinetics equation.


Introduction
Manganese is widely used in many fields, such as steel production, preparation of dietary additives, fertilizers, cells and fine chemicals.To meet the ever increasing demand for manganese, together with the gradual depletion of high-grade manganese ore, many efforts have been made to recover manganese from low-grade pyrolusite. 1 MnO 2 in pyrolusite is stable in both acid and alkaline oxidizing conditions, so the extraction of manganese from pyrolusite must be carried out under reducing conditions.Generally, there are two major technical routes to reduce pyrolusite: one is hydrometallurgical reduction, and the other is pyrometallurgical reduction.2] However, hydrometallurgical reduction method has not been widely applied in commercial practice because of its complicated purification process and serious water pollution. 2Pyrometallurgical reduction method with coal as reducing agent is a conventional technology to treat low-grade pyrolusite, 1,3 but this method produces a great deal of pollutants, including smoke dust, oxysulphides, nitrogen oxides, and so on. 1,4][7] Thus, it is of practical significance to develop a clean, energy-saving, and environmentally friendly technical route to reduce the low-grade pyrolusite.
][10][11][12][13][14] The results have shown that these biomass wastes can reduce manganese oxide of the ore at temperatures below 600 °C with a degree of reduction of more than 95 %.For example, Cheng et al. 8 has reported that low-grade manganese dioxide ores can be totally reduced by biomass cornstalk at 500 °C.Yang 11 and Long 9 found that low-grade pyrolusite can be reduced completely by bagasse at 450 °C.Zhou 14 has reported that pyrolusite can be reduced by bagasse pith at 350 °C.Moreover, biomass reduction is a zero emission processes, because the amount of CO 2 released during the reduction process is equal to that absorbed during biomass growth. 15ence, biomass can be considered an environmentally friendly, energy-saving, and low-cost reducing agent for roasting reduction of manganese dioxide ores.
Meanwhile, a comprehensible understanding of the thermal reduction process and kinetics of the reduction reactions is required for an efficient design of biomass reduction of pyrolusite.FT-IR (Fourier Transform Infrared Spectroscopy) and Py-GC-MS (Pyrolysis-gas chromatography-mass spectrometry) results on the thermal reduction process indicate that the reduction of pyrolusite is related to the pyrolysis of biomass. 14,16][20] However, there is a wide variety of biomass, meaning that it is impossible to investigate kinetics for each biomass.Therefore, it is necessary to look for a general kinetic equation to describe the reduction of pyrolusite by different kinds of biomass.
2] Each component of biomass has different molecular structure and nature so that the pyrolysis processes have different characteristics and generate different kinds of gas products, [23][24][25][26][27] which results in varied reduction behaviours of pyrolusite.Suggested is the superposition of the reduction behaviour of three major components.Thus, the elucidation of the reduction behaviour of individual components and their mixtures is essential for the better understanding of the reduction mechanism and obtaining the general kinetic equation for reduction of pyrolusite by different types of biomass.However, the research results concerning individual biomass and mixtures of biomass components during the reduction of pyrolusite are limited.
In this work, thermogravimetric analysis and differential thermogravimetry (TG, DTG) were used to investigate the overall kinetics for the reduction-roast of pyrolusite with the mixture of hemicellulose and lignin reduction, two major components of biomass.A new empirical model function based on the Johnson-Mehl-Avrami equation (JMA) was used to describe the kinetic behaviours of the respective reduction-roast processes using hemicellulose or lignin as reductants.On the assumption that the main components in biomass underwent individual thermo-reduction, a multicomponent kinetic model on the basis of the summation of the respective kinetics was developed to simulate the overall kinetics for the reduction of pyrolusite by the mixture of hemicellulose and lignin.The obtained information is necessary to understand the biomass reduction of pyrolusite and provide valuable assistance in the development of a general kinetics equation.

Experiments
Pyrolusite, collected from Guangxi, China, contained (by mass) 22.01 % Mn, 11.16 Fe, 27.62 % SiO 2 , 10.93 % Al 2 O 3 , 0.09 % CaO, 0.12 % MgO, 0.020 % S, and 0.181 % P. The ore samples were crushed to 0.147 mm (−100 meshes).Xylan extracted from beech wood was used as the representative of hemicelluloses.Lignin and hemicellulose were obtained from Sigma-Aldrich Corporation, USA.The hemicellulose and lignin were not further treated before mixing and testing.Table 1 gives the composition of the mixture of hemicellulose and lignin, where w is the mass fraction for lignin and hemicellulose.Pyrolusite was mixed with hemicellulose, lignin or their mixtures, respectively.Based on the reported results, 11 the mass ratio of pyrolusite to hemicellulose, lignin or their mixtures was 10 : 1.The reduction of pyrolusite by hemicellulose, lignin, and their mixture were named HP, LP, LH2, LH 4, LH 6, LH 7 and LH 8, respectively.Thermogravimetric analyser (Q50TGA) was applied to investigate the mass loss of the reduction of pyrolusite by lignin, hemicellulose and their mixture.In the TG experiments, the sample was heated from room temperature to 800 °C at 20 °C min −1 in nitrogen atmosphere.The flow rate of nitrogen was 40 ml min −1 to maintain an inert atmosphere for the decomposition.The mass of each sample was set at about 20 mg.
X-ray diffraction (XRD) patterns were obtained with a D8 Advance X-ray diffractometer using Cu-K α radiation with an accelerating voltage 40 kV, current 30 mA, and scan speed 10° min −1 at 2θ = 10° -70°.XRD was used to identify the mineralogical composition of the ore samples before roast and after roast at 400 °C for 30 minutes.The corresponding XRD patterns are shown in Fig. 2. The original ore was comprised of pyrolusite (MnO 2 ), hematite (Fe 2 O 3 ), silicon oxide (SiO 2 ), while its roasted products were mainly manganese oxide (MnO), hematite and kaolinite.Hence, the reduction processes of pyrolusite by hemicellulose and lignin were completed in the second stage of TG/DTG curves in Fig. 1.The mass loss behaviours of the pyrolusite reduction processes by the mixture of lignin and hemicellulose are also subdivided into three stages, which are similar to those of the respective reduction processes shown in Fig. 1.

Results and discussion
The main mass loss for the lignin-hemicellulose mixture reduction of pyrolusite is also concentrated in the second stage.
As shown in Fig. 3, the peaks related to lignin or hemicellulose reduction of pyrolusite cannot be distinguished, indicating that the respective reduction behaviours had overlapped in the pyrolusite reduction by the mixture of lignin and hemicellulose.Thus, the overall pyrolusite reduction processes by the mixture of lignin and hemicellulose are composed of respective reduction related to lignin and hemicellulose.

Respective kinetic analysis
According to the reported results, 14,16,18 various reductive volatiles (such as aldehydes, furans, ketones, alcohol, etc.) were produced during the pyrolysis of biomass, which directly reduced MnO 2 in the ore to MnO.The pyrolusite reduction processes by biomass are complex, and involve many kinds of reductive volatiles with different molecular structure and nature.Zhang 19 reported that reduction of low-grade manganese dioxide ore by wheat stalk obey the kinetic model (1) Bamboo, [18][19] sawdust, 19 wheat stalk 19 and straw [19][20] reduction of low-grade manganese dioxide ore can be described by: (2) Thus, biomass reduction of manganese dioxide ore under investigation cannot fully be described by the conventional kinetic model function f(α) because of the complexity of the reaction. 28In this case, it can be useful to find an empirical function containing the smallest possible number of constants, so that there is some flexibility sufficient to describe the real process as closely as possible.
The Johnson-Mehl-Avrami (JMA) equation, usually written in the following form: ( where k and n are constants with respect to time (t), and α is the degree of conversion.JMA equation has been used to describe the transformation kinetics of many solid Fig. 3 -TG-DTG curves for the reduction of pyrolusite by the mixture of lignin and hemicellulose at heating rate of 20 K min −1 in nitrogen Slika 3 -Krivulje TG-DTG za redukciju piroluzita mješavinom lignina i hemiceluloze pri brzini zagrijavanja 20 K min −1 u dušiku state processes under isothermal conditions. 29When the temperature increases at a constant rate and k can be expressed by Arrhenius equation, Eq. ( 3) can be rewritten as follows: ( where β is the heating rate, R is gas constant, E a is apparent activation energy, k 0 is constant and T 0 is the initial temperature for the reaction determined from the DTG data.Hence, Eq. ( 6) is an empirical model based on JMA equation.
As indicated by the results of TG/DTG in Fig. 1, the reduction processes of pyrolusite by lignin and hemicellulose are concentrated at the second stage.The degree of conversion (α) for the reduction of pyrolusite can be calculated from the TG/DTG data according to the formula: (7)   where m 0 and m F are the mass of the reaction mixture at initial and end stage, and m T is the mass of the reaction mixture at temperature T. The degree of conversion (α) for the reduction of pyrolusite by lignin and hemicellulose was calculated and is given in Fig. 4.
The most appropriate parameters, k 0 , E a , and n, of the respective reduction are calculated simultaneously by the nonlinear least square analysis to minimize the square sum of the residue when fitting the experimental curve of α exp versus temperature by the calculated curve of α cal versus temperature. ( The fitting results are presented in Table 2.
As seen in Fig. 4 and Table 2, the calculated conversion data agree well with the experimental data, and the correlation coefficients (R kor

2
) for the two reduction processes are greater than 0.99, indicating that these two reduction processes can be described by Eq. ( 6).The values of the apparent activation energy for the reduction of pyrolusite by lignin and hemicellulose are calculated to be 38.66 kJ mol −1 and 30.14 kJ mol −1 , respectively.

Overlapped kinetic analysis
The results of TG/DTG in Fig. 3 show that the reduction of pyrolusite by the mixture of lignin and hemicellulose is an overall process composed of two respective reduction processes.As long as the overall processes are composed of N independent kinetic processes, the overall kinetic behaviour can be expressed by the summation of the respective kinetic processes i by considering their mass-loss fractions, As seen in Table 3, initial temperatures (T 0 ) are different from each other.For instance, initial temperature of LH2 is lower than that of HP and LP, meaning that there are invalid conversion data calculated by Eq. ( 6).This problem can be avoided by using a unit step function μ as following form. ( After substitution of Eq. ( 10) into Eq.( 9), the following overall kinetic equation can be obtained: After substitution of Eq. ( 6) and the data in Table 2 into Eq.( 11), the overall kinetic equation can be expressed by the following equation: The experimental conversion data (α exp ) for the reduction of pyrolusite by the mixture of lignin and hemicellulose were calculated using Eq. ( 6) from the TG/DTG data in Fig. 3.The corresponding calculated conversion data (α cal ) were calculated using Eq.(12).The values of α exp and α cal are given in Fig. 5.The following Eq.( 13) was used to determine the adaptability of Eq. ( 12), and the corresponding results are given in Table 3. ( If the value of B in Eq. ( 13) is 1, the calculated data are the same as the experimental values.
As shown in Fig. 5, the calculated data are in agreement with the experimental data.From Table 3, it can be seen that values of B for six samples are close to 1, and the correlation coefficients (R kor

Conclusions
Kinetics of the overall reduction processes of pyrolusite by the mixture of lignin and hemicellulose, two major components of biomass, were investigated.The TG/DTG and XRD results reveal that the mass loss behaviours of the reduction processes are subdivided into three stages, and the main reduction processes are concentrated in the second stage.
The respective reduction processes of pyrolusite by lignin and hemicellulose are overlapped in the overall reduction processes.On the basis of JMA equation, an empirical equation was developed to describe the respective kinetics for the reduction of pyrolusite by lignin and hemicellulose.The apparent activation energy for the reduction of pyrolusite by hemicellulose and lignin was 30.14 kJ mol −1 and 38.91 kJ mol −1 , respectively.The overall reduction kinetics can be simulated by the summation of the respective kinetics by considering their mass-loss fractions.

3. 1 .
TG/DTG analysis TG/DTG experiments were carried out to measure the mass change during the reduction of pyrolusite by hemicellulose, lignin or their mixture under nitrogen at 20 °C min −1 heating rate.Fig.1presents the mass loss behaviours of the reduction of pyrolusite with hemicellulose or lignin, respectively.The mass loss behaviours of the reduction processes are also subdivided into three stages.The first stage (below 473 K) was a loss in mass upon drying the sample, and mainly moisture was released at this stage.The second stage contributes to the main mass loss.The temperature range of the main mass loss stages are 493 -663 K and 463 -683 K for HP and LP, respectively.The maximum values of the DTG curve (mass loss rate) for HP and LP are at about 570 K and 613 K, respectively.The last stage was high temperature charring of the residue and the mass loss was much lower.