Modeling of photolytic degradation of sulfamethoxazole using boosted regression tree (BRT), artificial neural network (ANN) and response surface methodology (RSM); energy consumption and intermediates study

Unveiling the role of artificial intelligence in tetracycline antibiotics removal using UV/sulfite/phenol advanced reduction process

UV/sulfite-based advanced reduction processes (ARP) have attracted increasing attention due to their high capability for removing a wide range of pollutants. Therefore, developing UV/sulfite ARP systems with assisted Artificial Intelligence (AI) models is considered an efficient strategy for sustainable pollutant removal. The present study delves into modeling and optimizing photodegradation of tetracycline (TC) antibiotics under UV/sulfite/рhenol reԁuсtion рroсess (UV/SPAP) using integrаteԁ Artifiсiаl Neurаl Networks (ANN), Suррort Veсtor Regression (SVR), аnԁ Genetiс Algorithm (GA). The сonсentrаtions of рhenol (X1) аnԁ sulfite (X2), рH (X3), reасtion time (X4), аnԁ TC сonсentrаtion (X5) in our exрerimentаl setuр were varied, аnԁ use the generаteԁ ԁаtа to trаin AI moԁels. The findings revealed that the AI-optimized performance is very effective in predicting and optimizing the removal of TC, thereby providing a sustainable water treatment approach. In general, SVR performed better based on scaling coefficients and ANN using different criteria indicated that X4 and X5 parameters were statistically significant. Oрtimаl rаnges for X1, X2, X3, X4, аnԁ X5 аre ԁetermineԁ to be 6.34, 3, 8.45, 80.13, аnԁ 1, resрeсtively. This аррroасh highlights the imрortаnсe of integrаting AI аnԁ ARP for sustаinаble environmentаl mаnаgement.

Catalytic wet air oxidation removal of tetracycline by La2O3 immobilized on recycled polyethylene terephthalate using the response surface methodology

This study investigated the removal of tetracycline from the aqueous solutions by lanthanum oxide nanoparticles covered with polyethylene terephthalate (PET) using a low-cost and facile co-precipitation method, via catalytic wet air oxidation process (CWAO) by response surface methodology (RSM). XRD, FTIR, SEM, and EDX-map techniques have been employed to investigate the crystal structure, functional groups on the surface, morphologic characteristics, and elemental composition, respectively. Under optimum conditions (pH= 9, initial TC concentration= 20 mg L-1, nanocomposite dosage= 1.5 g L-1, pressure= 4 bar, temperature= 70 °C, and time= 90 min), TC removal efficiency by La2O3-PET was achieved at about 99.9%. The environmental parameters were assessed to determine tetracycline catalytic wet air oxidation degradation rate, which included cleaning gases, hydrogen peroxide, type of organic compounds, anions, radical scavenger and reusability. The ANOVA results indicated that the polynomial model proves that the model is entirely meaningful (F-value> 0.001 and P-value< 0.0001) and has high coefficient values of adjusted R2 (0.7404) and predicted R2 (0.5940). The findings indicated that the variables of time, pH, temperature, dosage, and TC concentration have the greatest role in removing tetracycline, respectively. However, pressure as a factor does not have a considerable influence on the performance of the system. In general, due to the presence of the role of additional anionics, the effectiveness of this method for removing tetracycline from drinking water was 82.76%. The catalyst indicated pleasing stability and recycling power during eight testing cycles. Further, the estimated electrical energy per order consumption (EEO) for the CWAO/La2O3-PET system was calculated as 5.31 kWh m-3 with an operational cost (OC) utilization of 1.78 USD kg-1 and it has been shown that this process is feasible and economically comparable to other CWAO processes. The breakdown intermediate products of tetracycline in the CWAO were examined using gas chromatography/mass spectrometry (GC-MS) analysis. The toxicity analyses for the removal of TC were carried out using Daphnia magna and the CWAO process achieved a remarkable decrease in the presence of La2O3-PET nanocomposite (LC50 and toxicity unit (TU) 48 h equal to 0.634 and 157.72 vol percent).

Optimizing removal of antiretroviral drugs from tertiary wastewater using chlorination and AI-based prediction with response surface methodology

Chemical and pharmaceutical chemicals found in water sources create substantial risks to human health and the environment. The presence of pharmaceutical contaminants in water can cause antibiotic resistance development, toxicity to aquatic organisms, and endocrine disruption. Hence, the elimination of chemicals and other contaminants from wastewater prior to its release is a burgeoning concern in the domains of engineering and science. The use of treatment technologies in wastewater treatment plants can remove pharmaceutical contaminants through the oxidation process. However, many traditional wastewater treatment plants lack the advanced monitoring tools required to detect low concentrations of pharmaceuticals. Without the ability to detect these compounds, it's challenging to treat them effectively. The goal of this study was to use Response Surface Methodology (RSM) and Artificial Neural Networks (ANN) algorithms to model and improve how Nevirapine and Efavirenz break down in different chlorination conditions. The RSM analysis revealed statistically significant models (F-values: Nevirapine, pH-t: 108.15, T-t: 76.55, ICC-t: 110.84), indicating a strong correlation between operational parameters (pH, temperature, and initial chlorine concentration) and degradation behavior. The ANN model accurately predicted the degradation of both Nevirapine and Efavirenz under various chlorination conditions, as confirmed by analyzing actual-predicted graphs, residual plots, and Mean Squared Error (MSE) values. The ANN model using ICC-t achieved the highest MOD value of 31.31 % for Nevirapine. The ANN model based on ICC-t yielded a maximum MOD value of 16.06 % for Efavirenz. These findings provide valuable insights into optimizing chlorination processes for better removal of these pharmaceutical contaminants from water.

Biodegradation of ciprofloxacin using machine learning tools: Kinetics and modelling

Recently, the rampant administration of antibiotics and their synthetic organic constitutes have exacerbated adverse effects on ecosystems, affecting the health of animals, plants, and humans by promoting the emergence of extreme multidrug-resistant bacteria (XDR), antibiotic resistance bacterial variants (ARB), and genes (ARGs). The constraints, such as high costs, by-product formation, etc., associated with the physico-chemical treatment process limit their efficacy in achieving efficient wastewater remediation. Biodegradation is a cost-effective, energy-saving, sustainable alternative for removing emerging organic pollutants from environmental matrices. In view of the same, the current study aims to explore the biodegradation of ciprofloxacin using microbial consortia via metabolic pathways. The optimal parameters for biodegradation were assessed by employing machine learning tools, viz. Artificial Neural Network (ANN) and statistical optimization tool (Response Surface Methodology, RSM) using the Box-Behnken design (BBD). Under optimal culture conditions, the designed bacterial consortia degraded ciprofloxacin with 95.5% efficiency, aligning with model prediction results, i.e., 95.20% (RSM) and 94.53% (ANN), respectively. Thus, befitting amendments to the biodegradation process can augment efficiency and lead to a greener solution for antibiotic degradation from aqueous media.

A novel nanocomposite for photocatalytic rhodamine B dye removal from wastewater using visible light

Providing safe access to water and addressing the impact of waterborne diseases, which claim over two million lives annually, is a major contribution to water purification. The study introduces a novel nanocomposite, Ch/Fe3O4/α-MoO3, which exhibits outstanding photocatalytic efficacy under visible light. An in-depth investigation of the nanocomposite's synthesis, characterization, and photodegradation mechanisms reveals its outstanding capabilities. Photocatalytic activity is influenced by the catalytic dose, pH, dye concentration, and reaction time, according to the study. A response surface method is used to determine the optimal conditions for Rhodamine B degradation, which results in 96.3% removal efficiency at pH 8.5, dye concentration 25 mg/L, nanocomposite dose at 22 mg/L, and reaction time 50 min. As a result of its high surface area, biocompatibility, availability, and magnetization with iron compounds, Chitosan is an excellent substrate for enhancing the photocatalytic properties of MoO3 nanoparticles. A nanocomposite with an energy band of 3.18 eV exhibits improved visible light absorption. This study confirms the nanocomposite's recyclability and stability, affirming its practicality. Besides dye removal, it offers hope for the global quest for clean water sources by addressing a broader range of waterborne contaminants. By combining molybdenum and magnetite, nanocomposite materials facilitate the degradation of pollutant and bacteria, contributing positively to society's quest for clean and safe water. It emphasizes the role nanotechnology plays in preserving human health and well-being in combating waterborne diseases.

Integrating DFT and machine learning for the design and optimization of sodium alginate-based hydrogel adsorbents: Efficient removal of pollutants from wastewater

This study presents a novel approach to design and optimize a sodium alginate-based hydrogel (SAH) for efficient adsorption of the model water pollutant methylene blue (MB) dye. Utilizing density functional theory (DFT) calculations, sodium alginate-g-poly (acrylamide-co-itaconic acid) was identified with the lowest adsorption energy (Eads) for MB dye among 14 different clusters. SAHs were prepared using selected monomers and sodium alginate combinations through graft co-polymerization, and swelling studies were conducted to optimize grafting conditions. Advanced characterization techniques, including FTIR, XRD, XPS, SEM, EDS, and TGA, were employed, and the process was optimized using statistical and machine learning tools. Screening tests demonstrated that Eads serves as an effective predicting indicator for adsorption capacity (qe) and MB removal efficiency (RRMB,%), with reasonable agreement between Eads and both responses under given conditions. Process modeling and optimization revealed that 5 mg of selected SAH achieves a maximum qe of 3244 mg g-1 at 84.4% RRMB under pH 8.05, 98.8 min, and MB concentration of 383.3 mg L-1, as identified by the desirability function approach. Moreover, SAH effectively eliminated various contaminants from aqueous solutions, including sulfasalazine (SFZ) and dibenzothiophene (DBT). MB adsorption onto selected SAH was exothermic, spontaneous, and followed the pseudo-first-order and Langmuir-Freundlich isotherm models. The remarkable ability of SAH to adsorb MB is attributed to its well-designed structure predicted through DFT and optimal operational conditions achieved by AI-based parametric optimization. By integrating DFT-based computations and machine-learning tools, this study contributes to the efficient design of adsorbent materials and optimization of adsorption processes, also showcasing the potential of SAH as an efficient adsorbent for the abatement of aqueous pollution.

Insightful properties-performance study of Ti-Cu-O heterojunction sonochemically embedded in mesoporous silica matrix for efficient tetracycline adsorption and photodegradation: RSM and ANN-based modeling and optimization

This study aims to provide a comprehensive evaluation of the photocatalytic properties and performance of the Cu-Ti-O heterojunction sonochemically embedded in the mesoporous silica matrix. Various characterization analyses and adsorption/photodegradation experiments were performed to assess the potential of the sample for tetracycline (TC) removal. The characterization results indicated that sonication contributes to better dispersion of Ti-Cu-O species, resulting in more uniform particle sizes, stronger semiconductors-silica interaction, and less agglomeration. Furthermore, sonication significantly affected the optical nanocomposite features, leading to an improvement in charge carrier separation and a decrease in the band gap of Ti-Cu-Si (S) by approximately 2.6 eV. Based on the textural results, the ultrasound microjets increased the surface area and pore volume, which facilitate mass transfer and provide suitable adsorption sites for TC molecules. Accordingly, Cu-Ti-Si (S) demonstrated higher adsorption capacity (0.051 g TC/g adsorbent) and eliminated TC significantly faster (0.0054 L.mg-1.min-1) than a non-sonicated sample during 120 min of irradiation, resulting in 2.84 times improvement in the constant rate. In addition, experimental results were accurately modeled using a central composite design in combination with response surface methodology (RSM) and artificial neural networks (ANN) to predict and optimize TC photodegradation. Both RSM and ANN models revealed excellent predictability for TC degradation efficiency, with R2 = 99.47 and 99.71%, respectively. At optimal operational conditions (CTC = 20 ppm, photocatalyst dosage = 1.15 g.L-1, pH = 9, and irradiation time = 100 min), more than 95% and 87% of TC were degraded within the UV (375 W) and simulated solar light (400 W) irradiation periods, respectively. It was observed that the Cu-Ti-Si (S) nanocomposite maintained remarkable stability after four cycles with only a negligible 3% loss of activity, owing to the superior interaction between the bimetallic heterojunction and the silica matrix.

Comprehensive investigation of isotherm, RSM, and ANN modeling of CO2 capture by multi-walled carbon nanotube

Chemical vapor deposition was used to produce multi-walled carbon nanotubes (MWCNTs), which were modified by Fe-Ni/AC catalysts to enhance CO2 adsorption. In this study, a new realm of possibilities and potential advancements in CO2 capture technology is unveiled through the unique combination of cutting-edge modeling techniques and utilization of the recently synthesized Fe-Ni/AC catalyst adsorbent. SEM, BET, and FTIR were used to analyze their structure and morphology. The surface area of MWCNT was found to be 240 m2/g, but after modification, it was reduced to 11 m2/g. The modified MWCNT showed increased adsorption capacity with higher pressure and lower temperature, due to the introduction of new adsorption sites and favorable interactions at lower temperatures. At 25 °C and 10 bar, it reached a maximum adsorption capacity of 424.08 mg/g. The optimal values of the pressure, time, and temperature parameters were achieved at 7 bar, 2646 S and 313 K. The Freundlich and Hill models had the highest correlation with the experimental data. The Second-Order and Fractional Order kinetic models fit the adsorption results well. The adsorption process was found to be exothermic and spontaneous. The modified MWCNT has the potential for efficient gas adsorption in fields like gas storage or separation. The regenerated M-MWCNT adsorbent demonstrated the ability to be reused multiple times for the CO2 adsorption process, as evidenced by the study. In this study, a feed-forward MLP artificial neural network model was created using a back-propagation training approach to predict CO2 adsorption. The most suitable and efficient MLP network structure, selected for optimization, consisted of two hidden layers with 25 and 10 neurons, respectively. This network was trained using the Levenberg-Marquardt backpropagation algorithm. An MLP artificial neural network model was created, with a minimum MSE performance of 0.0004247 and an R2 value of 0.99904, indicating its accuracy. The experiment also utilized the blank spreadsheet design within the framework of response surface methodology to predict CO2 adsorption. The proximity between the Predicted R2 value of 0.8899 and the Adjusted R2 value of 0.9016, with a difference of less than 0.2, indicates a high level of similarity. This suggests that the model is exceptionally reliable in its ability to predict future observations, highlighting its robustness.

Enhancing sodium percarbonate catalytic wet peroxide oxidation with artificial intelligence-optimized swirl flow: Ni single atom sites on carbon nanotubes for improved reactivity and silicon resistance

H2O2 is widely used in the treatment of refractory organic pollutants.However, due to its explosive and corrosive chemical characteristics, H2O2 will bring great safety risks and troubles in transportation.So we chose sodium percarbonate(SPC) to be used in catalytic wet peroxide oxidation enhanced by swirl flow(SF-CWPO) and we designed carbon nanotubes with Ni single atom sites(Ni-NCNTs/AC) to activate SPC to treat an m-cresol wastewater containing Si.Meanwhile, artificial intelligence which used Artificial neural network (ANN) was used to optimize the conditions.Under the conditions of pH = 9.27, reaction time of 8.91 min, m-cresol concentration is 59.09 mg L-1, SPC dosage is 2.80 g L-1 and Na2SiO3·9H2O dosage is 77.27 mg L-1, the degradation rate of total organic carbon(TOC) and m-cresol reaches 94.37% and 100%, respectively.Finally, the applicability of Ni-NCNTs/AC-SPC-SF-CWPO technology was evaluated in a wastewater system of a sewage treatment enterprise and Fourier transform ion cyclotron resonance mass spectrum(FT-ICR MS) analysis and chemical oxygen demand(COD) analysis showed the great ability of Ni-NCNTs/AC-SPC-SF-CWPO technology to treat wastewater.It is believed that this paper is of great significance to the design and construction of the in-depth research and industrial application of SF-CWPO.

Modelling the effects of mixing ratio and temperature on the thermal conductivity of GNP-Alumina hybrid nanofluids: A comparison of ANN, RSM, and linear regression methods

This research aimed to evaluate and compare the efficacy of three distinct methods for forecasting the thermal conductivity of GNP-Alumina hybrid nanofluids. The methods under consideration were artificial neural network (ANN), response surface methodology (RSM), and linear regression (LR). The predictive performance of the ANN model was investigated in relation to the number of neurons in the hidden layer. The findings revealed that the optimal number of neurons was 7, which produced the best performance with an overall mean square error (MSE) of 1.08E-06. The correlation coefficient was also high at 0.99799. The RSM approach involved testing linear, quadratic, cubic, and quartic models, with the quadratic model showing the highest predicted R2 (0.9721) values, indicating that it provided the best fit to the data. Finally, the LR model was developed using a backward elimination approach, with temperature and volume fraction being the significant variables in the final model. Overall, the ANN model produced the most accurate predictions, followed by the RSM and LR models. These findings suggest that the ANN and RSM techniques can be effective tools for forecasting the thermal conductivity of nanofluids, and highlight the importance of selecting appropriate model parameters for optimal performance.

Modeling and optimization of chlorpyrifos and glyphosate biodegradation using RSM and ANN: Elucidating their degradation pathways by GC-MS based metabolomics

Ongoing and extensive use of pesticides negatively impact the environment and human health. Microbe-based remediation bears importance as it is an eco-friendly and cost-effective technique. The present study investigated chlorpyrifos (CHL) and glyphosate (GLY) degrading potential of Bacillus cereus AKAD 3-1, isolated from the soybean rhizosphere. Optimization and validation of different process variables were carried out by response surface methodology (RSM) and artificial neural network (ANN). Critical parameters which affect the degradation process are initial pesticide concentration, pH, and inoculum size. At optimum conditions, the bacterial strain demonstrated 94.52% and 83.58% removal of chlorpyrifos and glyphosate, respectively. Both Central-composite design (CCD-RSM) and ANN approaches proved to perform well in modeling and optimizing the growth conditions. The optimum ANN-GA model resulted in R² ≥ 0.99 for chlorpyrifos and glyphosate, while in the case of RSM, the obtained R² value was 0.96 and 0.95, respectively. Results indicated that the process variables significantly (p < 0.05) impact chlorpyrifos and glyphosate biodegradation. Moreover, the predicted RSM model had a "lack of fit p-value" of "0.8849" and "0.2502" for chlorpyrifos and glyphosate, respectively. GC-MS analysis revealed that the strain first converted chlorpyrifos into 3,5,6-trichloro pyridin-2-ol & O, O-diethyl O-hydrogen phosphorothiate. Later, these intermediate metabolites were broken and completely mineralized into non-toxic by-products. Similarly, glyphosate was first converted into 2-(methylamino) acetic acid and amino-oxyphosphonic acid, which were further mineralized without any toxic by-products. Taken together, the results of this study clarify the biodegradation pathways and highlights the promising potential of B. cereus AKAD 3-1 in the bioremediation of chlorpyrifos and glyphosate-polluted environments.

Surface facet Fe2O3-based visible light photocatalytic activation of persulfate for the removal of RR120 dye: nonlinear modeling and optimization

Photocatalytic activation of persulfate (PS) is recently emerged as an energy-efficient and environmentally sustainable approach for pollutants degradation, which enables to leverage the strengths of low-cost solar energy and heterogeneous catalysis. Herein, we investigated the photocatalytic decomposition of reactive red 120 (RR120) dye using PS-activated Fe₂O₃ nanoparticles and elucidated the effect of their facets, α-Fe₂O₃ (001), β-Fe₂O₃ (100), and γ-Fe₂O₃ (111). β-Fe₂O₃ not only boosted the charge carrier separation but also provided more active sites for PS activation resulting in 6- and 3.5-fold higher photocatalytic activities compared to α-Fe₂O₃ and γ-Fe₂O₃, respectively. Response surface methodology and artificial neural network coupled with genetic algorithm models were utilized to optimize and foresee Fe₂O₃/PS system under visible light. Almost 100% color removal and 82% organic removal were observed under the optimum conditions at 20 mg/L RR120, 22 mg/L β-Fe₂O₃, 18 mg/L PS, and pH: 3. Scavenger test indicated that both sulfate and hydroxyl radicals are responsible for the observed RR120 removal. Although cell viability test indicated that cytotoxicity of wastewater is not significantly reduced after treatment. All the results proposed that β-Fe₂O₃/PS at relatively low doses has a great potential to decompose and mineralize recalcitrant dyes in wastewater under invisible light.

Multi-object optimization of Navy-blue anodic oxidation via response surface models assisted with statistical and machine learning techniques

This study aims to model, analyze, and compare the electrochemical removal of Navy-blue dye (NB, %) and subsequent energy consumption (EC, Wh) using the integrated response surface modelling and optimization approaches. The Box-Behnken experimental design was exercised using current density, electrolyte concentration, pH and oxidation time as inputs, while NB removal and EC were recorded as responses for the implementation and analysis of multiple linear regression, support vector regression and artificial neural network models. The dual-response optimization using genetic algorithm generated multi-Pareto solutions for maximized NB removal at minimum energy cost, which were further ranked by employing the desirability function approach. The optimal parametric solution having total desirability of 0.804 is found when pH, current density, Na₂SO₄ concentration and electrolysis time were 6.4, 11.89 mA cm^-2, 0.055 M and 21.5 min, respectively. At these conditions, NB degradation and EC were 83.23% and 3.64 Wh, respectively. Sensitivity analyses revealed the influential patterns of variables on simultaneous optimization of NB removal and EC to be current density followed by treatment time and finally supporting electrolyte concentration. Statistical metrics of modeling and validation confirmed the accuracy of artificial neural network model followed by support vector regression and multiple linear regression anlaysis. The results revealed that statistical and computational modeling is an effective approach for the optimization of process variables of an electrochemical degradation process.

Modelling of transmembrane pressure using slot/pore blocking model, response surface and artificial intelligence approach

This work investigates the application of empirical, statistical and machine learning methods to appraise the prediction of transmembrane pressure (TMP) by oscillating slotted pore membrane for the treatment of two kinds of deformable oil drops. Here, we utilized the previous experimental runs with permeate flux, shear rate and filtration time as features, while TMP of crude oil and Tween-20 were two distinct targets. For 87 experimental runs, Response surface methodology (RSM) and Artificial Neural network (ANN) modelling were opted as statistical and machine learning tools, respectively, which were comprehensively compared with empirical slot-pore blocking model (SBM) considering accuracy and generalization. ANN with 10 neurons in the hidden layer could approximate the TMP of both oils better than RSM and SBM, which is reflected by computed performance metrics. Under the given conditions, almost similar analysis were predicted for TMP of both oils except changes in magnitude which were interpreted by (1) line plots, which showed that TMP of crude oil and Tween-20 were linearly related to flux rate and filtration time, and there was an inverse relationship between TMP and shear rate, (2) contour plots, which illustrated the strong interaction effect of flux rate and time on TMP, and (3)- sensitivity analysis, which revealed the influential sequence of variables on TMP as; flux rate > filtration time > shear rate, for both cases. The optimisation of the process showed that minimum TMP can be attained by maintaining higher shear rate and lower flux rate and time. Conclusively, the current findings indicate the utilization of ANN for the accurate assessment of TMP and can be helpful for the process designing and scale up.

Valorization of groundnut shell via pyrolysis: Product distribution, thermodynamic analysis, kinetic estimation, and artificial neural network modeling

Pyrolysis of agricultural biomass is a promising technique for producing renewable energy and effectively managing solid waste. In this study, groundnut shell (GNS) was processed at 500 °C in an inert gas atmosphere with a gas flow rate and a heating rate of 10 mL/min and 10 °C/min, respectively, in a custom-designed fluidized bed pyrolytic-reactor. Under optimal operating conditions, the GNS-derived pyrolytic-oil yield was 62.8 wt.%, with the corresponding biochar (19.5 wt.%) and biogas yields (17.7 wt.%). The GC-MS analysis of the GNS-based bio-oil confirmed the presence of (trifluoromethyl)pyridin-2-amine (18.814%), 2-Fluoroformyl-3,3,4,4-tetrafluoro-1,2-oxazetidine (16.23%), 5,7-dimethyl-1H-Indazole (11.613%), N-methyl-N-nitropropan-2-amine (6.5%) and butyl piperidino sulfone (5.668%) as major components, which are used as building blocks in the biofuel, pharmaceutical, and food industries. Furthermore, a 2 × 5 × 1 artificial neural network (ANN) architecture was developed to predict the decomposition behavior of GNS at heating rates of 5, 10, and 20 °C/min, while the thermodynamic and kinetic parameters were estimated using a non-isothermal model-free method. The Popescu method predicted activation energy (E_a) of GNS biomass ranging from 111 kJ/mol to 260 kJ/mol, with changes in enthalpy (ΔH), Gibbs-free energy (ΔG), and entropy (ΔS) ranging from 106 to 254 kJ/mol, 162-241 kJ/mol, and -0.0937 to 0.0598 kJ/mol/K, respectively. The extraction of high-quality precursors from GNS pyrolysis was demonstrated in this study, as well as the usefulness of the ANN technique for thermogravimetric analysis of biomass.