Skip to content
Study the hydrothermal reaction of cellulose and xylan using the catalyst Erbium Triflate (Er(OTf)3). Develop prediction models of hydrothermal reactions of agricultural residues using data from the hydrothermal reaction of cellulose and xylan.
Lactic acid concentration (ppm) = 5.0722 × (Peak area from HPLC) Succinic acid concentration (ppm) = 5.3031 × (Peak area from HPLC)Formic acid concentration (ppm) = 3.3734 × (Peak area from HPLC) 5-HMF concentration (ppm) = 1.0312 × (Peak area from HPLC)
Master's Project
“Study of Hydrothermal Reaction of Cellulose By Using Er(OTf)3 Catalyst with Design of Experiment”
PART-I INTRODUCTION
CE 2.1
During my Master’s degree in Chemical Engineering at King Mongkut's Institute of Technology Ladkrabang (KMITL), I conducted a research project titled "Study of Hydrothermal Reaction of Cellulose by using Er(OTf)3 Catalyst with Design of Experiment." The project spanned from 2016 to 2017 and was motivated by the need for sustainable chemical processes and the efficient conversion of biomass into value-added products. The primary goal was to synthesise lactic acid and other carboxylic acids from cellulose and xylan using a hydrothermal reaction in the presence of Er(OTf)3 catalyst. This project was part of my thesis work under the supervision of Assist. Prof. Dr. Tanawan Pinnarat.
In this project, I used several advanced chemical engineering techniques, including High-Performance Liquid Chromatography (HPLC), Design of Experiments (DOE), and regression analysis, to optimise the process and analyse the results.
PART-II BACKGROUND
CE 2.2
The primary objective of my project was to efficiently convert biomass into lactic acid, which is a valuable chemical used in the production of biodegradable plastics and other applications. I focused on two main types of substances: cellulose and xylan, both of which are abundant and renewable resources.
The main challenge was optimising the reaction conditions to achieve the highest yield of lactic acid. The experiments were conducted in an autoclave reactor, which allowed for high-pressure and high-temperature conditions necessary for the hydrothermal reaction.
I systematically varied the reaction parameters—including temperature, reaction time, and catalyst-to-reactant mass ratio—using Design of Experiments (DOE) to identify the optimal conditions.
My project aimed to:
The project steps are shows as Figure 1 as following;
Figure 1. Project steps
CE 2.3
I conducted this project independently as part of my Master's degree under the supervision of Assist. Prof. Dr. Tanawan Pinnarat, a faculty member in the Chemical Engineering Department at King Mongkut's Institute of Technology Ladkrabang (KMITL). Although I worked closely with my supervisor to receive guidance and feedback, I was solely responsible for the planning, execution, and analysis of the experiments. This independent work allowed me to fully immerse myself in the research and develop a deeper understanding of the chemical processes involved. Below is an organisational chart highlighting my position within the project:
Figure 2. Organisational chart
PART-III PERSONAL ENGINEERING ACTIVITY
CE 2.4
I began my project by researching the theory behind biomass and discovered that cellulose and hemicellulose (commonly found in the form of xylan), which are the main components of agricultural waste (illustrated as Figure 3 and 4), can be converted into lactic acid (a bioplastic derivative) and other carboxylic acids through hydrothermal reactions using a catalyst (Figure 5).
Figure 3. Agriculture waste component
Figure 4. Structure of cellulose, xylan, and lignin
Figure 5. Hydrothermal reaction of cellulose and xylan to produce lactic acid
CE 2.5
Since my goal was to optimise the reaction parameters, I applied the Design of Experiments (DOE) methodology. Specifically, I used the Box-Behnken Design with the Minitab software, which allowed me to systematically vary three key factors: temperature, reaction time, and catalyst-to-reactant mass ratio. The DOE approach enabled me to investigate the effects of these variables on lactic acid yield without having to test every possible combination of conditions, thereby saving both time and resources. Table 1 outlines the experimental conditions used for the DOE model.
Table 1: the experimental conditions for DOE
Factors
Unit
Levels
Levels
Levels
Factors
Unit
Low
Middle
High
Temperature,
°C
170
190
210
Reaction time
min
0
60
120
Catalyst to reactance mass ratio
-
0.10:1
0.25:1
0.40:1
There are no rows in this table
Figure 6. DOE model for hydrothermal experiment
CE 2.6
The overall steps I undertook for this project including reactant preparation, hydrothermal reaction in autoclave reactor, result analysis with HPLC, and then regression model to predic hydrothermal reaction.
Figure 7. Hydrothermal reaction process of cellulose and xylan
I began my experimental work by preparing the reactants, which included cellulose, xylan, and the Er(OTf)3 catalyst. The hydrothermal reactions were carried out in an autoclave reactor, which allowed me to perform batch reactions at temperatures ranging from 170°C to 210°C. The autoclave provided the necessary high-pressure environment to break down the reactance and convert it into lactic acid.
Setting up and operating the autoclave required a deep understanding of chemical reaction engineering principles. I carefully measured and mixed the reactants before loading them into the reactor. I controlled the temperature, pressure, and stirring rate throughout the reaction to ensure consistent conditions. Safety was a top priority, given the high-pressure environment, so I followed strict protocols to prevent any accidents. The autoclave reactor and its components are illustrated in Figure 8 below.
Figure 8. Components of batch reactor (autoclave)
CE 2.7
After each reaction, High-Performance Liquid Chromatography (HPLC) was utilsied to analyse the results from hytrotermal reaction. HPLC allowed me to separate and quantify the different components of the reaction mixture, including lactic acid and other carboxylic acids. This step was crucial in determining the effectiveness of the reaction conditions.
I prepared each sample by filtering the reaction mixture to remove any solid residues and then injected the liquid sample into the HPLC system. By comparing the retention times and peak areas of the components in the chromatograms, as demonstrated in Figure 9.
Figure 9. HPLC analytical method
To calculate the concentration of products from the hydrothermal reaction, I applied my knowledge of chemistry and chemical engineering principles to create and calculate the calibration graphs. These graphs allowed me to determine the concentration of each compound in the reaction mixture. The sample of calibration curves for glucose and xylose are illustrated in Figure 10.
Figure 10. Calibration curves for glucose and xylose
The following equations represent the calibration curves for the major products:
By using these calculations, I was able to determine the concentration of lactic acid and other products in each sample. This data was essential for evaluating the success of the experiments and for further optimisation of the reaction conditions.
From these knowledge and practice, I was able to calculate the concentration of lactic acid and other carboxylic acids in each sample. This data was essential for assessing the success of the experiments and for further prediction models. Yield of each product from hydrothermal reaction are shown in Figure 11.
Figure 11. Yield of each product from hydrothermal reaction
CE 2.8
Once I had gathered sufficient experimental data, I employed regression analysis to develop predictive models that quantified the relationship between the reaction conditions (temperature, time, and catalyst-to-reactant ratio) and the yield of lactic acid. To ensure the robustness of the model, I applied Response Surface Methodology (RSM), which allowed me to fit a quadratic model to the data. This provided a precise mathematical equation for predicting the lactic acid yield based on the input variables, giving me the ability to optimise the process systematically.
Figure 12. Regression models for hydrothermal reaction of cellulose and xylan
The regression model's quality was evaluated through Analysis of Variance (ANOVA), which tested the lack of fit of the model. The p-value of the lack-of-fit test was 0.065, which was greater than the significance level (α = 0.05). This indicated that there was no significant evidence that the regression model did not fit the data, confirming that the model was an adequate representation of the experimental results. Additionally, the model achieved an R² value of 97.37%, demonstrating that a high percentage of the variability in lactic acid yield could be explained by the model.
Figure 13. Analysis of Variance for cellulose hydrothermal reaction
This use of statistical methods, including ANOVA and p-value analysis, showcased my application of mathematical and statistical principles to engineering problems. These tools helped ensure that the model was statistically sound and that the conclusions drawn from the data were reliable.
While the regression model showed an excellent fit for cellulose-based reactions, some discrepancies were noted in the xylan-based reactions, indicating that further refinement was necessary. This highlighted the need to explore additional variables or interactions in future experiments to improve the predictive accuracy for different types of biomass.
In summary, my approach combined advanced statistical analysis and mathematical modeling to develop a reliable predictive model, emphasising the use of engineering principles to optimise the chemical reaction process effectively.
CE 2.9
To determine the optimum conditions for the cellulose hydrothermal reaction, I utilised the Response Optimisation function in the Minitab software. This tool allowed me to predict the optimal temperature, reaction time, and catalyst-to-reactant ratio to maximise the yield of lactic acid. After generating these predictions, I conducted real experiments to compare the predicted results with actual outcomes.
By systematically analysing the data, I identified that the actual experimental results showed an average deviation of 15.27% from the predicted values. This discrepancy highlighted the complexities involved in chemical reactions, such as unaccounted-for variables and reaction dynamics that are difficult to model precisely.
To address these challenges, I applied my chemical engineering skills in data analysis and Interpretation to analyse the experimental data and evaluate the discrepancies between predicted and actual results and considering potential sources of error.
Figure 14. Response Optimisation for cellulose hydrothermal
Table 2: Comparison of experimental results and prediction model at optimum condition.
CE 2.10
To achieve the goal of this project, then I applied the lactic regression model of cellulose and xylan to predic the total lactic acid form biomass (durian peel and corn cob) as demonstrated in Figure 15 and 16.
Figure 15. Lactic acid regression for biomass
Figure 16. Lactic acid yield of durian peel compared with the regression model
The results showed that the lactic acid yields from hydrothermal reactions using biomass (durian peel and corn cobs) differed from those predicted by the regression model. This discrepancy arose because the regression model was based on cellulose reactions, while biomass contains varying proportions of cellulose, hemicellulose, and lignin. The different compositions affect the reaction outcomes, leading to deviations from the model's predictions.
To address this, I applied my engineering skills in analysing the variability in biomass composition and recognising the need for further optimisation of the model. Specifically, I identified that additional studies on the hydrothermal reactions of hemicellulose would enhance the accuracy of the predictions for agricultural residues, showcasing my ability to adapt theoretical models to practical applications.
SUMMARY
In this project, I successfully applied advanced chemical engineering techniques to study the hydrothermal reaction of cellulose and xylan using an Er(OTf)3 catalyst. By utilising High-Performance Liquid Chromatography (HPLC), Design of Experiments (DOE), and regression analysis, I optimised the reaction conditions and developed a predictive model with a high degree of accuracy for cellulose hydrothermal reactions.
The model, validated through statistical methods such as Analysis of Variance (ANOVA) and p-value analysis, demonstrated a strong fit for cellulose-based reactions with an R² value of 97.37%, confirming its reliability in predicting lactic acid yield under various conditions. However, some discrepancies in xylan-based reactions indicated the need for further refinement and exploration of additional variables.
This project deepened my understanding of chemical reaction engineering and enhanced my ability to apply statistical analysis and mathematical modelling to solve complex engineering challenges. Overall, my work contributed to the advancement of sustainable chemical processes and demonstrated my capability to leverage engineering principles to optimise real-world chemical reactions.
Want to print your doc?
This is not the way.
This is not the way.
Try clicking the ··· in the right corner or using a keyboard shortcut (
CtrlP
) instead.