📄 chemistry

From Castor Seeds to Oil: Experimental Insights and Modeling of Oil Extraction

This study investigates the application of dynamic maceration for castor oil extraction, demonstrating that solvent polarity and higher temperatures significantly improve yield while validating three mathematical models for process optimization and scale-up.

Original authors: Mehdi LOUAER, Narimane LAMMARI, Abdeslam Hassan MENIAI

Published 2026-07-10

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Mehdi LOUAER, Narimane LAMMARI, Abdeslam Hassan MENIAI

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you have a bag of tiny, hard castor seeds, and your goal is to get the precious oil out of them without breaking the seeds into dust. This paper is like a detective story where the researchers, Mehdi Louaer and his team from the University of Constantine 3, tried to figure out the fastest and best way to squeeze that oil out, and then built a "virtual lab" to predict how it would happen.

The Great Solvent Showdown
First, the team tested three different "baths" (solvents) to see which one could pull the oil out of the seeds most effectively. Think of these solvents as different types of sponges:

Hexane: A non-polar sponge (like a dry, oily rag).
Acetone: A medium-polar sponge.
Ethanol: A highly polar sponge (like a super-absorbent cloth).

They tried two methods: the old-school "Soxhlet" method (which is like a slow, continuous drip-coffee machine that runs for hours) and a newer, faster method called Dynamic Maceration (which is like shaking the seeds in a jar of solvent with a magnetic stirrer).

Here is the big reveal: Polarity matters. The more "polar" the solvent, the more oil it grabbed. Ethanol was the superstar. When they used the slow Soxhlet method with Ethanol, they got a massive 49.938% yield. Even with the faster shaking method (Dynamic Maceration), Ethanol still won, pulling out 17.44% at room temperature (20°C) and 19.197% when they warmed it up to 50°C.

In contrast, the non-polar Hexane was much weaker. At 20°C, it only managed 9.787%. But, the team found that heat is a secret weapon. When they turned up the heat to 50°C, Hexane's performance jumped to 12.494%. It's like warming up honey; it flows better and gets into the seeds faster.

The Virtual Crystal Ball
Now, here is where it gets really cool. The researchers didn't just stop at mixing jars; they wanted to predict the future. They used a powerful computer program called Comsol Multiphysics to build three different mathematical "crystal balls" (models) to see if they could guess how fast the oil would come out.

The First Crystal Ball (Empirical Model 1): A simple formula that guessed the oil would come out quickly at first and then slow down.
The Second Crystal Ball (Empirical Model 2): A slightly different formula that also guessed the speed of extraction.
The Third Crystal Ball (The Single Sphere Model): This one was the most detailed. It imagined each seed particle as a tiny, perfect ball. It treated the oil moving out of the seed like heat cooling down from a hot ball into cold air. It assumed the oil was spread evenly inside the ball and only had to "diffuse" (wander) its way out.

Did the Crystal Balls Work?
The team compared their computer predictions to the real-life results they got in the lab. The results were surprisingly close! The average error between the computer and reality was very small (an Average Absolute Relative Deviation of 0.11).

This means the models were reliable. However, the "Single Sphere Model" (the one that treated seeds like cooling balls) seemed to be the most accurate of the three, though all three did a good job. The computer simulations also confirmed what the lab showed: as the temperature went up, the "effective diffusivity" (how fast the oil moves) increased. For example, with Hexane, the oil moved faster at 50°C (2.804E-18 m²/s) than at 20°C (1.897E-18 m²/s).

The Quality Check
Finally, they checked the "health" of the oil they extracted. They measured things like how much soap could be made from it (Saponification Value) and how much free acid was in it (Acid Value). The good news? All the oils, whether extracted by the fast shaking method or the slow drip method, passed the safety checks for food and cosmetics. They had low acid and peroxide values, meaning they were fresh and high quality.

The Bottom Line
The paper doesn't claim to have solved the world's oil problems, but it does show that Dynamic Maceration is a viable, fast way to get castor oil, especially if you use a polar solvent like Ethanol and maybe give it a little heat. It also proves that you can use computer simulations to predict exactly how this process works, which saves time and money in the lab. The "Single Sphere" model is the best tool they found to predict this dance between the seed and the solvent, but the other two models are still useful helpers.

Technical Summary: From Castor Seeds to Oil: Experimental Insights and Modeling of Oil Extraction

Problem Statement
Vegetable oils, including castor oil, possess significant nutritional and therapeutic value, yet their quality and efficacy are heavily dependent on the extraction method employed. While mechanical pressing and conventional solvent extraction (such as Soxhlet) are established techniques, there is a need to optimize extraction processes to improve efficiency, reduce costs, and preserve oil quality. Specifically, the application of dynamic maceration—a technique involving soaking plant material in a solvent under agitation to reduce mass transfer resistance—had not been previously applied to castor oil extraction. Furthermore, there is a lack of validated mathematical models capable of accurately predicting the kinetics of this specific process to facilitate scale-up and process optimization.

Methodology
The study employed a combined experimental and numerical approach to investigate castor oil extraction via dynamic maceration.

Experimental Setup: Ground castor seeds (mean diameter 0.991 mm, moisture 5.59%) were extracted using three solvents (Hexane, Acetone, and Ethanol) at two temperatures (20°C and 50°C) in a batch system with magnetic stirring (375 rpm). Extraction yields were monitored over 75 minutes. For comparative purposes, a conventional Soxhlet extraction was also performed.
Physicochemical Analysis: Extracted oils were characterized for Saponification Value (SV), Acid Value (AV), Peroxide Value (PV), and Ester Value (EV) to ensure compliance with food and cosmetic standards.
Modeling Approach: Three mathematical models were evaluated using COMSOL Multiphysics software to fit the experimental kinetic data:
1. First Empirical Model: Based on Naik et al., described by an exponential function ( $Y = Y_{\infty}(1 - e^{-kt})$ ).
2. Second Empirical Model: Based on Mafarat and Beliard, described by a rational function ( $Y = Y_{\infty}(\frac{t}{b+t})$ ).
3. Phenomenological Model (Single Sphere Model): A diffusion-based model treating the seed particle as a sphere where internal diffusion controls the process. This model utilized material balances and Fick's second law, solving for effective diffusivity ( $D_e$ ) and mass transfer coefficients ( $k_f$ ) derived from Sherwood, Reynolds, and Schmidt number correlations.

Key Contributions

First Application: This work represents the first reported application of dynamic maceration specifically for castor oil extraction.
Model Validation: The study successfully assessed and validated three distinct kinetic models (two empirical and one phenomenological) for this specific system using inverse modeling in COMSOL Multiphysics.
Parameter Optimization: The research determined optimal model parameters, including rate constants for empirical models and effective diffusivity values for the single sphere model, across varying solvent polarities and temperatures.

Results

Extraction Yields: The Soxhlet method yielded the highest oil recovery. Among dynamic maceration conditions, Ethanol (a polar solvent) produced the highest yields: 17.44% at 20°C and 19.197% at 50°C, compared to Hexane (non-polar) which yielded 9.787% and 12.494% respectively. Generally, higher extraction yields were observed with polar solvents (Ethanol, Acetone) compared to non-polar Hexane, attributed to the extraction of non-triglyceride constituents and coloring matter.
Temperature Effects: Increasing the temperature from 20°C to 50°C improved both the extraction yield and the rate for all solvents. For instance, Hexane yield increased from 9.787% to 12.494%. This is attributed to reduced solvent viscosity and surface tension, enhancing penetration and mass transfer.
Model Performance: All three models demonstrated reasonable accuracy in predicting experimental data, with an overall Average Absolute Relative Deviation (AARD) of 0.11.
- The Second Empirical Model showed an average AARD of 0.10.
- The Single Sphere Model also achieved an average AARD of 0.11.
- The First Empirical Model had a slightly higher average AARD of 0.11 (specifically 0.1210 for Ethanol at 20°C).
Diffusivity Trends: Effective diffusivity ( $D_e$ ) increased with temperature (e.g., Hexane $D_e$ rose from $1.897 \times 10^{-18}$ to $2.804 \times 10^{-18} m^2/s$ ) but decreased as solvent polarity increased.
Oil Quality: Physicochemical analysis indicated that the extracted oils met Codex Alimentarius and ASTM norms for food and cosmetic applications, characterized by low acid and peroxide values but high saponification and ester values.

Significance and Claims
The authors claim that the study demonstrates the viability of using a purely modeling approach to tackle maceration processes, offering valuable computational tools to minimize study time and costs. The research confirms that dynamic maceration is a feasible method for castor oil extraction, with process performance positively influenced by solvent polarity and temperature. While all three models were deemed reliable, the authors modestly suggest that the "Single sphere" model appeared slightly superior to the two empirical models in describing the extraction kinetics. The findings are presented as a contribution to the development of more efficient and cost-effective extraction methods for castor oil.

More like this