If the correlation coefficient (R2) of adsorption isotherm (langmuir, frendlich and Tempkin) are low ( around 0.4), how can i interpret this?
Hi Shaimaa,
Usually, the low coefficient for a modeling indicates that your data does not follow the model. But you should double-check the following issues firstly:
1) if the range of adsorbate concentration is too low or too high, it might not show the features of the isotherm models. If you can show your data here, it will be better;
2) the experimental data are reproducible (i.e., no experimental error occurred).
Besides the aforementioned explanation by Shi, considering the trend of the experimental data, you may try other isotherm models such as BET model.
Whether an Adsorption system is favourable or unfavorable can be predicted by the dimensionless constant, separation factor. The isotherm is considered unfavourable when the separation factor is greater than one......it is considered favourable when it is greater than zero and less than one......and the isotherm is said to be irreversible when it's separation factor is equal to zero.....
agree with Shi what you can also check is the statistical data. Did your adsorption reach equilibrium?
Hi Shaimaa,
Please also make sure that your reaction reach equilibrium by conducting a kinetic test firstly.
Good luck,
Qiantao
The low coefficient of determination (R2) indicates that the model is not representative of the data. You must check, the adsorption capacity and the concentration in the equilibrium. Also, there are several models of adsorption isotherms:
Isotherm with two parameters: Langmuir, Freundlich, Dubinin and Redushkevich, Isotherm of Temkin, Hill, Harkin-Jura, Jovanovic, Elovich.
Isotherm with three parameters: Redlich-Peterson, Sips, Pseudo 2nd order, Toth, Koble-Carrigan, Jossens.
Isotherms with four, five parameters.
Dear Shaimaa,
There are several types of adsorption isotherm models, i.e. one, two and three parameters, so your results may fit another model rather than Langumair and Freundlich models. You could refer to the following article to find the required answer:
"Modeling of adsorption isotherms of oil content through the electrocoagulation treatment of real oily wastewater, Conference Paper Modeling of adsorption isotherms of oil content through the ...
Good luck.
One measure of fit is usually not enough. It is also worth using the average relative error. Moreover, it is not known how you counted R2, whether for the data in the coordinate system qe vs Ce, or for the data in the coordinate system Ce/qe vs Ce (Langmuir isotherm)?
Low value of Coefficient of Determination mean that the applied isotherm (Langmuir, Freundlich and Temkin) couldn't fitted well for the adsorption process. You need to apply other isotherms also.
Dear researcher,
If the value of correlation coefficient (R2) is low, you can use some other isotherm models and can check which isotherm model will fit.
Or different error functions such as normalized standard deviation, the sum of error squares, average percentage error, mean squared error and sum of absolute error can be used. The model which is having lowest value of these error functions can be be considered to be fitting to the experimental data. The paper attached (error functions) can be referred for different error functions.
Otherwise the experimental and theoretical curves can be compared in the qe v/s ce curve.
All the best.
Although in such a situation it is best to see the equilibrium data. Otherwise it's guesswork.
The reason is that the isotherm you use is not the best choice for the adsorption process, you need other isotherm. Check your experiment for possible errors, you need to reach equilibrium and the reproducibility of the data.
langmuir and frendlich models apply to 90% of the mathematical relationships that govern the adsorption process.. I think there is an error in the drawing or extracting the values