For sure, you should use the non-linear format as the linear format causes some error in the result. There are many reports and papers in the literature on the accuracy of the non-linear format of the isotherms.

Dear, the best way is to use the non linear form

Il est préférable d'utiliser une forme linéaire de l'équation

You can initially use the analogue of Lineweaver-Burk plot (1/qe vs 1/Ce) to preevaluate the Langmuir isotherm parameters. The values of these parameters can then be used as a starting point in non-linear methods popularly called non-linear regression. Instead of the analogue of L-B plot you can also use e.g. the analogue of Hanes-Woolf plot (Ce/qe vs Ce). However, graphical methods have disadvantages: a change in the statistical distribution of original variables due to their non-linear transformation and a change in the error distribution in new coordinate systems.

Check this link

https://www.researchgate.net/post/What-do-you-prefer-to-determine-isotherm-models-linear-or-nonlinear

I would alway recomment the non-linear methode because you get a higher accuracy, if you also have collected data in the qm area, otherwise a non linear fitting can lead to errors.

About Mirosław Grzesik the Hanes Woolf plot, therer you get the problem the the y side is not indepenten for the x side, and this could lead to errors in the linearziation.

This problem is described in this papers:

Zhang, J.-Z. Avoiding spurious correlation in analysis of chemical kinetic data. Chem. Commun. 2011, 47 (24), 6861 DOI: 10.1039/c1cc11278c.

Badertscher, M.; Pretsch, E. Bad results from good data. TrAC Trends in Analytical Chemistry 2006, 25 (11), 1131–1138 DOI: 10.1016/j.trac.2006.09.003.

Using graphical methods, we change the position of the experimental points in such a way that they arrange themselves along a straight line. Regardless of whether it is a L-B like method, H-W like method, E-H like method, it is not a linearization but a non-linear transformation, just like taking logarithm on both sides of given function.. Moreover, R2 cannot be counted for nonlinear transformed data, only for original data, and then the differentiation of aproximation functions is clearer. And enzymologists wrote about errors in graphical methods as early as in the 1950s.

Dear Dr Andrej Saner . See the following useful RG link: Article Linearised and non-linearised isotherm models optimization a...

Kindly see also the following very good RG link: Article Statistical Analysis of Linear and Non-Linear Regression for...

Also check the following useful RG link: Article Comparison of linear and non-linear models for the adsorptio...

It will be good if we go through the nonlinear fitting to get more accuracy