Equilibrium and dynamic studies for adsorption of boron on calcium alginate gel beads using principal component analysis (PCA) and partial least squares (PLS). Issue 5 (29th May 2013)

• Record Type: Journal Article
• Authors: Ruiz, M.; Roset, L.; Demey, H.; Castro, S.; Sastre, A. M.; Pérez, J. J.
• Is Part Of: Materialwissenschaft und Werkstofftechnik. Volume 44:Issue 5(2013:May)
• Page Start: 410
• Page End: 415
• Publication Date: 2013-05-29
• Subjects: Materials -- Periodicals; Materials -- Testing -- Periodicals; 620.1
• DOI: 10.1002/mawe.201300144
• Languages: English
• ISSNs: 0933-5137

Abstract

<abstract abstract-type="main" xml:lang="en"> <title>Abstract</title> <p>In this paper, boron removal from aqueous solutions by adsorption on calcium alginate gel beads was investigated. Based on the batch tests, calcium alginate gel beads demonstrated a good capacity of adsorption; the maximum capacity obtained was 121 mg/g at pH 6. The aim of this paper is the study of analytical tools for extracting more analytical information from a few experiments. Principal component analysis (PCA) was employed to perform a comparative analysis between different isotherms obtained by varying the percentage of alginate on the adsorbent beads. The calculated fraction of the cumulative sum of squares for all the components (R2X(cum)) and the predictive reliability model based on the value of the correlation coefficient of cross validation (Q<sub>2</sub> (cum)) obtained were 0.9639 and 0.8523, respectively. Dynamic system study using 20 different columns was carried out. Partial least squares (PLS) modeling are applied to characterize the effect of the experimental variables on the adsorption process. Some of the descriptors studied were the initial boron concentration (<italic>C</italic><sub>o</sub>), adsorbent mass (<italic>m</italic>), <italic> pH</italic>, column diameter (<italic>d<sub>c</sub></italic>), column depth (<italic>Z</italic>) and superficial flow velocity (<italic>U</italic><sub>0</sub>). Experimental adsorptions were compared against predicted adsorptions and a correlation coefficient (r<sub>2</sub>) of 0.9927 was obtained.</p> </abstract>