Comparison of two adsorbents for simulated-moving-bed separation of galactose and levulinic acid in terms of throughput and desorbent usage

• Published in: Journal of industrial and engineering chemistry (Seoul, Korea) Vol. 97; pp. 337 - 348
• Main Authors: Park, Hangil, Kim, Jin Woo, Lee, Ki Bong, Mun, Sungyong
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 25.05.2021; 한국공업화학회
• Subjects: Galactose; Levulinic acid; Process optimization; Retention capacities; Simulated moving bed; 화학공학; Simulated moving bed; Process optimization; Levulinic acid; Galactose; Retention capacities
• ISSN: 1226-086X; 1876-794X
• DOI: 10.1016/j.jiec.2021.02.018

[Display omitted] •Our aim is to improve SMB performance for separation of galactose and levulinic acid.•Amberchrom-CG300C resin was proposed as alternative adsorbent to the previous one.•The use of the proposed adsorbent increased throughput and reduced desorbent usage.•The aforementioned improvements became greater as the target purity level was higher.•The relative merit of the proposed adsorbent was greater with increasing throughput. One of the key issues in the production of biofuels from agarose is to establish a highly-efficient process for separation of a target product (galactose) and a side-product (levulinic acid) that come from agarose hydrolysis. Regarding this issue, there has been a previous attempt at the separation of galactose (GA) and levulinic acid (LA) using a simulated-moving-bed (SMB) process based on a size-exclusion adsorbent, which, however, had limitations in increasing solute retention capacities and thus hindered a meaningful improvement in the SMB performances. To resolve such problem, we explored an alternative adsorbent that could compensate for the shortcoming of the previously adopted size-exclusion adsorbent. It was found first that an Amberchrom-CG300C resin could adsorb LA strongly while placing GA migration under the control of a size-exclusion mechanism. The resin was thus proposed as the adsorbent for the considered separation. To compare the proposed and the previous adsorbents in terms of the SMB performances for GA–LA separation, the SMB processes based on the two adsorbents were optimized each to maximize throughput and minimize desorbent usage while keeping the product purities above 99%. The resulting Pareto solutions demonstrated that the SMB based on the proposed adsorbent surpassed the SMB based on the previous adsorbent in throughput and desorbent usage by a wide margin. In addition, the replacement of adsorbent from the previous to the proposed one could lead to a significant reduction in the SMB pressure drop. Finally, it was confirmed that the relative superiority of the proposed adsorbent over the previous one in terms of the SMB performances became greater, as the required level of purity or throughput was elevated.