Overcoming fouling in desalting industrial organic solutions by electrodialysis
More Info
expand_more
Abstract
Lignocellulose is an abundant source of feedstock and consists of hemicellulose, cellulose, and lignin. Pre-treatment is done to fragment lignin and hydrolyse the hemicellulose and cellulose fractions into simple sugars which is then converted into bioplastics. The pre-treatment generates a mix of sugar solutions, lignin breakdown products and inorganic ions. However, the inorganic ions act an inhibitor for fermentation and needs to be removed. Electrodialysis is investigated as an alternative process to ion-exchange to desalt the sugar solutions for downstream processes. While the sugars are neutral compounds, the other organics are negatively charged and can cause fouling on the positively charged anion-exchange membrane.
The process waters are C5C6 and C5 sugar streams and the main ions present are sodium and sulphate. A desalting target of 60% is set for these two ions and the experiments are done using a lab-scale electrodialysis set-up. Tests are run to check the desalting rate, effect of filtration, powdered activated carbon (PAC) and acidification, persistence of fouling and the effect of cleaning-in-place. The increase in average resistance and run time is used as an indicator of membrane fouling.
The raw C5C6 sugar solution could not reach the desalting target but once the solution is filtered, a desalting of 75.5% (Na+) and 66.5% (SO42-) is achieved. The average resistance also drops significantly. The fouling is removed by a CIP and is deemed to be ‘reversible’ in nature. On the other hand, the C5 sugar solution did not improve in performance after filtration. Although the desalting target could be achieved after filtration, the average resistance rose to 130 ohms. A CIP did not have any effect and increased run-time and drop in desalting efficiency was observed. The main reason for fouling in both cases is attributed to the presence of lignin degradation compounds. Lignin is more concentrated in the C5 sugar as compared to the C5C6 sugar solution due to its production process. In C5 sugar, the irreversible fouling pattern is hypothesized to be due to adsorption whereas in case of C5C6 sugar the reversible fouling pattern is cake fouling. PAC was dosed to remove the organic foulants in C5 sugar and improved the average resistance to 48 ohms and significantly reduced the run time. Desalting of 61.4% (Na+) and 87.2% (SO42-) was reached. Acidification protonated the feed and reduced the charged interactions, thereby both reducing run time and the average resistance. Desalting of 65.79% (Na+) and 68.02% (SO42-) was reached after acidification.
To conclude, the studied ED system could desalt the organic C5C6 and C5 sugar streams after a treatment step. Lignin degradation products were found to heavily foul the AEM membranes as its increased presence in C5 sugar stream led to a more severe fouling which could not be restored even with a chemical CIP. However, mitigation of the degradation products in the pre-treatment step could ensure that the solution was desalted by the ED unit. Thus, quantifying the composition of organic streams is imperative to predict the propensity of fouling in an ED unit.