Date of Award

Fall 12-16-2017

Semester of Degree


Document Type

Restricted Access Dissertation

Degree Name

Ph.D. in Paper and Bioprocess Engineering


Paper & Bioprocess Engineering

Major Professor

Shijie Liu

Steering Committee Member

Bandaru Ramarao

Steering Committee Member

Charles Kroll

Steering Committee Member

Thomas Amidon

Steering Committee Member

Biljana Bujanovic

Steering Committee Member

Gary Scott


In this study, four hardwood hot-water extracts were studied by using one nanofiltration (NF) and one reverse osmosis (RO) membranes. The operating conditions used were: transmembrane pressure was gradually applied from 150 psi to 550 psi, while temperature was kept to no greater than 50°C. During the process, the change of the concentrations in both retentate and permeate flows were determined by 1H and 1H13C Heteronuclear Single Quantum Coherence (HSQC) nuclear magnetic resonance (NMR). For individual components, a simplistic model based on mass balance was established to evaluate the overall rejection of each species. With the overall rejections over 95% for all sugars, most of the desirable components, i.e. sugars, are preserved. However, most of the fermentation inhibitors, such as acetic acid, furfural, hydroxymethylfurfural, formic acid and methanol, are effectively removed with rejections ranging from 20% to 80%. What’s more, by calculating individual rejections for each species each point, it was found that acetic acid, furfural and methanol sometimes even had negative rejections. Besides, the NF membrane was shown to be the better choice in terms of removing inhibitors and enriching sugars, due to the fact that the NF membrane had a lower rejection for inhibitors compared to the RO ones as well as that both feed flux and permeate fluxes for NF membrane are higher than the RO ones. It was also found that the dilute alkali solution used to clean the membrane could slightly increase the permeate flux of the NF module while it couldn’t effectively reverse the fouling that the RO module suffers. By extending the Spiegler-Kedem model into n solutes scenario, an irreversible thermodynamic model (Extended Spiegler-Kedem model) was established and the corresponding parameters were estimated by means of non-linear parameter estimation function with Matlab. By taking the solute-solute interactions into consideration, the estimated parameters were proved to be useful for explaining the negative rejections found in acetic acid, furfural and methanol.