Addressing raw material variability: In-line FTIR sugar composition analysis of lignocellulosic process streams.

For a sustainable economy, biorefineries that use second-generation feedstocks to produce biochemicals and biofuels are essential. However, the exact composition of renewable feedstocks depends on the natural raw materials used and is therefore highly variable. In this contribution, a process analytical technique (PAT) strategy for determining the sugar composition of lignocellulosic process streams in real-time to enable better control of bioprocesses is presented. An in-line mid-IR probe was used to acquire spectra of ultra-filtered spent sulfite liquor (UF-SSL). Independent partial least squares models were developed for the most abundant sugars in the UF-SSL. Up to 5 sugars were quantified simultaneously to determine the sugar concentration and composition of the UF-SSL. The lowest root mean square errors of the predicted values obtained per analyte were 1.02 g/L arabinose, 1.25 g/L galactose, 0.50 g/L glucose, 1.60 g/L mannose, and 0.85 g/L xylose. Equipped with this novel PAT tool, new bioprocessing strategies can be developed for UF-SSL.

Bioprocess monitoring: minimizing sample matrix effects for total protein quantification with bicinchoninic acid assay.

Determining total protein content is a routine operation in many laboratories. Despite substantial work on assay optimization interferences, the widely used bicinchoninic acid (BCA) assay remains widely recognized for its robustness. Especially in the field of bioprocess engineering the inaccuracy caused by interfering substances remains hardly predictable and not well understood. Since the introduction of the assay, sample pre-treatment by trichloroacetic acid (TCA) precipitation has been indicated as necessary and sufficient to minimize interferences. However, the sample matrix in cultivation media is not only highly complex but also dynamically changing over process time in terms of qualitative and quantitative composition. A significant misestimation of the total protein concentration of bioprocess samples is often observed when following standard work-up schemes such as TCA precipitation, indicating that this step alone is not an adequate means to avoid measurement bias. Here, we propose a modification of the BCA assay, which is less influenced by sample complexity. The dynamically changing sample matrix composition of bioprocessing samples impairs the conventional approach of compensating for interfering substances via a static offset. Hence, we evaluated the use of a correction factor based on an internal spike measurement for the respective samples. Using protein spikes, the accuracy of the BCA protein quantification could be improved fivefold, taking the BCA protein quantification to a level of accuracy comparable to other, more expensive methods. This will allow reducing expensive iterations in bioprocess development to due inaccurate total protein analytics.