Potassium hydroxide solution

For the experiments recommended from the mDoE-toolbox, the initial dry cell weight (DCW) was adjusted to $c_{\mathrm{DCW}}=2{\mathrm{g}\mathrm{l}}^{-1}$ and the initial conditions of the complex culture medium were prepared, as shown in Table 1. The recommended feeding strategy for the mDoE experiments was a linearly rising feeding rate starting at $t_{\mathrm{start}}=1\mathrm{h}$ . The final feed volume flow at $t_{\mathrm{end}}=48\mathrm{h}$ was determined by the mDoE-toolbox for $F_{\mathrm{Glc}}$ and $F_{\mathrm{N}}$ separately. The feed rates were determined using the following function: $$\begin{array}{c}\hfill F_{\mathrm{k}}(t)=\frac{F_{\mathrm{k},\mathrm{end}}}{t_{\mathrm{end}}}\times (t-t_{\mathrm{start}}).\end{array}$$ For $t<t_{\mathrm{start}}$ , $F_{\mathrm{k}}(t)$ equals zero.
Linear feed strategies have been chosen instead of exponential strategies, as these are easier to handle and less risky. Slightly too large feed rates resulting from exponential strategies quickly result in over-feeding. The pH was held constant using 20 wt% potassium hydroxide solution or 20 wt% phosphoric acid (both VWR). The temperature was controlled at $30{}^{\circ }\mathrm{C}$ .

For the parameterization of the pH-related model part (see Section “Mathematical process model”), four cultivations with different pH were performed in 1 l baffled shake flasks (500 ml working volume, Schott, Germany), which were shaken at 170 rpm (1.9 cm shaking diameter, MaxQ4000, Thermo Fisher Scientific, USA) with initial $c_{\mathrm{DCW}}=10{\mathrm{g}\mathrm{l}}^{-1}$ . The temperature was controlled at ${30}^{\circ }\mathrm{C}$ . The pH was adjusted initially and maintained manually (pH = 3, 4, 5, 6, respectively) to the desired value. In all experiments, the pH was adjusted using 20 wt% potassium hydroxide solution or 20 wt% phosphoric acid (both VWR, Germany). One feed pulse (glucose and nitrogen source) of 50 ml was added to each flask after 24 h (feed concentration in Table 1).
After adjusting the pH part of the model, further model parameterization was done based on historical data of three bioreactor (2 l working volume, Biostat B, Sartorius, Germany) fed-batch cultivations with different initial concentrations and feed compositions (see Table 1). Gassing was manually set in relation to the state of the process between 1 and 2 vvm (max. 2 $\text{l}{\text{min}}^{-1}$ ), and stirring was held between 500 and 800 rpm to maintain a DO above 10%. The pH was automatically controlled at 5. Temperature was set to $30{}^{\circ }\mathrm{C}$ .