Chapter 2: A closed-loop biorefinery approach for polyhydroxybutyrate (PHB)
3.3 Results and discussions
3.3.1 Hydrodynamic studies of ABR
Tracer experiments were first carried out with the ABR operated under batch mode by circulating the liquid inside while the air sparged into the system was allowed to escape through a vent placed at the top flange of the bioreactor. The tracer injected in the stem of the bioreactor was evenly distributed within the bioreactor due to the continuous air sparging and liquid agitation in the bioreactor. Therefore, tracer concentration measured at the top of the bioreactor represented the hydrodynamic conditions and flow pattern within the bioreactor. The increase in tracer concentration as detected by conductivity meter during the bioreactor operation at various air sparge and agitation rates is shown in Fig.
3.3.
The overall pattern of tracer study conducted at different agitation and sparge rates resembles an exponential curve. However, variations in the mixing pattern due to the change in agitation and sparging rates are observed. For instance, a change in agitation rate from 250 to 500 rpm at a fixed sparge rate of 0.8 vvm resulted in the enhancement of KCl concentration from 55 to 70.5 mg/L. Similarly, a change in the sparge rate from 0.4 to 0.8 vvm at a fixed agitation rate of 250 rpm resulted in a hike in the KCl concentration from 52 to 55 mg/L. Similar tracer patterns were observed by Saravanan et al. (2010) in their study on the hydrodynamics of an internal loop airlift bioreactor. In contrary, Jamshidi et al. (2001) on their hydrodynamic study using downflow jet loop bioreactor reported an
initial sharp increase in the KCl concentration followed by a reduction in the value until a steady-state reached. This change in the tracer pattern might be due to the position of the conductivity probe placed in the bioreactor system. As mentioned earlier, the conductivity probe in the present study was placed on the top while the tracer was injected at the stem of the bioreactor. However, keeping the probe close to the injection area might have resulted in a steep increase in the KCl concentration. Therefore, it can be concluded from these tracer experiments that ABR operated at a sparge rate of 0.8 vvm and agitation rate of 500 rpm resulted in a better mixing than at other conditions. Since the ABR used in the present study was aimed at achieving a maximum growth of fragile bacterial cells, the agitation rate was not increased beyond this point. This observation is consistent with those of Qiao et al. (2014) for culturing of QM7 cells in a Taylor-Couette bioreactor and Ramezani et al. (2017) on using multiphase Taylor-Couette vortex bioreactor.
Fig. 3.3. Results of tracer (KCl) experiments with the ABR operated at varying operating conditions
In order to determine the capability of the bioreactor to efficiently deliver oxygen to the microorganism, KLa and dissolved oxygen measurements were carried out. Fig.
3.4(a) portrays that at the end of 600 s of operating the ABR at a sparge rate of 0.4 vvm and agitation rate of 250 rpm, the DO level in the liquid was 74% of the saturation efficiency. An increase of agitation rate from 250 to 350 rpm by keeping the sparge rate at 0.4 vvm resulted in 93% of the saturation efficiency. At all other operating conditions, the DO profile reached 100% saturation efficiency, i.e., a maximum DO concentration of 7.8 mg/L. A similar trend of DO profile (achieving the 100% saturation efficiency) was reported by Ramezani et al. (2015) in a narrow gap Taylor-Couette vortex bioreactor.
Fig. 3.4(b) shows the linearized form of DO profile for estimating the volumetric mass transfer coefficient (KLa) in ABR. KLa in the ABR increased with an increase in the agitation and sparge rates. For instance, an increase in the agitation rate from 250 to 500 rpm with a constant sparge rate of 0.4 vvm resulted in 5.6 times enhancement in KLa value from 8.12 to 45.2 h-1. The highest KLa of 70 h-1 was observed at a maximum agitation and sparge rate of 500 rpm and 0.8 vvm, respectively. The increased KLa with agitation rate is due to an efficient mixing in the ABR (Curran and Black, 2005; Qiao et al., 2018; Ramezani et al., 2015). KLa values obtained at various operating conditions of ABR were presented in Table 3.2. Air bubbles in the ABR were observed using a high-speed camera, and the images along with the bubble size distribution are shown in Fig. 3.5. It can be inferred from the figure that any effort to increase the turbulence in the ABR by varying the agitation rate resulted in a shift in the bubble size to a larger size. However, variation in the sparge rate at a given agitation rate displayed only a slight increase in the bubble size.
Table 3.2. Comparison of ABR performance with that of the STBR at varying operating conditions
Agitation rate (rpm)
250 250 350 350 500 500 250
(STBR) 300 (STBR) Air sparge
rate (vvm)
0.4 0.8 0.4 0.8 0.4 0.8 0.8
(STBR) 0.8 (STBR) OD600 (60h) 9.64 ±
0.1
27.96
± 1.47
21.9 ± 0.81
37.3 ± 0.89
53.16 ± 1.6
70.2 ± 2.0
50.16 ± 1.02
44.2 ± 1.0 PHB % (60 h) 70.43 64.5 ±
0.49
67.0 ± 0.70
60.1 ± 0.50
54.2 ± 1.00
50.1 ± 0.8
49.8 ± 0.52
47.5 ± 0.52
KLa (h-1) 8.12 24.1 16 32 45.2 70 42 45.2
Fig. 3.4. (a) DO profile and (b) linearised profile for KLa determination in the ABR operated at varying operating conditions
The increase in bubble size with an increase in the agitation rate is typically found only in such bioreactor having Taylor and Couette flow, but not in other conventional STBR or ALR (Saavanan et al., 2009). For instance, a narrow gap annular bioreactor having Taylor and Couette flow also resulted in a continuous shift in bubble size to a higher value with an increase in agitation/sparge rate (Ramezani et al., 2015). Similar results were observed by the same authors for a different study carried out using the same bioreactor with Taylor and Couette flow (Ramezani et al., 2017). Bubble size analysis in another horizontal bioreactor with Taylor and Couette flow was also found to match with the results obtained in this study (Hubacz and Wroński, 2004). The increase in bubble image and enhancement in oxygen KLa is attributed to the bubble coalescence (Fig. 3.5(b)) due to the Taylor flow and increased gas hold up in the Taylor vortices, respectively (Ramezani et al., 2015).
These results further confirm the presence of Taylor and Couette flow in the present ABR with a wide annular gap. In order to further ascertain the flow pattern in ABR, CFD simulation was performed on the bioreactor geometry and the results were discussed in the next section.
Fig. 3.5. Images showing (a) air bubbles originating from the sparger, (b) air bubbles at the stem of the ABR and (c) plot of bubble size distribution in the ABR operated at varying operating conditions