Granulation and enrichment performance
Initial reactor operations were aimed at complete retention of the biomass. The cycle length varied in the beginning, starting with a long cycle length of 24 h. Both a flat base addition profile and CO2 and H2 concentrations in the off gas below 1 % (m/m) were used as indicators of full conversion of glucose into organic acids. During the reaction time the biomass could grow and consume the glucose dosed. A short settling time (2-10 min) resulted in non-granular biomass at the start. After increasing the settling time to 3 h a visible improvement of biomass retention was achieved. After 6 d the settling time was manually lowered as shown in figure 1 to promote granule formation.
The granules formed during the start-up period were used as starting material for the first enrichment. An impression of the reactor operated at different SRT is depicted in figure 2. The start-up period served only as inoculum production for the first enrichments and thus day 0 was chosen to be the first day of the SRT research.
After the start-up phase the settling time was increased to 60 min and the cycle length to 6 h, to prolong the SRT. In this first period the SRT was uncontrolled, this meant that TSS was not actively removed from the system. TSS could still leave the system when a maximum sludge volume (SV) was reached (which meant that the sludge bed reached until the effluent point of the reactor which was half way the reactor height) and additional biomass production left the reactor with the discharge effluent. Using this approach an SRT of 40-50 d was achieved. After characterisation of the system, the SRT was manually controlled at 10-20 d SRT and 1-2 d SRT. During the start-up phase H2 and CO2 were produced and measured in the biogas. After the start-up period no H2 was measured in the biogas, the biogas consisted solely out of CO2 (20-40% m/m) and N2 (60-80% m/m) and no (<0.5%) H2 or CH4 were detected. An overview of the culture performance for the 3 SRT is shown in figure 3.
With increasing the SRT the concentration of biomass in the reactor increased from 6 gVSS·L-1 at 1-2 d SRT to 59 gVSS·L-1 at 40-50 d SRT. An high SRT could be achieved due to amongst others a very compact sludge bed with a SVI of 11 ± 2 ml·gTSS-1 (average ± standard deviation; n=8) and a biomass yield of 0.11 gCODX·gCOD-1 (X = biomass). The sludge bed became dense as it showed a strong compressive ability. After starting the settling phase an initial sludge bed was formed within 10 minutes. After the initial settling the sludge bed compressed for the remainder of the settling time (total settling time 60 minutes) resulting in a very low SVI60. An overview of system characteristics is given in table 1. In all experimental periods the VSS present in the effluent was low, 0.21 ± 0.18 (n=100) gVSS·L-1, compared to the VSS concentrations present in the reactor. The manual removed sludge was 3.1 gVSS·d-1, 1.9 gVSS·d-1 and 0 gVSS·d-1 for the 1-2 d SRT, 10-20 d SRT and 40-50 d SRT respectively. The biomass yield was estimated over multiple cycles for each SRT and was in the range of 0.09-0.17 gCOD·gCOD-1. The assumption was made that all VSS could be classified as biomass as described in the material and methods.