High-solids semi-continuous anaerobic digestion of corn silage in bag-type digester

2.1 Laboratory model

Anaerobic digestion was realized in a lab-scale model of a partially mixed bioreactor made of a triple-layer plastic bag. The bag was made of Sioen B6070 (Sioen Industries NV, Belgium, Ardooie) (polyvinyl chloride (PVC)-polyethersulfon (PES)-PVC foil with a specific weight of 1150 g m⁻² certified for gas holders of biogas plants), had a diameter of 0.5 m, a length of 3.6 m, and a total volume of 0.7 m³. The working volume (volume of the anaerobic batch) was set to 0.5 m³. The bag was stretched out on a short tubular adapter of circular steel cap with a diameter of 0.5 m. The outer casing surrounding the whole bag was built of polystyrene boards of 100 mm thickness, which served as heat insulation. The heating unit for recirculation of warm air around the bag was located at the back of the model. It consisted of two electric heating spirals (1×400 W and 1×400 W backup), an industrial standard tube axial fan (Vents VKMz100 type with a speed controller), flexible aluminum piping for air recirculation (2×4 m length), and a thermostat with an electricity consumption meter. The fan blew the warm air (75°C) on an arched end of the bag. Then, the warm air was flowing in the heat insulation casing along the bag towards the steel cap of the fermenter. The cooled air (40°C) was drawn from the cap via the internal aluminum piping back to the fan and the electric heating spirals heated it again. Air heating was chosen with regard to the large surface area heated, and due to construction simplicity, enabled rapid unfolding and folding of the reactor. No water circulating piping was used. The model bioreactor was located on the laboratory desk in the horizontal position.

To feed the liquid inoculum, a plastic vessel with a volume of 0.03 m³ was connected to a self-priming pump with an axial rubber impeller (NEREZ Blucina, DN50 type, power consumption 1.5 kW, Czech Republic). The liquid was fed to the steel cap of the bag. To feed the solid corn silage, the same vessel and pump were used after the silage was manually mixed into a small amount of digestate.

Batch stirring was performed in two ways: first of all by the means of recirculation by a pump that was switched on manually three to five tmes a day. The batch was drained by a hose with a 50 mm diameter (positioned at the bottom of the bag and led out at the rear) and returned to the opposite side of the bag (towards the cap), or vice versa. This recirculation was performed daily at 9:00 AM for 15 min before dosing and for 3 min after each dosing of the substrate. The batch was also partially stirred by a continuously running stirrer with two rectangular tinny blades mounted horizontally in the steel cap slightly above its axis. This stirrer with a rotation speed of 23 min⁻¹ prevented the foam from leaking into the biogas collecting tank placed above the steel cap. The stirrer was driven by an asynchronous motor of 0.18 kW power input.

The constant batch volume was ensured by an overflow of the digestate through the bent tube (50 mm in diameter) at a constant height (slightly above the stirrer axis). Digestate sampling was done through a valve and hose (60 mm in diameter) beneath the rear of the bag. Once each working day (before morning silage dosing) 2.5 l of digestate was drained out and un-analyzed excess was returned to the bag. Biogas overpressure (10–100 Pa) was checked with a U-type pressure gauge and secured by a liquid-seal. The biogas was collected in the vertical rectangular biogas trap (tinny box of 0.02 m³ volume connected with the circular steel cap). The biogas production was continuously measured with a drum-type gas flowmeter and the composition was measured with a portable IR/electrochemical analyzer. Figure 1 shows a schematic section of the model bioreactor.

Figure 1:

Model bioreactor schema – longitudinal section.

2.2 Feedstock

Liquid digestate (reacting slurry) from the first reactor of the agricultural biogas station Pustejov II, where cattle slurry and agricultural substrates like corn silage and lucerne haylage are processed, was used as inoculum. The inoculum temperature dropped during the transport from 40°C to 28°C. This inoculum with a pH value of 7.82 contained in average 6.0% of the total solids (TS), 79.9% of volatile solids in TS (VS_TS) and the volatile fatty acid (VFA)/total inorganic carbon (TIC) ratio was 0.22. The C:N ratio was extremely low (5.1:1).

Solid fibrous substrate in the form of corn silage of KWS Atletico cultivar was obtained from the same biogas station. This corn silage was made of the whole crop in the period when the solids content in the chopped material was 30%–35%. Silage was delivered every second week and stored in plastic barrels with internal lids lying on the silage surface to avoid mold growing. Before dosing into the bioreactor, silage was not grinded so the particle size was in the range of 15–20 mm (sometimes up to 50 mm). The silage with a pH of 4.50 contained 30.6% of TS, and 95.6% VS_TS. The average parameters of the inoculum and substrate are shown in Table 1. Methods and standards of individual analyses are shown in Table 2.

2.3 Modelling method

The mesophilic (40±3°C) anaerobic digestion was conducted in the semi-continuous mode. The bioreactor was filled with 500 kg of liquid inoculum. Within a couple of hours, an average temperature of 38°C was reached and on the 3rd day, the temperature reached 40°C, as required. At first, the daily dose of corn silage was specified with respect to the known load of the first stage bioreactor in the biogas station Pustejov, which corresponded to about 2.5–3.5 kg d⁻¹. Daily dosing was performed after dilution with digestate recirculated from the bag. For trouble-free pumping with the self-priming pump via a short hose, a ratio in the range of 3:1 to 5:1 (five parts by weight of digestate) was sufficient. Gradually, with concurrent control of pH, lower fatty acids, ratio of volatile organic acids concentration and a total concentration of inorganic carbonate (VFA/TIC ratio), and ammonia nitrogen, the daily dose of silage was increased. Dosing up to 3.5 kg of substrate was performed once a day; larger daily quantities were divided into two to three sub-doses that were pumped into the fermenter with an interval of approximately 4 h. Effort was taken to increase the TS content in the reactor above 10 wt% during the shortest time. Thereafter, the TS content was increased slowly up to the known limit for a safe and reliable operation of the pump (approximately 15 wt%). In general, effort was taken to intensify the methane production. Due to the high inlet solids, a minimum daily volume of digestate was produced.

2.4 Analysis

Batch temperature was measured continuously using eight sheathed thermocouples arranged in pairs (left/right) located close to the bottom of the bag at four distances from the cap (0.5 m, 1.0 m, 2.0 m, 3.0 m). The temperature was also checked on a thermometer located in the center of the circular steel cap. Batch suspension (digestate) samples were analyzed using the portable pH-meter EUTECH pH6+. Manual titration of the VFA/TIC ratio using 0.05 M H₂SO₄ to two end-points (pH 5.0 and pH 4.4) was performed on 20 g samples. The substrate and digestate TS content was measured using an OHAUS MB23 infrared moisture analyzer (10±1 g at 105°C to constant weight). Then,the samples were stored in a refrigerator at a temperature of 2°C–4°C. Once a week, all the batch suspension samples and a fresh substrate sample were subjected to thermogravimetric analysis on a LECO TGA 701 analyzer (LECO, Saint Joseph, USA). There, the TS content and volatile solids content (VS as a loss on ignition at 550°C to constant weight) were specified on 1–2 g samples with an accuracy of ±0.02% relative standard deviation.

Continuous recording of the biogas flow and total volume was done using a drum-type gas flowmeter Ritter TG1/5 (RITTER Apparatebau GmbH & Co. KG, Bochum, Germany) (0.002–0.120 m³ h⁻¹, ±0.5% accuracy) made of transparent PVC and connected to a PC. Daily biogas production was recorded also manually each morning and it was recalculated to normal conditions, dry gas (0°C, 101.325 kPa).

The raw biogas composition was measured sequentially, every working day (before dosing of the substrate at 9:00 AM) using a portable IR/electrochemical analyzer (Biogas5000: Geotechnical Instruments Ltd, Queensway, UK) type “Biogas” with accuracy of ±3% over 15% CH₄, ±3% over 15% CO₂ and of ±1% below 5% O₂, after calibration. This portable analyzer does not contain H₂ and H₂S sensors.

Once a week, the biogas composition was verified with an Agilent 7890A gas chromatograph (Agilent Technologies, Santa Clara, CA, USA) with a flame-ionization detector. Once a month, the biogas composition was measured every hour for 24 h to correct an average content of methane. The average daily content of methane in biogas corresponded to 0.9×CH₄ value measured in the morning. On completion of the test, an average value of the biogas and methane specific production was calculated. On the last day, the final digestate samples were drained for a complete analysis, especially for checking risk elements (Cd, Pb, Hg, As, Cr, Mo, Ni, Zn, and Cu) contents according to the CZ Regulation no. 271/2009 Coll., as amended.

3.1 Anaerobic digestion test

Digestion of corn silage was conducted for 120 days. The maximum dose of substrate was 12 kg and the average organic loading rate reached 4.27 kg_VS m⁻³ d⁻¹. This loading is at the normal upper limit for the wet process [25], [26]. In the long term, it seemed practical to not exceed 5.0 kg_VS m⁻³ d⁻¹ which is still lower than some others use [27]. Higher doses caused noticeable overloading of the fermenter within the short test period. The hydraulic retention time was lowered from 280 days to 64 days with an average value of 85 days, which is a little longer than is common but should decrease after a longer period. The typical hydraulic retention time for corn silage is 30–90 days [28]. The pH value was stable; it never decreased below 7.65. For the first 45 days, the VFA/TIC ratio was low (up to 0.25) which means almost no overloading, then it increased slowly over 0.50 up to 0.75, which indicated overloading. During the long-term mono-digestion of corn silage (after about 8 months) it would probably be necessary to dose micronutrients, especially Co, Mo, and Se [29].

Due to the horizontal bag fermenter configuration, a two-phase reaction took place, since the paddle stirrer and recirculation pump were not able to maintain the batch fully suspended. There was always a crust in the top part of the bag with the solids content approximately 15–20 wt% TS and a liquid phase with solids content up to 13 wt% TS in the lower part. The process was high-solids only in the crust. During the 120 day test period, an average TS content of 10.3 wt% and average VS content of 77.5 wt%_TS was achieved in the recirculated liquid digestate.

The biogas production intensity calculated from the working volume of the bag increased relatively linearly from 1.52 m_N³ m⁻³ d⁻¹ to 4.86 m_N³ m⁻³ d⁻¹. The average value was 3.42 m_N³ m⁻³ d⁻¹. This intensity corresponds to the semi-dry process [30], [31]. Themethane content in raw biogas fluctuated from 48 vol% to 59 vol%, but from the 40th day it was relatively stable between 50 vol% and 55 vol%. The average value was 52.5 vol%. This is the normal value of methane content from corn silage [32]. The specific methane production stabilized at 0.123 m_N³ kg⁻¹ (0.419 m_N³ kg_VS⁻¹) and the specific biogas production was 0.234 m_N³ kg⁻¹ (0.800 m_N³ kg_VS⁻¹), which is approximately 2.63 times lower than the biogas production measured by Koutný et al. [32]. The authors state a specific biogas production of 0.617 m_N³ kg_VS⁻¹. However, this is the value from a 26 days batch test, not a long-term average. It is, however, evident that in the bag, the methanization was inhibited by ammonia reaching 4–5 kg m⁻³.

The progress of the digestion temperature, pH and VFA/TIC ratio is shown in Figure 2. The progress of TS content in the reactor, digestate solids loss on ignition at 550°C, organic loading rate, the daily production of biogas and the average dry methane content are shown in Figure 3. Average process parameters for the entire period of the test are shown in Table 3.

Figure 2:

The progress of the temperature, FOS/TAC and pH.

Figure 3:

The progress of organic loading and biogas production.

The average power consumption of the air heating system was 7.980 kWh d⁻¹ (0.666 kWh m⁻³ of the batch and day) and the average net calorific value of the biogas was 8.958 kWh d⁻¹. The ratio of produced and consumed energy was low, but we can expect a better result by an operating device with a volume of 20–50 m³. The average heat input of biogas (calculated from biogas net calorific value and batch volume in the reactor) amounted to 0.746 kW_th m⁻³. Assuming a 30% electrical efficiency in eventual biogas cogeneration, the electric power could be 0.22 kW_el m⁻³ at least. This value is higher than has been verified, for example, by Besgen [33] at four small agricultural biogas stations with two-stage wet processes (max. 0.13 kW_el m⁻³). By contrast, it is significantly less than that published from pilot studies (0.33–0.42 kW_el m⁻³) by Kayhanian [34] and Kayhanian and Tchobanoglous [35], or by Banks [36] from full-scale high-solids digestion in the DRANCO system (0.5–0.6 kW_el m⁻³).

3.2 High-solids digestate

The average daily production of liquid digestate was approximately 4.8 kg which formed approximately 66 wt% of substrate fed. The samples acquired on the last day contained 13.5 wt% of TS and 82.4 wt% of VS in TS in average. The average parameters of high-solids digestate are shown in the last column in Table 1. Approximately 0.15 wt% was formed of volatile fatty acids. The total carbon content (40.6 wt%_TS) was almost completely formed by organic carbon. The C:N ratio was very low from the process point of view (7:1) which was caused by low C:N of the inoculum. The optimum C:N ratio is 15–20:1 [37]. More than half of the nitrogen content (5.8 wt%_TS) was formed of ammonia nitrogen. Nitrate nitrogen content was relatively negligible. The solids were not analyzed completely, but the crude fiber content was 10.6 wt%_TS, and lipids content was 3.1 wt%_TS. Simple carbohydrates and starch contents were low. Approximately 1.7 wt% of TS was formed by phosphorus, which was higher than necessary (C:P ratio was only 24:1). Kayhanian and Rich [38] recommend a C:P ratio in the range 120–150:1. Potassium content was also excessive; the C:K ratio was 12:1. Kayhanian and Rich [38] recommend a C:P ratio in the range 45–100:1. Iron content was at the upper limit (0.65 wt%_TS) and the content of nickel, as one of the most important micro-nutrients, was 1.5×higher than the concentration needed, which is 0.002 wt%_TS. Co and Mo were present.

The corn silage digestate met legislative limits for risk elements (Cd, Pb, Hg, As, Cr, Mo, Cu, and Zn) with TS content above 13 wt%_TS according to the Regulation no. 271/2009 Coll., as amended, as a result of the input material. Only Cd, Ni and Zn contents were significantly closer to the limits.

Digestate quality parameters set for land reclamation according to the Regulation no. 341/2008 Coll., Addendum no. 5, table no. 5.3, as amended (maximal humidity 98.0 wt%, minimum total nitrogen as N re-calculated for a dried sample 0.3 wt%, and pH in the range of 6.5–9.0) were also met. On the basis of risk element contents, it was able to classify this digestate into Group 2 and Class I (most valuable class) of digestate for land reclamation (table no. 5.1 – maximum limit of 170 mg kg_TS⁻¹ for copper and maximal limit of 65 mg kg_TS⁻¹ for nickel). This means that this digestate could be used outside agricultural or forest lands – on sport and recreational grounds and green lands, as well as in urban areas, apart from outside playing fields, on urban green fields, parks, forest parks, industrial zones, waste dumps, and sludge beds.

The smell was mostly ammonia. After cooling down to 20°C, the pH value increased from 7.8 to 8.2 and the smell dropped down at a level comparable to liquid digestate from well operated agricultural biogas stations.