Waste Management xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Waste Management journal homepage: www.elsevier.com/locate/wasman

Rapid digestion of shredded MSW by sequentially flooding and draining small landfill cells William P. Clarke ⇑, Sihuang Xie, Miheka Patel Centre for Solid Waste Bioprocessing, Schools of Civil and Chemical Engineering, The University of Queensland, Brisbane 4072, Australia

a r t i c l e

i n f o

Article history: Received 22 July 2015 Revised 29 November 2015 Accepted 30 November 2015 Available online xxxx Keywords: Anaerobic digestion Flood and drain Landfill cell Leachate Municipal solid waste

a b s t r a c t This paper compares the digestion of a packed bed of shredded municipal waste using a flood and drain regime against a control digestion of similarly prepared material using a trickle flow regime. All trials were performed on shallow (2 m) beds of the sub-8 cm fraction of shredded mixed MSW, encapsulated in a polyethylene bladder. The control cell (Cell 1) was loaded with 1974 tonnes shredded municipal waste and produced 76 ± 9 m3 CH4 dry t
1 (45 ± 2 m3 CH4 ‘as received’ t
1) over 200 days in response to a daily recirculation of the leachate inventory which was maintained at 60 m3. The flood and drain operation was performed on two co-located cells (Cell 2 and Cell 3) that were loaded simultaneously with 1026 and 915 tonnes of the sub-8 cm fraction of shredded mixed MSW, with a third empty cell used as a reservoir for 275 m3 of mature landfill leachate. Cell 2 was first digested in isolation by flooding and draining once per week to avoid excessive souring. Gas production from Cell 2 peaked and declined to a steady residual level in 150 days. Cell 3 was flooded and drained for the first time 186 days after the commencement of Cell 2, using the same inventory of leachate which was now exchanged between the cells, such that each cell was flooded and drained twice per week. Biogas production from Cell 3 commenced immediately with flooding, peaking and reducing to a residual level within 100 days. The average CH4 yield from Cells 2 and 3 was 123 ± 15 m3 dry t
1 (92 ± 2 m3 ‘as received’ t
1, equal to 95% of the long term (2 month) BMP yield. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction A widely practised method of accelerating the digestion of municipal solid waste (MSW) in landfills is to irrigate the waste with leachate to overcome moisture limitations and to distribute buffered leachate flushed from stabilised layers to fresh, reactive waste. Landfills with this practice are commonly referred to as bioreactor landfills. The goal of bioreactor landfill operation is to achieve more rapid stabilisation of MSW rather than high rate digestion, primarily because digestate is left in-situ in bioreactor landfills while the space in high rate anaerobic digesters is re-used. Minimum time to arrive at peak CH4 generation in a bioreactor landfill is predicted to be 1.5–2 years (Pohland and Al-Yousfi, 1994). In contrast, it is well established that the batch digestion of static MSW beds can be completed within one month using in-vessel processes, by recirculating and trickling leachate over the waste bed. For example, Chynoweth et al. (1992) developed a sequential batch digestion method where leachate from a mature bed was ⇑ Corresponding author. E-mail address: [email protected] (W.P. Clarke).

used to flush acidic leachate from a fresh bed, to accelerate the establishment of methanogenic conditions. CH4 yields as much as 200 m3 CH4 t
1 VS (equivalent to 200 m3 biogas t
1 for their waste) were achieved in 21 days at 55 °C in 675 L batches of OFMSW (organic fraction of MSW) that was hand sorted and hammermilled to less than 8 cm particle size (Table 1). Chugh et al. (1998) verified this process in repeated trials on shredded MSW in 200 L reactors operated at 38 °C and found the rate of digestion increased with the volume of leachate used in the recirculation circuit (Table 1). At a commercial in-vessel scale, ten Brummeler et al. (1992) and ten Brummeler (2000) degraded source separated household organic waste, achieving 70 m3 of biogas per tonne in 21 days in 480 m3 cells using the BIOCEL process, where digestate from previous batches was mixed with fresh waste, optimally at a 1:1 ratio and leachate was recirculated through the bed, nominally at a rate relative to the working volume of 0.3 m3 m
3 d
1. Similar commercialised static leach-bed processes such as the BEKON process have produced higher yields from source separated OFMSW, with typical CH4 yields of 80 m3 t
1 in 28 to 35 days. Complete biogas records from the batch digestion of waste in landfill settings are rare. Barlaz et al. (2010) conducted a survey

http://dx.doi.org/10.1016/j.wasman.2015.11.050 0956-053X/Ó 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Clarke, W.P., et al. Rapid digestion of shredded MSW by sequentially flooding and draining small landfill cells. Waste Management (2015), http://dx.doi.org/10.1016/j.wasman.2015.11.050

2

W.P. Clarke et al. / Waste Management xxx (2015) xxx–xxx

Table 1 Performance of some pilot and full scale landfill bioreactors. Feedstock composition and mass

Inoculation

Bioreactor

Flow regime/rate

Operational parameters

Methane yield

References

 Organic fraction of MSW (less than 10 cm particle size)  (180 kg per batch)

Sequencing with mature, digested beds of MSW

675 L batch laboratory reactor

Recirculating, trickling leachate over the waste bed Rate: unknown

SRT 21 days T = 55 °C

200 m3 t
1 VS (200 m3 biogas t
1)

Chynoweth et al. (1992)

 Shredded MSW, average particles size  10 cm  50 kg per batch

Sequencing with mature, digested beds of MSW

200 L batch laboratory reactor

Recirculating, trickling leachate over the waste bed, at rates of 0.02, 0.1, 0.3 m3 m
3 d
1

SRT 60 days T = 38 °C

120, 160 and 190 m3 t
1 VS (from lowest to highest recirculation rates)

Chugh et al. (1998)

 Source separated household org. waste  400t of total waste (fresh + digestate)/batch

Mixing of fresh waste and digestate at a ratio of between 1:1 and 1:1.5.

480 m3 concrete batch digester

Recirculating, trickling leachate at a rate of 0.3 m3 m
3 d
1

SRT 21–30 days T = 35 °C

70 m3 biogas t
1

ten Brummeler et al. (1992), ten Brummeler (2000)

 Source separated OFMSW

Mixing of fresh waste and digestate at unknown ratio

Concrete containers

Recirculating, trickling leachate Rate: unknown

SRT 28–35 days T = 37 °C

80 m3 t
1

BEKON process

 MSW  0.1–15.6 million t

No Inoculation

Full-scale landfill based bioreactors with area between 1.4–40.0 ha

Recirculating, trickling leachate at a rate of 0.015–0.4 m yr
1

SRT 6–20 years No temperature control

59–120 m3 t
1

Barlaz et al. (2010)

Green waste (1718t) and aged horse manure (118t)

Mixing with aged horse manure

3500 m3 truncated pyramid shaped polyethylene lined landfill cell

Recirculating, trickling leachate at a rate of 2  10
3 m3 m3 d
1, or 1.8 m yr
1

SRT 366 days No temperature control, average temperature T = 46.5 °C

27.3 m3 t
1 (60 m3 biogas t
1)

Yazdani et al. (2012)

 370 t per batch

of biogas production and leachate composition trends in 5 bioreactor landfills across the USA, including a landfill cell (69,240t) that was rapidly filled and sealed. That cell was operated with a leachate recirculation rate of approximately 0.4 m yr
1. The waste degradation rate, according to a fit of the US EPA LandGem model to biogas data and an assumed ultimate methane yield of 59 m3 t
1 was 0.35 yr
1, which equates to 80% methane yield in 4.6 years. Most recently, Yazdani et al. (2012) digested green waste (1718t) and aged horse manure (118t) in a truncated pyramid shaped polyethylene lined cell that was approximately 3500 m3. They applied trickle flow irrigation and recirculation, at a normalised rate of 2  10
3 m3 m
3 d
1, equal to 1.8 m yr
1 for their cell geometry. This is orders of magnitude lower than that applied in the above mentioned laboratory studies and commercial leachbed processes. Without any heating, the cell maintained an average temperature of 46 °C during digestion with a cumulative CH4 yield after 366 days of 27.3 m3 CH4 t
1 (60 m3 biogas t
1). A further 7.2 m3 CH4 t
1 was produced from anaerobic pockets that persisted over a subsequent 66 day aeration phase. Subsequent BMP assays by Yazdani et al. on digestate at the end of the 366 day period indicated 60% of the ultimate BMP yield was realised during the anaerobic phase. A conclusion cannot be drawn from these few trials of the potential to match in-vessel digestion rates in landfill cells. There are numerous effects such waste type, particle size reduction, leachate application and temperature that can be manipulated to enhance degradation. The aim of this paper is to isolate the effect of leachate application. It is hypothesised in this paper that digestion in field scale leach-beds is slow primarily because of poor distribution of leachate through the bed. Infiltrating flow will only distribute evenly if the waste particles are fine enough to retard and distribute flow by capillary and friction forces. In reality, waste particles are not uniformly sized or shaped. Particles can block flow and can be rigid and large enough for trickle flow to occur. This is less significant at the laboratory scale where high flow rates can be applied to narrow columns. In contrast, flow can deviate and by-pass pockets of waste in broad beds such as landfill cells.

Poor liquid–solid contact can be overcome by completely submerging the waste. Flooding and draining cycles have been previously used by Rees-White et al. (2011) on a 25,000 tonne 5 m deep bed of MSW for the purpose of flushing leachable metals and other hazardous components from the waste bed. The superficial velocity applied during flooding was 0.6 mm hr
1. The extent of leachate hold-up in the bed increased with the number of flood and drain cycles. In contrast, complete and rapid exchange of leachate is required to implement strategies such as sequential batch digestion where the time to flood and drain should be insignificant compared to the overall degradation time. This paper presents flood and drain digestion trials on small landfill based 2 m deep beds of the sub-8 cm fraction of shredded MSW. A control bed of the same depth and similarly sourced material was operated in a trickle flow regime. 2. Material and methods Three landfill cells were constructed at the Swanbank landfill, Ipswich, Queensland. Cell 1 was operated between October 2010 and April 2011 using a trickle flow regime. Cell 2 and Cell 3 were operated between August 2012 and May 2013 using a flood and drain flow regime. The exact dates of operation are shown in Table 2. 2.1. Description of the waste and leachate used in the cells All cells were loaded with the sub-8 cm fraction of shredded kerbside collected MSW. A 750 DK Hammel shredder and a Komptech Farwick trommel with an 80 mm mesh was used to shred and screen the waste for all cells. The chosen mesh size was estimated to be large enough to pass organic particles but fine enough to reject materials that survived the shredding processes such as metal, timber and plastic sheets. The sub-8 cm fraction ranged between 46% (Cell 3) and 51% (Cell 2) of the total MSW fed to the shredder. Cell 1 was loaded almost 2 years prior to the flood and drain cells (Table 2), which brings into question the comparability of the waste

Please cite this article in press as: Clarke, W.P., et al. Rapid digestion of shredded MSW by sequentially flooding and draining small landfill cells. Waste Management (2015), http://dx.doi.org/10.1016/j.wasman.2015.11.050

3

W.P. Clarke et al. / Waste Management xxx (2015) xxx–xxx Table 2 Characteristics of the sub-8 cm fraction of shredded MSW loaded to the Cells (±std dev).

Mode of operation Operational period Leachate first applied Finished operation Total days Characteristics fresh waste (±S.D.) Total waste (t) Total solids (%) Volatile solids (%, dry basis) BMP Yield (N m3 t
1 VS) Visual classification (%, dry basis) Organics Plastic Inerts (Metal, glass, mineral) Cell CH4 yields (±S.D.) N m3 (‘as received’ t)
1 N m3 dry t
1 N m3 t
1 VS Characteristics digestate (±S.D.) Collection of digestate Volatile solids (%, dry basis) BMP yield (N m3 t
1 VS)

Cell 1

Cell 2

Cell 3

Trickle flow

Flood and drain

Flood and drain, sequenced with Cell 2

17-Oct-10 05-May-11 200

9-Aug-12 26-May-13 290

11-Feb-13 26-May-13 104

1974 59 (12) 62 (13) N.D.

1026

N.D. N.D. N.D.

66 5 29

45 (2) 76 (9) 123 (21) 09-May-11 48 (5) 62 (4)

loaded into the trickle flow cell and the flood and drain cells. However, the average VS of sub-8 cm samples loaded to the flood and drain cells was within 4% of that loaded into Cell 1, well within the sample variation observed within each cell (Table 2). The VS of the sub-8 cm fraction was within the range observed by others for source separated OFMSW, suggesting that mechanical processing produced a feedstock enriched towards organic material. Bolzonella et al., 2006 found the VS for source separated OFMSW to be within the range of 55–90%. The VS of source separated rural waste used by De Baere (2000), comprised of 10% food waste, 75% garden waste and 15% paper on a wet basis, was 55% (w/w), close to the VS content of the sub-8 cm waste used in this study, although 5% (dry basis) of the sub-8 cm fraction in this current study was plastic according to visual classification (Table 2). The start-up of Cells1 and 2 was achieved by using mature landfill leachate. Cell 3 was inoculated with enriched leachate generated in Cell 2, as described below. The COD and pH of the landfill leachate added throughout the operation of Cell 1 was 3200 ± 730 mg L
1 with a well buffered pH of 7.4 ± 0.1. In contrast, the COD and pH of the landfill leachate used to fill the reservoir at the start of the flood and drain trails was 1360 mg L
1 and 7.0 respectively. 2.2. Cell design Cell 1 was designed to operate as a leach-bed (Fig. 1). The cell was built in-ground. The excavation was 76 m  21 m at the ground surface and approximately 2.5 m deep, with all walls sloping at a grade of 3:1 (horizontal:vertical). The excavation was lined with high density poly-ethylene (HDPE) sheets, which were welded as required to create a continuous liner that extended approximately 2 m beyond the periphery of the excavation. The liner was anchored by compacting soil over this skirt. Geofabric was placed over the HDPE liner, to protect the liner when filling the cell with waste. A 300 mm layer of 20 mm gravel was placed in bottom of the cell, as a leachate collection sump. Collected leachate was cleared daily using a 200 mm diameter submersible pump positioned in a pit in the gravel layer and maintained via a 250 mm access tube. Leachate was pumped to polyethylene tanks at ground level with a total holding capacity of 60 m3. Landfill leachate was added as required to maintain a leachate inventory of 60 m3. The sub-8 cm waste was dropped into the cell from the outlet conveyor of the trommel. The waste was distributed across the cell

915 75 (12) 58 (13) 223 (16)

92 (2) 123 (15) 211 (38) 13-Aug-13 44 (9) 7 (1)

1-Jun-13 48 (5) 66 (9)

using a long armed excavator such that the final waste profile was mounded slightly (approximately 300 mm) to enhance rainfall runoff. A total of 1974t of sub-8 cm waste was loaded into Cell 1 (Table 2) over a period of 4 weeks to achieve an average bulk density of 0.87 t m
3 on a wet basis. This is within the range of bulk densities found by Reddy et al. (2009) for shredded MSW (