Environmental Geotechnics Environmental Geotechnics Issue EGXXXX

Volume XXXX Issue EGXXXX Pages 1–XX http://dx.doi.org/10.1680/jenge.15.00078

Paper 15.00078
Applying MICP by denitrification in soils: Received 16/10/2015; accepted 05/10/2016
Keywords: calcium carbonate/geochemistry/strength & testing of materials
a process analysis
Pham, Nakano, van der Star, Heimovaara and ICE Publishing: All rights reserved
van Paassen

1
2 Applying MICP by 1
2
3 3
4
5
denitrification in soils: 4
5
6
7
a process analysis 6
7
8 Q1 1 Vinh P. Pham Ir. 3 Wouter R. L. van der Star Dr. Ir. 8
9 Q2 PhD Student, Department of Geoscience and Engineering, Delft Researcher, Subsurface and Groundwater Systems, Deltares, Delft,
9
University of Technology, Delft, Netherlands (corresponding author: Netherlands
10 p.v.pham-1@tudelft.nl) 10
4 Timo J. Heimovaara Dr. Ir.
11 2 Akiko Nakano Dr. Professor, Department of Geoscience and Engineering, Delft University 11
12 Assistant Professor, Faculty of Agriculture, Kyushu University, Fukuoka, of Technology, Delft, Netherlands 12
Japan 5 Leon A. van Paassen Dr. Ir.
13 13
Assistant Professor, Department of Geoscience and Engineering, Delft
14 14
University of Technology, Delft, Netherlands
15 15
16 16
17 17
18 18
19 19
20 20
21 1 2 3 4 5
21
22 22
23 23
24 In this study, the process of microbially induced carbonate precipitation by denitrification was investigated in 24
25 relation to its potential use as ground improvement method. Liquid batch experiments indicated that the substrate 25
26 26
solution had an optimum carbon–nitrogen ratio of 1·6 and confirmed that combining nitrate reduction and calcium
27 27
carbonate precipitation leads to an efficient conversion, at which the pH is buffered slightly below 7 and
28 28
29 accumulation of toxic intermediate nitrogen compounds is limited. Sand column experiments confirmed that the 29
30 volume and distribution of the gas phase strongly depend on the stress conditions. The produced gas volume is 30
31 inversely related to the pore pressure and can be predicted based on a mass balance analysis, assuming conservation 31
32 of mass and using theoretical laws of physics. At low pore pressure, the gas formed and accumulated at the top of 32
33 33
the column, whereas the calcium carbonate precipitation occurred mostly at the bottom near the substrate inlet; an
34 34
excess amount of gas was produced, which vented from the sand columns and induced cracks in the sand at low
35 35
36 confining pressures, which negatively affected the sand-stabilising effect of the calcium carbonate minerals. 36
37 37
Introduction removal of ammonium chloride, which is formed as a by-product
38 38
Microbially induced carbonate precipitation (MICP) has attracted of the process. MICP by denitrification has been considered as a
39 39
increased interest in recent years for its potential in geotechnical potential alternative biomediated ground improvement process, as
40 40
and environmental applications (DeJong et al., 2013). Most studies it has several advantages over urea hydrolysis (Karatas et al.,
41 41
on MICP were based on the hydrolysis of urea. Hydrolysis of urea 2008; Kavazanjian et al., 2015; van Paassen et al., 2010b). When
42 42
is one of the first processes associated with MICP in the late nitrate (NO3−) is completely reduced to nitrogen gas, no adverse
43 43
nineteenth century (Ehrlich and Newman, 1996). Due to the by-products are formed and removal effort is not required. The
44 44
limited abundance of urea in nature, MICP by urea is now required substrates for denitrification are sufficiently soluble to
45 45
considered to be one of the least important sources of naturally limit the number of injections to reach the target amount of
46 46
occurring biogenic carbonate. However, it has been demonstrated calcium carbonate (CaCO3). Substrates can even be produced from
47 47
at laboratory scale (Chu et al., 2012; DeJong et al., 2006; Harkes waste streams, and the conversion does not require cultivation of
48 48
Q3 et al., 2010; Montoya et al., 2013; Whiffin et al., 2007) and field very specific organisms. In fact, indigenous populations of
49 49
scale (Burbank et al., 2011; DeJong et al., 2009; van Paassen, denitrifying bacteria can be used (Martin et al., 2013; van der Star
50 50
2011; van Paassen et al., 2010a) that MICP by urea hydrolysis can et al., 2009; van Paassen, 2009a; van Paassen et al., 2010b). Q4
51 51
significantly strengthen granular soils and shows a wide range of
52 52
potential applications (Phillips et al., 2013). Commercial Denitrification is one of the main biological processes in the
53 53
applications of MICP by urea hydrolysis are still limited, partly global nitrogen cycle, in which a genetically diverse group of
54 54
due to the costs of cultivation of ureolytic bacteria and the required microorganisms is able to reduce nitrate through the intermediate
55 55

envgeo1500078.pdf 1 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 products of nitrite (NO2−), nitric oxide (NO) and nitrous oxide For zero growth conditions, the overall metabolic stoichiometry 1
2 (N2O) to nitrogen (N2) gas (Robertson and Groffman, 2015). is equal to catabolic reaction I. When following this approach, 2
3 Denitrification has been intensively studied since the late the actual metabolic stoichiometry ranges between reactions I 3
4 nineteenth century both in natural systems and for industrial and III. In other words, the amount of required acetate at zero 4
5 applications (Archna and Sobti, 2012; Ferguson, 1994; Knowles, growth is 0·6 mol/mol nitrate, while the acetate-to-nitrate (Ac/N) 5
6 1982; Kuenen and Robertson, 1988; Payne, 1981; Soares, 2000; ratio at maximum growth is 1·25 (corresponding to carbon-to- 6
7 Voorhees, 1902; Wang et al., 1995). nitrogen (C/N) ratio range of 1·2–2·5). Similarly, the amount of 7
8 the produced nitrogen gas ranges between 0·4 and 0·5 mol/mol 8
9 The stoichiometry of the metabolic reactions involved in nitrate and the inorganic carbon (carbonic acid (H2CO3), 9
10 denitrification can be calculated using a method suggested by bicarbonate (HCO3−), carbon dioxide (CO2)) ranges from 1·25 to 10
11 Heijnen et al. (1992) and Heijnen and Kleerebezem (2010). 1·45 mol/mol nitrate. 11
12 Metabolic reactions can be divided in an anabolic reaction, which 12
13 describes the production of biomass, and a catabolic reaction which If calcium ions (Ca2+) are present, the inorganic carbon produced 13
14 generates the energy for the cells to produce new biomass in the during the metabolism of the denitrifiers can precipitate to form 14
15 anabolic reaction (Haynie, 2008). The stoichiometry of the anabolic calcium carbonate 15
16 and catabolic redox reactions are determined separately by solving 16
17 the mass and electron balance for each reaction. The ratio between IV. Ca2þ þ HCO3− → CaCO3 þ Hþ 17
18 the catabolic and anabolic reactions is determined by solving the 18
19 energy balance – that is, the produced energy from the catabolic 19
20 reaction is equal to the energy required for biomass production and At neutral pH, precipitation of calcium carbonate produces 20
21 cell maintenance. The actual ratio depends on the growth rate of acid, which buffers the alkalinity production associated with 21
22 the microorganisms, which can range from maximum growth denitrification. The complete set of coupled reactions related to 22
23 conditions, where microorganisms grow exponentially, to zero MICP by denitrification, together with the potential consequences 23
24 growth conditions, where the total amount of microorganisms does for the physical behaviour of granular soils, is illustrated in 24
25 not increase, but maintaining the population still requires energy. Figure 1. MICP by denitrification also results in gas production 25
26 The actual growth rate is controlled by the process and and biomass growth. Biofilms can also increase strength and 26
27 environmental conditions such as the availability of substrates and stiffness of soils and reduce hydraulic conductivity (e.g. Stal, 27
28 nutrients and the presence of inhibiting compounds (the condition 2010). Biogenic gas production in soil displaces water, which 28
29 of negative growth rate or cell decay, e.g. in the case where there may decrease the hydraulic conductivity. The presence of gas in 29
30 are no substrates available is not considered here). the pores can also increase shear strength by generating capillary 30
31 suction forces (Fredlund and Rahardjo, 1993) and increase the 31
32 Using acetate (C2H3O2−) as the electron and carbon donor and resistance of the soil to cyclic loading (He et al., 2013; 32
33 nitrate as electron acceptor leads to the following reaction Kavazanjian et al., 2015; Yegian et al., 2007). Unsaturated 33
34 stoichiometry for the catabolic reaction conditions may also improve the strengthening effect of MICP, as 34
35 the precipitation is considered to take place more efficiently 35
36 at the particle contacts (Cheng and Cord-Ruwisch, 2012; 36
C2 H3 O2− þ 1   6NO3− þ 0   6Hþ Cheng et al., 2013).
37 37
I. → 0   8N2 þ 2HCO3− þ 0   8H2 O
38 38
39 Consequently, MICP by denitrification is considered to have 39
40 significant potential as a ground improvement method, through 40
Taking CH1·8O0·5N0·2 as a representative molecular formula for
41 biogenic gas production and MICP (Hamdan et al., 2016; 41
bacterial biomass, the anabolic reaction equals
42 Kavazanjian et al., 2015). Still, the demonstrated reaction rate of 42
43 this process is low compared with that of urea hydrolysis. Selecting 43
44 0   725C2 H3 O2− þ 0   2NO3− þ 0   475Hþ the right substrate composition is essential. Too much nitrate may 44
45 II. → CH1  8 O0  5 N0  2 þ 0   45HCO3− þ 0   2H2 O lead to accumulation of intermediate compounds, which can be 45
46 toxic for the bacteria and inhibit growth (Almeida et al., 1995) or 46
47 increased emissions of nitrous oxide, which is a very strong 47
48 Solving the energy balance for maximum growth the stoichiometry greenhouse gas (GHG) (Chung and Chung, 2000; Hanaki et al., 48
49 of the metabolic reaction (pH 7, temperature 298 K) becomes 1992), while leaving a large excess of acetate would be inefficient. 49
50 Finally, the formation, distribution and mobility of gas in porous 50
51
1   21C2 H3 O2− þ 0   97NO3− þ 0   76Hþ media strongly depend on the soil characteristics and environmental 51
52 conditions. Pore pressure in the underground strongly affects the 52
53 → CH1  8 O0  5 N0  2 þ 1   41HCO3− þ 0   39N2 þ 0   59H2 O produced gas volume and solubility (e.g. Rebata-Landa and 53
54 Santamarina, 2011), and the accumulation and storage of gas is 54
III.
55 affected by the substrate supply regime (Soares et al., 1991). If too 55

envgeo1500078.pdf 2 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 1
Representative elementary volume
2 Gas phase 2
N2 CO 2 Reduces permeability
3 3
and water saturation,
4 increases the pore 4
CO 2 Acid-base
5 Liquid phase compressibility 5
H+ equilibrium
6 6
7 aNO3– + bC2H3O2– + cH + → biomass + dN2 + eHCO3– + fH2O 7
8 Denitrification
8
9 9
10 H+ 10
11 11
CO32– + Ca2+ ↔ CaCO3
12 12
13 13
14 Soil particles Biofilm CaCO3 14
Precipitation
15 Solid phase 15
16 16
17 Clogging, contributes to cohesion Increases cohesion and friction, reduces pore size 17
18 18
19 Figure 1. MICP induced by denitrification and potential effects in 19
20 a representative elementary volume of soil or a soil specimen. 20
21 Detailed stoichiometry of the denitrification reaction is given in the 21
22 text 22
23 23
24 24
25 much gas is produced, it could trigger liquefaction on very loose Liquid batch experiments 25
26 sand instead of mitigating it (Grozic et al., 1998). As inoculum for the first batch experiment, a soil sample was 26
27 taken from the Botanical Garden of the Delft University of 27
28 For these reasons, the authors present in this paper the results of Technology. Denitrifying organisms typically occur in soils, 28
29 three sets of experiments which were executed to evaluate the which are wet and anaerobic and have sufficiently high organic 29
30 process performance of MICP through denitrification for a content and a source of nitrate, such as wetlands or agricultural 30
31 selection of process variables and test conditions related to its fields (Keddy, 2000). The sample was taken from a depth of 2 m 31
32 application as ground improvement method by using a hand auger. The black colour and wet condition of the 32
33 sample indicated that favourable conditions for denitrifiers were 33
34 (a) batch experiments in a mixed liquid culture in which substrate met. 34
35 composition was varied to study the preferable substrate ratio 35
36 and understand the process mechanisms The experiments were performed in 250-ml glass bottles (Duran 36
37 (b) sand column experiments in a triaxial cell, operated at GLS80). Electrodes were inserted through the cap to measure the 37
38 different pore and confining pressure conditions, to study the pH and electrical conductivity (CE), and tubes were connected to 38
39 Q5 gas formation at different water levels and soil depths the sample liquid to monitor the gas volume. In the first 39
40 (c) sand column experiments at ambient pressure and low incubation, 30 g of soil was suspended in 250 ml of growth 40
41 confinement with multiple sequential batches of substrates to medium. After 3 d of incubation, all nitrate and nitrite were 41
42 investigate the precipitation efficiency of the treatment. consumed and the bottle was shaken to stimulate the detachment 42
43 of microorganisms from the soil particles. Afterwards the liquid 43
44 containing free denitrifying cells (without the solid fraction which 44
Materials and methods
45 had been allowed to settle) was collected and transferred into a 45
46 Substrate solutions new bottle for further experiments. All five subsequent 46
47 The substrate solutions which were used throughout the experiments incubations used the inoculum from the previous incubation with 47
48 contained calcium acetate (Ca(C2H3O2)2) and calcium nitrate the concentrations as shown in Figure 2. Substrates which were 48
49 (Ca(NO3)2) (Sigma-Aldrich) at varying concentrations (Figure 2). used in these batches were 30 mM calcium acetate and 25 mM 49
50 Besides those main substrates, all media further contained the calcium nitrate, corresponding to calcium, acetate and nitrate 50
51 following nutrients: 0·003 mM ammonium sulfate ((NH4)2SO4), (Ca–Ac–N) concentrations of 55, 60 and 50 mM respectively 51
52 0·0024 mM magnesium sulfate (MgSO4), 0·006 mM monopotassium (55:60:50). In the fifth batch, calcium salts in one of the two 52
53 phosphate (KH2PO4), 0·014 mM dipotassium phosphate (K2HPO4) incubations were replaced by sodium salts (110:60:50 53
54 and 1 ml/l trace element solution SL12B (Overmann et al., 1992) to sodium–acetate–nitrate (Na–Ac–N)). In the sixth batch, the 54
55 avoid nutrient limitation during bacterial growth. acetate-to-nitrate ratio was varied. Explicit concentrations of 55

envgeo1500078.pdf 3 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 Ca–Ac–N 55:60:50: Ca(C2H3O2)2 30 mM – Ca(NO3)2 25 mM [Ac/N = 1·2] 1

Ca–Ac–N 45:40:50: Ca(C2H3O2)2 20 mM – Ca(NO3)2 25 mM [Ac/N = 0·8]
2 Ca–Ac–N 40:30:50: Sixth incubation:
2
Ca(C2H3O2)2 15 mM – Ca(NO3)2 25 mM [Ac/N = 0·6]
3 Ca–Ac–N 110:120:100: Ca(C2H3O2)2 60 mM – Ca(NO3)2 50 mM [Ac/N = 1·2] Previous suspension 3
4 Na–Ac–N 110:60:50: NaC2H3O2 60 mM – NaNO3 50 mM [Ac/N = 1·2] 0·25 l/l, 4
All the suspensions contained the same concentrations of micronutrients and Ca–Ac–N 40:30:50.
5 trace elements. 5
6 Fifth incubation: Sixth incubation: 6
7 Previous suspension Previous suspension 7
0·25 l/l, 0·25 l/l,
8 Na–Ac–N 110:60:50. Ca–Ac–N 45:40:50.
8
9 9
10 First incubation: Second to fourth Fifth incubation: Sixth incubation: 10
11 30 g soil in 250 ml incubation: Previous Previous suspension Previous suspension 11
liquid. suspension 0·5 l/l, 0·25 l/l, 0·25 l/l,
12 Ca–Ac–N 55:60:50. Ca–Ac–N 55:60:50. Ca–Ac–N 55:60:50. Ca–Ac–N 55:60:50.
12
13 13
Experiments in the liquid environment
14 14
15 Experiments in the sand columns 15
16 16
17 Incubation of activated Sand column experiment at ambient Triaxial tests, 17
sludge, pressure, using growth medium Previous suspension
18 Ca–Ac–N 110:120:100. First to third batch Ca–Ac–N 130:160:100 0·25l/l,
18
19 Fourth to ninth batch Ca–Ac–N 120:130:110 Ca–Ac–N 55:60:50. 19
20 At ambient pressure With pressure control 20
21 21
22 22
Figure 2. Schematic overview of the sequence of liquid batch
23 23
experiments, ambient pressure column experiment and triaxial cell
24 24
experiments. Arrows illustrate the sequence of incubations and
25 25
indicate the source of the inoculum for the test
26 26
27 27
28 substrates and inoculums of these experiments are presented in gas volume by using the ideal gas law at 1 atm and 298 K, 28
29 Figure 2. assuming nitrogen gas was the only produced gas. 29
30 30
31 Liquid from the bottles was regularly sampled to analyse the Sand column experiments in a triaxial cell at varying 31
32 concentration of different solutes, the electrical conductivity (CE) pressure conditions 32
33 and the pH. All liquid samples were filtered through a 0·45-µm Sand column experiments were performed at varying pressure 33
34 membrane and subsequently diluted to the measureable conditions, using a triaxial test set-up described in the ISO/TS 34
35 range before measurements. Nitrate, nitrite and calcium were 17892-9:2004 standard (CEN, 2004). A uniform fine-grained 35
36 determined spectrophotometrically (Lasa 100, Hach Lange) with siliceous sand (0·125–0·355 mm) was used (Sibelco, S60) with a 36
37 standard test kits LCK339, LCK341 and LCK327 respectively). silica content of 99·5% and a specific gravity of 2·65 g/cm3. The 37
38 Total carbon (CT) was measured with a Dohrmann DC190 minimum and maximum dry bulk densities (ASTM D 4253 and Q6 38
39 chromatograph. Acetate consumption was determined indirectly ASTM D 4254) range from 1·57 ± 0·01 to 1·66 ± 0·02 g/cm3. 39
40 with total organic carbon (TOC), which is the result from the CT Sand columns were prepared using a split mould with a rubber 40
41 measurements but pre-treated with hydrochloric acid (HCl) to membrane inside. The mould was filled in three layers. For each 41
42 remove the inorganic carbon (Schumacher, 2002). All incubations layer, first, the suspension containing the bacteria and substrates 42
43 were performed in a climate room at 25°C. The pH and CE were was poured into the mould, followed by dry sand, ensuring that 43
44 measured on all samples and continuously in the incubating the sand always remained below the fluid level. The sample was 44
45 bottles. not tamped, thus resulting in a very loosely packed sand with an 45
46 average dry bulk density of 1·50 ± 0·1 g/cm3. 46
47 The gas produced during the experiment was captured using a 47
48 water clock, made by placing a graduated cylinder upside down in The suspension was prepared by mixing a 0·25 l/l inoculum 48
49 a water bath. Before starting the incubation, nitrogen gas was harvested at the end of the fifth incubation of the liquid batch 49
50 flushed through the bottle to ensure anoxic conditions and the experiments, with concentrated substrate solution and demineralised 50
51 partial pressure of nitrogen in the gas phase was set at 100%. The water. The overall substrate concentration of the suspension was 51
52 pressure of the gas was kept constant at ambient pressure by 30 mM calcium acetate and 25 mM calcium nitrate (Ca–Ac–N 52
53 manually adjusting the height of the cylinder. Liquid sampling 55:60:50). This corresponds to an Ac–N ratio of 1·2, which was 53
54 and gas volume measurements were done on a daily basis. The selected to ensure that there would be a sufficient carbon source in 54
55 amount of nitrogen gas (NN2) was calculated from the measured case the conditions in the sand would be favourable for bacterial 55

envgeo1500078.pdf 4 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 growth, changing the metabolic stoichiometry towards maximum expelled and flushed liquids to determine the concentrations of 1
2 growth. nitrate, nitrite and calcium ions and CT. During the first 2
3 experiment, biomass growth was observed in the cell water. 3
4 The tests were performed at three different pressure conditions. Consequently, for the other two tests also the volume and 4
5 First, two back pressure conditions were selected to evaluate the concentrations in the cell water were determined. 5
6 effect of pore pressure on gas volume, while the effective 6
7 confining pressure was kept constant: a low back pressure of Sand column experiments with multiple batches 7
8 50 kPa with a cell pressure of 100 kPa and a high back pressure of of substrate solution 8
9 300 kPa with a cell pressure of 350 kPa. Next, a third test was Sand column experiments were performed at ambient pressure 9
10 performed in which the back pressure was set at 345 kPa, while conditions in which multiple batches of substrate solution were 10
11 the cell pressure was kept at 350 kPa to evaluate the process at a flushed through the columns for the total period of 65 d. 11
12 low confinement pressure. A control test was performed at a back 12
13 pressure of 50 kPa and a cell pressure of 100 kPa by using the The sand columns were prepared using plastic (polyvinyl 13
14 growth medium without bacteria. chloride (PVC)) columns with a length of 180 mm and inner 14
15 diameter of 65 mm as described by Harkes et al. (2010). 15
16 Once the cell and back pressures were set at the designated stress The column was filled with a uniform fine siliceous sand (d50: 16
17 levels, the sample was left to react while the volume change in 0·166 mm; Itterbeck fine, Smals IKW, SZI 0032). The top and 17
18 the controllers was monitored. The pore pressure coefficient (or B bottom of the column were filled with 1 cm of fine gravel (d50: 18
19 factor), which is the ratio of the change in pore pressure over the 2·5 mm). The sand and gravel were packed by tamping under 19
20 change in cell pressure and which is related to the degree of water water to an average dry density of 1·57 g/cm3. Before treatment, 20
21 saturation of the sample (Skempton, 1954), was measured at the the sand column was flushed with several pore volumes of 21
22 beginning and at the end of the treatment. At the end of each distilled water. 22
23 experiment, the sample was flushed from bottom to top with 23
24 deaired water by using a third pressure controller while the back The inoculum for this experiment was obtained from activated 24
25 pressure was lowered to obtain a constant head difference of sludge collected at the Harnaschpolder municipal waste water 25
26 20 kPa. While flushing, the flow rate was measured to calculate treatment plant (Delft, Netherlands). The sludge was suspended in 26
27 the hydraulic conductivity of the sample. substrate solution containing 60 mM calcium acetate and 50 mM 27
28 calcium nitrate (110:120:100 Ca–Ac–N), and the suspension 28
29 After flushing the high-pressure samples, the volume of the was incubated for 6 d, in which nitrate was fully consumed. The 29
30 expelled and flushed liquids was determined by emptying the suspension containing the biomass, without the sludge (which 30
31 back pressure controller. For the low-pressure conditions, the back was allowed to settle), was used as the inoculum for the column 31
32 pressure controller was emptied before flushing the sample to set-up. For the column experiment, 250 ml of the suspension 32
33 collect the ‘expelled’ water and after flushing it was emptied containing the inoculum 0·5 l/l was mixed with substrate 33
34 again to collect the ‘flushed’ water. Samples were taken from the solutions, reaching concentrations as indicated in Figure 2, and 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 Collect every 46
Pump 5–10 min
47 for analysis
47
48 48
Balance
49 49
50 50
Flushing under constant head Gas and liquid collection between two
51 flushing times
51
52 52
53 Figure 3. Set-up of the sand column with multiple substrate 53
54 flushes at ambient pressure 54
55 55

envgeo1500078.pdf 5 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 Ca-based Na-based 1
2 140 140 2
3 120 120 3
Concentration: mM

4 100 100
4
5 5
80 80
6 6
7 60 60 7
8 40 40 CT 8
9 20 20 [Ca2 + ] 9
10 10
0 0
11 0 5 10 15 20 25 0 5 10 15 20 25 11
12 (a) (d) 12
13 NTotal 13
80 80 NNO3–
14 14
NNO2–
15 15
Concentration: mM

60 60 NN2
16 16
17 17
18 40 40 18
19 19
20 20 20 20
21 21
22 0 0 22
23 0 5 10 15 20 25 0 5 10 15 20 25 23
(b) (e)
24 24
10 10
25 25
26 26
8 8
pH and CE: mS/cm

27 pH 27
28 CE 28
6 6
29 29
30 4 4 30
31 31
32 2 2 32
33 33
0 5 10 15 20 25 0 5 10 15 20 25
34 34
Time: d Time: d
35 35
(c) (f)
36 36
37 Q7 Figure 4. Comparison of changes in chemical composition during 37
38 denitrification in the presence (left column) and absence (right 38
39 column) of calcium ions during the fifth incubation. Legends 39
40 indicate chemical compounds (CT is total dissolved carbon) 40
41 41
42 42
43 flushed into the column. Subsequently, the column was flushed During each flush, the flow rate was determined by collecting the 43
44 nine times with 250 ml of substrate solution, after which the effluent from the top of the set-up at regular time intervals and 44
45 column was flushed with one pore volume of distilled water to the hydraulic conductivity was determined using Darcy’s law. The 45
46 wash the remaining solutes from the column. For the first to pH, CE and solute concentrations of these samples were 46
47 third flushes, the acetate-to-nitrate ratio was kept high (about 1·6) measured. After each flush, the inlet at the bottom was closed and 47
48 in order to stimulate bacterial growth and prevent accumulation the column was left for 7–9 d to react. During this period, the 48
49 of toxic intermediate nitrogen compounds. For all subsequent produced gas and the expelled liquid were collected. The mass of 49
50 flushes, the acetate-to-nitrate ratio was about 1·2, which the expelled liquid was measured after the reaction period before 50
51 corresponds to the reaction stoichiometry at maximum growth the next flush. After treatment, the sand column was first analysed 51
52 (reaction III). All liquids were flushed in upward flow under a using X-ray computed tomography (CT) scanning, then the PVC 52
53 constant head difference of 20 cm, and the effluent replaced by tube was cut open over the length of the sample and removed. 53
54 newly flushed medium was collected from the top during each The sample was sliced into nine parts and dried in an oven at 54
55 flush. The set-up is presented in Figure 3. 105°C. The dry samples before and after treatment were analysed 55

envgeo1500078.pdf 6 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 with an environmental scanning electron microscope (ESEM) Figure 5 shows the results of the incubations where the acetate-to- 1
2 (Philips ESEM XL30). The calcium carbonate concentration was nitrate ratio was varied. The experiment using the growth medium 2
3 determined for 1·5–2·0 g of the dried samples at different with an Ac–N ratio of 1·2, which is close to the reaction 3
4 locations in the column by using the acid dissolution method stoichiometry for maximum growth of the bacteria (reaction III), 4
5 described by Whiffin et al. (2007). resulted in an excess of calcium and acetate at the end of the 5
6 experiment. The Ac–N ratio of 0·6, which corresponds to the 6
7 stoichiometry of the catabolic reaction (zero growth; reaction I), 7
Results
8 resulted in an excess of nitrate and accumulation of nitrite. For the 8
9 Liquid batch experiments Ac–N ratio of 0·8, all substrates are converted most efficiently. 9
10 Figure 4 shows the difference between incubation with calcium- 10
11 based substrates and sodium-based substrates. The reduction in Triaxial tests at different pressure conditions 11
12 CT and calcium ion concentrations in Figure 4(a) and the drop in The volume change in the pressure controllers during the 12
13 CE in Figure 4(c) indicate that calcium carbonate precipitation incubation experiments in the triaxial cell is shown in Figure 6. 13
14 took place. Due to the precipitation of carbonate, the pH is The total volume change is the sum of the volume change in the 14
15 buffered between 6 and 7, which reduces the accumulation of cell and back pressure controllers and represents the change in the 15
16 nitrite to a maximum level of 10 mM during the first week, sample volume. All tests showed that the volume in the back 16
17 compared to 30 mM of nitrite for the sodium-based substrate pressure controller increased rapidly during the first 3 d. At the 17
18 solution. In the sodium-based system (Figure 4(d)), no same time, however, the volume of the cell pressure controller 18
19 precipitation took place; hence, CT remains constant. The small decreased in all experiments. As a result, the sample volume 19
20 amount of calcium ions was due to the small fraction which was decreased before it started to rise. After 4 d, the volume in the 20
21 still present in the inoculum obtained from the fourth incubation. back pressure controller of the experiment with a back pressure of 21
22 Due to MICP, this amount was quickly depleted. The decreases in 50 kPa continued to rise slowly, while for the two tests with 22
23 nitrate and nitrite and the increase in nitrogen gas (Figures 4(b) higher back pressures, the volume started to decrease from the 23
24 and 4(e)) are evidence that denitrification still took place in both third day onwards, resulting in a significantly lower volume 24
25 systems. In the absence of calcium ions, denitrification leads to an change at the end of the experiment. The change in sample 25
26 alkaline environment, as shown in Figure 4(f), where the pH rose volume even became negative at the end of the experiment. The 26
27 within 3 d to a level between 9 and 9·5. resulting volume changes and B factors are presented in Table 1. 27
28 28
29 29
30 30
31 Ac–N = 1·2 Ac–N = 0·8 Ac–N = 0·6 31
32 32
120 120 120
Concentration: mM

33 TOC 33
34 90 90 90 [Ca2+] 34
35 35
60 60 60
36 36
37 30 30 30 37
38 38
0 0 0
39 0 5 10 15 20 25 30 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40 39
40 NTotal 40
41 80 80 80 NNO3– 41
42 42
Concentration: mM

NNO2–
43 60 60 60 43
NN2
44 44
40 40 40
45 45
46 20 20 20 46
47 47
48 0 0 0 48
0 5 10 15 20 25 30 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
49 Time: d Time: d Time: d 49
50 50
51 Figure 5. Comparison of changes in chemical composition during 45:40:50, Ac–N = 0·8; right column: Ca–Ac–N = 40:30:50, 51
52 denitrification in the presence of calcium at three different Ac–N = 0·6. Legends indicate chemical compounds (TOC is total 52
53 acetate–nitrate ratios in the sixth incubation. Left column: organic carbon) 53
54 Ca–Ac–N = 55:60:50, Ac–N = 1·2; middle column: Ca–Ac–N = 54
55 55

envgeo1500078.pdf 7 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 50–100 kPa a relatively constant value of 4–5 mS/cm over the height of the 1
2 60 column, indicating that there were no differences in solute 2
Volume change: ml

3 concentration in the whole column. From the fourth flush onwards, 3

40
4 the CE profiles started to show a decrease from top to bottom, 4
5 reaching maximum values ranging from about 8 mS/cm at the top 5
20
6 to about 3 mS/cm at the bottom, which indicates that during the 6
7 final flushes, a large part of the substrates in the upper half of the 7
0
8 column was not yet converted when the next flush was applied. 8
9 0 5 10 15 20 9
10 Time: d The solute concentrations show a similar distribution. Nitrate was 10
11 (a) almost completely consumed over the full height of the column in 11
300–350 kPa
12 10 the first four flushes, as only small concentrations of nitrite 12
13 (<1·3 mM) was measured and about 20 mM calcium was still 13
Volume change: ml

14 5 present. The gradient of the substrate distribution in the column 14

15 was very clear in the last two flushes. During these flushes, the 15
0
16 remaining nitrate at the top of the column was 20–30 mM 16
17 –5 (indicating a conversion of only about 50%), while at the bottom it 17
18 was completely consumed. The calcium and acetate concentrations 18
19 –10 showed a similar distribution. Incomplete nitrate conversion in the 19
0 5 10 15 20
20 last flushes caused a considerable amount of nitrite accumulation. 20
Time: d
21 (b)
21
22 345–350 kPa During the reaction period between each flush, 31–53 ml of fluid 22
Total volume
23 10 was expelled from the pores due to gas production inside the 23
Back volume
Volume change: ml

24 Cell volume column. The hydraulic conductivity, which was calculated from the 24
5
25 measured flow rate during flushing, showed a significant difference 25
26 0 between the first flush and all other flushes. In the second and later 26
27 flushes, initially no liquid flowed out of the column. The injected 27
28 –5 liquid first filled up the unsaturated pore space before flowing out 28
29 of the column. During each flush, the flow rate gradually 29
–10
30 0 5 10 15 20 25 increased. The hydraulic conductivity at the end of the second and 30
31 Time: d all later flushes was 50–70% lower that than during the first flush. 31
(c)
32 The increase in hydraulic conductivity can be explained by an 32
33 increase in saturation as the trapped gas was gradually flushed out 33
Q8 Figure 6. Volume changes in the triaxial tests at cell pressure –
34 together with the liquid (increasing the relative permeability). 34
back pressure values of (a) 50–100, (b) 300–350 and
35 35
(c) 345–350 kPa. The total volume change is the difference
36 The images produced by X-ray CT scanning (Figure 8) show that 36
between the measured volume changes in the pressure cells
37 the gas produced during the experiment led to the development of 37
38 cracks and open voids, which remained present after the column 38
39 The volumes and solute concentrations of the different liquid was flushed with water. The constructed 3D image in Figure 8(b) 39
40 fractions are presented in Table 2. At low pressure, the back shows that the air was mostly present in the pores in the gravel 40
41 pressure controller indicated that 55 ml was expelled from the layers and in the voids, which were developed mostly at the top 41
42 sand column. However, the controller contained only 35 ml of of the column by the air bubbles themselves. However, as the 42
43 liquid. Similarly, during flushing, 212 ml was flushed out, of resolution of the CT scan is limited to approximately 200 mm, 43
44 which only 207 ml was collected. It was assumed that the rest of small gas bubbles could not be visualised. 44
45 the controller volume was filled with gas. 45
46 Calcium carbonate measurements showed an average calcium 46
47 Sand column experiment with multiple substrate flushes carbonate content of about 1·1%. However, as shown in Figure 9, 47
48 at ambient pressure the calcium carbonate was not homogeneously distributed 48
49 The breakthrough curves of CE, pH and solute concentrations, throughout the column. The distribution of calcium carbonate is 49
50 which were measured during each flush, are shown in Figure 7. largest at the bottom of the sand layer and decreased to the top. 50
51 The initial part of each breakthrough curve can be considered as a Both gravel layers at the top and the bottom and the upper part of 51
52 vertical profile from the top to the bottom through the sand the column had a relatively low calcium carbonate content. The 52
53 column. The flush was continued until complete breakthrough, presence of calcium carbonate was also confirmed by ESEM 53
54 indicated by sharp increases in the CE and substrate concentrations. analysis (Figure 10). The crystals showed a dendritic texture with 54
55 In the first two flushes of the fresh medium, the CE profile showed a size of up to 200 mm. 55

envgeo1500078.pdf 8 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 1
2 Experiment Controller volume: ml B factor Hydraulic conductivity: 2
3 10−5 m/s 3
Day 3 Day 15 Initial End
4 4
5 5
Back Cell Total Back Cell Total
6 6
7 Control: 50–100 kPa 0 −2·2 −2·2 1·3 −4·5 −3·2 0·98 0·98 5·9 7
8 300–350 kPa 6·4 −3·0 3·4 3·7 −5·5 −1·8 0·98 0·60 5·9 8
9 345–350 kPa 7·4 −1·6 5·8 1·4 −3·5 −2·1 0·96 0·86 5·7 9
10 50–100 kPa 42 −5·5 36·5 50 −7·9 42·1 0·58 0·04 5·8 10
11 11
12 12
13
Q9 Table 1. Volume changes, B factors and hydraulic conductivities at
13
the end of the reaction in the triaxial tests
14 14
15 15
16 16
Discussion sum of all nitrogen compounds (Ntotal) should be constant and equal
17 17
In order to assess the relevance of these results for ground to the initial nitrate concentration. Nevertheless, during the liquid
18 18
improvement applications, first, the accuracy of the experiments batch experiments, the sum of nitrogen compounds ranged between
19 19
are discussed by analysing the mass balance and, second, the 40 and 70 mM, whereas the initial nitrate concentration was 55 mM
20 20
effects of different process variables and test conditions on the in each of the experiments, which corresponds to a deficiency in the
21 21
process performance are evaluated. mass balance of ±25%.
22 22
23 23
Mass balance For the triaxial tests, the total amounts of nitrate and nitrite were
24 24
Concentration measurements during and after the reaction in each of calculated by multiplying the measured concentration in each of
25 25
the experiments are used to evaluate the mass balance of the the liquid fractions with the liquid volume. The total amount of
26 26
different nitrogen compounds. Figures 4 and 5 show the measured nitrogen gas was calculated using the ideal gas law, assuming that
27 27
concentration of the different nitrogen compounds. The nitrogen gas the volume of nitrogen gas is equal to the maximum volume
28 28
concentration is calculated using the ideal gas law assuming ambient change in the back pressure controller and that the partial pressure
29 29
pressure (1 atm) and temperature (298 K) and assuming that of nitrogen gas is equal to the back pressure. Assuming that the
30 30
nitrogen gas was the only gas produced and all inorganic carbon gas phase is in equilibrium with the dissolved nitrogen gas, its
31 31
either precipitated or remained in solution due to the relatively high concentration was calculated using Henry’s law. The resulting
32 32
solubility at neutral to high pH. Following these assumptions, the nitrogen balance in the triaxial test is shown in Table 3.
33 33
34 34
35 35
36 Experiment Liquid Liquid volume: ml Nitrate: mM Nitrite: mM Calcium ions: mM CT: mM 36
37 37
38 300–350 kPa Initial suspension 97 50·0 0 51·9 103·7 38
39 Expelled liquid ND ND ND ND ND 39
40 Flushed liquid 206 2·9 0·04 0·69 9·25 40
41 Cell water ND ND ND ND ND 41
42 345–350 kPa Initial suspension 97 52·7 0 61·6 104 42
43 Expelled liquid ND ND ND ND ND 43
44 Flushed liquid 206 0 0 1·5 7·7 44
45 Cell water 3630 0·15 0·15 0·59 1·73 45
46 50–100 kPa Initial suspension 97 51·8 0 57·6 126 46
47 Expelled liquid 35 (55)a 1·44 0·01 3·42 2·2 47
48 Flushed liquid 207 (212)a 0·11 0 1·83 4·08 48
49 Cell water 3633 0·1 0·158 0·66 0·92 49
50 a
The numbers between parentheses are the values recorded in the back pressure controller 50
51 ND, not determined 51
52 52
53 Q10 Table 2. Volumes and solute concentrations of different liquids in 53
54 the triaxial tests 54
55 55

envgeo1500078.pdf 9 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 20 The mass balance calculations in Table 3 show that for the two 1
2 experiments under high-pressure conditions, 26 and 33% of the 2
15
CE: mS/cm

3 nitrogen compounds in the nitrate added are unaccounted for and 3

4 10 therefore lost, while for the experiment at (near) ambient pressure, 4
5 5 the final nitrogen level is 25% higher than initially added. Further 5
6 investigation is required to validate the assumptions and explain 6
0
7 0 1 2 3 4 5 6 7 8 9 10 the observed discrepancies in the mass balance. 7
8 (a)
8
9 7·0 The effect of substrate composition on process efficiency 9
10 6·5 The results of the liquid batch experiments show that the substrate 10
11 6·0 composition significantly affects the process performance. When 11
pH

12 5·5 denitrification is combined with MICP, the pH is buffered at values 12

13 5·0 which are slightly lower than the optimal range for denitrification 13
4·5
14 reported in literature (Knowles, 1982; Wang et al., 1995). A similar 14
4·0
15 0 1 2 3 4 5 6 7 8 9 10 pH buffering effect by MICP was observed by Burbank et al. 15
16 (2011). In their field experiments, in which they stimulated MICP 16
(b)
17 120
by a combination of urea hydrolysis and denitrification, they 17
NO3– concentration:

18 100 observed that the pH remained around neutral. The results of this 18
19 80 study are partly in agreement with the observations of Glass and 19
mM

20 60 Silverstein (1998), showing that a high pH – which occurs when 20

21 40 sodium-based media are used – favours the reduction of nitrate to 21
22 20 nitrite but can cause nitrite to accumulate. On the other hand, the 22
23 0 solutions containing calcium showed complete reduction of nitrate, 23
0 1 2 3 4 5 6 7 8 9 10
24 with limited accumulation of nitrite, even though the pH was 24
(c)
25 20 buffered at a value below 7. This contradicts the observations by 25
NO2– concentration:

26 Glass and Silverstein (1988), who found that nitrate reduction at pH Q11 26
15
27 7 and below is completely inhibited. Still, the observed pH values 27
mM

28 10 are within the range at which active denitrification has been 28

29 5 reported by others and far from the lower limit that could cause 29
30 nitrous oxide accumulation (Knowles, 1982). Hence, the combined 30
0
31 0 1 2 3 4 5 6 7 8 9 10 process of nitrate reduction and calcium carbonate precipitation 31
32 promotes efficient conversion of nitrate to nitrogen gas and prevents 32
(d)
33 150 the accumulation of toxic intermediates and potential GHGs. 33
concentration: mM

34 34
35 100 Also, the C/N ratio in the growth medium significantly affects the 35
Ca2+

36 conversion efficiency. As similarly observed by Chung and Chung 36

37 50 (2000), this study found that high C/N ratios favour the complete 37
38 conversion of nitrate and limit nitrite accumulation but lead to an 38
0
39 excess of acetate and calcium at the end of the experiment. A low 39
0 1 2 3 4 5 6 7 8 9 10
40 (e) C/N ratio, on the other hand, leads to full acetate consumption, 40
41 0·020 but at the end of the experiment, significant concentrations of 41
conductivity:

42 residual nitrate and accumulated nitrite remain. In this study’s 42

Hydraulic

0·015
43 experiments, the optimum Ac/N ratio appears to be around 0·8, 43
cm/s

44 0·010 which is slightly lower than the optimum ratio found by Chung 44
45 0·005 and Chung (2000), but is significantly lower than the ratio at 45
46 0 maximum growth according to reaction III. 46
47 0 1 2 3 4 5 6 7 8 9 10 47
48 Pore volume (1 pore volume = 230 ml) The effect of pressure conditions on gas formation and 48
49 (f) distribution 49
50 When stimulating MICP by denitrification in sand, the pressure 50
51 Figure 7. Measured values of (a) CE, (b) pH, (c) nitrate conditions significantly affect the formation and distribution of 51
52 concentration, (d) nitrite concentration, (e) calcium ion gas in the pore space. For a given amount of substrate, the 52
53 concentration and (f) hydraulic conductivity for the sand column volume of gas is much lower at high pore pressures than at low 53
54 experiment at ambient pressure pore pressures. This is expected considering firstly that according 54
55 to the ideal gas law, the volume and pressure of gas are inversely 55

10

envgeo1500078.pdf 10 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 proportional. Secondly, according to Henry’s law, gas solubility is confinement pressure is influences the potential of MICP by 1
2 higher at high pressures. Thirdly, a higher gas solubility leads to denitrification for ground reinforcement applications. At low 2
3 higher dissolved gas concentrations, which may lead to larger confinement conditions (shallow depth), the gas may form cracks 3
4 diffusive fluxes to surrounding water as observed in the case of and disturb the sand structure and reduce the stabilising effect of 4
5 300–350 and 345–350 kPa in Figure 6. the calcium carbonate minerals, as shown in Figure 8. 5
6 6
7 The total gas volume in the pore space significantly influences the The relationship between the distribution of gas and 7
8 geotechnical properties of sand, including strength, permeability calcium carbonate 8
9 and stiffness (DeJong et al., 2013). After a specific amount of gas The average calcium carbonate content in the sand column which 9
10 is produced, single gas bubbles will coagulate into gas pockets or was flushed nine times with substrate solution reached 1·1%. 10
11 form a continuous gas phase, which tends to migrate upwards in Although this amount of calcium carbonate is relatively low 11
12 Q12 irregular patterns (Haines, 1930; van Paassen, 2010a). Also, the compared with those from other studies on MICP based on 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 (a)
40
41 41
42 42
43 43
44 44
45 45
46 46
47 47
48 48
49 (b) 49
50 50
51 Figure 8. Result of the gas distribution in the sand column treated 51
52 at ambient pressure with multiple substrate flushes: (a) the X-ray 52
53 CT scanned images and (b) the constructed three-dimensional 53
54 (3D) image 54
55 55

11

envgeo1500078.pdf 11 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 al., 2014; Kavazanjian et al., 2015). The relationship between 1

2 6·5 cm strength and calcium carbonate content depends on many factors, 2
3 including the density, mineral type and grain size distribution of 3
4 the treated sand (Cheng et al., 2013; van Paassen et al., 2010a) Q13 4
5 0·40% and substrate concentrations, reaction rate and environmental 5
6 conditions of the treatment process (Al Qabany et al., 2012; 6
7 Mortensen et al., 2011; van Paassen, 2009a). Q14 7
8 0·74% 8
9 Gas production can have a positive effect on the strength gain due 9
10 to MICP. Cheng and Cord-Ruwisch (2013) showed that columns Q15 10
11 which have been treated under partially saturated conditions require 11
12 0·89% lower amounts of calcium carbonate to obtain a similar increase in 12
13 strength for fully saturated conditions. They considered that in 13
14 partially saturated conditions, substrate solutions preferably fill the 14
18 cm
15 pore throats due to suction effects, resulting in a more efficient 15
16 1·05% distribution of cementing calcium carbonate minerals. Similarly, 16
17 Kavazanjian et al. (2015) showed that MICP by denitrification 17
18 results in more efficient cementation than MICP by urea hydrolysis. 18
19 The present study showed that within the sand column treated at 19
1·02%
20 ambient pressure, calcium carbonate was predominantly formed at 20
21 the bottom of the column, while gas was predominantly formed 21
22 at the top. This seems obvious because there is less substrate 22
23 1·00%
available to react when gas displaces the substrate solution. As a 23
24 result, the potential improvement in the cementation efficiency due 24
25 1·09% to desaturation is counteracted by a lower conversion efficiency. 25
26 1·53% Secondly, the cracks which were formed at low confining pressure 26
27 disturb the sand-stabilising effect of the calcium carbonate minerals. 27
28 0·88% The low number, large size and dendritic structure of the calcium 28
29 carbonate crystals can be due to the fact that the crystals have 29
30 Figure 9. Calcium carbonate distribution in the sand column grown in multiple phases at a relatively low reaction rate (van 30
31 treated at ambient pressure Paassen, 2009b) and in the presence of impurities such as organic 31
32 polymers from the growing biomass. 32
33 33
34 hydrolysis of urea, such as that of van Paassen et al. (2010a), it Conversion rate 34
35 may still be sufficient to increase significantly cyclic shear The overall reaction rate of denitrification and precipitation is an 35
36 resistance to mitigate liquefaction of loose sandy soils (DeJong et important process parameter, which depends on many factors, 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
47 47
48 48
49 49
50 Acc.V Spot Magn Det WD Exp 200 µm Acc.V Spot Magn Det WD Exp 20 µm 50
51 16·0 kV 5·0 100x BSE 9·7 1 1·0 Torr sand treated 16·0 kV 5·0 100x BSE 9·9 1 1·0 Torr sand treated 51
52 52
53 Figure 10. Presence of calcium carbonate in the sand treated with 53
54 multiple substrate flushes at ambient pressure in ESEM images 54
55 55

12

envgeo1500078.pdf 12 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 1
2 300–350 kPa 345–350 kPa 50–100 kPa 2
3 3
Initial: mmol nitrogen Nitrate in the suspension 4·85 5·11 5·03
4 4
Final: mmol nitrogen Nitrate in cell water ND 0·55 0·36
5 5
Nitrite in cell water ND 0·54 0·57
6 6
Nitrate in [flushed + expelled] water 0·6 0 0·1
7 7
Nitrite in [flushed + expelled] water 0·01 0 0
8 8
Nitrogen gas 2·14 2·13 5·18
9 9
Nitrogen gas, aqueous 0·51 0·57 0·19
10 10
Total 3·26 3·78 6·40
11 11
Ngap: mmol 1·59 1·33 −1·37
12 12
Ngap: % 33 26 −25
13 13
14 ND, not determined 14
15 15
16 Q16 Table 3. Nitrogen balance for the triaxial tests 16
17 17
18 18
19 including the amount and type of denitrifying organisms, substrate thermodynamic approach, which solves the mass and charge 19
20 and product concentrations and environmental effects, such as balance to determine the stoichiometry of the metabolic reaction, 20
21 temperature, pH, salinity and the presence of other catalysing or and making some simplifying assumptions to determine the 21
22 Q17 inhibiting compounds (van Paassen, 2009a). Reported reaction rates distribution of gas over the dissolved and gas phases. Further 22
23 range from a few millimoles of nitrate per litre per day up to investigation is required to explain the observed deficiency 23
24 170 mmol nitrate/(l d) (Glass and Silverstein, 1998; Martin et al., between measured and expected values. Under high-pore pressure 24
25 Q18 2013; Soares et al., 1991; van Paassen et al., 2010a). This study conditions in the triaxial cell, the produced gas initially expelled 25
26 showed that the reaction rate in the triaxial cell was about the liquid phase but later on seemed to dissipate from the sand 26
27 Q19 10–15 mmol nitrate/(l d). The liquid batch experiments in which the sample, probably by diffusion through the latex membrane. At 27
28 same substrate concentrations and inoculum size were used showed low pore pressures, the produced gas was higher than the gas 28
29 a rate of 1–2 mmol nitrate/(l d), about five to eight times lower than storage capacity of the sand column and escaped to the back 29
30 those for the triaxial cell. The average reaction rate in the sand pressure controller column while expelling part of the liquid 30
31 column at ambient pressure was about 6 mmol nitrate/(l d). phase. In the sand column which was treated with multiple Q20 31
32 However, the remaining substrates in the final flushes indicated that batches of substrate solution, the conversion rate and cumulative 32
33 the reaction rate in the unsaturated top part of the column was amount of calcium carbonate appeared to be inversely related to 33
34 lower than that in the saturated bottom part. Considering that at the gas distribution because low confinement cracks were 34
35 least a small mass percentage of calcium carbonate needs to be induced, which may negatively influence the sand-stabilising 35
36 formed, it is expected that based on the current results, several effect of the calcium carbonate minerals. 36
37 weeks to several months of treatment is still required in order to 37
38 obtain sufficient strength improvement when using MICP through Acknowledgements 38
39 denitrification. Further investigation is required to explain the This study was supported by the International Training Program 39
40 observed differences in reaction rate. and Institutional Program of Young Researcher Overseas Visits 40
41 launched by the Japan Society for the Promotion of Science 41
42 Conclusions (JSPS) and the Dutch Technology Foundation STW, which is part 42
43 This study showed that the combined process of biological of the Netherlands Organisation for Scientific Research, and was 43
44 denitrification and MICP leads to a more efficient conversion than partly funded by the Dutch Ministry of Economic Affairs, through 44
45 denitrification without precipitation. The precipitation of calcium STW perspective program BioGeoCivil (11337) and JSPS Grant- 45
46 carbonate buffers the pH and prevents the accumulation of toxic in-Aid for Young Scientists (B) Grant Number 25850166. 46
47 intermediates. The optimum acetate-to-nitrate ratio, at which all 47
48 substrates are consumed most efficiently, is about 0·8 (which REFERENCES Q21 48
49 corresponds to a C/N ratio of 1·6). Lower Ac/N ratios lead to the Al Qabany AA, Soga K and Santamarina C (2012) Factors 49
50 accumulation of toxic intermediate nitrogen compounds, while affecting efficiency of microbially induced calcite 50
51 higher ratios leave significant amounts of residual calcium and precipitation. Journal of Geotechnical and Geoenvironmental 51
52 acetate. Sand column experiments confirmed that the volume and Engineering 138: 992–1001. Q22 52
53 distribution of the gas phase strongly depend on the pressure Almeida JS, Julio SM, Reis MAM and Carrondo MJT (1995) Nitrite 53
54 conditions. The produced gas volume is inversely related to the inhibition of denitrification by Pseudomonas fluorescens. 54
55 pore pressure and can be reasonably predicted using a generic Biotechnology and Bioengineering 46(3): 194–201. 55

13

envgeo1500078.pdf 13 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 Archna SKS and Sobti RC (2012) Nitrate removal from ground Hamdan N, Kavazanjian E, Rittmann BE and Karatas I (2016) 1
2 water: a review. E-Journal of Chemistry 9(4): 1667–1675. Carbonate mineral precipitation for soil improvement through 2
3 Burbank MB, Weaver TJ, Green TL, Williams BC and Crawford RL microbial denitrification. Geomicrobiology Journal, http://dx. Q24 3
4 (2011) Precipitation of calcite by indigenous microorganisms doi.org/10.1080/01490451.2016.1154117. 4
5 to strengthen liquefiable soils. Geomicrobiology Journal 28(4): Hanaki K, Hong Z and Mastsuo T (1992) Production of nitrous 5
6 301–312. oxide gas during denitrification of wastewater. Water Science 6
7 CEN (Comité Européen de Normalisation) (2004) ISO/TS 17892- and Technology 26(5–6): 1027–1036. 7
8 9:2004: Geotechnical investigation and testing – laboratory Harkes MP, van Paassen LA, Booster JL, Whiffen VS and van 8
9 testing of soil – part 9: consolidated triaxial compression tests Loosdrecht MCM (2010) Fixation and distribution of 9
10 on water saturated soil. CEN, Brussels, Belgium. bacterial activity in sand to induce carbonate precipitation 10
11 Cheng L and Cord-Ruwisch R (2012) In situ soil cementation with for ground reinforcement. Ecological Engineering 36: 11
12 ureolytic bacteria by surface percolation. Ecological 112–117. 12
13 Engineering 42: 64–72. Haynie DT (2008) Biological Thermodynamics. Cambridge 13
14 Cheng L, Cord-Ruwisch R and Shahin MA (2013) Cementation of University Press, Cambridge, UK. 14
15 sand soil by microbially induced calcite precipitation at He J, Chu J and Ivanov V (2013) Mitigation of liquefaction of 15
16 various degrees of saturation. Canadian Geotechnical Journal saturated sand using biogas. Géotechnique 63(4): 267–275, 16
17 50(1): 81–90. http://dx.doi.org/10.1680/bcmpge.60531.011. 17
18 Chu J, Stabnikov V and Ivanov V (2012) Microbially induced Heijnen JJ and Kleerebezem R (2010) Bioenergetics of microbial 18
19 calcium carbonate precipitation on surface or in the bulk of growth. In Encyclopedia of Industrial Biotechnology: 19
20 soil. Geomicrobiology Journal 29(6): 544–549. Bioprocess, Bioseparation and Cell Technology (Flickinger 20
21 Chung YC and Chung MS (2000) BNP test to evaluate the influence MC (ed.)). Wiley. New York, NY, USA, vol. 1, pp. 594–617. 21
22 of C/N ratio on N2O. Water Science and Technology 42: 23–27. Heijnen JJ, van Loosdrecht MCM and Tijhuis L (1992) A black 22
23 DeJong JT, Fritzges MB and Nusslein K (2006) Microbially box mathematical model to calculate auto- and heterotrophic 23
24 induced cementation to control sand response to undrained biomass yields based on Gibbs energy dissipation. 24
25 shear. Journal of Geotechnical and Geoenvironmental Biotechnology and Bioengineering 40: 1139–1154. 25
26 Engineering 132(11): 1381–1392. Karatas I, Kavazanjian E and Rittmann B (2008) Microbially 26
27 DeJong JT, Martinez BC, Mortensen BM et al. (2009) Upscaling induced precipitation of calcite using Pseudomonas 27
28 of bio-mediated soil improvement. In 17th International denitrificans. Proceedings of 1st International Conference on 28
29 Conference on Soil Mechanics and Geotechnical Engineering Bio-geo-civil Engineering, Delft, Netherlands, pp. 58–66. Q25 29
30 Q23 (Hamza M, Shahien M and El-Mossallamy Y (eds)). IOS Kauzmann W (2013) Kinetic Theory of Gases. Dover, 30
31 Press, Egypt. Newburyport, MA, USA. 31
32 DeJong JT, Soga K, Kavazanjian E et al. (2013) Biogeochemical Kavazanjian E, O’Donnell ST and Hamdan N (2015) 32
33 process and geotechnical applications: progress, opportunities Biogeotechnical mitigation of earthquake-induced soil 33
34 and challenges. Géotechnique 63(4): 287–301, http://dx.doi. liquefaction by denitrification: a two-stage process. 6th 34
35 org/10.1680/geot.SIP13.P.017. International Conference on Earthquake Geotechnical 35
36 DeJong JT, Martinez BC, Ginn TR et al. (2014) Development of a Engineering, Christchurch, New Zealand. Q26 36
37 scaled repeated five spot treatment model for examining Keddy PA (2000) Wetland Ecology – Principle and Conservation. 37
38 microbial induced calcite precipitation feasibility in field Cambridge University Press, Cambridge, UK. 38
39 applications. Geotechnical Testing Journal 37(3): 424–435. Knowles R (1982) Denitrification. Microbiological Reviews 46(1): 39
40 Ehrlich HL and Newman DK (1996) Geomicrobiology, 3rd edn. 43. Q27 40
41 CRC Boca Raton, FL, USA. Kuenen JG and Robertson LA (1988) Ecology of nitrification and 41
42 Ferguson SJ (1994) Denitrification and its control. Antonie van denitrification. In SSGM 42: the Nitrogen and Sulpher (Society Q28 42
43 Leeuwenhoek 66(1–3): 89–110. for General Microbiology Symposia) (Cole JA and Ferguson 43
44 Fredlund DG and Rahardjo H (1993) Soil Mechanics for SJ (eds)). Cambridge University Press, Cambridge, UK, 44
45 Unsaturated Soils. Wiley, New York, NY, USA. pp. 161–218. 45
46 Glass C and Silverstein J (1998) Denitrification kinetics of high Martin D, Dodds K, Butler IB and Ngwenya BT (2013) Carbonate 46
47 nitrate concentration water: pH effect on inhibition and nitrite precipitation under pressure for bioengineering in the 47
48 accumulation. Water Research 32(3): 831–839. anaerobic subsurface via denitrification. Environmental 48
49 Grozic JL, Robertson PK and Morgenstern NR (1998) The Science and Technology 47(15): 8692–8699. 49
50 behavior of loose gassy sand. Canadian Geotechnical Journal Mortensen BM, Haber MJ, Dejong JT, Caslake LF and Nelson DC 50
51 36: 482–492. (2011) Effects of environmental factors on microbial induced 51
52 Haines WB (1930) Studies in the physical properties of soils: V. calcium carbonate precipitation. Journal of Applied 52
53 the hysteresis effect in capillary properties, and the modes of Microbiology 111(2): 338–349. 53
54 water distribution associated therewith. Journal of Agricultural Overmann J, Fischer U and Pfennig N (1992) A new purple sulfur 54
55 Sciences 20(1): 97–116. bacterium from saline littoral sediments, Thiorhodovibrio 55

14

envgeo1500078.pdf 14 Manila Typesetting Company 10/18/2016 9:42pm

 Environmental Geotechnics Applying MICP by denitrification in soils:
Volume XXXX Issue EGXXXX a process analysis
Pham, Nakano, van der Star, Heimovaara and van Paassen

1 winogradskyi gen. nov. and sp. nov. Archives of Microbiology Publishing Services, Singapore, pp. 177–182, http://dx.doi.org/ 1
2 157(4): 329–335. 10.3850/GI126. 2
3 Q29 Ozgumus T, Mobedi M and Ozkol U (2014) Determination of van Paassen LA (2009a) Biogrout – Ground Improvement by 3
4 Kozeny constant based on porosity and pore to throat size ratio Microbially Induced Carbonate Precipitation. PhD thesis, 4
5 in porous medium with rectangular rods. Engineering Delft University of Technology, Delft, Netherlands. 5
6 Applications of Computational Fluid Mechanics 8(2): 308–318. van Paassen LA (2009b) Microbes turning sand into sandstone, 6
7 Q30 Payne WJ (1981) Denitrification. Wiley-Interscience, Canada. using waste as cement. 4 iYGCE’09 – 4th International Young 7
8 Phillips AJ, Gerlach R, Lauchnor E et al. (2013) Engineered Geotechnical Engineers’ Conference, Alexandria, Egypt. Q31 8
9 applications of ureolytic biomineralization: a review. van Paassen LA (2011) Bio-mediated ground improvement: from 9
10 Biofouling 29(6): 715–733. laboratory experiment to pilot applications. In Geo-frontiers 10
11 Rebata-Landa V and Santamarina JC (2011) Mechanical effects of Congress 2011: Advances in Geotechnical Engineering (Han J Q32 11
12 biogenic nitrogen gas bubbles in soils. Journal of Geotechnical and Alzamora DE (eds)). American Society of Civil 12
13 and Geoenvironmental Engineering 138(2): 128–137. Engineers, Reston, VA, USA, pp. 4099–4108. 13
14 Robertson GP and Groffman PM (2015) Nitrogen van Paassen LA, Ghose R, van der Linden TJM, van der Star WRL 14
15 transformations. In Soil Microbiology, Ecology and and Loosdrecht MCM (2010a) Quantifying biomediated 15
16 Biochemistry, 4th edn. (Paul EA (ed.)). Academic, Boston, ground improvement by ureolysis: large-scale biogrout 16
17 MA, USA, pp. 421–446. experiment. Journal of Geotechnical and Geoenvironmental 17
18 Schumacher BA (2002) Methods for the Determination of Total Engineering 136: 1721–1728. 18
19 Organic Carbon (TOC) in and Sediments. US Environmental van Paassen LA, Daza CM, Staal M et al. (2010b) Potential soil 19
20 Protection Agency, Washington, DC, USA. reinforcement by biological denitrification. Ecological 20
21 Skempton AW (1954) The pore-pressure coefficients A and B. Engineering 36: 168–175. 21
22 Géotechnique 4(4): 143–147, http://dx.doi.org/10.1680/geot. Voorhees EB (1902) Studies in denitrification. Journal of the 22
23 1954.4.4.143. American Chemical Society 24(9): 785–823. 23
24 Soares MIM (2000) Biological denitrification of groundwater. Wang JH, Baltzis BC and Lewandowski GA (1995) 24
25 Water, Air, and Soil Pollution 123(1–4): 183–193. Fundamental denitrification kinetic studies with Pseudomonas 25
26 Soares MI, Braester C, Shimshon B and Abeliovich A (1991) denitrificans. Biotechnology and Bioengineering 47(1): 26
27 Denitrification in laboratory sand columns: carbon regime, gas 26–41. 27
28 accumulation and hydraulic properties. Water Research 25(3): Whiffin VS, van Paassen LA and Harkes MP (2007) Microbial 28
29 325–332. carbonate precipitation as a soil improvement technique. 29
30 Stal LJ (2010) Microphytobenthos as a biogeomorphological force Geomicrobiology Journal 24(5): 417–423. 30
31 in intertidal sediment stabilization. Ecological Engineering 36: Yaws CL (2012) Yaws’ Handbook of Properties for Aqueous Q33 31
32 236–245. Systems. Knovel, New York, NY, USA. 32
33 van der Star WRL, Taher E, Harkes MP et al. (2009) Use of waste Yegian MK, Eseller-Bayat E and Alshawabkeh A (2007) Induced- 33
34 stream and microbes for in situ transformation of sand into partial saturation for liquefaction mitigation: experimental 34
35 sandstone. In Ground Improvement Technologies and Case investigation. Journal of Geotechnical and Geoenvironmental 35
36 Histories (Leung CF, Chu J and Shen RF (eds)). Research Engineering 133: 372–380. 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
47 47
48 48
49 49
50 HOW CAN YOU CONTRIBUTE? 50
51 To discuss this paper, please submit up to 500 words 51
52 to the editor at journals@ice.org.uk. Your contribution 52
53 will be forwarded to the author(s) for a reply and, if 53
54 considered appropriate by the editorial board, it will be 54
55 published as a discussion in a future issue of the journal. 55

15

envgeo1500078.pdf 15 Manila Typesetting Company 10/18/2016 9:42pm

 Author Queries

Q1: Please replace titles “Ir.”, ”Dr.” and “Dr. Ir.” with the corresponding post-nominal degrees.
Q2: Please check that the affiliations are presented correctly.
Q3: Please add Montoya et al., 2013 to the reference list.
Q4: Please verify citation “van Paassen, 2009a”.
Q5: Please check change to “water levels”.
Q6: Please add ASTM D 4253 and ASTM D 4254 to the reference list so that these can be cited properly.
Q7: Please provide descriptions for panels (a) to (f) of Figure 4.
Q8: Please check changes in Figure 6 caption and panel labels.
Q9: Please check changes in the first column of Table 1. Also, please consider standardizing the number of decimal places in
the second column.
Q10: Please check changes in the first column of Table 2. Also, please check the definition provided for ND in Tables 2 and 3.
Q11: Please check if “Glass and Silverstein (1988)” should be “Glass and Silverstein (1998)”. If not, please add Glass and
Silverstein (1988) to the reference list.
Q12: Please verify citation “van Paassen, 2010a”.
Q13: Please verify citation “van Paassen, 2010a”.
Q14: Please verify citation “van Paassen, 2009a”.
Q15: Please check if “Cheng and Cord-Ruwisch (2013)” should be “Cheng and Cord-Ruwisch (2012)”. If not, please add
Cheng and Cord-Ruwisch (2013) to the reference list.
Q16: Please check changes in Table 3.
Q17: Please verify citation “van Paassen, 2009a”.
Q18: Please verify citation “van Paassen, 2010a”.
Q19: Please verify that changes in the sentence “The liquid batch experiments in which the same ...” have preserved your
meaning.
Q20: Please verify that changes in the sentence “In the sand column which was treated ...” have preserved your meaning.
Q21: Please provide DOIs for all journal and proceedings references where available and verify the provided DOIs.
Q22: Please provide issue nos. for Al Qabany et al. (2012),Cheng L and Cord-Ruwisch R (2012), Chung YC and Chung MS
(2000), Grozic et al. (1998), Harkes et al. (2010), Heijnen et al. (1992), Stal LJ (2010), van Paassen et al. (2010a), van
Paassen et al. (2010b), and Yegian et al. (2007).
Q23: Please provide city/town location of IOS Press and page nos. for DeJong et al. (2009).
Q24: Please provide volume, issue, and page nos. for Hamdan et al. (2016).
Q25: Kauzmann W (2013) is not cited in the text. Please insert a citation or delete the reference.
Q26: Please provide page nos. for Kavazanjian et al. (2015).
Q27: Please verify that Knowles (1982) is a one-page article. If it is not, please provide the complete page range.
Q28: Please verify title of conference in Kuenen JG and Robertson LA (1988).
Q29: Ozgumus et al. (2014) is not cited in the text. Please insert a citation or delete the reference.
Q30: Please provide city/town and province location of Wiley-Interscience.
Q31: Please provide page nos. for van Paassen LA (2009b).
Q32: Please verify full publication title and editor list in van Paassen LA (2011).
Q33: Yaws CL (2012) is not cited in the text. Please insert a citation or delete the reference.