Enhanced Retention of Chemotactic Bacteria in a Pore Network with

Dec 3, 2015 - NAPL droplets (toluene) were trapped within a fine pore network, and ... Bacteria in Granular Media with Distributed Contaminant Sources...
2 downloads 0 Views 2MB Size
Subscriber access provided by UNIV OF NEBRASKA - LINCOLN

Article

Enhanced retention of chemotactic bacteria in a pore network with residual NAPL contamination Xiaopu Wang, Larry M. Lanning, and Roseanne Marie Ford Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.5b03872 • Publication Date (Web): 03 Dec 2015 Downloaded from http://pubs.acs.org on December 8, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31

Environmental Science & Technology

1

Enhanced Retention of Chemotactic Bacteria in a Pore

2

Network with Residual NAPL Contamination

3 4

Xiaopu Wang§, †, Larry M. Lanning † and Roseanne M. Ford †, *

5

§ School of Petroleum Engineering, China University of Petroleum, Qingdao, Shandong,

6

China 266580

7

† Department of Chemical Engineering, School of Engineering and Applied Science,

8

University of Virginia, Charlottesville, VA, USA 22904

9

* Corresponding author phone: +1-434-924-6283; email: [email protected]

10 11 12

Running title:

13

Chemotaxis in heterogeneous porous microfluidics

14

Keywords:

15

Chemotaxis, Pseudomonas putida, heterogeneity, porous media, microfluidics,

16

bioremediation, biodegradation

17 18

Contact grant sponsor: NSF, National Natural Science Foundation of China, the

19

Fundamental Research Funds for the Central Universities

20

Contact grant number: EAR-0711377, EAR-1141400, 51509260 and 15CX02008A

21

Additional Supporting Information may be found in the online version of this article.

1

ACS Paragon Plus Environment

Environmental Science & Technology

22

Abstract

23

Nonaqueous phase liquid (NAPL) contaminants are difficult to eliminate from natural

24

aquifers due, in part, to the heterogeneous structure of the soil. Chemotaxis enhances the

25

mixing of bacteria with contaminant sources in low permeable regions, which may not be

26

readily accessible by advection and dispersion alone. A microfluidic device was designed

27

to mimic heterogeneous features of a contaminated groundwater aquifer. NAPL droplets

28

(toluene) were trapped within a fine pore network, and bacteria were injected through a

29

highly conductive adjacent macrochannel. Chemotactic bacteria (Pseudomonas putida

30

F1) exhibited greater accumulation near the pore network at 0.5 m/d than both the

31

nonchemotactic control and the chemotactic bacteria at a higher groundwater velocity of

32

5 m/d. Chemotactic bacteria accumulated in the vicinity of NAPL droplets, and the

33

accumulation was 15% greater than a nonchemotactic mutant. Indirect evidence showed

34

that chemotactic bacteria were retained within the contaminated low permeable region

35

longer than nonchemotactic bacteria at 0.25 m/d. This retention was diminished at 5 m/d.

36

Numerical solutions of the bacterial transport equations were consistent with the

37

experimental results. Because toluene is degraded by P. putida F1, the accumulation of

38

chemotactic bacteria around NAPL sources is expected to increase contaminant

39

consumption and improve the efficiency of bioremediation.

2

ACS Paragon Plus Environment

Page 2 of 31

Page 3 of 31

40

Environmental Science & Technology

Introduction

41

Nonaqueous phase liquids (NAPLs) are the most common industrial contaminants

42

in polluted groundwater.1 NAPL contaminants, including BTEX (benzene, toluene,

43

ethylbenzene, and xylenes), trichloroethylene (TCE), petroleum and chlorinated

44

hydrocarbons,2 are difficult to eliminate from natural aquifers due to their low solubility,

45

low volatility, low intrinsic reactivity and low release rates from soil or sediments.3,4 As a

46

consequence, they have become a major threat to groundwater resources for our daily

47

usage. In-situ bioremediation provides an effective and economical solution to remove

48

the

49

microorganisms or other forms of biomass are injected into the contaminated sites and

50

directly degrade or transform contaminant chemicals into nontoxic byproducts;8 at the

51

same time, microorganisms may increase their population by growth upon the

52

contaminant sources, which can further accelerate the biodegradation process. However,

53

in-situ bioremediation has its own limitations. Because of the heterogeneity of the natural

54

groundwater systems, the NAPL contaminant sources, which are trapped in regions

55

where the conductivity is fairly low, are not readily carried away by the advective flow in

56

zones with high hydraulic conductivity.9-11 One prominent example is the oil spill from

57

the tanker Exxon Valdez in 1989. Although soluble contaminants were removed from the

58

high-permeability upper layer of the beach by conventional remediation, residual oil

59

persists in the low-permeability lower layer, and continues to leach slowly into the upper

60

layer.12 Therefore, a more effective method is needed to remediate the low permeable

61

regions, in order to ensure complete removal of oil.

organic

pollutants

from

groundwater.5-7

3

Generally

ACS Paragon Plus Environment

speaking,

indigenous

Environmental Science & Technology

62

Chemotaxis, the microbial property by which microorganisms sense the

63

concentration gradient of chemicals and migrate towards the preferential regions for their

64

survival and growth, may be a key factor to achieve more efficient bioremediation.13,14

65

Chemotaxis phenomena were associated with the biodegradation of many environmental

66

chemicals, including NAPL and various aromatic compounds; because of the ability to

67

locate and degrade the pollutants, such bacteria have a selective advantage to survive and

68

grow in various contaminated sites, and as a consequence, reduce the pollutants in these

69

areas.10,13,15-23 An indigenous bacterial strain P. putida F1 was studied, because it showed

70

strong chemotactic responses to and degraded many chlorinated hydrocarbons, such as

71

toluene, benzene, trichloroethylene etc.,16,20 and was widely applied to study bacterial

72

chemotaxis.

73

Microfluidic devices have proven to be effective tools in a wide range of

74

scientific fields,24 and have some ideal features for the study of bacterial chemotaxis.

75

Microfluidic devices allow accurate fabrication of microscopic scale features and precise

76

manipulation of very small amounts of fluid, which is ideal for the study of NAPL

77

contaminants in pore-scale geometries that are representative of the natural aquifer

78

environment. Moreover, the microscale channel operates under laminar flow in most

79

scenarios, so that the concentration gradient of attractant is smooth and mathematically

80

predictable. Its absolute transparency and small size allow direct and quantitative

81

measurement of the bacterial distributions inside the device under a microscope.

82

Amongst the earlier studies, microfluidic devices with pore-scale features were

83

created to study colloid transport.25-27 More recently, a variety of microfluidics designs

84

were applied to the study of bacterial chemotaxis with advective flow. Under flow

4

ACS Paragon Plus Environment

Page 4 of 31

Page 5 of 31

Environmental Science & Technology

85

conditions, parallel-flow devices were among the first approaches to generate steep and

86

steady concentration gradients and assess chemotaxis by the extent of transverse bacterial

87

movement towards the chemoeffector stream.28-30 Long and Ford observed bacterial

88

chemotaxis in a direction transverse to flow under fluid velocities typical of groundwater

89

flow rates in a T-shaped homogeneous porous channel with a staggered array of

90

cylindrical posts.31 Singh and Olson created a heterogeneous porous microfluidic device

91

to study transverse chemotaxis, and an increase of chemotactic bacterial population was

92

detected in the low permeable regions containing attractant.32 All the previous studies

93

successfully detected the effect of bacterial chemotaxis in porous structure under typical

94

groundwater flow rates; however, they all evaluated contaminants in the aqueous phase

95

rather than the organic phase. A two-phase heterogeneous porous system is more

96

representative of the actual NAPL contaminated sites. Wan and coworkers published

97

several studies of bacterial transport in two-phase systems with water and air in the

98

microfluidic device,25,26 and several researchers continued to study unsaturated

99

(liquid/gas interfaced) porous media at different scales.33-37 Wang and co-workers

100

developed a single-pore microfluidic device to quantify the chemotaxis effect in the

101

vicinity of a NAPL/water interface.38

102

In this study, a heterogeneous porous microfluidic device was designed to trap

103

organic phase chemicals in the low permeable regions in order to test the following

104

hypothesis: In well-defined heterogeneous porous media, chemotaxis facilitates bacterial

105

migration preferentially towards NAPL attractant sources trapped in regions with lower

106

hydraulic conductivity, which leads to accumulation of bacteria at chemically favorable

5

ACS Paragon Plus Environment

Environmental Science & Technology

107

locations, and, furthermore, an increase in the groundwater velocity beyond typical rates

108

diminishes the impact of chemotaxis.

109

Materials and Methods

110

Preparation of the Bacteria/Attractant System. C. S. Harwood provided the

111

wild type Pseudomonas putida F1,16 and R. E. Parales constructed the nonchemotactic

112

mutant P. putida F1 CheA.39 The bacterial cultivation, harvest, resuspension and staining

113

were thoroughly described in the previous work.40 The two bacterial suspensions with

114

different dyes (the wild type with FM 4-64 and the mutant with DAPI) were then mixed

115

for later experimental use, together with 1.2 mM sodium nitrate (Sigma-Aldrich, ACS

116

reagent). 7 mL pure toluene (Sigma-Aldrich, HPLC grade) was stained with 1 mg Oil

117

Red O (FisherBiotech, CAT#: BP112-10) to show observable red color.

118

Microfluidic Design, Fabrication and Operation. The design used for the

119

microfluidic device is presented in Figure 1 (a), in which the higher permeable region is

120

represented by the horizontal channel (macrochannel) above and the lower permeable

121

region with contaminant sources is represented below by the fine cross-hatched pore

122

network with aligned round NAPL source zones. Figure 1c shows a closer view of the

123

fine pore network with one round NAPL source zone. The dimensions of the

124

microchannels and grids are typical of pores and soil grains within a low permeable layer.

125

The round NAPL source zones were etched into a separate borofloat glass layer, which

126

was distinct from the pore network, as shown in Figure 1d. The channel depth was

127

chosen to be 20 µm, so that the channel ceiling did not interfere with bacterial motility.41

128

Note that while Molaei and coworkers found that the cell-boundary interactions can trap

129

E. coli on channel surfaces, and thus suppress its chemotactic ability;42 P. putida tends to

6

ACS Paragon Plus Environment

Page 6 of 31

Page 7 of 31

Environmental Science & Technology

130

reverse direction between runs and, unlike E. coli, it exhibits relatively fast association

131

and dissociation with solid surfaces.43 Thus, any interaction with the surface was

132

considered to be negligible for P. putida. The NAPL source zones had a depth of 100 µm

133

to retain the toluene droplets. The microfabrication process and the thermal bonding

134

method to assemble the two etched glass layers were well described by Roper and

135

coworkers.44

136

DI water was injected into the microfluidic device from both the bacterial and

137

NAPL inlets, in order to saturate the system. Air bubbles were eliminated by continuous

138

high-pressure water flow and the microfluidic system was kept saturated during the entire

139

experiment. Toluene stained with Oil-Red-O was injected from the NAPL inlet to replace

140

the water in the fine pore network, and then DI water was injected again to wash out the

141

toluene in the fine pore network, while toluene droplets remained in all the NAPL source

142

zones, as shown in Figure 1b. The entrapment of toluene droplets was facilitated by the

143

different physical properties between glass that is hydrophilic and toluene that is

144

hydrophobic. Therefore, toluene exhibits a high contact angle in the water-saturated

145

spaces with glass walls and keeps a round intact shape. When water pushes toluene out of

146

the pore network, it flows along the glass surface because both water and glass are

147

hydrophilic, so the residual toluene in the pores can be carried away easily by water flow

148

due to the their poor attachment to the glass surface; however, the toluene droplets are

149

difficult to wash out, because surface tension at the openings to the pore network

150

prevents the toluene droplets from floating out of NAPL source zones, and also there is

151

not enough pressure from the flow to push the droplets out. More details about the

152

mechanism of NAPL trapping are available in the Supporting Information. After NAPL

7

ACS Paragon Plus Environment

Environmental Science & Technology

153

droplets were trapped in designated locations, the NAPL inlet was closed by a solid pin,

154

and saturated with 10% RMB to prevent the introduction of air bubbles into the system

155

from the inlet.

156

The mixed bacterial suspension was injected into the macrochannel using 100 µL

157

point-style syringes (Spectrum Laboratory Products Inc), and then the syringes were

158

loaded onto a syringe pump (PHD 2000 Infusion, Harvard Apparatus) and pushed at a

159

constant flow rate of 100 µL/h, which was comparable to 40 m/d in the macrochannel.

160

After 30 min of continuous injection, the experimental system was saturated with the

161

bacterial suspension.

162

The Wide-Field epifluorescence microscope (Olympus IX-70) was used for all

163

experimental observation and image acquisition with the 2×/0.05 (Olympus, Plan) and

164

20×/0.40 objectives (Olympus, PlanFI). An arc lamp and dichroic filter (Chroma

165

51001bs, double coated for TRITC and DAPI) were used to look at the FM 4-64

166

(excitation at 515 nm, emission at 640 nm) and DAPI (excitation at 345 nm, emission at

167

455 nm) stained bacterial strains.

168

Data Acquisition and Processing. Three independent studies were carried out

169

under different operating conditions, including steady, transient, and no flow conditions

170

with continuous or step change inputs, and at different observation locations, as shown in

171

Figure 1b. The first study was conducted at steady state to examine the bacterial

172

distribution at the interface between the macrochannel and the pore network (which

173

contained toluene droplets) at different flow rates of 0.5 and 5 m/d. These two values of

174

flow velocity were chosen because they were within the typical range of groundwater

175

flow rates from 0.1 to 10 m/d,45 and were representative of slow and fast flows,

8

ACS Paragon Plus Environment

Page 8 of 31

Page 9 of 31

Environmental Science & Technology

176

respectively. Images of bacterial accumulation based on gray-level light intensity (gray

177

scale) were taken at eight evenly distributed positions in the macrochannel near the

178

adjacent porous network, as shown by the red dashed box marked with “I” in Figure 1b.

179

Each of the eight observation positions covered an area with a distance range of 0 to 100

180

µm from the junction and a width of 100 µm; the eight positions were numbered from 1

181

to 8 along the flow direction to facilitate identification of the plotted data. The details of

182

image processing were explained in our previous study.38 The second study took place

183

under no-flow conditions; after 10 minutes of the quiescent status, observations of

184

bacterial distributions were recorded in the vicinity of randomly selected toluene droplets.

185

Thirty images were taken in different microchannels of the pore network near the edge of

186

the droplets, and then were averaged via the function of Image Calculator in ImageJ to

187

reduce data fluctuation and produce a more continuous bacterial distribution. The third

188

study, conducted under transient flow conditions, was to compare the washout of a

189

mixture of chemotactic and nonchemotactic bacteria at two different flow rates, 0.25 and

190

5 m/d. Once the pore network was saturated with the bacterial mixture, the injectate fluid

191

was changed from bacteria to 10% random motility buffer (RMB), and the speed was

192

reduced to either 0.25 or 5 m/day to wash the bacteria out of the microfluidic device over

193

a period of 18 continuous hours. Bacterial density, which was correlated as fluorescence

194

intensity and normalized with initial intensity in the macrochannel at a certain speed, was

195

recorded at the outlet of the macrochannel at different time points. Image analysis for all

196

of the three studies followed the same protocol in Wang et al., 2012.

197

Two-way ANOVA tests were used to determine statistical differences in bacterial

198

distribution among various experimental data. Averages and standard deviations were

9

ACS Paragon Plus Environment

Environmental Science & Technology

Page 10 of 31

199

calculated based on three independent experimental runs for each data set. The tests were

200

evaluated in Prism 5.0a (GraphPad Software Inc., San Diego, CA).

201

Mathematical Modeling. Because the height of the macrochannel was much less

202

than its width and the fluid velocity in the pore network was slower than the bacterial

203

swimming speed (i.e. bacterial movement dominated transport within the pore network) a

204

two-dimensional Brinkman equation (Equation 1),  u + ∇ ∙ − (∇u+(∇u) ) + I = − u     ∇∙u=0

(1) (2)

205

was suitable to describe the flow profile in the porous microfluidic device where u

206

denotes the flow velocity vector (m/s), ρ the fluid density (kg/m3), εp the porosity

207

(dimensionless), κbr the permeability of the porous medium (m2), µ the dynamic viscosity

208

(Pa.s), and p the pressure (Pa). The boundary conditions are no slip conditions except at

209

the inlet which is laminar flow at a certain flow rate and at the outlet which has no

210

viscous stress (pout = 0). Equation 2 is the continuity equation under the assumption of an

211

incompressible fluid. A two-dimensional form of the advection-dispersion governing

212

equation was adopted to represent bacterial transport including the effect of chemotaxis.

213

The advection-dispersion governing equations that follow were modified from Olson et

214

al., 46           = ,  + ,  − −     

(3)

  !"       !"  = ,  + ,  − − − −        215

where

10

ACS Paragon Plus Environment

(4)

Page 11 of 31

Environmental Science & Technology

, =  

(5)

!" =

(6)

2 +, -.   tanh * / 3 2 (-. + ) 

2 +, -.  !" =  tanh * / 3 2 (-. + ) 

(7)

216

where a refers to the attractant concentration, b the bacterial concentration, vx and vy the

217

longitudinal and transverse fluid velocity solved from the two-dimensional Brinkman

218

equation, Da the diffusion coefficient for attractant, Db the bacterial random motility, De,a

219

and De,b are their effective values. vChx and vChy are the longitudinal and transverse

220

chemotactic velocities describing the advective transport of bacteria caused by

221

chemotaxis toward attractant gradients, χ0 the chemotactic sensitivity coefficient and Kc

222

the chemotactic receptor constant. Bacterial adaptation is not accounted for explicitly in

223

Equations 6 and 7 because of the shallow and steady state chemical gradients in the

224

experimental system. We chose “no flux” as the overall boundary conditions, and then

225

applied a fixed concentration of attractant/bacteria to the inlets, as well as the NAPL

226

source zones, and “Convective Flux” to the outlet.

227

The commercial finite-element solver COMSOL Multiphysics 4.3 (COMSOL,

228

Burlington, MA) was used to solve the system of differential equations for the fluid

229

velocity and the attractant and bacterial concentrations. Three two-dimensional modules

230

were included in this simulation: one Incompressible Free and Porous Media Flow

231

module for the advective flow profile, and two Species Transport in Porous Media

232

modules for attractant and bacteria. Input parameters included the initial toluene

233

concentration a0 =5.65 mol/m3, attractant diffusion coefficient Da = 9.5 × 10-10 m2/s,

234

initial bacterial density b0 =1 mol/m3, bacterial swimming speed vb = 44 µm/s,14 bacterial

11

ACS Paragon Plus Environment

Environmental Science & Technology

235

random motility coefficient Db = 13 × 10-10 m2/s,14 chemotactic sensitivity coefficient χ0

236

= 1.8 × 10-8 m2/s,40 chemotactic receptor constant Kc = 1.0 mM,40 the calculated pore

237

network porosity εp = 0.43 and permeability κbr = 1.0 × 10-11 m2. κbr is the only fitting

238

parameter, while all other parameters are either calculated in this study or reported in the

239

literature.

240

Results and Discussion

241

The impact of chemotaxis was observed in three facets of the heterogeneous

242

environment of the microfluidic device. First, chemotactic bacteria in the macrochannel

243

accumulated at higher density near the connections to the pore network, which contained

244

the toluene droplets. Second, chemotactic bacteria within the pore network accumulated

245

in greater number near the vicinity of the toluene droplets. Third, chemotactic bacteria

246

were retained to a greater extent within the pore network than nonchemotactic bacteria at

247

a fluid velocity of 0.25 m/d. This effect was reduced at a higher fluid velocity of 5.0 m/d.

248

Bacterial Accumulation in the Macrochannel near the Pore Network

249

In Figure 2 bacterial abundance at the interface of the macrochannel with the pore

250

network is plotted at eight positions along the length of the macrochannel. The bacterial

251

abundance is normalized with the bulk concentration near the inlet in the macrochannel;

252

for nonchemotactic bacteria, as expected, the normalized concentration was near unity

253

(open triangles and dashed line) at 0.5 m/d as shown in Figure 2a. At a fluid velocity

254

typical of natural groundwater flow rates 0.5 m/d, we observed an average 7.3% increase

255

in the chemotactic bacterial accumulation at the eight positions (black squares and black

256

line). When we increased the fluid velocity by an order of magnitude, the accumulation

257

of chemotactic bacteria decreased (gray circles and gray line in Figure 2b) and was

12

ACS Paragon Plus Environment

Page 12 of 31

Page 13 of 31

Environmental Science & Technology

258

comparable to the control case with nonchemotactic bacteria. The solid lines connect

259

simulation results from solving the governing equations for mass and momentum transfer

260

using COMSOL software packages.

261

The effect of fluid velocity on chemotactic accumulation in Figure 2b was

262

consistent with the findings for a microfluidic device with a macrochannel connected to a

263

single pore filled with toluene;38 Wang and coworkers reported that bacterial

264

accumulation transverse to flow was observed in the vicinity of a NAPL interface at a

265

velocity of 0.5 m/d, but the accumulation decreased dramatically when the flow rate was

266

increased to 5 m/d. The slower advective flow rate resulted in a greater retention time,

267

which allowed the chemotactic bacteria time to respond to the attractant source in the

268

transverse direction.

269

The experimental data of chemotactic distribution at 0.5 m/d were compared

270

statistically with the nonchemotactic group at the same flow rate and chemotactic group

271

at 5 m/d by a two-way ANOVA test. The results showed that the chemotaxis group at 0.5

272

m/d was significantly different from either the nonchemotaxis control or the chemotaxis

273

group at 5 m/d, with a p-value less than 0.0001 in both cases. Such a difference was not

274

found when the nonchemotaxis group was compared with the 5 m/d group with p =

275

0.1492, which implies that transport by advection dominated microbial motility and

276

chemotaxis at the higher flow rate.

277

These results suggested that the chemotactic bacteria carried along with the flow

278

in the macrochannel sensed the concentration gradient from the NAPL attractant trapped

279

in the low permeable region (pore network), and accumulated in the vicinity of the

280

positions between these two regions, which facilitated the movement of bacteria into the

13

ACS Paragon Plus Environment

Environmental Science & Technology

281

pore network, to facilitate contact with the contaminant sources. Because the rate of

282

biodegradation is proportional to the bacterial population density,47 we expect this

283

accumulation of bacteria will ultimately enhance the biodegradation rate.

284

Bacterial Accumulation near the Toluene Sources

285

Figures 3 and 4 show the bacterial distribution within pores at the edge of toluene

286

droplets under quiescent conditions. The schematic in Figure 3a illustrates the expected

287

bacterial distribution in a microchannel near one droplet. Figure 3b shows a single field-

288

of-view image cropped for a single channel of the actual bacterial distribution and Figure

289

3c represents an average of 30 images from multiple fields-of-view taken at different

290

locations around different droplets under 20×/0.40 objective. The values of bacterial

291

density in Figure 4 were normalized to the initial value in the macrochannel at the point

292

of injection. Accumulation of P. putida F1 increased as the distance from the toluene

293

source decreased, while nonchemotactic bacteria were distributed evenly over the same

294

distances and no accumulation was evident. This indicated that chemotactic bacteria

295

responded to the concentration gradient of the attractant and migrated toward the regions

296

where the toluene concentration was higher, and conversely, the nonchemotactic mutant

297

did not present any observable response to the attractant gradient. Approximately 15%

298

more chemotactic bacteria accumulated at the visual edge of the toluene sources than at a

299

distance of 500 µm away from the sources. The differences in the distribution between

300

the wild type and the mutant were statistically significant in the range of 0-300 µm, with

301

p < 0.0001 by the two-way ANOVA. The results of both chemotactic accumulation and

302

statistical analysis were consistent with what was reported from a previous study using a

303

single-pore device.38 Note that the previous study measured the chemotaxis effect near

14

ACS Paragon Plus Environment

Page 14 of 31

Page 15 of 31

Environmental Science & Technology

304

one single pore slit, while this study successfully captured similar results in a porous

305

structure that is more representative of a complex groundwater system with distributed

306

contaminant sources.

307

Bacterial Retention within the Pore Network

308

In the third experiment we monitored bacterial concentration in the effluent from

309

the macrochannel as a breakthrough curve. To conceptualize the expected output, we

310

compared it to an ideal scenario in which the heterogeneous µchip was represented as

311

two flow paths analogous to electrical resistances in parallel. The macrochannel

312

represented the lower resistance path, essentially a plug flow situation, with the buffer

313

solution replacing the bacterial solution to yield a step-change in concentration. Within

314

the fine pore network the resistance to flow was greater and the time frame over which

315

the bacterial suspension in the pore network was replaced by buffer was much longer

316

with a more gradual change in concentration.

317

Bacterial concentration in the effluent was monitored as a function of time and

318

was scaled with the steady state concentration prior to the step-change of the influent to

319

buffer. For the case at the lower fluid velocity of 0.25 m/d, we observed a sharp drop in

320

concentration around the 2-hour time point and then a more gradual decrease in the

321

concentration over time with measurable concentrations continuing up to 18 hours for

322

both the chemotactic and nonchemotactic bacteria (Figure 5). The 2-hour time point

323

corresponded to the time needed to replace the bacterial suspension in the 2-cm long

324

macrochannel with buffer flowing at a velocity of 0.25 m/d. For the chemotactic bacteria,

325

the concentration was significantly less than the nonchemotactic control at all the time

326

points greater than 2 hours. The difference in concentration between chemotactic and

15

ACS Paragon Plus Environment

Environmental Science & Technology

327

nonchemotactic bacteria in the effluent was comparable to the simulations from the

328

mathematical models and was shown to be significant by a two-way ANOVA with p