Deposition Dynamics of Rod-shaped Colloids during Transport in

Feb 5, 2018 - A 3D computational modeling study of the deposition dynamics of rod-shaped colloids during transport in porous media under favorable con...
0 downloads 7 Views 2MB Size
Subscriber access provided by READING UNIV

Article

Deposition Dynamics of Rod-shaped Colloids during Transport in Porous Media under Favorable Conditions ke li, and Huilian Ma Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.7b03983 • Publication Date (Web): 05 Feb 2018 Downloaded from http://pubs.acs.org on February 6, 2018

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 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 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.

Langmuir 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 56 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

1

Deposition Dynamics of Rod-shaped Colloids during Transport in

2

Porous Media under Favorable Conditions

3 4

Ke Li, Huilian Ma∗

5

Department of Geology and Geophysics

6

University of Utah, Salt Lake City, UT 84112

7 8

Abstract

9

A 3D computational modeling study of the deposition dynamics of rod-shaped colloids during

10

transport in porous media under favorable conditions (no energy barrier to deposition) was

11

presented. The objective was to explore the influences of particle shape on colloid transport

12

and retention. During the simulation, both translation and rotation of ellipsoidal particles were

13

tracked and evaluated based on an analysis of all the forces and torques acting on the particle.

14

We observed that shape was a key factor affecting colloid transport and attachment. Rod

15

particles exhibited enhanced retention compared with spheres of equivalent volume at size

16

range greater than ~ 200 nm. The shape effect was most pronounced for particles around 200

17

nm to 1 µm under simulated conditions. Shape effect was also strongly dependent upon fluid

18

velocity; it was most significant at high velocity, but not so at very low velocity. The above-

19

described shape effect on retention was directly related to particle rotation dynamics, due to

20

the coupled effects from rotational diffusion and flow hydrodynamics. Rotational diffusion

21

changed particle orientation randomly, which caused rod particles to drift considerably across



Corresponding author. E-mail: huilian. [email protected]; Tel: (801)585-5976; Fax: (801)581-7065.

1 ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

22

flow streamlines for attachment at size range 200 nm ~ 1µm. The hydrodynamic effect

23

induced periodic particle rotation and oscillation, which rendered large size rod particles to

24

behave like “spinning bodies” prescribed by their long axes so to easily intercept with

25

collector surface for retention. Our findings demonstrated that the practice of using equivalent

26

spheres to approximate rods is inadequate in predicting the transport fate and adhesion

27

dynamics of rod-shaped colloids in porous media.

28 29

1. Introduction

30

The transport and deposition of colloids in saturated porous media is essential to manage and

31

remediate various environmental processes. For example, an accurate understanding of

32

colloid transport is necessary to assess surfactant loss in enhanced oil recovery,1 to track

33

contaminant fate,2 and especially to produce safe drinking water.3 Almost all colloids in the

34

environment are non-spherical in shape, such as clay particles, different types of bacteria (e.g.

35

Bacilli), human blood cells, polymer molecules, agglomerates and asbestos particles.

36

However, existing filtration theories often simplify natural colloids as spheres of certain

37

effective diameters, which have been demonstrated by increasing evidence4-9 to be insufficient

38

in describing their transport and retention behaviors.

39 40

Unlike spheres in a force field, for each different orientation, the forces acting on non-

41

spherical particles are different. Various researchers have proved that ellipsoidal particles

42

have different transport and retention behaviors compared with spheres.10-11 For example, rod-

43

shaped colloids exhibited elevated retention in column experiments compared to their 2 ACS Paragon Plus Environment

Page 2 of 56

Page 3 of 56 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

44

equivalent-volume spheres under otherwise identical conditions.12-13 A few studies also

45

showed that the non-spherical colloid shape affected not only the extent, but the pattern of

46

colloidal retention.10,

47

straining kinetics and found that as those particles approached pore-space constrictions, a

48

preferred orientation was adopted and theirs major axes tended to align with the local flow

49

direction. Yoo et al.17 used synthesized PLGA particles with spherical and elliptical disk shape

50

to investigate the rate of particle endocytosis and their distribution in endothelial cells. They

51

found that elliptical disks were endocytosed slower compared to spheres, and that disks

52

oriented tangentially along the nuclear membrane. Zhao et al.10,

53

simulation work of ellipsoids that those particles aligned in the stream-wise direction in the

54

near-wall region during channel flow, and that the particles affected the flow by altering the

55

turbulence structure near and around the kinetic energy peak. In the process of particles

56

depositing at the edge of evaporating suspension, Yunker et al.15 found that spheres deposited

57

randomly with surface diffusion, while ellipsoids induced strong capillary attraction on the

58

air-water interface and the deposition appeared to be ‘sticky’, i.e., clustering and creating void

59

regions and particle-rich regions. The contrasting behaviors between non-spherical and

60

spherical particles reflect different mechanisms by which non-spherical and spherical particles

61

transport and attach.

14-17

Xu et al.16 explored the effects of peanut-shaped particles on

14

concluded in their

62

63

Jeffrey (1922) predicted that rigid ellipsoidal particles tended to rotate with a certain period in

64

a linear shear field.18 This prediction was validated by Zia via his experiments,19 and has been

65

generally accepted. Although spheres also rotated in Jeffrey’s prediction, homogeneous 3 ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

66

spheres tended to follow streamlines under conditions when external forces or Brownian

67

diffusion were absent.18,

68

rotation of non-spherical particles, especially inertial particles, may also affect significantly

69

its translational motion.20 Broday21 calculated that inertial particles in vertical shear flow

70

migrated across the streamlines. Huang and Joseph22 also showed that elliptical cylinders

71

settling vertically in a channel drifted laterally due to the orientation-dependent hydrodynamic

72

interactions. These pioneer studies into shape effect introduced a valuable perspective on non-

73

spherical particle transport and retention. However, the above-mentioned studies did not

74

consider diffusion process, which is especially important for micro- and nano-meter size

75

ranged particles.

20

However, this is not the case for non-spherical particles. The

76 77

With respect to homogeneous sphere transport, Brownian rotation may be irrelevant; hence, it

78

is generally overlooked in general filtration theory. In contrast, rotational diffusion is a

79

fundamental characteristic describing the tumbling of natural anisotropic (non-spherical)

80

particles in solution. Rotational diffusion (or angular diffusion) influences non-spherical

81

particle transport through the orientation dependency of drag forces.23-24 Dresher et al.25

82

observed in their study that rotational diffusion diminished the hydrodynamic interactions

83

between two E. coli bacteria and altered their travelling paths. Kuzhir et al.11 found that the

84

magnetically induced rotational diffusion process caused the misalignment of particle

85

aggregates from the streamlines, which further illustrated that the long-range non-

86

hydrodynamic interactions of rotary diffusion process was also indispensable in studying

87

spheroid transport phenomenon. However, most of these studies were limited to particle sizes 4 ACS Paragon Plus Environment

Page 4 of 56

Page 5 of 56 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

88

in the Brownian regime, and they did not consider the influences of hydrodynamic

89

interactions induced by non-spherical shape.

90 91

Realizing the unique physical properties of non-spherical particles that can regulate transport

92

process, increasing attention has been attracted to the utilization of such particles in

93

biomedical field.26-28 For instances, many studies in drug-delivery systems showed that non-

94

spherical particles (e.g., rod-shaped or disc-shaped) exhibited better migration propensity or

95

adhesion onto blood vessel walls than spheres (including nano-spheres).29-35 But most of these

96

investigations, experimental or simulation, were carried out in parallel plate flow chambers

97

(PPFC).32-35 In contrast, studies on the influences of particle shape on transport and retention

98

in colloidal filtration field have been very limited.5, 36-37 Those few existing studies were all

99

from experimental investigations; and to the best knowledge of the authors, there is no

100

simulation work available yet from literature on non-spherical particle transport in porous

101

media.

102 103

In this work, we develop a three-dimensional particle transport model that incorporates the

104

non-spherical shape (e.g., rod-shaped particles as our model colloids here). This model will

105

simulate both the translation and orientation of rod-shaped particles in representative flow

106

fields based on an analysis of all the forces and torques acting the particle. From this

107

trajectory model analysis, we will obtain: i) 3D translation and rotation trajectory of

108

individual particles; ii) retention probability and distribution; iii) attachment orientation; iv)

109

transient particle position distribution; v) transient linear and angular velocities. Simulating 5 ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

110

both translation and rotation is very challenging, because non-spherical colloids may have

111

several different axes of rotation and also because its translation may depend upon its real-

112

time orientation. That is probably why quite a number of existing numerical studies on non-

113

spherical particle transport from biomedical field chose to simplify the problem with a 2D

114

model.20, 38-40 However, due to the complexities involved, we have two main assumptions in

115

our model that we believe are reasonable: i) since what we obtain here was clean single

116

collector efficiency, particles injected into our model systems were of dilute concentration

117

such that particle-particle collisions and aggregate formations could be neglected; ii) fluid

118

field disturbance due to the presence of particles was ignored. The model thus developed for

119

non-spherical particle transport can be modified and applied to various different geometric

120

flow models, because the essential simulation methodology stays the same. In this work, we

121

will present results obtained during our systematic investigations of rod-shaped colloids in

122

Happel sphere-in-cell model (i.e., a well-known model to represent porous media in classic

123

colloid filtration theory) under favorable attachment conditions (i.e., particle-collector

124

repulsion is absent). The simulation parameters spanned a wide range of ellipsoidal particle

125

sizes and shapes flowing under different pore water velocity. The main goals from these

126

simulations are to answer the following questions: i) under what conditions, non-spherical

127

particles can be treated as “effective spheres” and under what conditions, they can’t; ii) how

128

the shape of a particle affects its transport and retention? By what mechanisms?

129

130

2. Simulating Transport of Ellipsoidal Particles

131

Particle Transport Model. A Lagrangian trajectory approach was developed in this work to 6 ACS Paragon Plus Environment

Page 6 of 56

Page 7 of 56 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

132

simulate the time evolution of particle translation and rotation in representative porous media,

133

e.g., the well-known Happel sphere-in-cell model. The Happel model represents porous media

134

as a collection of spherical collectors each enclosed in a concentric spherical fluid shell. The

135

flow field within the Happel model can be described analytically.41-42 The general equations

136

of motion to account for the translation and rotation of a rigid particle are given by

137

138

139 140

  = ∑ 

 



=  

 ∙ = ∑      =   

where t is time, m is the mass of the particle,  is the moment of inertia of the particle; x and u

are the linear position and velocity vectors of the particle, respectively; Ω and  are the

141

respective angular position and velocity vectors of the particle; and ∑  and ∑  represent the

142

total forces and torques acting on the particle, respectively. Our previous work simulated the

143

linear translations (i.e., eqs 1-2) of spherical particles in representative porous media models

144

extensively,4, 43-47 but did not explicitly account for particle orientation excepting converting

145

rotational motion of a sphere to linear translation.48 Here, we will focus on how to incorporate

146

both the rotational and linear motions of particles into simulation, using prolate ellipsoid

147

particles as model colloids. Due to symmetry for ellipsoids, the coupling tensor between

148

translation and rotation in eqs 1 and 3 vanishes based on Brenner’s research.49 Also in this work,

149

we treat Brownian motions differently from our previous work, where Brownian forces were

150

included into eq 1 to account for Brownian translation but Brownian rotation was neglected.

151

Here, we first determined the linear and rotational displacements from the total forces and

152

torques in eqs 1 and 3 that included only the deterministic forces and torques (i.e., excluding 7 ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 56

153

those from Brownian motions). We then added Brownian translation and rotation, which were

154

modeled in each degree of freedom as a stochastic process using Einstein’s equations, to those

155

resulted from the deterministic motions, respectively. Below, we shall describe in details how

156

to numerically solve the above equations of motion for an ellipsoid moving in shear flow.

157 158

Coordinate Systems. In order to describe the three-dimensional rotation and translation of

159

ellipsoidal particles, the following coordinate systems were employed (Figure 1): i) an inertial

160

frame of reference, = x, y, z, with its origin located at the collector (e.g., the origin of the

161

Happel sphere-in-cell model); ii) a particle frame,  = x, y, z, with its origin at the center-of-

162

mass of the particle and its axes being the particle principal axes; and iii) a co-moving frame,

163

 = x, , z, with its origin at the center-of-mass of the particle and its axes parallel to the

164

corresponding axes of the inertial frame. The translational motions of the ellipsoid were

165

described by tracking the center-of-mass of the particle in the inertial frame. The rotational

166

motions were described by following the three Euler angles (Figure 1) or the four Euler

167

parameters (i.e., quaternions) in the particle frame.

168 169

Transformation of the coordinates between the particle and co-moving frames is given by 50-51

 =  ,

170

(5)

171

where the orthonormal transformation matrix A can be expressed in terms of either Euler

172

angles or the four Euler parameters (i.e., q0, q1, q2, q3). Below, matrix A is given using Euler

173

parameters as

8 ACS Paragon Plus Environment

Page 9 of 56 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

1 − 2 $%% + $'% 2 $( $% + $' $) 2 $( $' − $% $)  = 2 $% $( − $' $) 1 − 2 $'% + $(% 2 $% $' + $( $) *. 2 $' $( + $% $) 2 $' $% − $( $) 1 − 2 $(% + $%%

174

(6)

175

The four Euler parameters are subject to the following constraint to preserve the three degrees

176

of freedom for a rotating rigid body: $)% + $(% + $%% + $'% = 1.

177 178 179 180

(7)

Euler angles are related to the Euler parameters as $) = cos .

/01 3 2 cos . % 2 , $( %

/01 3 2 cos . % 2. %

sin .

= cos .

/41 3 2 sin .% 2, $% %

= sin .

/41 3 2 sin .% 2, $' %

=

(8)

181

The ranges for Euler angles used in this study are ϕ ∈[-π,π], θ ∈[0,π], and ψ ∈[-π,π].

182

Conversion from Euler parameters to Euler angles is provided in the Supporting Information.

183 184

Rotation Dynamics of an Ellipsoid in Shear Flow. In this work, we chose the Euler four

185

parameters to describe the time evolution of particle angular displacements, because the use

186

of Euler angles would involve a large number of trigonometric calculations, which are

187

computationally expensive, and also because there is a singularity in Euler angles (see

188

Supporting Information). However, Euler angles were used to assign the initial orientation of

189

a particle (most of simulation data here with random initial orientations). The initial Euler

190

parameters were then determined from eq 8. At the subsequent time steps, the Euler

191

parameters evolved with time as51

192

193 194

−$(