Elbing 2016

Subscriber access provided by CMU Libraries - http://library.cmich.
edu
Review
Review of Bubble Column Reactors with Vibration
Brian R. Elbing, Adam L. Still, and Afshin Ghajar
Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.5b02535 • Publication Date (Web): 22 Dec 2015
Downloaded from http://pubs.acs.org on December 26, 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.
Industrial & Engineering Chemistry Research 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 59 Industrial & Engineering Chemistry Research
1
2
3
4 Review of Bubble Column Reactors with Vibration
5
6 Brian R. Elbing*, Adam L. Still† & Afshin J. Ghajar
7
8 Mechanical & Aerospace Engineering, Oklahoma State University, 218 Engineering
9 North, Stillwater, OK, 74078, United States of America
10
11
12
13
Abstract
14
15
16 Vibrating a bubble column reactor can increase the gas holdup (void fraction) as well the
17
18 mass transfer rate. Since the seminal work in the 1960’s, there has been minimal effort focused
19
20 on this topic until the early 2000’s. Currently there are several groups studying this problem
21
22
23 making advancements in our fundamental understanding of the process with detailed
24
25 experiments, theoretical analyses and physics based models. However, throughout the literature
26
27 there are inconsistencies with both experimental results and proposed scaling of the fundamental
28
29
30 properties as well as minimal data spanning the parameter space. This review serves as an
31
32 overview of key works from the 1960’s and the 2000’s as well as to identify these
33
34
35
inconsistencies between key studies. Recommendations for how to proceed with future work is
36
37 provided with an emphasis on defining the parameter space in terms of the Reynolds number and
38
39 Froude number.
40
41
42 Keywords: bubble column, multiphase flow, fluid mechanics, void fraction, mass transfer
43
44
45
46
47
48
49
50
51
52
53
54
55
56 *
57 Author to whom correspondence should be addressed: elbing@okstate.edu
†
58 A.L. Still is currently at Sandia National Laboratory, Albuquerque, NM, USA.
59
60
ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research Page 2 of 59
Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

1
2
3
4 1 Introduction
5
6
7
8 1.1 Motivation
9
10
11 Numerous real-world systems involve multiphase flow, which produce complex flow
12
13 patterns that are dependent on the balance of forces and phase distribution. The control of these
14
15 flows for engineered systems, such as piping networks for the petroleum industry, require a
16
17
18 complete understanding of the phase interactions with the boundary conditions and the other
19
20 phase(s). However, for many of these flow-fields there is relative velocity between phases and
21
22
nonhomogeneous distributions that prevent the governing equations from being solved as a
23
24
25 mixture with average fluid properties. Consequently, each phase has to be solved individually,
26
27 which requires an understanding of the mass, force and energy interaction terms between the
28
29
30 phases. Currently these relationships are unknown, which forces researchers to form heuristic
31
32 models for a specific flow pattern.1 For this reason, the study of multiphase flows are typically
33
34 confined to specific flow configurations and rely heavily on experimental data and empirical
35
36
37 modeling. The current review focuses on a bubble column reactor (BCR) with vibration, which
38
39 involves two phases interacting with each other to dissolve the dispersed phase (gas) into the
40
41 continuous phase (liquid mixture).
42
43
44 BCRs are widely used throughout the chemical, biochemical and petrochemical
45
46 industries due to their simple design, low cost, compactness, ease of operation and high
47
48
49
heat/mass transfer rates. Some BCR applications include waste-water treatment, aeration of
50
51 organic organisms in bio-reactors, solvent gasification for chemical reactions and hydrogenation
52
53 of coal-slurries to produce synthetic fuels. The last example has gained significant attention
54
55
56 recently due to several military and civilian aircraft being qualified to fly with synthetic fuel
57
58
59
60 2

1
2
3
blends, which are produced with the Fischer-Tropsch process. United Airlines ran the first flight
4
5
6 demonstration of a commercial airliner in the United States using synthetic fuel blends with an
7
8 Airbus A319 in April of 2010.2 Currently there are several United States Air Force fighters and
9
10
11
aircraft (e.g. A-10, C-17, KC-135 and F-22) approved to fly with synthetic fuel.3-6 This example,
12
13 as well as numerous others related to chemical processing, highlight the ever increasing use and
14
15 application of BCRs in our society. Thus there is also an increasing interest in the ability to
16
17
18 control the physical processes within the reactor. While the primary limiting factor in Fischer-
19
20 Tropsch process example is not the mass transfer rate,7 mass transfer rate and gas holdup are
21
22 critical parameters for any BCR application.
23
24
25 It was discovered in the early 1960’s that vibrating a BCR could increase the mass
26
27 transfer rate and gas holdup. While some additional research expanded the theory, minimal
28
29
research effort focused on this phenomena until the early 2000’s. A summary of the primary
30
31
32 studies spanning this range are provided in Table 1 along with the primary measurements from
33
34 each study.8-23 Recent research has gone so far as to develop theoretical, physics based models to
35
36
37 predict mass transfer and void fraction in these systems. While there is a nontrivial body of work
38
39 on vibrating BCRs, to date the parameter space has been explored in a rather haphazard fashion.
40
41 With numerous studies available there is a need for a systematic, dimensionally reasoned
42
43
44 analysis of the available data to identify specific flow regimes and the flow physics that
45
46 dominate the given range of conditions. The available models have been tested against limited
47
48 data, but there are inconsistencies with both experimental results as well as the proposed scaling
49
50
51 of the fundamental properties. In addition, the large gap in time between studies provide
52
53 motivation to reexamine these past studies due to the significant advancements in experimental
54
55
56
and computational tools. This review will provide a broad overview of previous studies, identify
57
58
59
60 3

1
2
3
key findings that have established our current knowledge and note inconsistencies in the
4
5
6 literature that need further examination.
7
8
9 1.2 Background
10
11
12 It is beneficial to first review some basic physical properties and common definitions
13
14
15 used studying vibrating BCRs. From an industrial point of view, the chemical reaction rate is
16
17 typically the most important property. It is often assumed to be proportional to the volumetric
18
19 mass transfer coefficient (kLa), which is the product of the mass transfer coefficient (kL) and the
20
21
22 interfacial surface area per unit mixture volume (a). For simplicity, the volumetric mass transfer
23
24 coefficient (kLa) is generally referred to as the mass transfer rate. For BCR’s as well as other
25
26 aerated systems, kLa is used to relate the time rate of change of the concentration (dC/dt) to a
27
28
29 concentration potential,
30
( )
31 dC
32 = kLa C * − C , (1)
33 dt
34
35 where C is the concentration of gas dissolved in the liquid and C* is the saturation concentration.
36
37 If spherical bubbles are assumed, then the interfacial bubble area per unit volume (a) can
38
39
40 be determined from the mean bubble diameter (db) and the void fraction or gas holdup (ε),
41
42 a = 6ε d b . Here void fraction is defined as the ratio of gas phase volume (Vg) to the total
43
44
45 volume (Vg + Vl) in a multiphase system,
46
47 ε = V g (V l + V g ) .
48
49
50 It should be noted that this estimate of a is an approximation used by some to simplify large
51
52 reactor systems or for the purpose of modeling.20,22,24 The validity of this approximation is
53
54
55 primarily dependent on the bubble size, which when db < 2 mm it is general assumed valid.25 At
56
57 this size capillary effects typically dominate the governing dynamics. Experimental observations
58
59
60 4

1
2
3
4
of bubbles in vibrating BCR show that typically db < 5 mm due to hydrodynamic breakup,8
5
6 which is smaller than that observed without vibration. While the average diameter indicates that
7
8
9
the bubbles are nearly spherical, in most instances the bubble surface geometry constantly
10
11 changes as it is formed and rises.26-28 These surface oscillations affect both the bubble volume
12
13 and shape, but it is generally noted that the shape oscillations primarily influence mass transfer at
14
15
16 the interface.29
17
18 While it is desirable to use a single characteristic length scale for the bubble diameter,
19
20 there are inherent variations in the distribution of sizes and shapes that prevent a universally
21
22
23 accepted length scale. One method measures the bubble chord length probability distribution and
24
25 use the resulting mean diameter as the characteristic length.30-32 The advantage of this method is
26
27
that it does not rely on the spherical bubble assumption, and it produces an accurate measure of
28
29
30 the bubble size variation. However, many researchers16,22,33 prefer to use the Sauter mean
31
32 diameter (d32), which is the ratio of the representative bubble volume to the bubble surface area,
33
34
35 n
36 ∑n d i
3
b ,i
37 d 32 = i =1
n . (2)
38
39 ∑n d
i =1
i
2
b ,i
40
41 Here ni is the number of bubbles with size db,i. The Sauter mean diameter is commonly used
42
43
44 when bubble size is measured from 2D bubble images. Here the projected bubble area (Aproj) in
45
46 each 2D image is used to estimate an equivalent bubble diameter (deq) assuming a spherical
47
48
49 bubble,16 d eq = 4 Aproj π . The equivalent diameter is used as the estimated bubble diameter
50
51
52
(db,i) in Eq. (2) to calculate the mean Sauter bubble diameter for a given condition.
53
54 The gas phase velocity is also a critical measurement, as this determines the bubble
55
56 residence time. It is typically reported as the bubble velocity (Ub), gas phase velocity (Ug) and/or
57
58
59
60 5

1
2
3
superficial gas velocity (Usg). While there is some variation in the use of each term in the
4
5
6 literature, Ub and Ug refer to the actual measured phase velocity and Usg is a representation of the
7
8 area or volume averaged phased velocity. Thus the superficial gas velocity is equal to the
9
10
11
volumetric gas flux (Qg) divided by the column cross-sectional area (Acs), which can be related to
12
13 the average gas phase velocity using control volume analysis, U sg = Q g Acs = ε U g . Here the
14
15
16 brackets denote averaged quantities. It should be noted that Ub is often used in the literature as a
17
18 specific description of a particular condition in which individual bubbles or clouds are measured.
19
20
21 In stationary bubble columns with no flow Ub is simply the rise velocity. However, when there is
22
23 flow in/out of the column, or when the fluid column is moving, Ub can be related to the slip
24
25 velocity and the void fraction typically through a correlation.34 Specifically in the case of an
26
27
28 oscillating BCR, Ub is an important indication of the forces acting on the bubble as will be
29
30 discussed subsequently.
31
32
Therefore, a fundamental understanding of the multiphase flow properties such as void
33
34
35 fraction and bubble size distribution as well as the related mass transfer properties are crucial to
36
37 understanding and thereby improving the operation of BCRs. A review of the relevant literature
38
39
40
pertaining to these measurements in a vibrating or oscillating BCR is provided here. This review
41
42 is divided into four sections. First, a review of the most pertinent work with an emphasis on the
43
44 fundamental physics governing the observed phenomena. Second, a review of common
45
46
47 experimental methods including testing procedures, facilities and instrumentation. The emphasis
48
49 on experimental techniques is required since, as previously stated, this research area is heavily
50
51 dependent on experimental data due to the need for heuristic models specific to a given flow
52
53
54 regime. Third, there is a brief review of other relevant literature as well as computational efforts,
55
56 which provide additional perspective as well as direction for future work. Finally, the manuscript
57
58
59
60 6

1
2
3
concludes with an overview of the topic with an emphasis on what is the required knowledge to
4
5
6 advance and implement vibrating bubble columns.
7
8
9
10
11 2 History of Vibrating BCR Research
12
13
14 2.1 Houghton and Harbaum (1960-1963)
15
16
17 Harbaum & Houghton8 used Minnaert26 and Smith35 to reason that vibrating a bubbly
18
19 flow could impact mass and heat transfer. This was confirmed by measuring CO2 absorption rate
20
21
in a water column and its dependence on vibration amplitude and frequency. Subsequently,
22
23
24 Harbaum & Houghton8 showed that the mass transfer rate was not necessarily linked to the
25
26 amplitude, but attributed the increase to “resonance effects” that were only frequency (f)
27
28
29
dependent. Then they attempted to differentiate the frequency effects on kL and a individually.
30
31 The results (shown in Figure 1) indicate that an increase in mass transfer was primarily due to an
32
33 increase in a associate with an increase in ε at specific frequencies rather than an increase in kL,
34
35
36 which showed a marked dip at the same frequency. Experimental observations indicated that
37
38 vibrations cause the bubbles to become smaller and more spherical, though elongated in the
39
40 direction of motion. However, no quantitative discussion of bubble size was offered, only bubble
41
42
43 count per volume.
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 7

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23 Figure 1. Data from Harbaum & Houghton8 showing the effect of vibration frequency on
24 interfacial area per unit volume (a), bubble number density (n), bubble superficial gas velocity
25
26
(Usg) and mass transfer (kL) for a water column with CO2. Adapted with permission from Journal
27 of Chemical Technology & Biotechnology, vol. 12, K.L. Harbaum and G. Houghton, “Effects of
28 sonic vibrations on the rate of absorption of carbon dioxide in gas bubble-beds,” 234-240.
29 Copyright 1962 John Wiley and Sons.
30
31
32 Houghton36 studied particles suspended in an oscillating velocity field by applying a
33
34 force balance to a particle (gas phase in bubbly flow) in a velocity field u = Aω cos (ωt ) . Here A
35
36
37 is the amplitude, ω is the radian frequency and more details about this field are provided in Still37.
38
39 This analysis decomposes to a linearized steady state solution similar to a Mathieu equation
40
41
42 when the transient drag effects approach zero. Solutions of the Mathieu equation can be readily
43
44 interpreted using a Mathieu stability diagram (see Houghton36), with regions of the parameter
45
46 space having stable solutions. The important implication of this analysis is that careful selection
47
48
49 of the frequency and amplitude will result in a motionless bubble (with respect to a fixed
50
51 reference). Houghton36 concluded that the particle behavior within an oscillating flow should be
52
53
predictable based on drag coefficients obtained from terminal velocity measurements of the
54
55
56 particles.
57
58
59
60 8

1
2
3
4 2.2 Buchanan et al. (1962)
5
6
During the same period, Buchanan et al.9 explored the effect of vibration on bubble
7
8
9 migration in a liquid column. This was the first report of unified bubble migration, which acts
10
11 against buoyancy at specific frequencies. Using electromagnetic-hydrodynamic analogies,38 a
12
13
14
frequency was derived at which the bubbles are stable (stationary). Imagine a bubble placed at a
15
16 location h below the free surface of a fluid body undergoing oscillatory motion z = A sin(ωt ) ,
17
18
19 where z is the vertical position relative to a fixed coordinate system as illustrated in Figure 2.
20
21
22
23
24
25
26
27
28
29
30
31
32
33 Figure 2. Diagram of a single bubble in an oscillating BCR (prime denotes varying quantity).
34
35
The following analysis makes several assumptions; (1) the bubble’s resonant pulsation
36
37
38 frequency is greater than the vibration frequency, (2) the bubble is sufficiently large that surface
39
40 tension can be neglected, (3) the bubble expands and contracts isothermally, (4) the bubble
41
42
43
internal pressure follows Boyle’s law and (5) spherical momentum effects caused by the bubble
44
45 surface oscillation are negligible. Here the instantaneous total pressure can be divided into three
46
47 components pT = pe + ps + pv , where pe is the ambient pressure, ps is the hydrostatic pressure (=
48
49
50 ρgh), g is the gravitational acceleration and pv is the pressure due to vibration. The pressure field
51
52 can be solved for by using the unsteady, incompressible Navier-Stokes equation and applying the
53
54
55 boundary condition that pT = p e when h = 0. From this solution,
56
57
58
59
60 9

1
2
pT = pe + ρgh − ρhω 2 A sin(ωt ) , it is apparent that the contributions due to the column vibrating
3
4
5
6
7
is pv = − ρhω 2 A sin ωt .
8
9 The instantaneous bubble volume can be written as V = Vo + ∆V , where Vo is the initial
10
11
12 bubble volume (i.e. when pv = 0) and ∆V is the deviation from Vo. Inserting the total pressure and
13
14
15
instantaneous bubble volume into Boyle’s law ( p oVo = pV ) and rearranging produces
16
17 ∆V ρhω 2 A sin (ωt )
18 = . (3)
19 Vo + ∆V po
20
21
22
Here p0 is the total pressure without oscillation (or at the initial condition) p 0 = p e + ρgh . The
23
24 maximum volume displacement is achieved at the top of the stroke when sin(ωt) = 1. Therefore
25
26
27
Eq. (3) can be used to establish a relationship for the maximum volume displacement,
28
29 ∆Vmax ρhω 2 A
30 = . (4)
31 V0 p0 − ρhω 2 A
32
33 Bjerknes38 states that a body in translational motion within a fluid is subject to a
34
35
36
buoyancy force that is the product of the body’s acceleration and the mass of the fluid displaced
37
38 by the body. Applying this principle to a force balance on the bubble results in a temporally
39
40 evolving force,
41
( )
42
43 F (t ) = ρ (V0 + ∆V ) g − ω 2 A sin(ωt ) .
44
45
46
It is important to note that Buchanan et al.9 erroneously neglected the liquid density term in their
47
48 derivation. Addition of the term is required to properly account for the mass of the fluid.
49
50 Integration of the instantaneous force over a period of oscillation (T) gives the average force on
51
52
53 the bubble,
54
55 T
1  1 
56 Favg = ∫ F (t )dt = ρ V0 g − ω 2 A∆Vmax  (5)
57 T0  2 
58
59
60 10

1
2
3
When the average force is zero, the bubble motion relative to the fluid is zero. Thus, setting the
4
5
6 average force to zero and applying Eq. (4) produces a quadratic equation that can be solved for
7
8 the stabilization frequency (ωs) required for a stationary bubble,
9
10
11 2 gpe
ω s 2 A = − g ± 3g 2 + .
12 ρh
13
14
15 It is interesting to note that this frequency is independent of the volume, which is due to
16
17 the fact that in this analysis gravitational acceleration is balanced with the unsteady acceleration.
18
19
20 Buchanan et al.9 performed this analysis and compared the predicted stabilization frequency to
21
22 that of an experimentally determined “cut-out” frequency (ωc). The cut-out (or critical)
23
24
frequency was defined as the minimum frequency where bubble cyclic migration was observed,
25
26
27 which corresponds to when bubbles move against gravity. If the assumptions used in the
28
29 analytical analysis were valid then the stabilization frequency should be nearly equal to the cut-
30
31
32 out frequency, since the zero bubble velocity condition would be the boundary between
33
34 ascending and descending bubbles. The cut-out frequency results9 are plotted versus the
35
36 stabilization frequency in Figure 3, which shows excellent agreement noting that the solid line
37
38
corresponds to ωc = ω s .
2 2
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 11

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22 Figure 3. Data for different column inner diameters (ID) from Buchanan et al.9 of the cut-out
23 frequency (ωc) plotted versus the stabilization frequency (ωs) for a water column with air
24 injection. Data was acquired at atmospheric pressure and room temperature. Adapted with
25
26
permission from Industrial & Engineering Chemistry Fundamentals, vol. 1, R. H. Buchanan, G.
27 Jameson and D. Oedjoe, “Cyclic migration of bubbles in vertically vibrating liquid columns,”
28 82-86. Copyright 1962 American Chemical Society.
29
30
31 2.3 Baird et al. (1962-1963)
32
33
34 Around this same period, Baird & Davidson39 investigated gas absorption by singularly
35
36 rising CO2 bubbles in a stationary BCR. This work provided two key insights into gas absorption
37
38 (mass transfer) in stationary liquid columns, which are critical to understanding the more
39
40
41 complex condition with vibrations. First, the mass transfer coefficient (kL) of smaller bubbles (deq
42
43 < 25 mm) rising at a steady rate within a stationary water column is not (or weakly) time
44
45
dependent. This is demonstrated in Figure 4 with mass transfer coefficients for various bubbles
46
47
48 sizes39 plotted as a function of time. Here the smallest bubbles have negligible variation in the
49
50 mass transfer coefficient over time while the larger bubbles decay proportional to t-0.5.
51
52
53
54
55
56
57
58
59
60 12

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21 Figure 4. Mass transfer coefficient as a function of time for CO2 bubbles in a quiescent water
22 column. Data and curves are from Baird & Davidson.39 Adapted from Chemical Engineering
23
24
Science, 17, M.H.I. Baird & J.F. Davidson, “Gas absorption by large rising bubbles,” 87-93,
25 Copyright 1962, with permission from Elsevier.
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45 Figure 5. Surface tension effects on mass transfer coefficient of CO2 in solution. Figure has been
46 adapted from Baird & Davidson39 with each data point representing the average of several single
47
48
bubble measurements. Adapted from Chemical Engineering Science, 17, M.H.I. Baird & J.F.
49 Davidson, “Gas absorption by large rising bubbles,” 87-93, Copyright 1962, with permission
50 from Elsevier.
51
52
53 The second key insight from this work is a corollary to the first, which proposed that
54
55 circulation and renewal effects from the bubble wake might be responsible for establishing a
56
57 steady-state condition. It was noted that a stagnation layer was established for bubbles rising in
58
59
60 13

1
2
3
n-hexanol and n-butanol solutions, where surface tension between CO2-solution was lower than
4
5
6 CO2-water. In essence, the reduction of the turbulent wake behind the spherical bubble due to
7
8 surface tension “smoothing” the interface reduces the concentration gradient. Thus the gas phase
9
10
11
is insulated from the liquid by a thin boundary layer of partially dissolved gas-solution, which
12
13 would be mixed with the unsaturated liquid if there were more turbulent mixing. This was
14
15 demonstrated in Baird & Davidson39 (and shown in Figure 5) noting the decrease in the wake
16
17
18 profile and subsequent decrease in kL with the addition of surfactant (Lissapol). This theory could
19
20 explain the increase in kLa with vibration, the oscillating shear forces cause the bubble boundary
21
22 layer to detach, which prevents the buildup of the high-concentrated layer around small bubbles.
23
24
25 Baird11 acquired measurements of the resonant frequency of stationary singular bubbles
26
27 and slugs within a vibrating liquid column. These results were compared with established
28
29
theoretical estimates of bubble resonant frequency (fr).26 The theoretical estimates matched
30
31
32 experiments between 5-85%. Consequently, Baird11 proposed a modified relationship that
33
34 included secondary effects (e.g. confinement effects and bubble depth),
35
36 −1 / 2
1/ 2
37 1  3 γ po    ravg  4h  
fr =   1 +   − 1  ,
38
2π ravg  ρ    R  
39    R
40
41
which matched observations to within ±20%. Here ravg is the mean radius of the bubble, γ is the
42
43
44 specific heat ratio of gas (= 1.4 for air), ρ is the fluid density and R is the column radius. In
45
46 addition, stroboscopic photography was used to confirm that bubbles do reach a maximum
47
48
49 volume at the top of the stroke, which supports the analysis that produced Eq. (5). In addition, it
50
51 was observed that the larger and more visible slugs experienced larger expansion amplitudes
52
53 than contraction amplitudes, which leads to an amplitude disparity. This disparity in slug motion
54
55
56 could be applied to smaller bubbles, and can be explained by Bjerknes38 who notes that “the
57
58
59
60 14

1
2
3
light…body [bubble] will in the two extreme positions be in different masses of [liquid], and if
4
5
6 these have not exactly the same motion, it will be subject in these two positions to kinetic
7
8 buoyancies not exactly equal and not exactly opposite in direction.” Since these pioneering
9
10
11
efforts, Baird and coworkers have continued advancing these concepts and related applications.
12
13 Additional details from this group’s work will be discussed in the experimental methods section,
14
15 but the interested reader is directed to Ni et al.44 for an excellent review of recent activity with an
16
17
18 emphasis on baffled columns and their application to producing enhancements for the chemical
19
20 and process industries.
21
22
23
24 2.4 Krishna, Ellenberger and coworkers (2000-2007)
25
26 There was minimal research activity on vibrating and pulsing BCRs between the 1960’s
27
28
29 and the start of the 21st century. Krishna et al.33 provided data on bubble break-up within a
30
31 bubble column vibrating at 100-200 Hz and the associated energy requirements. Krishna &
32
33
34
Ellenberger17 provided a more detailed investigation reexamining measurements of void fraction
35
36 and mass transfer in a vibrating BCR. Here a marked improvement in void fraction was observed
37
38 with vibration (see Figure 6), which had a local maxima ε at specific “critical” frequencies
39
40
41 (Figure 7). Here Ho is the water column height measured above the injection location with no air
42
43 injection and H is the column height for the given condition.
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 15

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Figure 6. Images illustrating typical improvement in the void fraction (gas holdup) with
19 vibrations from Krishna & Ellenberger;17 (a) no gas injection, (b) gas injection with no vibration
20 and (c) gas injection and vibration. Reprinted from International Journal of Multiphase Flow, 28,
21 R. Krishna & J. Ellenberger, “Improving gas-liquid contacting in bubble columns by vibration
22 excitement,” 1223-1234, Copyright 2002, with permission from Elsevier.
23
24
25 Optimum combinations of frequency and amplitude were found to double the void
26
27 fraction as the gas superficial velocity was increased. The results showed that kLa increased 1.5-2
28
29
30 times more than the increase in void fraction at higher frequencies (see Figure 7). However,
31
32 contrary to Harbaum & Houghton,8 Krishna & Ellenberger17 proposed that the mass transfer rate
33
34
increase was not solely due to an increase in interfacial area per unit volume (a). They claimed
35
36
37 that due to an increase in turbulence the mass transfer coefficient (kL) could have also increased.
38
39 However, Krishna & Ellenberger17 did not independently measure kL and a as Harbaum &
40
41
42
Houghton8 did to base their conclusions upon. It is possible for both assessments to be valid if
43
44 two mechanisms are active with one or the other dominating for a given flow regime.
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 16

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Figure 7. Vibration induced improvement of volumetric mass transfer coefficient and void
24 fraction (gas holdup) with Usg = 1.0 cm/s and A = 0.5 mm. Mass transfer and void fraction data
25 from Krishna & Ellenberger17 scaled versus the mass transfer coefficient (kLa)o and void fraction
26 εo with no vibration, respectively. Adapted from International Journal of Multiphase Flow, 28, R.
27 Krishna & J. Ellenberger, “Improving gas-liquid contacting in bubble columns by vibration
29
30
31 Ellenberger & Krishna34 also attempted to distinguish between the effects of vibration
32
33
34
frequency and the amplitude on void fraction. Their results show increased void fraction for both
35
36 increasing frequency and amplitude, but the general trends differ in shape, as illustrated in Figure
37
38 8. Here increasing the vibration amplitude has a more abrupt impact on the void fraction
39
40
41 compared to increasing the frequency. For example, examining the condition of 1 cm/s there is a
42
43 17% increase in void fraction between 40 and 60 Hz whereas a 50% increase in amplitude (0.5
44
45 mm to 0.75 mm) increases the void fraction 36%. This example is illustrated in Figure 8 with the
46
47
48 overlaid dashed lines. Since a rise in ε is expected to directly increase a it suggests a greater
49
50 increase in kLa can be achieved through higher amplitude as well.
51
52
53
54
55
56
57
58
59
60 17

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 Figure 8. Data from Ellenberger & Krishna34 showing the influence of (a) vibration frequency
27 and (b) vibration amplitude on void fraction for air injected into a vibrating water column.
28 Dashed lines are used for reference to compare behavior at Usg = 1 cm/s. Adapted from Chemical
29
30
Engineering Science, 58, J. Ellenberger & R. Krishna, “Shaken, not stirred, bubble column
31 reactors: Enhancement of mass transfer by vibration excitement,” 705-710, Copyright 2003, with
32 permission from Elsevier.
33
34
35 While the experimental setup of Ellenberger & Krishna34 was limited to smaller
36
37 amplitudes (A ≤ 1.2 mm), these results are plotted in Figure 9 and provides the relationships
38
39 between the vibration amplitude and the void fraction (ε) or the single bubble rise velocity (Ub).
40
41
42 These have inverse trends with the void fraction increasing with amplitude while the bubble
43
44 velocity decreases. Thus the increase in void fraction was attributed to the decrease in rise
45
46
47
velocity, which was suspected of being due to the generation of standing waves in the column. In
48
49 addition, measurements of bubbles produced from a single capillary tube showed a 40-50%
50
51 reduction in size due to the vibration.
52
53
54
55
56
57
58
59
60 18

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22 Figure 9. Dependence of the rise velocity of a single bubble (Ub) and the void fraction (gas
23 holdup) at regime transition (εtrans) on the vibration amplitude. Data from Ellenberger &
24 Krishna.34 Adapted from Chemical Engineering Science, 58, J. Ellenberger & R. Krishna,
25
26
“Shaken, not stirred, bubble column reactors: Enhancement of mass transfer by vibration
28
29 Ellenberger et al.18 produced a theoretical estimate of the local void fraction and bubble
30
31
32 velocity as a function of liquid column height using the Rayleigh-Plesset equation,
33
34
3 ∂r   
3
∂2r σ  r0  σ µ ∂r
35 = − +  p + 2    − 2 − 4 − p − p ( z , t ) ,
2 r ∂t   r0  r  
e e
36 ∂t 2 r r ∂t 
37
38
39 and a force balance on a bubble,
40
41 ∂U b dp ( z , t )
V + (ρ − ρ g )V g − C D ρ U b U b πr 2 .
1
42 ρ gV =−
43 ∂t dz 2
44
45 Here r is the bubble radius, ro is the initial bubble radius, ρg is the gas density, p(z,t) is the local
46
47
pressure, σ is surface tension, µ is the liquid phase viscosity and CD is the coefficient of drag that
48
49
50 includes viscous and pressure effects. The Rayleigh-Plesset equation describes the change in
51
52 bubble radius (r) as a function of time and vertical height (z). The force balance on a bubble
53
54
55 includes the Bjerknes force as a pressure gradient. It should be noted that bubble force balance
56
57 above is not exactly the equation presented in Ellenberger et al.,18 which omitted the mass term
58
59
60 19

1
2
3
on the left hand side. The correction has been made here to make the equation dimensionally
4
5
6 consistent. It appears that this error was merely a misprint and that subsequent calculations were
7
8 not impacted, but the omission would only impact the magnitudes not the trends. The solution to
9
10
11
this set of equations, illustrated in Figure 10, suggest that local void fraction increases at
12
13 harmonic wavelengths, which is most prevalent at the pressure antinodes and higher harmonic
14
15 modes. Thus supports observations of increased mass transfer at specific frequencies.
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36 Figure 10. Results from the analytical analysis of Ellenberger et al.,18 which shows the influence
37 of harmonic modes (HM) on local void fraction within a water column with air injection.
38 Reprinted from Chemical Engineering Science, 60(22), J. Ellenberger, J.M. van Baten, & R.
39 Krishna, “Exploiting the Bjerknes force in bubble column reactors,” 5962-5970, Copyright 2005,
40
with permission from Elsevier.
41
42
43 The validity of these theoretical predictions was assessed with local void fraction
44
45
46 measurements at various column heights with an electrical conductivity meter.18 Measurements
47
48 of the local void fraction allows for the examination of spatial effects such as the impact of phase
49
50 on the bubble distribution, which cannot be assessed with global measurements. The results from
51
52
53 these measurements are provided in Figure 11, which show that the local void fraction (gas
54
55 holdup) peaks at corresponding nodes for each harmonic mode (HM). The identification of peaks
56
57 corresponding to the number of nodes for each condition is striking. However, the slope of the
58
59
60 20

1
2
3
axis of symmetry (dashed lines) typically has either a positive or negative slope that is not
4
5
6 predicted from the analysis. This is suspected to be the product of an unaccounted force in the
7
8 analytical analysis or experimental technique, but further investigation is required to deduce the
9
10
11
primary cause.
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34 Figure 11. Local void fraction (gas holdup) experimental results from Ellenberger et al.18 plotted
35 versus the location within the column height for various harmonic modes (HM). All the
36 conditions were acquired at Usg = 0.01 m/s and A = 0.5 mm, except for HM-5 that was tested at A
37 = 0.4 mm. The Ho = 0.8 and 1.1 m for the top and bottom row, respectively. Dashed lines are the
38
linear fit to the data in order to show trends. Adapted from Chemical Engineering Science,
39
40 60(22), J. Ellenberger, J.M. van Baten, & R. Krishna, “Exploiting the Bjerknes force in bubble
41 column reactors,” 5962-5970, Copyright 2005, with permission from Elsevier.
42
43
44
This group’s most recent work45,46 provides experimental data demonstrating that single
45
46 gas bubbles and slugs can be forced to levitate (i.e. held stationary) within a vibrating bubble
47
48 column. The columns were oscillated at frequencies between 50 and 400 Hz in order to levitate
49
50
51 the bubbles over a range of column diameters, fill heights, liquid densities and viscosities,
52
53 vibration amplitudes and operating pressure. The experimentally determined conditions for
54
55 levitation were in good agreement with the theoretical model of Baird.11
56
57
58
59
60 21

1
2
3
4 2.5 Knopf and coworkers (2006-2009)
5
6
This group’s first main contribution to oscillating bubble column performance focused on
7
8
9 low gas rates20 and jetting gas rates.21 They expanded upon previous research, which focused
10
11 primarily on the effect of frequency, in three key areas. First, a unified model to predict void
12
13
14
fraction and mass transfer coefficient based on correlations and theory was proposed, which
15
16 compared well with experimental results.47 Second, they provided experimental results showing
17
18 the effect of viscosity on kLa and a theoretical relationship that predicts the viscosity
19
20
21 dependence.22 Finally, a model to predict the bubble size distribution at varying column heights
22
23 and vibration frequencies based on population-balance-modeling, which was compared with
24
25 experimental results with some success.48 Each of these key results are discussed in greater detail
26
27
28 here, but the interested reader is directed to Waghmare49 as well as the individual papers for
29
30 more details.
31
32
33
Waghmare et al.47 proposed a model to predict void fraction that builds upon the analysis
34
35 of Buchanan et al.9 and is based primarily on bubble breakage scaling with the power input. Here
36
37 it is assumed that the column static pressure is large compared to the sum of the hydrostatic
38
39
40 pressure and vibration pressure, which allows Eq. (4) to be rewritten as
41
42 ∆Vmax ρhω 2 A
43 = .
44 V0 pe
45
46 This assumption obviously limits the universality of the analysis due to the fact that for many test
47
48
49 conditions the magnitude of the vibrational pressure is on the same order as the external
50
51 (atmospheric) pressure. For example, it is not uncommon to oscillate a 78 cm column of water
52
53 with air injection at atmospheric pressure with a frequency of 40 Hz and an amplitude of 2.5 mm.
54
55
56 This condition results in a magnitude equal to ρ gH − ρ H ω 2 A = 115,000 Pa, which is
57
58
59
60 22

1
2
3
4
comparable to atmospheric pressure and violates the requirement p e >> ρ gH − ρ H ω 2 A . While
5
6 this assumption might limit the applicability of the analysis, it offers physical insights into this
7
8
9 flow regime.
10
11 This analysis produces a relationship for void fraction based on input parameters,
12
( )
13
14 1
H U gU + 0.5ω 3 A 2 2 / 5 
ε = ∫ ε (h)dh = 2.25 E ( Bj ) ,
sg sg
(6)
15
16
H 0  (σ ρ )3 / 5 g ν

2/3

 ( )
17
18
19 where the bracket on ε indicates that it is the average value, ν is the kinematic viscosity of the
20
21
22
[ ]
liquid phase, E ( Bj) = 3 1 − (1 − Bj)1 / 3 Bj and Bj is the Bjerknes number ( = ρH ω 4 A 2 (2 gp e ) ).
23
24 Waghmare et al.47 couples this relationship with a modified penetration theory,
25
26
27
k L a = 4D πt c (6ε d 32 ) , where D is the molecular-diffusion coefficient of the species (units of
28
29
30
length2/time) and tc is the contact time for mass transfer estimated as d 32 U b . This combined
31
32 relationship produces an equation for the average mass transfer as a function of input parameters,
33
34
35 U
k L a = 4.58
(
sg D gU sg + 0.5ω A
3 2
)
4/5

G ( Bj ) , (7)
( )
36
(σ ρ )6 / 5 g ν
1/ 3
37  
 
38
39
40 [
where G( Bj) = 3 1 − (1 − Bj) 2 / 3 ] (2Bj) . The derivation of Eqs. (6) and (7) was based upon several
41
42
43 additional assumptions, including (i) gas concentration is dilute ( ε << 1 ), (ii) fluid amplitude is
44
45 continuous, (iii) coalescence is small, (iv) bubble breakage occurs primarily through shear and
46
47 (v) bubble expansion and contraction has a negligible effect on mass transfer.
48
49
50 Waghmare et al.47 also provided experimental measurements of void fraction and mass
51
52 transfer from air injected into a 78 cm cylindrical column of water. The column was oscillated at
53
54
amplitudes of 1.66 and 2.46 mm over a range of frequencies (0-30 Hz). These frequencies were
55
56
57 lower than most of the previous available data in the literature.9,18 The frequency and amplitude
58
59
60 23

1
2
3
combinations were selected such that Bj < 1 to avoid flooding at the injector tip.22,47 Air was
4
5
6 injected into the column via a 0.75 mm diameter stainless steel capillary tube at Usg = 0.1, 0.25,
7
8 0.5, 0.75, 1.0, 1.25 and 1.5 cm/s. A high speed camera and image processing was used to find
9
10
11
bubble number density and projected surface area, which was used to determine the d32. The
12
13 bubble diameter results plotted versus the maximum power input per unit mass,
14
15
Pm = gU sg + 0.5 A 2ω 3 , are provided in Figure 12 (left) and are in good agreement with the Hinze
16
17
18 breakup model.50 Figure 12 (right) shows the overall void fraction as a function of Sauter bubble
19
20
21
diameter for varying frequencies. These results demonstrate how increasing the frequency
22
23 (power input) reduces the average bubble size and produces a more homogeneous distribution
24
25 (spike narrows).
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42 Figure 12. Experimental results from Waghmare et al.47 showing (left) agreement with Hinze
43
44 correlation50 at column location z/H = 0.69 and (right) void fraction and bubble size distribution
45 with varying frequency for A = 1.66 mm and Usg = 0.16 cm/s. Adapted with permission from
46 AICHE Journal, vol. 53, Y.G. Waghmare, F.C. Knopf & R.G. Rice, “The Bjerknes effect:
47 Explaining pulsed-flow behavior in bubble columns,” 1678-1686. Copyright 2007 John Wiley
48 and Sons.
49
50
51 The experimental results showed a linear relationship between the void fraction and the
52
53
54 superficial gas velocity, ε = 0.04284U sg with Usg in cm/s, which is in agreement with Eq. (6)
55
56 for the given test frequency (15 Hz), amplitude (2.46 mm) and liquid properties. This result is
57
58
59
60 24

1
2
3
not surprising since it implies that the void fraction is dependent on the amount of gas being
4
5
6 injected into the system. However, the results were also plotted as a function of input power,
7
8 which was also in agreement with Eq. (6) and exhibits a power-law trend, ε = 6.65 × 10 −3 Pm0.4 .
9
10
11 Specific power provides a more useful tool as an input parameter, since it accounts for all input
12
13
parameters of a vibrating BCR; superficial gas velocity, frequency and amplitude.
14
15
16 Similarly, the experimental mass transfer results were in agreement with the predicted
17
18 rates of Eq. (7) for both the superficial gas velocity ( k L a = 0.0193U sg ) and specific power input
19
20
21 ( k L a = 0.004Pm0.8 ). However, unlike the relationship between void fraction and superficial gas
22
23
24 velocity, the effect of Usg on mass transfer is reported to be more significant. Here, if one was to
25
26 conjecture that improvement of kLa from superficial gas velocity was mostly due to an increase
27
28
29 in a through void fraction, then d ε dU sg would be comparable to d (k L a ) dU sg . However,
30
31
32 the d (k L a ) dU sg slope was half that of d ε dU sg , which suggests the kLa improvement was
33
34
caused by an improvement to the mass transfer coefficient as well as the interfacial surface area
35
36
37 per unit volume. In fact, the increased exponential dependence of kLa on the input power relative
38
39 to ε implies that vibration has an impact on kL independent of a.
40
41
42 An interesting result of the Waghmare et al.47 model for predicting void fraction and
43
44
45
mass transfer is the contribution of fluid properties. Both relationships, Eqs. (6) and (7), are
46
47 inversely proportional to surface tension and directly proportional to kinematic viscosity and
48
49 density (raised to a given power) of the surrounding fluid. While Buchanan et al.9 mentioned
50
51
52 testing fluids other than water (including slurries), the reported results were limited. Additionally,
53
54 only one other research group had investigated the effect of viscosity on bubbles in oscillating
55
56 fluid columns, but the work was limited to a kinematic investigation of singular bubbles.12,51
57
58
59
60 25

1
2
3
Furthermore, all of the previous research focusing on mass transfer in a vibrating BCR have used
4
5
6 water as the continuous media, and no experimental results have been reported on the effects of
7
8 liquid properties on void fraction (ε) or particularly, mass transfer (kLa) in vibrating BCRs.
9
10
11
Waghmare et al.22 followed this work and produced a relationship similar to Eq. (7), but
12
13 determines kLa based on fluid viscosity in a vibrating column reactor. In the low-Reynolds
14
15 number range, the primary physical mechanism through which viscosity impacts mass transfer is
16
17
18 via molecular diffusivity. The Stokes-Einstein equation at constant temperature and bubble
19
20 diameter, D1 D2 = ν 2 ν 1 , can be used to relate the molecular diffusivity and the kinematic
21
22
23 viscosity. The resulting analysis produces
24
U sg Doν o Pm
25 4/5
26 k L a = 4.58 G (B j ), (8)
27 (gν )1 / 3 (σ ρ )6 / 5
28
29
30
where D0 and ν0 are the diffusivity and kinematic viscosity of a reference fluid (typically tap
31
32 water). The significance of Eq. (8) is that kLa can be predicted given vibration (amplitude and
33
34
35
frequency), superficial gas velocity and fluid properties and it predicts k L a ∝ U sgν −1 / 3 . Note
36
37 that this proportionality of viscosity is the complete opposite of that predicted from Eq. (7),
38
39
40 which will be discussed subsequently.
41
42 With the same apparatus used in Waghmare et al.47, Waghmare et al.22 tested water, a
43
44 2%wt solution of water and carboxymethylcellulose (CMC) and a 3%wt CMC solution, which
45
46
47 had viscosities of 1, 11 and 62 cP, respectively. These results agreed with the Stokes-Einstein
48
49 equation showing that diffusivity, which is inversely proportional to viscosity, has a direct effect
50
51 on mass transfer. Thus as fluid viscosity is increased the mass transfer should decrease, which
52
53
54 was confirmed as shown in Figure 13. Therefore, the mass transfer rate can still be increased
55
56 with vibration in the practical case when the continuous phase has a viscosity higher than water,
57
58
59
60 26

1
2
3
but the increase will be dampened compared to a lower viscosity fluid. Waghmare et al.22 notes
4
5
6 that the bubble column literature has a wide variation in the predicted dependence of kLa on
7
8 viscosity. The analysis and experimental data from Waghmare et al.22 best matches with the
9
10
11 correlation of Nakanoh & Yoshida,52 k L a ∝ U sgν −0.28 , which was formulated for a non-
12
13 vibrating column. Waghmare et al.48 provides experimental data of bubble size distributions over
14
15
16 a similar range of operation within the oscillating bubble column.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37 Figure 13. Experimental results from Waghmare et al.22 of mass transfer in a vibrating BCR with
38 increasing fluid viscosity for air-water (1 cP), air-2%CMC solution (11 cP) and air-3%CMC
39 solution (62 cP). Adapted with permission from Industrial & Engineering Chemistry Research,
40 47, Y.G. Waghmare, R.G. Rice and F.C. Knopf, “Mass transfer in a viscous bubble column with
41 forced oscillations,” 5386-5394. Copyright 2008 American Chemical Society.
42
43
44 The mechanism for increasing kLa arising from the “Bjerknes effect” does not seem to be
45
46
hindered by viscosity which aligns with Jameson’s conclusion51 that the motionless condition for
47
48
49 a single bubble in an oscillating fluid is not affected by viscosity as long as the fluid is not
50
51 extremely viscous (i.e. sufficiently high Reynolds number). However, when comparing the
52
53
54
experimental results in a unified fashion as a function of frequency there is still considerable
55
56 scatter and ambiguity (Figure 14). It is still difficult to distinguish viscous effects at higher
57
58
59
60 27

1
2
3
frequencies or higher amplitudes than the experiments above could achieve. Further investigation
4
5
6 and improvement of the relationships between mass transfer and viscosity in a vibrating BCR is
7
8 certainly warranted, especially at higher amplitude, frequency and pressure. Note that in Figure
9
10
11
14 fc is the critical frequency at which there is no net force on the bubbles resulting in stationary
12
13 bubbles. This leads to a “flooding condition” with bubbles not traversing through the column.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30 Figure 14. Comparison between experimental results from Waghmare et al.22 and Eq. (8) (the
31 solid curve) with frequency and superficial gas velocity in (white) air-water, (gray) air-2%CMC
32
solution and (black) air-3%CMC solution at an amplitude of (left) 1.66 mm and (right) 2.46 mm.
33
34 Adapted with permission from Industrial & Engineering Chemistry Research, 47, Y.G.
35 Waghmare, R.G. Rice and F.C. Knopf, “Mass transfer in a viscous bubble column with forced
36 oscillations,” 5386-5394. Copyright 2008 American Chemical Society.
37
38
39 3 Review of Common Experimental Methods
40
41
42
43 3.1 Measurement of void fraction (ε) and bubble Size
44
45
46 The primary measurement for most multiphase flow experiments is the bubble size (mean
47
48 and distribution) and void fraction (local and global). Most studies focus on measurement of
49
50
51 volumetrically averaged (global) void fraction because local void fraction measurements are
52
53 extremely challenging. There have been numerous experiments that have measured local void
54
55 fraction in bubbly flows with advanced techniques,53-58 but to the authors’ knowledge there has
56
57
58 been only one case with a vibrating BCR.18 Bubble size, shape and void fraction are especially
59
60 28

1
2
3
important parameters in BCRs (stationary and vibrating) because of their direct impact on the
4
5
6 interfacial area (a). However, excluding Ellenberger et al.,18 all of the research to date in a
7
8 vibrating BCR has relied on manometry or volume displacement to give a column averaged
9
10
11
value, which does not provide information on spatially varying properties such as the Bjerknes
12
13 force.
14
15 In manometry, two pressure taps are created in the column wall separated by a known
16
17
18 column height H, as illustrated in Figure 15. A different fluid (e.g. Meriam Red 295) is used as
19
20 the manometer fluid and the volume averaged void fraction between the two pressure taps can be
21
22 determined from
23
24
25  ρ m − ρ mix  ∆h
26 ε =   ,
27  ρ mix  H
28
29 where ρm is the manometer fluid density, ρmix is the mixture average density and Δh is the
30
31
32 measured height difference. One considerable drawback to this method when applied to a
33
34 vibrating BCR is that vibration induce pressure fluctuations across the column lead to erroneous
35
36
37
manometer readings. Waghmare49 notes that this issue was a significant source of error,
38
39 particularly when frequency is high and in cases exhibiting “flooding” or aggregation of the
40
41 bubble swarm at the injector site.
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 29

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19 Figure 15. Illustration of the manometric void fraction method.
20
21
22 Another common means of measuring the global void fraction in vibrating (and
23
24 stationary) BCRs is by measuring the change in liquid height (see Figure 6). Given a constant
25
26 cross-sectional area, any changes in the column height are directly related to changes in the
27
28
29 volume averaged void fraction, ε = 1 − H o H . Here Ho is the column height without air
30
31
32
injection (i.e. all liquid) and H is the final mixture column height for the given test condition.
33
34 This method has the advantage of directly assessing the impact of vibration on the bulk void
35
36 fraction since the column height can be measured without air injection, stationary with air
37
38
39 injection and vibrating with air injection. These three measurements would provide a direct
40
41 measurement of the bulk void fraction at each condition as well as an accurate measurement of
42
43 the increase associated with vibration. The accuracy of this method is primarily fixed based on
44
45
46 the ability to accurately measure the column height, which can be difficult if the interface
47
48 exhibits significant turbulence or sloshing. Vibrating BCR are particularly sensitive to this issue
49
50
51
because the vibrations generate potentially significant disturbances on the free-surface.
52
53 Consequently, it is common to float a closed-cell polystyrene foam (e.g. Styrofoam) disk on the
54
55 free-surface to dampen the interface and make measurement of the column height easier.20,22,23
56
57
58
59
60 30

1
2
3
While there is a dearth of experimental data focused on measuring local void fraction
4
5
6 within a vibrating BCR, void fraction measurements for local cross sectional areas have been
7
8 successfully acquired in both vibrating and non-vibrating BCRs. Ellenberger et al.18 used a
9
10
11
conductivity meter to measure the average void fraction at a given column height in a vibrating
12
13 BCR. The conductivity meter had two conductive rods separated by a known distance. A voltage
14
15 was applied to the one rod and the received signal at the other rod is related to the conductivity
16
17
18 of the medium between the rods (i.e. the bubbly mixture). The local void fraction was
19
20 determined by ε area
= 1 − φ m φ o , where φm is the conductivity of the mixture and φo is the
21
22
23 conductivity for the no flow (quiescent) condition at each location. This approach assumes a
24
25 linear relationship between the conductivity and the local void fraction, but this relationship is
26
27
28 unlikely to be linear. If we assume three-dimensional inclusions of an insulating phase (air)
29
30 dispersed in a conducting phase (water), the Maxwell-Hewitt relation59,60 can be used to relate
31
32 the mixture conductivity and the liquid conductivity to void fraction in a non-linear relationship
33
34
35 (for a more detailed discussion the interested reader is directed to George et al.61).
36
37 This method requires the liquid to be electrically conductive, and the signal resolution is
38
39
40
fixed based on the conductivity of the liquid. The addition of salt to water is a simple method to
41
42 increase the conductivity of the liquid, but cannot be used for all applications. Consequently, it is
43
44 limited to conductive fluids, and organic liquids are often non-conductive or weakly conductive.
45
46
47 Thus this method is suitable for laboratory experiments, but not directly to field applications
48
49 where a BCR is often used in a reaction process with organic liquids such as hydrogenation of
50
51 coil-oil slurries in the Fisher-Tropsch process.62
52
53
54 A related method, electrical-impedance tomography, uses multiple electrodes that enclose
55
56 the cross-section to reconstruct the entire impedance distribution. This impedance distribution
57
58
59
60 31

1
2
3
can then be used to determine the spatial void fraction distribution at a given column height. This
4
5
6 method was validated against the more established gamma-densitometry tomography (GDT)
7
8 technique in George et al.61 Huang et al.63 used capacitance tomography in a non-vibrating
9
10
11
bubbly flow to determine the local void fraction. Dudukovic et al.64 provides a detailed analysis
12
13 of other novel techniques used to measure local void fraction in non-vibrating bubbly flows
14
15 including X-ray tomography or Computed Tomography (CT) (see Figure 16) and Computer
16
17
18 Automated Radioactive Particle Tracking (CARPT). These methods are advantageous for
19
20 imaging the flow and local void fraction in cases where the system excludes more visual or direct
21
22 probe methods such as in high pressure or corrosive environments. However, tomography
23
24
25 methods are generally computationally and financially prohibitive due to the expensive
26
27 equipment and complex algorithms used to reconstruct the flow-field.
28
29
30
31
32
33
34
35
36
37
38
39 Figure 16. CT images showing local void fraction with superficial gas velocity at (a) pe = 1 atm,
40
41
(b) pe = 3 atm and (c) pe = 7 atm. Adapted from U.S. Department of Energy final technical report
42 by Dudukovic, Fan & Chang.64
43
44
Optical based measurements (point probes and photographic images) have been used as a
45
46
47 relatively inexpensive and simple alternative to these flow-field reconstruction methods.
48
49 Multiphase flow properties such as interfacial area, void fraction, bubble size and bubble speed
50
51
52
have been measured with single and dual point optical probes.65-67 These optical probes are
53
54 typically relatively simple to produce and cost effective.55 Recently, a 4-point optical probe has
55
56 been used quite extensively to measure local void fraction in BCRs with better accuracy than the
57
58
59
60 32

1
2
3
single and dual point versions.30,32 The 4-point optical probe measures bubble sizes as they flow
4
5
6 past an inspection point and gives the chord length distribution as a probability distribution for a
7
8 large sample size (see an example distribution from Youssef & Al-Dahhan68 in Figure 17). The
9
10
11
4-point optical probe has been used to determine bubble size distributions as a function of chord
12
13 length in slurry flows30 and in a non-vibrating BCR with internals68 and without internals.31,32
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Figure 17. Example of the bubble chord length probability density function acquired with a 4-
31
32 point optical probe. Data obtained from Youssef & Al-Dahhan68 with a superficial gas velocity
33 of 20 cm/s. Adapted with permission from Industrial & Engineering Chemistry Research, 48,
34 A.A. Youssef and M.H. Al-Dahhan, “Impact of internals on the gas holdup and bubble properties
35 of a bubble column,” 8007-8013. Copyright 2009 American Chemical Society.
36
37
38 The vast majority of all vibrating BCR studies has used photography or video to
39
40 determine the bubble size distributions. Current methods utilize charge coupled device (CCD) or
41
42
43 complementary metal-oxide semiconductor (CMOS) digital cameras to capture images of a
44
45 vertical column cross section. The two dimensional bubble images are typically post-processed
46
47 to generate binary images, which are used to calculate the projected area for each bubble (see
48
49
50 Figure 18). This method usually requires the assumption of a spherical bubble to ascribe an
51
52 equivalent diameter to each identified bubble. The Sauter mean diameter can then be readily
53
54
obtained from Eq. (2). Oliveira & Ni16 provide an excellent description for this process with a
55
56
57 baffled vibrating BCR.
58
59
60 33

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Figure 18. Photographic processing of the bubble distribution with software; (a) pre-processed
21 image and (b) post-processed image using ImageJ software.48 Reprinted with permission from
22 Industrial & Engineering Chemistry Research, vol. 48(4), Y. Waghmare, C.A. Dorao, H.A.
23 Jakobsen, F.C. Knopf & R.G. Rice, “Bubble size distribution for a bubble column reactor
24 undergoing forced oscillations,” 1786-1796. Copyright 2009 American Chemical Society.
25
26
27 The advantages of using photographic methods include; cost, availability and versatility.
28
29 Relatively low cost cameras are typically adequate to capture high resolution images needed for
30
31
32 the calculation of the projected cross sectional area. The method is also versatile enough to
33
34 capture bubble distributions at varying locations along the column height, which gives
35
36 qualitative insight to the mechanics of the flow.22 However, photographic techniques are limited
37
38
39 by the focal area, depth of field and the optical path at higher void fractions. Thus measurements
40
41 are typically limited to the region of the flow near the column wall. Consequently, bubble size
42
43
44
distributions produced from photographic methods must be used with caution since it fails to
45
46 capture bubbles near the center of the column where larger bubbles are typically located.
47
48 Therefore the method is only inclusive for homogeneous flow or low superficial gas velocity
49
50
51 (Usg < 5 cm/s).69 Water boxes have been used to decrease the distortion due to the change in
52
53 refractive index for flow visualization.16,47,70
54
55
56
57
58
59
60 34

1
2
3
4 3.2 Measurement of volumetric mass transfer coefficient (kLa)
5
6
The primary industrial motivation for BCR research is for optimizing the mass transfer of
7
8
9 one species to another, typically gas absorption, for chemical processing. Historically,
10
11 measurements of the volumetric mass transfer coefficient (kLa) has been used as a reference for
12
13
14
the overall reaction efficiency and speed. The method to measure mass transfer (kLa) in a BCR
15
16 involves determining the concentration of the absorbed gas in solution. Early studies
17
18 accomplished this by completing the reaction of a sample through titration,8 but since those
19
20
21 studies polarographic electrode dissolved oxygen (DO) probes have been invented. DO probes
22
23 are now widely used and an accepted form of measuring the mass transfer (absorption) of
24
25 oxygen in aqueous solutions.15,20,22,34
26
27
28 DO probes work by measuring the current flow between two electrodes generated by an
29
30 oxidation-reduction reaction. The DO probe measurements are representative of dissolved
31
32
33
oxygen concentration in the sample at any time. The measured oxygen concentration can be used
34
35 to determine the volumetric mass transfer (kLa) by applying Eq. (1), a mass balance.15
36
37 Integration of Eq. (1) and applying the boundary conditions C (t = 0) = C o and limt →∞ C = C *
38
39
40 results in
41
42
 C * −C 
43 ln  = −k L at .
44  C * −Co 
45
46
47 Note that this method is analogous to the lumped capacitance method for transient
48
49 conduction heat transfer.71 As for the case of the lumped capacitance method, some assumptions
50
51
52
about the bulk fluid properties must be made to use this approach. The oxygen concentration of
53
54 the fluid batch must be uniform at any instant, which is typically an appropriate assumption for
55
56 well mixed systems like that of a highly turbulent vibrating BCR. Additionally, the integration
57
58
59
60 35

1
2
3
was performed assuming that kLa is independent of time. It is also assumed that there is
4
5
6 negligible time lag between the probe measurement and the flow-condition (i.e. the sensor
7
8 response time is small compared to changes in the flow-field). Benedek & Heideger72 and Baird
9
10
11
& Garstang15 give a reasonable account of the sources of error with using an oxygen probe in
12
13 unsteady state measurements. The uncertainty in the kLa measurement with a DO probe was
14
15 estimated to be ≤15%. Furthermore, the polarographic probes used in the literature are all of a
16
17
18 similar type, response times and accuracies.20,22,34
19
20 This process can also be applied in reverse by tracking the desorption rate of oxygen.
21
22 Here a fluid is initially saturated with oxygen and nitrogen is injected into the column, which
23
24
25 nitrogen causes oxygen to come out of solution. Fan & Cui19 used the desorption method with a
26
27 novel optical method to measure the oxygen concentration. The optical probe emitted 470 nm
28
29
light to excite fluorescent dye coated on the tip of the probe, which causes it to emit a 590 nm
30
31
32 light. Energy is transferred to oxygen molecules when they collide with the fluorescing tip
33
34 (fluorescence quenching). The amount of energy transferred to the oxygen molecules is
35
36
37 proportional to the collision-number frequency, which results in the intensity of the fluoresced
38
39 light being proportional to the partial pressure of oxygen. The partial pressure (pi) can be related
40
41 to the oxygen concentration through Henry’s Law,73 pi yi = hT xi , where yi and xi are the mole
42
43
44 fractions of the gas and liquid phases, respectively, and hT is Henry’s constant that is temperature
45
46 dependent for a gas species i.
47
48
49 An alternative method to measure the mass transfer rate of a reactive solution is to
50
51 monitor the color change of the solution over time. This method typically uses an indicator
52
53
54
similar to titration with starch and triiodide. Fan & Cui19 used this method by injecting ozone
55
56 into a solution of potassium iodide producing the reaction,
57
58
59
60 36

1
2
3
4
2 KI + O3 + H 2O → I 2 + 2 KOH + O2 .
5
6 When the starch reacts with the iodine it turns purple, which serves as an indicator. The color
7
8
9
change indicates a concentration change of potassium hydroxide (dissolved oxygen in potassium
10
11 solution), or conversely the absorption of ozone (bubbles) in solution. This technique is
12
13 demonstrated in Figure 19 as well as showing how vibrating the column causes the concentration
14
15
16 change to occur faster than the quiescent reaction. The primary disadvantage to both physical
17
18 and visual titration methods is that they require chemical processing, which is both more
19
20 laborious and difficult than using a DO probe.
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 Figure 19. Example of using color indication (a) without and (b) with vibration to observe mass
47 transfer. Reprinted with permission from Industrial & Engineering Chemistry Research, vol. 44,
48
49
J.M. Fan & Z. Cui, “Effect of acoustic standing wave in a bubble column,” 7010-7018.
50 Copyright 2005 American Chemical Society.
51
52
53 3.3 Measurement of bubble velocity (Ub)
54
55
While bubble velocity is a fundamental property that determines the bubble residence for
56
57
58 a given condition, there has been minimal effort to measure the bubble velocity within vibrating
59
60 37

1
2
3
BCR.34 There is significantly more work that has been done measuring bubble velocity in non-
4
5
6 vibrating conditions. Redfield & Houghton74 examined the mass transfer and drag coefficients
7
8 for singular bubbles rising in a quiescent fluid by measuring the rise velocity of bubbles using a
9
10
11
series of light beams and photocell detectors. Additionally, bubble rise velocity has been
12
13 measured in bubbly flows and column reactors directly using 2-point conductivity probes,75
14
15 single point optical probes55 and 4-point optical probes.54,65,76-78 While there was some variation
16
17
18 in the accuracy of each method, in all cases the bubble velocity uncertainty was less than 20%.
19
20 Bubble velocity measurements in non-vibrating BCRs have led to considerable
21
22 improvements in understanding the overall flow physics. Bubble velocity measurements in a
23
24
25 BCR with a 4-point optical probe has allowed researchers to relate reactor flows to bubbly pipe
26
27 flows.54 The success of using bubble velocity measurements within non-vibrating columns
28
29
suggest that there is a distinct need to measure the bubble velocity in vibrating BCRs to improve
30
31
32 the current understanding of the flow physics. Changes in rise velocity are indicators of the
33
34 dominate forces acting on the bubble, including the Bjerknes force that has already been
35
36
37 observed qualitatively in the literature. Quantitative measurements would give a more precise
38
39 assessment of the forces acting on the bubble for various conditions. In addition, measurements
40
41 of the liquid phase combined with bubble velocities would give a measure of the slip velocity,
42
43
44 which is fundamental to understanding how the conservation laws for each phase are related.
45
46
47 3.4 Vibrating apparatuses: advantages and disadvantages
48
49
50 There are generally two common methods to oscillate a bubble column in the literature.
51
52
The first method involves shaking the entire column as illustrated in Figure 20 (left) and similar
53
54
55 designs have been used by many researchers.9,11,12,23,79 The alternative method applies vibration
56
57 directly to the fluid with the column stationary through a piston type device10,16,17,49 as shown in
58
59
60 38

1
2
3
Figure 20 (right), a resonating device placed in the fluid19 or pulsing air.14 There are advantages
4
5
6 and disadvantages to using either method that will be briefly discussed here.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29 Figure 20. Sketches of a typical (left) “whole cylinder” shaking device adapted from Buchanan
30 et al.9 and (right) “piston pulsing” device adapted from Waghmare et al.22
31
32
33 In terms of the “quality” of the amplitude and frequency input, vibrating the entire
34
35 column is typically considered the preferred method. There is some skepticism as to whether
36
37 piston inputs create unpredictable wall effects.9 One could argue that the pressure pulse being
38
39
40 applied to the fluid column may not be uniformly distributed (or symmetric) based solely on the
41
42 geometric shape of the piston as it moves through its stroke. As a corollary, the velocity profiles
43
44
developing in the liquid could be significantly different about the centerline or at the walls.
45
46
47 Knopf et al.20 also observed that the fluid amplitude can be different from the mechanical
48
49 (piston) amplitude depending on the piston membrane thickness.
50
51
52
One obvious advantage to pulsing the liquid directly is that it requires lower power input
53
54 and has improved scalability. Pulsing provides better scalability for larger size systems since the
55
56 fluid mass alone is being driven, and not the container or whole system mass as in the case of
57
58
59
60 39

1
2
3
cylinder shaking. Consequently, these types of systems can typically achieve much higher
4
5
6 frequencies, where mass transfer is greatly increased.18 However, since amplitude and frequency
7
8 have not been clearly disconnected in the previous mass transfer experiments it could be possible
9
10
11
to generate higher mass transfer rates at lower frequency and higher amplitude leading to lower
12
13 structural and power requirements. This possibility is an open research question.
14
15
16
17 4 Related Works
18
19
20 In this section several works from the literature will be discussed that can provide
21
22 insights into methods and applications of vibrating BCRs. One related method used to improve
23
24
mixing in reactors is the oscillation of baffles within the reactor15 or oscillating the reactor with
25
26
27 fixed baffles.14,16 These baffles impart turbulence to the flow through the creation of periodic
28
29 vortices, which enhances mixing. In both cases, mass transfer improvements have been observed
30
31
32
in connection with the frequency and amplitude of these induced velocity fluctuations. Oliveira
33
34 & Ni16 demonstrated that an oscillating baffled column reactor provides improvements in column
35
36 averaged void fraction that are correlated directly to the specific power input, similar to
37
38
39 Waghmare et al.22
40
41 Using a photographic technique, a correlation was produced that related the Sauter mean
42
43
44
diameter and the time averaged power density (PV), which has units of W/m3, d 32 = c1U sgc2 PVc3 .
45
46 Here c1, c2 and c3 are correlation constants that are dependent on the experimental setup. In the
47
48
49 baffled column experimental setup used in Oliveira & Ni,16 PV (= Pm /V) was defined
50
51 as PV = U sg ρg + {2 ρN b (1 − Ab2 ) 3πD 2 Ab2 }A 3ω 3 . Here Nb is the number of baffles per unit length
52
53
54 (m-1), Ab is the baffle free cross-sectional area and D is the discharge coefficient. It is interesting
55
56
57
58
59
60 40

1
2
3
to note that the time averaged power density varies from Waghmare et al.,47 particularly in
4
5
6 regard to the vibration and amplitude exponents.
7
8 The same approach used for the diameter measurements was applied to void fraction,
9
10
which Oliveira & Ni16 produced the following correlation ε = 0.1U sg PV . This computed void
0.4 0.2
11
12
13
14
fraction was in good agreement with the experimental results. Combining the correlations from
15
16 Oliveira & Ni16 with a mass transfer correlation,80 a relationship for the mass transfer within an
17
18 oscillating baffled column reactor that depends on the Sauter mean bubble diameter and void
19
20
21 fraction,
22
23 ε 1 .5
24 k L a = 0.284 0 .6
.
25 d 32
26
27 Remarkably, this function is only dependent on two derived parameters; void fraction and bubble
28
29
30 size, which are in turn, functions of superficial gas velocity and volume specific power input.
31
32 However, these results were derived from analysis specific to oscillating baffled columns, but it
33
34
is informative to note the similarity to the approach of Waghmare et al.22 Indeed, the advantage
35
36
37 of relating two oscillating flows even separated by physical internals through an input power
38
39 parameter may be a useful tool in relating all oscillating flows regardless of setup.
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 41

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23 Figure 21. Experimental results from Fan & Cui19 showing the improved kLa with superficial gas
24 velocity and acoustic power for nitrogen-oxygen saturated water. Adapted with permission from
25
26
Industrial & Engineering Chemistry Research, vol. 44, J.M. Fan & Z. Cui, “Effect of acoustic
27 standing wave in a bubble column,” 7010-7018. Copyright 2005 American Chemical Society.
28
29
Fan & Cui19 provides a closely related study focused on studying the Bjerknes effect.
30
31
32 Contrary to the studies discussed previously, which studied the Bjerknes effects at low
33
34 frequencies (f < 120 Hz), Fan & Cui19 studied the effects at ultrasonic frequencies (f = 16 and 20
35
36
37
kHz). This was achieved using an oscillator placed inside the fluid at the top and bottom of the
38
39 column. This study clearly showed the concentration of bubbles at the antinodes as reported in
40
41 Ellenberger et al.18 Experimental results showed the migration of bubbles opposite to the
42
43
44 pressure field as predicted by Bjerknes,38 which is indicated as a delay in the rise time or a
45
46 decrease in the rise velocity of approximately 20%.19 Mass transfer rates19 are provided in Figure
47
48 21 that includes f = 16 kHz resulting in a 120% improvement, which is slightly smaller than that
49
50
51 observed in Ellenberger et al.18 for f = 60 Hz. Of particular interest, the increased mass transfer
52
53 rate is proportional to the input power and not necessarily the frequency. This finding is in
54
55
56
agreement with both Waghmare et al.22 and Oliveira & Ni16 that mass transfer is strongly
57
58
59
60 42

1
2
3
correlated to power input. However, this could also be indicating that the amplitude is the more
4
5
6 dominant parameter influencing mass transfer within this regime.
7
8 Drag has been included as an important force in each analysis described above, and
9
10
11
rightfully so as it opposes both the buoyancy and the Bjerknes forces. Only Houghton36 and
12
13 Jameson51 have treated viscous forces with anything more complex than a simple drag model.
14
15 Jameson51 analyzed the linear Navier-Stokes equation to determine the viscous forces on a
16
17
18 bubble at the condition when there is zero net force on the bubble. This analysis predicts that the
19
20 Bjerknes number (Bj) should be unity. Experimental results, shown in Figure 22, are in excellent
21
22 agreement for low Reynolds number (Re ≤ 2). Here the “vibration” Reynolds number was
23
24
25 defined as Re = Aω d b ν , where db is the mean bubble diameter.
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48 Figure 22. Experimental results from Jameson51 with correlation between dimensionless
49 Bjerknes number (Bj) and vibration based Reynolds number. The solid line corresponds to the
50 theoretically predicted value. Adapted from Chemical Engineering Science, 21, G.J. Jameson,
51
52 “The motion of a bubble in a vertically oscillating viscous liquid,” 35-48, Copyright 1966, with
53 permission from Elsevier.
54
55
56
57
58
59
60 43

1
2
3
While experimental efforts are critical for advancing the fundamental understanding of
4
5
6 such complex flows, significant advancements in numerical techniques and computational power
7
8 has made the use of computational fluid dynamics (CFD) much more valuable. CFD offers
9
10
11
greater spatial and temporal resolution than experiments can produce as well as flexibility in test
12
13 conditions. This facilitates correlations between local gas/liquid interactions with the resulting
14
15 global behavior (e.g. overall mass transfer, gas holdup, bubble sizes), which will likely lead to
16
17
18 the unraveling of the underlying physics of this complex flow-field. Consequently, we offer a
19
20 brief overview of recent CFD efforts, though not limited to vibrating columns since modeling of
21
22 the flow in a stationary column remains an open research topic. Key CFD research questions
23
24
25 involve model formulation,81-86 turbulence model selection (e.g. k-ε, LES, Reynolds Stress
26
27 Model),87-89 suitable interfacial forces,89-90 transient effects,89,85 regime transitions,91
28
29
computational efficiency,92 and scaling up results for industrial applications.93 Krishna & van
30
31
32 Baten93 offer a multi-tiered approach to BCR scale up that leverages both experiments and CFD.
33
34 First a single bubble morphology and rise dynamics are identified with experimental
35
36
37 observations and a Volume-of-Fluid (VOF) simulation, which is also used to quantify the
38
39 bubble-bubble interactions. Then the bubble swarm behavior is described and interfacial
40
41 momentum exchange relations between the bubbles and liquid are developed. Finally, CFD
42
43
44 simulations in an Eulerian framework is used to extrapolate the laboratory scale information to
45
46 full-scale commercial reactors. The interested reader is directed to recent reviews94-96 of
47
48 computational methods for more in depth discussion on each of the CFD research topics.
49
50
51
52
53
54
55
56
57
58
59
60 44

1
2
3
4 5 Conclusions
5
6
7
8 5.1 Summary
9
10
11 The history of vibrating BCR research dates back to the 1960’s with the early work of
12
13 Harbaum & Houghton,8,97 which found that mass transfer in a bubble column increased in the
14
15 presence of vibration. Buchanan et al.9 studied the general motion of bubbles in a column of
16
17
18 varying height, head pressure, frequency and amplitude noting a critical frequency that produced
19
20 stationary bubbles. This was ascribed to an inviscid force balance relating kinetic buoyancy or
21
22
“Bjerknes force” to natural buoyancy. Further study of the Bjerknes force and attempts to model
23
24
25 it for a vibrating liquid column were carried out by Harbaum & Houghton,8 Houghton,36
26
27 Jameson & Davidson,12 and Jameson51 with varying and conflicting results.
28
29
30 Ellenberger & Krishna34 returned to the subject of mass transfer in a vibrating BCR
31
32 showing that particular frequency and amplitude combinations provided optimum mass transfer.
33
34 Furthermore, an extended study of frequency effects (0 < f < 100) showed particular increases in
35
36
37 local void fraction (gas holdup) corresponding to standing pressure wave nodes.18 However, the
38
39 results of both studies were limited to a single fluid (water), a single head pressure (atmospheric)
40
41 and low amplitudes (A ≤ 1 mm). Knopf et al.,21 also studied the effects of vibration on a BCR
42
43
44 and found that, contrary to Harbaum & Houghton,8 amplitude had more of an impact on
45
46 improving void fraction and mass transfer. Previous researchers have opted to apply piston
47
48
49
pulsing to impart vibration to the fluid, but a distinct non-linearity has been found in the
50
51 measured amplitude input based on the stiffness of the piston’s rubber membrane raising doubt
52
53 about the accuracy of the method.20 Furthermore, some of the previous research using piston
54
55
56
57
58
59
60 45

1
2
3
pulsing was unable to fully separate frequency and amplitude based upon the use of an electro-
4
5
6 dynamic shaker.17,18,34
7
8 Combining a bubble breakage theory by Hinze,50 Stokes-Einstein relationship for
9
10
11
viscosity and diffusivity, and penetration theory, Waghmare et al.22 offered a semi-empirical
12
13 correlation to predict void fraction and mass transfer of a vibrating BCR based upon fluid
14
15 properties and power input. The correlation is shown to give moderately good results based on
16
17
18 experiments performed with a low frequency piston pulsing set-up at two amplitudes with
19
20 varying fluids at atmospheric pressure.22
21
22
23
24 5.2 Future work
25
26 The path forward can be broadly divided between applied work focused on adapting
27
28
29 findings to specific applications and fundamental work focused on determining the underlying
30
31 physical mechanisms. Here the focus is on fundamental work as it will ultimately advance
32
33
34
specific applications and other researchers with expertise on the specific applications are better
35
36 suited for the applied discussion. To facilitate a fundamental discussion it is helpful to clearly
37
38 identify the problem, which the experimental decisions are outlined with the illustration in Figure
39
40
41 23.
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 46

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 Figure 23. Schematic illustrating the decisions made related to experimental setup, selecting test
28 conditions (independent parameters) and the primary measurements (dependent variables).
29
30 Bubble distribution from Ferreira et al.98 and gas holdup images from Krishna & Ellenberger.17
31 Adapted from Chemical Engineering Journal, 180, A. Ferreira, G. Pereira, J.A. Teixeira & F.
32 Rocha, “Statistical tool combined with image analysis to characterize hydrodynamics and mass
33 transfer in a bubble column,” 216-228, Copyright 2012, with permission from Elsevier. Adapted
34 from International Journal of Multiphase Flow, 28, R. Krishna & J. Ellenberger, “Improving
35
36
gas-liquid contacting in bubble columns by vibration excitement,” 1223-1234, Copyright 2002,
37 with permission from Elsevier.
38
39 From this illustration it is apparent that the first issue that needs to be addressed is to
40
41
42 determine whether or not it is appropriate to compare results when the whole column is vibrated
43
44 versus when the fluid is directly pulsed. While it is important for both methods to be studied, if
45
46
47
the underlying physics is going to be identified it will require numerous researchers reporting
48
49 data from their unique experimental setups. Ultimately, the research community then needs to
50
51 know which datasets are proper to compare when assessing a given phenomenon. Unfortunately,
52
53
54 this will likely be specific to each experimental setup, but a detailed comparative study between
55
56 vibration methods could identify criteria that must be met in order to compare data between
57
58
59
60 47

1
2
3
experimental facilities. To a lesser extent, there is also a need to do a parametric study on the
4
5
6 column geometry to identify limiting cases when confinement is no longer a secondary effect.
7
8 In terms of selecting the test conditions, there has been several studies discussed in this
9
10
11
review that have provided a wide range of operation conditions. In spite of this data, there
12
13 remains fundamental questions about the impact of basic parameters such as vibration amplitude,
14
15 vibration frequency and viscosity on mass transfer. Typically when numerous researchers
16
17
18 producing conflicting results it indicates that multiple fluid mechanisms are active within the
19
20 flow. For a vibrating BCR, there are three primary forces that are active; buoyancy, unsteady
21
22 loading and viscous forces. This suggests that at a minimum the parameter space needs to be
23
24
25 divided with two dimensionless parameters, which we suggest using the vibration based
26
27 Reynolds number ( Re = Aω d b ν ) and Froude number ( Fr = A ω gd b ). The Reynolds
28
29
30 number is the ratio of inertial (unsteady) forces to viscous forces, and the Froude number is the
31
32
ratio of inertial (unsteady) forces to gravitational forces. Consequently, it is recommended that
33
34
35 while the data in the literature continues to be expanded, the results and associated analysis be
36
37 defined in terms of its Reynolds and Froude numbers.
38
39
40
Characterizing the flows with these parameters will likely clarify inconsistencies in the
41
42 literature. For example, it was previously noted that Eq. (7) predicts that k L a ∝ ν 1 / 3 while Eq.
43
44
45 (8) predicts k L a ∝ ν −1 / 3 . These seemingly contradictory findings are actually due to complex
46
47 role that viscosity plays in the fluid motion and each relationship being valid for separate
48
49
50 Reynolds number ranges. Viscosity acts to resist motion, which reduces the bubble rise velocity
51
52 and suppresses turbulent mixing. Decreasing the bubble velocity increases bubble residence time
53
54 (i.e. promotes mass transfer), but the decrease in mixing produces a layer of high concentrated
55
56
57 liquid around the bubble minimizing mass transfer. In the low Reynolds number limit, the
58
59
60 48

1
2
3
viscosity acts to stabilize the bubble wake and boundary layer resulting in mass transfer
4
5
6 primarily occurring via molecular diffusion (i.e. Stokes-Einstein equation is valid). Conversely,
7
8 with sufficiently high Reynolds number the flow is always turbulent and viscosity primarily acts
9
10
11
to increase the bubble residence time.
12
13 Finally with regards to reporting of results, there is an obvious complication created from
14
15 the above suggested parameters for characterizing the experimental conditions, which is both
16
17
18 parameters are scaled in terms of the mean bubble diameter. The mean bubble diameter is a
19
20 dependent variable that is a function of the fluid properties (e.g. surface tension, viscosity and
21
22 density), operating condition (vibration, volumetric flow rate and operating pressure),
23
24
25 confinement effects and gravitational forces. Ultimately the bubble dynamics (e.g. size,
26
27 distribution, velocity) are fixing the mass transfer and gas holdup for a given liquid-gas
28
29
combination, which is why there is much confusion when trying to assess the bulk properties
30
31
32 (gas holdup and mass transfer) purely in terms of the independent variables. Thus there is a
33
34 strong need for a study focused purely on the behavior of the local properties (bubble dynamics
35
36
37 and phase velocities) on the independent variables. Once a fundamental scaling law for bubble
38
39 size (mean, fluctuating and distribution), shape and spatial distribution (local void fraction), the
40
41 proper scaling of the mass transfer for a given flow regime should be apparent. A related open
42
43
44 question is separating the influence of kL and a in the mass transfer improvements, but this is
45
46 likely unable to be fully understood until the flow regimes are well defined.
47
48
49
50 Acknowledgements
51
52
53 The M.S. thesis work of author Mr. Adam Still, which produced the initial draft of this
54
55 review, was partially funded by Sandia National Laboratories, Albuquerque, New Mexico
56
57
58
through Dr. Timothy O’Hern. Sandia National Laboratories is a multi-program laboratory
59
60 49

1
2
3
managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin
4
5
6 Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration
7
8 under contract DE-AC04-94AL85000. The authors would also like to thank Shahrouz
9
10
11
Mohagheghian for providing the bubble image in Figure 23.
12
13
14 References
15
16
17 (1) Brennen, C. E. Fundamentals of Multiphase Flow; Cambridge University Press: 2005.
18
19
(2) United Airlines conducts first commercial flight in U.S. using synthetic jet fuel. United
20 Airlines: Investor Relations News, 2010;
21 http://ir.united.com/phoenix.zhtml?c=83680&p=irol-newsArticle&ID=1420753. Accessed
22 on May 6, 2015.
23 (3) Drinnon, R. C-17 uses synthetic fuel blend on transcontinental flight. U.S. Air Force News,
24
2007; http://www.af.mil/News/ArticleDisplay/tabid/223/Article/124820/c-17-uses-synthetic-
25
26 fuel-blend-on-transcontinental-flight.aspx. Accessed on May 6, 2015.
27 (4) King, S. Jr. A-10 flies on synthetic fuel blend. U.S. Department of Defense News, 2010;
28 http://www.defense.gov/news/newsarticle.aspx?id=58488. Accessed on May 6, 2015.
29 (5) Vogelaar, R. F-22 Raptor flown on synthetic biofuel. AVIATIONNEWS.EU, March 24, 2011;
30 http://www.aviationnews.eu/2011/03/24/f-22-raptor-flown-on-synthetic-biofuel. Accessed on
31
32
May 6, 2015.
33 (6) Vogelaar, R. Lufthansa and Airbus launch worlds first passenger bio-fuel flights.
34 AVIATIONNEWS.EU, Nov. 29, 2010; http://www.aviationnews.eu/2010/11/29/lufthansa-
35 and-airbus-launch-worlds-first-passenger-bio-fuel-flights. Accessed on May 6, 2015.
36 (7) Vik, C.B.; Solsvik, J.; Hillestad, M.; Jakobsen, H. A. Modeling of a slurry bubble column
37
reactor for the production of biofuels via the Fischer-Tropsch synthesis. Chem. Eng. Technol.
38
39 2015, 38, 690-700.
40 (8) Harbaum, K. L.; Houghton, G. Effects of sonic vibrations on the rate of absorption of carbon
41 dioxide in gas bubble-beds. J. Appl. Chem. 1962, 12, 234-240.
42 (9) Buchanan, R. H.; Jameson, G.; Oedjoe, D. Cyclic migration of bubbles in vertically vibrating
43 liquid columns. Ind. Eng. Chem. Fundam. 1962, 1, 82-86.
44
45
(10) Bretsznajder, S.; Jaszczak, M.; Pasiuk, W. Increasing the rate of certain industrial
46 chemical processes by the use of vibration. Int. Chem. Eng. 1963, 3, 496-502.
47 (11) Baird, M. H. I. Resonant bubbles in a vertically vibrating liquid column. Can. J. Chem.
48 Eng. 1963, 41, 52-55.
49 (12) Jameson, G. J.; Davidson, J. F. The motion of a bubble in a vertically oscillating liquid:
50
Theory for an inviscid liquid, and experimental results. Chem. Eng. Sci. 1966, 21, 29-34.
51
52 (13) Nyborg, W. L.; Rodgers, A. The motion of liquid inside a closed vibrating vessel.
53 Biotechnol. Bioeng. 1967, 9, 235-256.
54 (14) Baird, M. H. I.; Garstang, J. H. Power consumption and gas hold-up in a pulsed column.
55 Chem. Eng. Sci. 1967, 22, 1663-1673.
56 (15) Baird, M. H. I.; Garstang, J. H. Gas absorption in a pulsed bubble column. Chem. Eng. Sci.
57
58
1972, 27, 823-833.
59
60 50

1
2
3
(16) Oliveira, M. S. N.; Ni, X. Gas hold-up and bubble diameters in a gassed oscillatory baffled
4
5 column. Chem. Eng. Sci. 2001, 56, 6143-6148.
6 (17) Krishna, R.; Ellengerger, J. Improving gas-liquid contacting in bubble columns by
7 vibration excitement. Int. J. Multiphase Flow 2002, 28, 1223-1234.
8 (18) Ellenberger, J.; van Baten, J.M; Krishna, R. Exploiting the Bjerknes force in bubble
9 column reactors. Chem. Eng. Sci. 2005, 60, 5962-5970.
10
11
(19) Fan, J. M.; Cui, Z. Effect of acoustic standing wave in a bubble column. Ind. Eng. Chem.
12 Res. 2005, 44, 7010-7018.
13 (20) Knopf, F. C.; Ma, J.; Rice, R. G.; Nikitopoulos, D. Pulsing to Improve Bubble Column
14 Performance: I. Low Gas Rates. AIChE J. 2006, 52, 1103-1115.
15 (21) Knopf, F. C.; Waghmare, Y.; Ma, J.; Rice, R. G. Pulsing to improve bubble column
16
performance: II. Jetting gas rates. AIChE J. 2006, 52, 1116-1126.
17
18 (22) Waghmare, Y. G.; Rice, R.G.; Knopf, F. C. Mass transfer in a viscous bubble column with
19 forced oscillations. Ind. Eng. Chem. Res. 2008, 47, 5386-5394.
20 (23) Still, A. L.; Ghajar, A. J.; O’Hern, T. J. Effect of amplitude on mass transport, void
21 fraction and bubble size in a vertically vibrating liquid-gas bubble column reactor.
22 Proceedings of the ASME Fluids Engineering Summer Meeting, FEDSM2013, Incline
23
24
Village, Nevada, USA, July 7-11, 2013.
25 (24) Xue, J. Bubble velocity, size and interfacial area measurements in bubble columns; Ph.D.
26 Dissertation, Washington University, St. Louis, MO, USA, 2004.
27 (25) Raju, K. S. N. Fluid Mechanics, Heat Transfer, and Mass Transfer: Chemical Engineering
28 Practice; Wiley-AIChE: Hoboken, NJ, 455-473, 2011.
29
(26) Minnaert, M. On musical air-bubbles and the sounds of running water. Philos. Mag.:
30
31 Series 7 1933, 16, 235-248.
32 (27) Manasseh, R.; LaFontaine, R. F.; Davy, J.; Shepherd, I.; Zhu, Y. -G. Passive acoustic
33 bubble sizing in sparged systems. Exp. Fluids 2001, 30, 672-682.
34 (28) Kulkarni, A. A.; Joshi, J. B. Bubble formation and bubble rise velocity in gas-liquid
35 systems: A review. Ind. Eng. Chem. Res. 2005, 44, 5873-5931.
36
37 (29) Ghiaasiaan, S. M. Two Phase Flow, Boiling, and Condensation: In Conventional and
38 Miniature Systems; Cambridge University Press: Ithaca, NY, 2007.
39 (30) Wu, C.; Suddard, K.; Al-Dahhan, M. H. Bubble dynamics investigation in a slurry bubble
40 column. AIChE J. 2008, 54, 1203-1212.
41 (31) Xue, J.; Al-Dahhan, M. H.; Dudukovic, M. P. Bubble velocity, size, and interfacial area
42
43
measurements in a bubble column by four-point optical probe. AIChE J. 2008, 54, 350-363.
44 (32) Xue, J. L.; Al-Dahhan, M.; Dudukovic, M. P.; Mudde, R. F. Four-point optical probe for
45 measurement of bubble dynamics: Validation of the technique. Flow Meas. Instrum. 2008, 19,
46 293-300.
47 (33) Krishna, R.; Ellenberger, J.; Urseanu, M. I.; Keil, F. J. Utilisation of bubble resonance
48 phenomena to improve gas-liquid contact. Naturwissenschaften 2000, 87, 455-459.
49
50 (34) Ellenberger, J.; Krishna, R. Shaken, not stirred, bubble column reactors: Enhancement of
51 mass transfer by vibration excitement. Chem. Eng. Sci. 2003, 58, 705-710.
52 (35) Smith, F. D. On the destructive mechanical effects of the gas bubbles liberated by the
53 passage of intense sound through a liquid. Philos. Mag. 1935, 19, 1147-1151.
54 (36) Houghton, G. The behaviour of particles in a sinusoidal velocity field. Proc. R. Soc.
55
56
London, Ser. A 1963, 272, 33-43.
57
58
59
60 51

1
2
3
(37) Still, A. Multiphase phenomena in a vibrating bubble column reactor; M.Sc. Thesis,
4
5 Mechanical & Aerospace Engineering, Oklahoma State University, Stillwater, OK, USA,
6 2010.
7 (38) Bjerknes, V. F. K. Fields of force: Supplementary lectures, applications to meteorology;
8 Reproduced by Cornell University Library, Ithaca, New York, Original by The Columbia
9 University Press/The Macmillan Company, London, 1906.
10
11
(39) Baird, M. H. I.; Davidson, J. F. Gas absorption by large rising bubbles. Chem. Eng. Sci.
12 1962, 17, 87-93.
13 (40) Guyer, A.; Pfister, X. Über die Absorption von Gasblasen I. Helv. Chim. Acta 1946, 29,
14 1173-1183.
15 (41) Coppock, P. D.; Meiklejohn, G. T. The behavior of gas bubbles in relation to mass transfer.
16
Trans. Inst. Chem. Eng. 1951, 29, 75-86.
17
18 (42) Hammerton, D.; Garner, F. H. Gas absorption from single bubbles. Trans. Inst. Chem. Eng.
19 1954, 32, S18-24.
20 (43) von Bogdandy, L.; Rutsch W.; Stranski, I. N. Gasaustausch zwischen Blasen und
21 gaslösenden Flüssigkeiten. Chem. Ing. Tech. 1959, 31, 580-582.
22 (44) Ni, X.; Mackley, M. R.; Harvey, A. P.; Stonestreet, P.; Baird, M. H. I.; Rama Rao, N. V.
23
24
Mixing through oscillations and pulsations – A guide to achieving process enhancements in
25 the chemical and process industries. Chem. Eng. Res. Des. 2003, 81, 373-383.
26 (45) Ellenberger, J.; Krishna, R. Levitation of air bubbles in liquid under low frequency
27 vibration excitement. Chem. Eng. Sci. 2007, 62, 5669-5673.
28 (46) Ellenberger, J.; Krishna, R. Levitation of air bubbles and slugs in liquid under low
29
frequency vibration excitement. Chem. Eng. Sci. 2007, 62, 7548-7553.
30
31 (47) Waghmare, Y. G.; Knopf, F. C.; Rice, R. G. The Bjerknes effect: Explaining pulsed-flow
32 behavior in bubble columns. AIChE J. 2007, 53, 1678-1686.
33 (48) Waghmare, Y.; Dorao, C. A.; Jakobsen, H. A.; Knopf, F. C.; Rice, R. G. Bubble size
34 distribution for a bubble column reactor undergoing forced oscillations. Ind. Eng. Chem. Res.
35 2009, 48, 1786-1796.
36
37 (49) Waghmare, Y. G. Vibrations for improving multiphase contact; Ph.D. Dissertation,
38 Louisiana State University, Baton Rouge, LA, USA, 2008.
39 (50) Hinze, J. O. Fundamentals of the hydrodynamic mechanism of splitting in dispersion
40 processes. AIChE J. 1955, 1, 289-295.
41 (51) Jameson, G. J. The motion of a bubble in a vertically oscillating viscous liquid. Chem. Eng.
42
43
Sci. 1966, 21, 35-48.
44 (52) Nakanoh, M.; Yoshida, F. Gas absorption by Newtonian and non-Newtonian liquids in a
45 bubble column. Ind. Eng. Chem. Process Des. Dev. 1980, 19, 190-195.
46 (53) George, D. L.; Shollenberger, K. A.; Torczynski, J. R. Sparger effects on gas volume
47 fraction distributions in vertical bubble-column flows as measured by gamma-densitometry
48 tomography. Proceedings of 2000 ASME Fluids Engineering Division Summer Meeting,
49
50 Boston, MA, USA, June 11-15, 2000.
51 (54) Mudde, R. F.; Saito, T. Hydrodynamical similarities between bubble column and bubbly
52 pipe flow. J. Fluid Mech. 2001, 437, 203-228.
53 (55) Cartellier, A. Optical probes for multiphase flow characterization: Some recent
54 improvements. Chem. Eng. Technol. 2001, 24, 535-538.
55
56
57
58
59
60 52

1
2
3
(56) Elbing, B. R.; Winkel, E. S.; Lay, K. A.; Ceccio, S. L.; Dowling, D. R.; Perlin, M. Bubble-
4
5 induced skin-friction drag reduction and the abrupt transition to air-layer drag reduction. J.
6 Fluid Mech. 2008, 612, 201-236.
7 (57) Elbing, B. R.; Mäkiharju, S.; Wiggins, A.; Perlin, M.; Dowling, D. R.; Ceccio, S.L. On the
8 scaling of air layer drag reduction. J. Fluid Mech. 2013, 717, 484-513.
9 (58) Mäkiharju, S. A.; Gabillet, C.; Paik, B. -G.; Change, N. A.; Perlin, M.; Ceccio, S. L. Time-
10
11
resolved two-dimensional X-ray densitometry of a two-phase flow downstream of a
12 ventilated cavity. Exp. Fluids 2013, 54, 1561:1-21.
13 (59) Maxwell, J. C. A Treatise on Electricity and Magnetism; Clarendon Press: Oxford,
14 England, 1881.
15 (60) Hewitt, G. F. Measurement of Two-Phase Flow Parameters; Academic Press: London,
16
1978.
17
18 (61) George, D. L.; Torczynski, J. R.; Shollenberger, K. A.; O’Hern, T. J.; Ceccio, S. L.
19 Validation of electrical-impedance tomography for measurements of material distribution in
20 two-phase flows. Int. J. Multiphase Flow 2000, 26, 549-581.
21 (62) Fan, L.-S. Gas-Liquid-Solid Fluidization Engineering 1st Ed.; Butterworth Publishers:
22 Stoneham, MA, 1989.
23
24
(63) Huang, Z.; Wang, B.; Li, H. Application of electrical capacitance tomography to the void
25 fraction measurement of two-phase flow. IEEE Trans. Instrum. Meas. 2003, 52, 7-12.
26 (64) Dudukovic, M. P.; Fan, L. -S.; Chang, M. Novel techniques for slurry bubble column
27 hydrodynamics: Final Technical Report – 1999. Department of Energy, DOE-FG 22-95 PC
28 95212, 1999.
29
(65) Guet, S.; Fortunati, R. V.; Mudde, R. F.; Ooms, G. Bubble velocity and size measurement
30
31 with a four-point optical fiber probe. Part. Part. Syst. Charact. 2003, 20, 219-230.
32 (66) Guet, S.; Luther, S.; Ooms, G. Bubble shape and orientation determination with a four-
33 point optical fibre probe. Exp. Therm. Fluid Sci. 2005, 29, 803-812.
34 (67) Julia, J. E.; Hartveld, W. K.; Mudde, R. F.; Van den Akker, H. E. A. On the accuracy of
35 the void fraction measurements using optical probes in bubbly flows. Rev. Sci. Instrum. 2005,
36
37 76, 1-13.
38 (68) Youssef, A. A.; Al-Dahhan, M. H. Impact of internals on the gas holdup and bubble
39 properties of a bubble column. Ind. Eng. Chem. Res. 2009, 48, 8007-8013.
40 (69) Shah, Y. T.; Kelkar, B. G.; Godbole, S. P. Design parameters estimations for bubble
41 column reactors. AIChE J. 1982, 28, 353-379.
42
43
(70) Zhu, Y.; Wu, J.; Manasseh, R. Rapid measurement of bubble size in gas-liquid flows using
44 a bubble detection technique. Proceedings of the 14th Australasian Fluid Mechanics
45 Conference, Adelaide University, Adelaide, Australia, December 10-14, 2001.
46 (71) Çengel, Y.; Ghajar, A. Fundamentals of Heat and Mass Transfer, 5th ed.; McGraw-Hill:
47 New York, NY, 2015.
48 (72) Benedek, A. A.; Heideger, W. J. Polarographic oxygen analyzer response: The effect of
49
50 instrument lag in the non-steady state reaeration test. Water Res. 1970, 4, 627-640.
51 (73) Tapley, B. D. Eshbach’s Handbook of Engineering Fundamentals, 4th ed.; John Wiley &
52 Sons: New York, NY, 14.33-34, 1990.
53 (74) Redfield, J. A.; Houghton, G. Mass transfer and drag coefficients for single bubbles at
54 Reynolds numbers of 0.02-5000. Chem. Eng. Sci. 1965, 20, 131-139.
55
56
(75) Lewis, D. A.; Davidson, J. F. Bubble sizes produced by shear and turbulence in a bubble
57 column. Chem. Eng. Sci. 1983, 38, 161-167.
58
59
60 53

1
2
3
(76) Frijlink, J. J. Physical aspects of gassed suspension reactors; Ph.D. Dissertation, Delft
4
5 University, Delft, Netherlands, 1987.
6 (77) Luther, S.; Rensen, J.; Guet, S. Bubble aspect ratio and velocity measurement using a four-
7 point fiber-optical probe. Exp. Fluids 2004, 36, 326-333.
8 (78) Bai, W.; Deen, N. G.; Mudde, R. F.; Kuipers, J. A. M. Accuracy of bubble velocity
9 measurement with a four-point optical fiber probe. Proceedings of the 6th International
10
11
Conference on CFD in Oil & Gas, Metallurgical and Process Industries SINTEF/NTNU,
12 Trondheim, Norway, June 10-12, 2008.
13 (79) Rodgers, A.; Hughes, D. E. The disintegration of micro-organisms by shaking with glass
14 beads. J. Biochem. Microbiol. Technol. Eng. 1960, 2, 49-70.
15 (80) Ni, X.; Gao, S. Scale-up correlation for mass transfer coefficients in pulsed baffled reactors.
16
Chem. Eng. J. 1996, 63, 157-166.
17
18 (81) Krishna, R.; van Baten, J. M. Eulerian simulations of bubble columns operating at elevated
19 pressures in the churn turbulent flow regime. Chem. Eng. Sci. 2001, 56, 6249-6258.
20 (82) Monahan, S. M.; Fox, R.O. Effect of model formulation on flow-regime predictions for
21 bubble columns. AICHE J. 2007, 53, 9-18.
22 (83) Monahan, S. M.; Fox, R.O. Linear stability analysis of a two-fluid model for air-water
23
24
bubble columns. Chem. Eng. Sci. 2007, 62, 3159-3177.
25 (84) Monahan, S. M.; Fox, R.O. Validation of two-fluid simulations of a pseudo-two-
26 dimensional bubble column with uniform and nonuniform aeration. Ind. Eng. Chem. Res.
27 2009, 48, 8134-8147.
28 (85) Jia, X.;Yuan, Q.;Wen, J.; Feng, W. Fluid flow modeling of a gas-induced pulsating flow
29
bubble column. Chem. Biochem. Eng. Q. 2011, 25, 27-36.
30
31 (86) Masood, R. M. A.; Delgado, A. Numerical Investigation of Three-Dimensional Bubble
32 Column Flows: A Detached Eddy Simulation Approach. Chem. Eng. Technol. 2014, 37, 1-9.
33 (87) Dhotre, M. T.; Deen, N. G.; Niceno, B.; Khan, Z.; Joshi, J. B. Large eddy simulation for
34 dispersed bubbly flows: A review. Int. J. Chem. Eng. 2013, 343276.
35 (88) Masood, R. M. A.; Rauh, C.; Delgado, A. CFD simulation of bubble column flows: An
36
37 explicit algebraic Reynolds stress model approach. Int. J. Multiphase Flow 2014, 66, 11-25.
38 (89) Tabib, M. V.; Roy, S. A.; Joshi, J. B. CFD simulation of a bubble column – An analysis of
39 interphase forces and turbulence models. Chem. Eng. J. 2008, 139, 589-614.
40 (90) Masood, R. M. A.; Delgado, A. Numerical investigation of the interphase forces and
41 turbulence closure in 3D square bubble columns. Chem. Eng. Sci. 2014, 108, 154-168.
42
43
(91) Monahan, S. M.; Fox, R. O. CFD predictions for flow-regime transitions in bubble
44 columns. AICHE J. 2005, 51, 1897-1923.
45 (92) Pourtousi, M.; Sahu, J. N.; Ganesan, P.;Shamshirband, S.; Redzwan, G. A combination of
46 computational fluid dynamics (CFD) and adaptive neuro-fuzzy system (ANFIS) for
47 prediction of the bubble column hydrodynamics. Powder Technol. 2015, 274, 466-481.
48 (93) Krishna, R.; van Baten, J. M. Scaling up bubble column reactors with the aid of CFD.
49
50 Chem. Eng. Res. Des. 2001, 79, 283-309.
51 (94) Kantarci, N.; Borak, F.; Ulgen, K. O. Bubble column reactors. Process Biochem. 2005, 40,
52 2263-2283.
53 (95) Monahan, S. M. Computational fluid dynamics analysis of air-water bubble columns; Ph.D.
54 Dissertation, Chemical Engineering, Iowa State University, Ames, IA, USA, 2007.
55
56
57
58
59
60 54

1
2
3
(96) Pourtousi, M.; Sahu, J. N.; Ganesan, P. Effect of interfacial forces and turbulence models
4
5 on predicting flow pattern inside the bubble column. Chem. Eng. Process: Process
6 Intensification. 2014, 75, 38-47.
7 (97) Harbaum, K. L.; Houghton, G. Effects of sonic vibrations on the rate of absorption of
8 gases from bubble beds. Chem. Eng. Sci. 1960, 18, 90-92.
9 (98) Ferreira, A.; Pereira, G.; Teixeira, J. A.; Rocha, F. Statistical tool combined with image
10
11
analysis to characterize hydrodynamics and mass transfer in a bubble column. Chem. Eng. J.
12 2012, 180, 216-228.
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 55

1
2
3
4 Table
5
6
7
8
9 Column Fluids
10 Source Type depth dia CP DP Usg Amp freq Primary
11 (cm) (cm) (cm/s) (cm) (Hz) Measurements
12
13
14
Harbaum & Houghton8 PD 61 7.6 H2O CO2 0.13-1.4 10-4-0.1 20-2000 MT VF BS
15 Buchanan et al.9 WC 15-100 5-15 varieda air (?) (-) 0.05-1 20-48
16 Bretsznajder et al.10 PD 84 20 H2O CO2 NA 0.05-1.05 5-67 MT
17 Baird11 WC 23 5.1/6.4 variedb air NA NA 0-1050 BS
18 Jameson & Davidson12 WC 39-42 1.9 variedc air single bubble 0.4-0.8 20-50
19 Nyborg & Rogers13 WC 12 [Sq] H2O air NA 0-0.3 25-200
20
21 Baird & Garstang14,15 * AP 305 7.6 H2O air 0-2.6 0-15.2 0.9-1.4 MT VF
22 Oliveira & Ni16* MB 150 5 H2O air 0.1-0.64 0.2-0.8 1-5 VF BS
23 Krishna & Ellenberger17 PD 200 10 H2O air 0.6-1.4 0.01-0.14 0-120 MT VF
24 Ellenberger et al.18 PD 200 10 variedd air 0.4-1 0.03-0.05 0-95 VF
25 Fan & Cui19 AT 80 10.3 H2O,KI variede 0.4-7.5 NA 16,000-20,000 MT BS
26
27
Knopf et al.20,21 PD 82 8.9 H2O air 0.01-0.15 0.01-0.25 0-30 MT VF
28 Waghmare et al.22 PD 78 8.9 variedf air 0.5-1.5 0.17-0.25 0-25 MT BS
29 Still et al.23 PD 102 10.2 H2O air 0.1-1 0.15-0.95 0-22.5 MT VF BS
30
31 Types: AP = air pulsed, AT = acoustic transducer, MB = metal bellows, PD = piston-diaphragm, WC = whole column;
32 Fluid Phases: CP = continuous phase, DP = dispersed phase
33 Varied Fluids: (a) = Hg, oil, glycerol, ethanol, butanol, slurry; (b) = H2O, glycerol, paint resin, hydroxyl-ethyl cellulose; (c) = H2O, dilute glycerol,
34 glycerol; (d) = H2O with 0.03% weight NaCl; (e) = N2, O2, O3; (f) = H2O, 2% CMC, 3% CMC
35
Measurements: MT = mass transfer coefficient; VF = void fraction; BS = bubble size
36
37 Notes: [Sq] = square cross-section of 2 cm width by 0.3 or 0.6 cm; * = baffled
38
39 Table 1. Summary of parameter space explored by key studies and the primarily measurements for the given study.
40
41
42
43
44 56
45
46 ACS Paragon Plus Environment
47
48

1
2
3
4 Nomenclature
5
6
7 a Interfacial surface area per unit volume p Local pressure
8 A Amplitude of vibration pe Ambient pressure
9 Ab Baffle free cross-sectional area pi Partial pressure
10 Acs Column cross-sectional area Pm Maximum power input per unit mass,
11
12
Aproj Projected bubble area Pm = gU sg + 0.5 A 2ω 3
13 Bj Bjerknes number, ρHω 4 A 2 (2 gp e ) po Total pressure without oscillation,
14 C Dissolved gas concentration in the p 0 = p e + ρgh
15 liquid
16 ps Hydrostatic pressure (= ρgh)
17 C* Saturation concentration pT Total pressure
18 c1 Correlation constant for the setup pv Pressure due to vibration
19 c2 Correlation constant for the setup PV Time average power density (W/m3)
20 c3 Correlation constant for the setup
21 Qg Volumetric flux of gas into the column
CD Coefficient of drag r Bubble radius
22
23
D Discharge coefficient R Radius of the bubble column
24 D Molecular-diffusion coefficient ravg Mean radius of the bubble
25 D0 Diffusivity of a reference fluid Re Unsteady Reynolds number, Aωd b ν
26 d32 Sauter mean diameter
27 ro Initial bubble radius
db Mean bubble diameter
28 T Period of oscillation
29 deq Equivalent bubble diameter based on
t Time
30 project area, 4 A proj π tc Contact time for mass transfer,
31
32 E(Bj) [ ]
3 1 − (1 − Bj)1 / 3 Bj d 32 U b
33 f Vibration frequency u Liquid velocity field
34 F(t) Temporally evolving force on a bubble
35
Ub Bubble velocity, often measurement of
36 Favg Average force acting on the bubble individually bubbles
37 fc Critical frequency at which there is no Ug Gas phase velocity
38 net force on the bubble Usg Average superficial gas velocity,
39 fr Bubble resonant frequency
40
Q g Acs = ε U g
41 Fr Froude number, Aω gd b
V Instantaneous bubble volume
42 g Gravitational acceleration Vg Gas phase volume
43
44
G(Bj) [ ]
3 1 − (1 − Bj) 2 / 3 (2Bj ) Vl Liquid phase volume
45 h Bubble location below the free surface Vo Initial bubble volume
46 H Water column height xi Mole fraction of the liquid phase
47 Ho Water column height without air yi Mole fraction of the gas phase
48 z Vertical position relative to a fixed
injection.
49
hT Henry’s constant coordinate system
50
51 kL Mass transfer coefficient γ Specific heat ratio of gas (= 1.4 for air).
52 kLa Volumetric mass transfer coefficient (or Δh Measured manometer height difference
53 mass transfer rate) ∆V Deviation from Vo
54 (kLa)o Mass transfer coefficient without ∆Vmax Maximum volume deviation from Vo
55 ε Void fraction (gas holdup),
vibration
V g (V l + V g )
56
57 n Bubble count within a given volume
58 Nb Number of baffles per unit length (m-1), εo Void fraction with no vibration
59
60 57

1
2
3
εtrans Void fraction at regime transition
4 µ Liquid phase viscosity
5 ν Kinematic viscosity of the liquid phase
6 ν0 Kinematic viscosity of a reference fluid
7 ρ Liquid density
8
ρg Gas density
9
10 ρm Manometer fluid density
11 ρmix Mixture average density
12 σ Surface tension
13 φm Mixture conductivity
14
15
φo Conductivity for the no flow condition
16 ω Radian frequency
17 ωc Cut-out or critical frequency
18 ωs Stabilization frequency
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 58
1
2
3 For Table of Contents Only
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

Elbing 2016

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Elbing 2016

Uploaded by

Copyright:

Available Formats

Subscriber access provided by CMU Libraries - http://library.cmich.

Industrial & Engineering Chemistry Research is published by the American Chemical

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

Review of Vibrating Bubble Column Reactor Elbing, Still and Ghajar

You might also like