Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Patrick F. Dunn, Flint O. Thomas, James B. Leighton, Dayu LV

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Fuel 90 (2011) 12571263

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Determination of Henrys law constant and the diffusion and polytropic


coefcients of air in aviation fuel
Patrick F. Dunn , Flint O. Thomas, James B. Leighton, Dayu Lv
Department of Aerospace and Mechanical Engineering, University of Notre Dame, Notre Dame, IN 46556, United States

a r t i c l e

i n f o

Article history:
Received 19 May 2010
Received in revised form 24 November 2010
Accepted 26 November 2010
Available online 10 December 2010
Keywords:
Aviation fuel
Polytropic constant
Diffusion coefcient
Henrys law constant
Bubble

a b s t r a c t
The volume change of air microbubbles on surface nucleation sites was studied experimentally and compared with predictions. Measurements were used to determine the polytropic constant, the diffusion
coefcient, and the Henrys law constant of air in distilled water, dodecane, and JP-8 aviation fuel. The
liquids were exposed to sub-atmospheric pressures, but above their vapor pressures. In one type of
experiment, bubble size reduction was recorded as the liquids ambient pressure was increased from a
low pressure to atmospheric pressure though a series of step increases. The results were used to determine the polytropic constant. In another type of experiment, bubble growth was monitored in time following a sudden reduction in the liquids pressure from ambient. The EpsteinPlesset model of mass
diffusion was coupled with a Lipschitzian optimization technique to determine the values of the diffusion
coefcient and Henrys law constant.
2010 Elsevier Ltd. All rights reserved.

1. Introduction
This paper presents the results of an investigation to determine
the polytropic constant, the diffusion coefcient, and Henrys law
constant for air in JP-8 aviation fuel. The study was undertaken because of a lack of this information in the open literature. Such property information is a necessary part of the larger effort to
understand cavitation behavior in modern aircraft-fuel systems
[1].
Cavitation in an aircraft-fuel system can lead to unexpected
degradation in system performance and/or damage to fuel system
components. These systems often are characterized by complex
internal ow geometries that involve very narrow ow path
restrictions and sharp turns. Such geometries can lead to localized
regions of high uid velocity and low static pressure. For some scenarios in which the static pressure is close to but above the liquid
vapor pressure, gaseous cavitation occurs. For other scenarios, the
static pressure can go below the liquid vapor pressure, leading to
both gaseous and vaporous cavitation. The present study focused
on bubble growth related to gaseous cavitation.
It is well known that nucleation sites on surfaces are necessary
to catalyze the bubble growth that produces cavitation. The nucleation of bubbles into gaseous or vaporous cavities have been
classied into four different types of events [2]. Two types of nucle-

Corresponding author. Tel.: +1 574 631 6089.


E-mail address: pdunn@nd.edu (P.F. Dunn).
0016-2361/$ - see front matter 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.fuel.2010.11.039

ation events can occur in ow devices such as in aircraft-fuel


systems at pressures at or above the vapor pressure of the liquid.
This is attributed to pre-existing gas and/or vapor-containing sites,
which primarily are stable on the solid surfaces that conne the liquid or are present within the liquid (microparticles) [3]. Solid
microparticles in the form of silicates and iron oxides typically
are present in aviation fuel and harbor numerous nucleation sites
for cavitation inception. In addition, fuel containment surfaces
can have many nucleation sites.
JP-8 is the aviation fuel used most commonly by the United
States military [4]. Approximately 60 billion gallons of JP-8 and
its commercial equivalent, Jet A-1, are used annually worldwide
[5]. JP-8 is a complex mixture of over 228 hydrocarbons and various additives to meet military specication MIL-DTL-83133 [6,7]. It
is comprised of 18% (by volume) aromatics, 20% naphthenes, 60%
parans, 2% olens [8]. The primary constituent of JP-8 is kerosene,
which consists of hydrocarbons, mainly in the C9C16 range [9]. Details regarding the bulk uid properties and the chemical composition of JP-8 can be found in [7].
Modelling the ow of JP-8 through complex geometries requires accurate knowledge of the properties of air dissolved in
JP-8. Little such information is available in the open literature for
the properties that govern polytropic and diffusional processes.
These include the polytropic constant, the diffusion coefcient,
and Henrys law constant. Further, there is no information in the
open literature about the effect of sub-atmospheric pressure on
these three properties. Addressing this paucity of information
through benchmark experiments was the primary reason for the
present investigation.

1258

P.F. Dunn et al. / Fuel 90 (2011) 12571263

2. Experimental approach
Two different series of experiments were performed using the
same experimental apparatus and diagnostics. These characterized
the volumetric change of air microbubbles on surface nucleation
sites immersed in a static, constant-temperature liquid. The liquid
was exposed to sub-atmospheric pressures, but above its vapor
pressure. In the rst series, the relatively instantaneous volumetric change was recorded as the liquids ambient pressure was increased from a low pressure to atmospheric pressure though a
series of step increases. This determined the microbubbles pVk
(pressurevolume) behavior, as characterized by its polytropic
constant, k. In the second series, the microbubbles longer-term
(diffusional) growth was monitored in time following a sudden
reduction in the liquids ambient pressure. This information was
used to determine values of the diffusion coefcient, D, and
Henrys law constant, H. Similar techniques have been employed
in the past to determine the diffusion coefcients for a gas in a liquid, such as O2, N2, and He in water and in organic liquids [10].
A schematic is shown in Fig. 1. Further details of the experimental apparatus and diagnostics used for the present experiments are
described [11]. The test cell, mounted on a three-axis microtraverse, consisted of a hollow steel cube (5 cm-length side) with
six circular ports used for either viewing windows or instrument
feed-through. The bottom port held a pedestal that extended into
the near center of the cell. Three side ports were used for windows
to view the interior of the test cell. The fourth side port was used as
a feed-through for a type-K thermocouple that monitored the liquid temperature. The top port supported an adapter that connected the test cell to a vacuum system.
The test cell was illuminated using two ber-optic light guides
(positioned at each of the side-port windows) connected to a halogen lamp. A CCD camera with a macro lens and 6 extension tube
was used to monitor the microbubbles volume change. This magnication produced a resolution of 3.7 lm (a zero-order uncertainty
for a length measurement of 1.9 lm at the 95% condence level
[12]). The cameras framing rate was 33 Hz. The output of the camera was recorded by a personal computer. Individual frames of the
captured video subsequently were analyzed using MATLAB.
A pressure transducer was used to measure the static pressure
within the test cell. A ball valve isolated the test cell from a 3.8-L
reservoir. This reservoir was maintained at a desired pressure (to
as low as 0.5 psia for up to 2 h with only a 0.02 psia decrease in
pressure). The overall uncertainty in the measured pressure was

0.25 kPa (0.04 psi), estimated at the 95% condence level. The liquid temperature was monitored and remained at 295 1 K during
all experiments.
The top surface of a stainless steel pedestal contained the
microbubble nucleation site. This was a 1 mm-deep hole drilled
into the surface with 319 lm-diameter drill bit. This is shown
schematically in Fig. 2.
Three uids were used: aviation fuel (JP-8), dodecane (C12H26),
and distilled water. Dodecane served as a single-component hydrocarbon surrogate of JP-8, being its largest mass fraction component
(23%) and having a well dened vapor pressure. Distilled water
was used to validate the experimental technique and to provide
a baseline for comparison with the other uids.
The properties of the three uids used are summarized in Table
1. It is important to note that the property values of JP-8 can differ
because of the variability in its mixture, as well as its storage
history and location. For example, the vapor pressure of JP-8 gathered from seven different locations in the US ranged from 74 Pa to
430 Pa (at 294 K [70 F]), with an average vapor pressure of 240 Pa
[14]. The static pressures used in the present investigation were
above this range of JP-8s vapor pressures and those of dodecane
and water.
JP-8 in its natural, unsettled state contains solid microparticles
in the form of iron oxides and silicates. Data provided by the Honeywell Corporation indicated a particle mass concentration of
2 mg/L fuel and a number concentration of 107 particles/L fuel.
Each of the liquids used in the present experiments was essentially
free of particles. The distilled/deionized water was obtained from a
ltered laboratory water-purication system. The dodecane was
purchased commercially in 5-L containers. While the JP-8 did contain microparticles, as described previously, these were effectively
ltered from the JP-8 over time by gravitational settling in the storage tanks.
In the present experiments, JP-8 was stored at ambient temperature and pressure (295 K and 101.3 kPa, respectively) in
closed, 3.8-L containers until an experiment was performed. Under
these ambient conditions, the dissolved content of air in JP-8 is
0.139 mm3 air/mm3 fuel (based upon 78% of N2 and 21% O2, having
solubilities of 0.120 mm3 N2/mm3 fuel and 0.215 mm3 O2/mm3
fuel, respectively) [7]. The fuel was obtained commercially and
was not degassed prior to experiment. This was done to examine
the behavior of JP-8 stored under typical conditions without any
special treatment or handling such that the results were more relevant for typical situations.

Voltmeter
Analog Pressure Gauge

Digital Pressure
Transducer
Ball Valve 1
Astrovid 2000
CCD Camera

Vivitar Pentax
Macro Lens
w/ 6x Extension

Ball Valve 2
Reservoir

Test Cell
Thermocouple
and Readout
Pressure
Release Valve

Computer

3-Axis Traverse
Fig. 1. Schematic of the experimental apparatus.

Vacuum Pump

1259

P.F. Dunn et al. / Fuel 90 (2011) 12571263

that occurs across the microbubbleliquid interface. In the rst series, the microbubble radius is measured before and after a step increase in pressure, both at steady state. In the second series, the
microbubble radius is measured over the time following a step decrease in pressure.
For all three liquids examined in these experiments, the liquid
pressure, either before or after the imposed change in pressure,
was on the order of 101000 greater than the liquid vapor pressure and the interface pressure change. Thus, for these experiments, pg(t)pl(t) and the microbubbles content is primarily air.
For this situation, the change in the microbubble volume is governed by the mass diffusion of primarily air (mainly N2 and O2).

Vb
Rb

Rc

Vh

3.1. First experimental series


In this series of experiments, the microbubbles radius was recorded as the liquids ambient pressure was increased from a low
pressure to atmospheric pressure though a series of step increases
in pressure. For the range of pressures and times investigated,
microbubbles of air in either water, dodecane or JP-8 were found
to respond nearly instantaneously (<1 ms) and isothermally to step
changes in pressure.
The experimental results are presented in Figs. 35 for water,
dodecane and JP-8, respectively. In these gures, the dimensionless
volume, V, is plotted versus dimensionless pressure (the pressure
ratio), p. In this manner, the polytropic relation becomes kV = p.
These dimensionless variables are dened as

Fig. 2. Schematic of a bubble and its drilled nucleation site.

Table 1
Properties of uids used in the experiments. Values obtained from [1315].

Molecular weight, MW (g/mol)


Density, q (kg/m3)
Absolute viscosity, l (mPa s)
Surface tension, r (mN/m)
Vapor pressure, pv (Pa)

Water

C12H26

JP-8

18.015
997
0.99
72
2123

170.34
752
1.34
25
16

173
796
1.62
23
240

V  log10 V i =V f 

3. Results
and
For both series of experiments described in the following, a
change in microbubble volume occurs as a result of an imposed
change in liquid pressure. The pressure of the gas inside the microbubble, pg(t), is given by

pg t pl t  pv 2r=Rt;

where pl(t) is the liquid pressure, pv the vapor pressure of the liquid,
r is the surface tension of the liquid, and R(t) is the radius of the
microbubble. The term 2r/R(t) represents the change in pressure

p log10 pgi =pgf :

Here, V is the entire volume of the gas microbubble (Vg + Vh), where
Vg is the volume of the microbubble per se and Vh is the constant
volume of the nucleation site hole (refer to Fig. 2). The subscripts
i and f denote initial and nal steady-state values, respectively. In
each case, the solid line depicts the least-squares linear regression
ts of the data and the dashed lines indicate the lower and upper
95% condence interval limits.

2.8

Linear Regression Fit


(k = 0.97)
2.4

dimensionless volume

Isothermal Model
1.6

1.2

0.8

0.4

0
0

0.4

0.8

1.2

1.6

2.4

2.8

dimensionless pressure
Fig. 3. Dimensionless volume versus dimensionless pressure for water. The solid line is the linear best-t of the 39 data pairs; the dashed lines are the lower and upper 95%
condence limits. The regression coefcient is 0.9826 for 39 data values.

1260

P.F. Dunn et al. / Fuel 90 (2011) 12571263

1
0.9
Isothermal Model

dimensionless volume

0.8
0.7
0.6
0.5
Linear Regression Fit
(k = 1.01)

0.4
0.3
0.2
0.1
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

dimensionless pressure
Fig. 4. Dimensionless volume versus dimensionless pressure for dodecane. The solid line is the linear best-t of the 39 data pairs; the dashed lines are the lower and upper
95% condence limits. The regression coefcient is 0.9624 for 39 data values.

1.6

1.4
Linear Regression Fit
(k = 0.96)

dimensionless volume

1.2

1
Isothermal Model

0.8

0.6

0.4

0.2

0.2

0.4

0.6

0.8

1.2

1.4

1.6

dimensionless pressure
Fig. 5. Dimensionless volume versus dimensionless pressure for JP-8. The solid line is the linear best-t of the 49 data pairs; the dashed lines are the lower and upper 95%
condence limits. The regression coefcient is 0.9887 for 49 data values.

For all three liquids, the slope, k, is unity to within experimental


uncertainty (water: 0.967; dodecane: 1.012; JP-8: 0.952). Further,
the condence that the data are described by a linear t is greater
than 99% (see [12]). For water, the slope agrees with that reported
by Ran and Katz [16] to within 0.004, or 0.4%. These results imply
that the response of an air bubble to a sudden pressure change is
isothermal in each liquid, for which k is unity.
3.2. Second experimental series
In this series of experiments, the microbubbles growth was
monitored following a sudden reduction in the liquids pressure

from ambient to a sub-atmospheric value. The resulting measured


microbubble radius versus time for each of the three liquids at various pressures are presented in Figs. 68. For each liquid and pressure case, the microbubble radius increased with time. As the
liquid pressure was reduced, the microbubble radius became relatively larger and its growth rate greater as compared the previous
higher-pressure case. This data was compared with the predictions
of the EpsteinPlesset model [17], which are shown as solid curves
in the gures.
The EpsteinPlesset model considers the diffusion of a gas bubble in a liquidgas solution. It considers the change of the radius of
a center-stationary, spherical bubble in time for either under-satu-

1261

P.F. Dunn et al. / Fuel 90 (2011) 12571263

1000

900

800
16.55 kPa

radius, m

700

600

500
14.48 kPa
400

300

200

100
0

50

100

150

200

250

300

350

400

time, s
Fig. 6. Bubble radius versus time for two different-pressure water cases. Solid curve depicts model; solid circles data.

650
5.52 kPa
600
13.79 kPa

radius, m

550

500

450

400

350

300
0

10

15

20

25

30

35

40

45

time, s
Fig. 7. Bubble radius versus time for two different-pressure dodecane cases. Solid curve depicts model; solid circles data.

rated or over-saturated liquidgas solutions. The time rate of


change of the bubble radius, dR/dt, results from diffusional mass
transfer and can be expressed as





dR D pi  pRu T=MW 1
1
R0

p 1  p 2 ;
dt H
p 2r=R
R
pDt
2R

in which R = R(t) is the radius, R0 the initial radius, Ru the universal


gas constant equal to 8313.3 J/(kg mole K), MW the molecular
weight (=28.966 kg/kg mole for air), T the temperature, pi the initial
pressure of the liquid, p the pressure of the saturated liquid, D the
diffusion coefcient of the gas in the liquid, and H Henrys law constant for the gas.

The last term in brackets on the right hand side of Eq. (4) is a
present modication to the original equation to account for the reduced surface area of the bubble being in contact with a solid surface. The actual surface area of the bubble in contact with the
liquid, Ab, equals the surface area of a bubble of radius R minus that
of a truncated spherical cap of radius Rc (refer to Fig. 2). This difference can be nondimensionalized
by the surface area of a bubble to
p
become v 1  Rc = 2R2 . In the experiments, v was initially
0.5. As time progressed, v rapidly approached unity, reaching a
value of 0.93 at R = 10Rc.
Using the initial condition R(0) = R0, which is determined experimentally, Eq. (4) can be integrated numerically using a 4th-order
RungaKutta technique to nd R(t). For the present situation, this

1262

P.F. Dunn et al. / Fuel 90 (2011) 12571263

300
4.14 kPa
13.79 kPa

22.07 kPa

250

26.89 kPa

radius, m

200

150

33.79 kPa

100

50

0
0

10

15

20

25

30

35

40

45

time, s
Fig. 8. Bubble radius versus time for ve different-pressure JP-8 cases. Solid curve depicts model; solid circles data.

approach was coupled with an optimization technique to nd the


values of D and H that gave the best-t with data. The deterministic optimal search method called DIRECT [18,19], was implemented to nd global optimum values. One of its advantages is
the ability of search globally in a multi-dimension parameters
space with simple boundaries.
The numerical search was conducted in two steps. First, D and H
were allowed to vary. The initial radius of the bubble was assumed
to be of the order of that observed in the experiments (typically
several micrometers). This initial residual-minimization search
yielded values of H that were close to the known or estimated values presented in Table 2. The search supported that H did not depend upon the pressure, as is known for pressures up to approx
5 atm [20]. The converged values of D, however, were found to vary
with pressure. In the second, subsequent step, H was held xed for
each liquid. Values of D for each liquid were found to increase with
decreasing pressure. The nal, optimum values of H and D for all
experimental cases are presented in Table 3.
The resulting pressure-diffusion coefcient pair values could be
tted best by the relation pD = constant. This relation is similar to
that obtained for a binary mixture of gases [20,21]. A comparison

Table 2
Reference values for H and D. Units are kPa for pressure, MPa m3/kg for Hcr, and (m2/
s)  109 for Dcr. Values are for atmospheric pressure. Subscript cr for H and D denotes
values obtained computation and/or published reference. Superscripts denote sources
for values:

Liquid

Hcr

Dcr

JP-8
Dodecane
Water

4.34a
4.33c
3.975.90d

0.872.50b
1.203.50b
2.104.73d; 2.03e; 1.63.0f

Upper limit from [22].


Using StokesEinstein diffusion equation and range of D values for water from
1.63 to 4.73.
c
Average of ve values from [23].
d
From [24] from 17 C to 47.3 C.
e
From [25].
f
From [26] for stationary bubble from 8 C to 27 C.
b

Table 3
Determined values of H and D. Units are kPa for pressure, MPa m3/kg for H, and (m2/
s)  109 for D. Subscript m for H and D denotes values obtained from minimization
search.
Fluid

Pressure

Hm

Dm

JP-8
JP-8
JP-8
JP-8
JP-8
Dodecane
Dodecane
Water
Water

33.79
26.89
22.07
13.79
4.14
13.79
5.52
14.48
16.55

4.80
4.80
4.80
4.80
4.80
4.40
4.40
4.70
4.70

7.5
12.7
13.6
19.5
40.2
20.8
33.1
13.2
8.4

of this pD relation with the optimum values of H and D and data


is shown in Fig. 9 with pressure as a parameter for the ve JP-8
pressure cases examined. Similar results were obtained for the
dodecane and water cases. The values of D for all ve JP-8 cases
agree with the pD = constant relation to within 30% (at 95%
condence).
Extrapolation of the pD = constant relation to atmospheric conditions, as shown in Fig. 9 for the JP-8 cases, yielded values of the
diffusion coefcient at standard atmospheric conditions of
2.63  109 m2/s, 2.32  109 m2/s, and 1.63  109 m2/s, for air
into JP-8, dodecane, and water, respectively. Previously obtained
values at atmospheric conditions for air into water (stationary bubble diffusion) range from approximately 1.6  109 m2/s [26] to
4.7  109 m2/s [24]. Also, a value of 2.5  109 m2/s at 293 K
was reported for air into water [27]. The relatively lower value of
D obtained in the present experiments can be explained by the effect on diffusion of a plane surface tangent to a sphere [26] (refer
to Fig. 2), which reduces the diffusivity by a factor of ln 2 (=0.69).
The values of D for dodecane and JP-8 are within the ranges of
the estimated reference values (see Table 2). The values of H for
all three liquids also are very close to those estimated. The overall
uncertainties, estimated at the 95% condence level, are 5% for H
and 10% for D. Note that the only published values listed in Table
2 were those of D and H for air in water and H for air in dodecane.

1263

P.F. Dunn et al. / Fuel 90 (2011) 12571263


5

10

D, m2/s

10

10

Datm = 2.63 x
109 m2/s

10

10

10

20

30

40

50

60

70

80

90

100

110

120

pressure, kPa
Fig. 9. Diffusion coefcient versus pressure for ve JP-8 cases. Solid curve depicts Dp1 = constant best-t; solid circles optimized data values. The dashed and dotted curves
are 10% and 30% deviations from the solid curve, respectively.

4. Summary and conclusions


Values of the polytropic constant, the diffusion coefcient, and
Henrys law constant for air in JP-8, dodecane and water were
determined through combined experimental and numerical-optimization approach. The values obtained for water were consistent
with published reference values. Those for dodecane and JP-8 were
similar to estimated values.
The polytropic constant of air in either JP-8, dodecane or water
was found to be unity. This implies that the volume change of air
microbubbles in solution can be modelled as an isothermal process. Henrys law constant was found to be independent of pressure and different for all three liquids. The diffusion coefcient
was determined to increase with decreasing pressure for air in
each of the three liquids and described by the relation pD = constant for each liquid. This behavior is similar to that found for a
mixture of binary gases in which the diffusion coefcient is inversely proportional to the pressure.
These ndings, which are summarized in Table 3, provide much
needed property information for air in JP-8. They can be used in the
modelling of ow through geometries representative of aircraftfuel systems.
Acknowledgements
The experimental results presented in this paper were part of
the MS thesis research of J. Leighton. D. Lv performed the optimization calculations. This research was supported by the Honeywell
Corporation. We especially thank Steve Emo, Corey Bourassa, and
Abigail Parsons for their many technical interactions and support
throughout this research.
References
[1] Dunn PF, Thomas FO, Davis MP, Dorofeeva IE. Experimental characterization of
aviation-fuel cavitation. Phys Fluids 2010;22:117102-1117102-17.
[2] Jones SF, Evans GM, Galvin KP. Bubble nucleation from gas cavities a review.
Adv Colloid Interface Sci 1999;80:2750.
[3] Apfel RE. The role of impurities in cavitation threshold determination. J
Acoust Soc Am 1970;48:117986.

[4] Smith BL, Bruno TJ. Composition-explicit distillation curves of aviation fuel JP8 and coal-based jet fuel. Energy Fuels 2007;21:285362.
[5] Ritchie GD, Bekkedal MYV, Bobb AJ, Still KR. Biological and health effects of JP8 exposure. Report no. TOXDET 01-01, Naval Health Research Center
Detachment (Toxicology), Wright-Patterson Air Force Base, OH; 2001.
[6] Mayeld HT. JP-8 composition and variability. Report no. AL/EQ-TR-19960006, Armstrong Laboratory, Environmental Research Division, Tyndall Air
Force Base, FL; 1996.
[7] Handbook of aviation fuel properties, 3rd ed., Coordinating Research Council,
Inc.; 2004.
[8] Edwards T, Maurice LQ. Surrogate mixtures to represent complex aviation and
fuel rockets. J Propul Power 2001;17:4616.
[9] Violi S, Yan EG. Experimental formulation and kinetic model for JP-8 surrogate
mixtures. Combust Sci Technol 2002;174:399417.
[10] Krieger IM, Mulholland GW, Dickey CS. Diffusion coefcients for gases in
liquids from rates of solution of small bubbles. J Phys Chem 1967;71:11239.
[11] Leighton JB. Pressure and diffusion-driven bubble growth on a static surface.
MS thesis, University of Notre Dame, Notre Dame, IN; November 2006.
[12] Dunn PF. Measurement and data analysis for engineering and science. 2nd
ed. Boca Raton (FL): Taylor & Francis/CRC Press; 2010.
[13] Mackay D, Shiu WY, Ma KC. Illustrated handbook of physicalchemical
properties and environmental fate for organic chemicals. Volatile organic
chemicals, vol. III. Ann Arbor (MI): Lewis Publishers; 1993.
[14] JP-8 volatility study. Report no. IERA-RS-BR-SR-2001-0002, Southwest
Research Institute, San Antonio, TX; 2001.
[15] White FM. Fluid mechanics. 6th ed. New York (NY): McGraw-Hill; 2006.
[16] Ran B, Katz J. The response of microscopic bubbles to sudden changes in
ambient pressure. J Fluid Mech 1991;224:91115.
[17] Epstein PS, Plesset MS. On the stability of gas bubbles in liquidgas solutions. J
Chem Phys 1950;18:15059.
[18] He J, Watson LT, Ramakrishnan N, Schafer CA, Verstak A, Jiang J, et al. Dynamic
data structures for a direct search algorithm. Comput Optim Appl
2002;23:525.
[19] Jones JR, Perttunen CD, Stuckman BE. Lipschitzian optimization without the
Lipschitz constant. J Optim Theory Appl 1993;79:15781.
[20] Incropera FP, Dewitt DP. Fundamentals of heat and mass transfer. 5th ed. New
York: John Wiley & Sons, Inc.; 2006.
[21] Byrd RB, Stewart WE, Lightfoot EN. Transport phenomena. 2nd ed. New
York: John Wiley & Sons, Inc.; 2002.
[22] The installation restoration program toxicology guide, vol. 4, Oak Ridge
National Laboratory, report no. A200512; 1989.
[23] Mackay D, Shiu WY. A critical review of Henrys law constnats for chemicals of
environmental interest. J Phys Chem Ref Data 1981;10:117599.
[24] Shimiya Y, Yano T. Diffusion-controlled shrinkage and growth of an air bubble
entrained in water and in wheat our particles. Agric Biol Chem
1987;51:193540.
[25] Green DW, Perry RH. Perrys chemical engineers handbook. 8th ed. New
York: McGraw-Hill; 2008.
[26] Liebermann L. Air bubbles in water. J Appl Phys 1957;28:20511.
[27] Wise DL, Houghton G. The diffusion coefcients of ten slightly soluble gases in
water at 1060 C. Chem Eng Sci 1966;21:9991010.

You might also like