Upload
harikeshmn09
View
233
Download
0
Embed Size (px)
Citation preview
8/3/2019 Capillarity and Its Thermodynamics
1/31
8/3/2019 Capillarity and Its Thermodynamics
2/31
Even the problem of defining an equilibrium pressure relationfor a fluid in a porous medium such as would be
found in a capillary rise experiment is difficult to do interms of readily measurable parameters. Work along these
lines has been based largely on analyses that define pressurevia changes in the free energy of the fluid/media system with
respect to its volume. Interfacial area has been shown to beparticularly relevant e.g., 11 13, and a recent paper by
Cheng et al. 14 has demonstrated the connection betweenfluid interfacial area in a porous system and capillary pressure
by measuring both directly
8/3/2019 Capillarity and Its Thermodynamics
3/31
In the present work, we extend the thermodynamicdiscussion
of equilibrium capillary pressure to a fractal mediumand consider the conditions under which interfacial area willbe the dominant determining factor. We derive such a
relation
using the first law of thermodynamics applied to a fluidelement that distributes itself within a region of a porousmedium through a quasistatic process. By describing thepore space as a fractal, we are able to express the change
in
8/3/2019 Capillarity and Its Thermodynamics
4/31
fluid interfacial area with respect to volume and achieve aclosed-form result for capillary pressure that can be fit todata obtained from a capillary rise experiment conducted in
acoarse-grained SiO2 sand. The pore space of such media
isknown to exhibit fractal properties e.g., 15. The analysisalso shows that nonequilibrium capillary pressures will
occurwhen fluid-solid contact lines move within the region of
interestand indicates a dependence on the velocity of movement
as well as the degree of media saturation.
8/3/2019 Capillarity and Its Thermodynamics
5/31
Equilibrium fluid pressure inpourous medium
Consider an element of wetting fluid that enters a regionof porous media under pressure pw and distributes itself bydisplacing a nonwetting fluid that is at pressure pn . Let A bethe interfacial area of the fluid m2, Ap be the projectedinterfacial area of the fluid within a representative elementaryvolume m2, F be the force on the fluid-solid contactline per unit of its length per unit of its velocity kg/m s, F be the magnitude of the force on the fluid-solid contact lineper unit of
itslength per unit of its velocity kg/m s, g be thegravitational constantm/s2, L be the primary fractal length
scale m , Pc be the capillary pressure kg/m s2, pn be thepressure in the nonwetting fluid kg/m s2, pw be the pressurein the wetting fluid kg/m s2, r be the pore radius m, S be the entropy of fluid J/K , sv be the volume shape factor
8/3/2019 Capillarity and Its Thermodynamics
6/31
for wetting fluid in contact with the respective nonwettingphase, sv be the volume shape factor for wetting fluid in
contact with the respective nonwetting phase, T be the fluidtemperature K, U be the internal energy of the fluid J ,REV be the representative elementary volume over whichpressure is defined, Vw be the volume of the wetting fluid
m 3, Vn be the volume of the nonwetting fluid m 3, Vp bethe pore volume within a REV m3, V REV be the volume of
8/3/2019 Capillarity and Its Thermodynamics
7/31
8/3/2019 Capillarity and Its Thermodynamics
8/31
FIG. 1. Schematic of fluid-solid contact line, velocity, and resultant. a Schematic of the fluid-solid contact line for awetting fluid. b Schematic of the contact line and resultant surfaceforce acting on the solid. For equilibrium situations, Fns ,wn ,ws aregiven by Youngs equation 17 and FR is the resultant force exertedby the solid matrix e.g., 18. When the fluid-solid contact line isin motion, viscous drag on the fluid-solid contact line will also bepresent e.g., 8.
8/3/2019 Capillarity and Its Thermodynamics
9/31
a representative elementary volume REV m3, be thevelocity of the fluid-solid contact line in a REV m/s, bethe magnitude of
the contact line velocity within a REVm/s, be the total contact line length per unit of projectedarea Ap in a REV 1/m, be the fractal scaling factor,
water,air be the interfacial energy of the water/air boundaryJ/m2, water,solid be the interfacial energy of the water/solidboundary J/m2, w ,i be the interfacial energy of the wettingfluid in contact with the i th nonwetting phase J/m2, waterbe the density of water kg/m3, and be the wetting fluid
content of the media m fluid3 / m media3 .
8/3/2019 Capillarity and Its Thermodynamics
10/31
We adopt the convention that fluid pressure is defined as aquantity averaged over a representative elementary volumeREV within the soil matrix 16. Assuming that the surfacetensions of the fluid interfaces are functions of temperatureonly and that the system is taken to be the region occupiedby the wetting fluid, then for a constant porosity medium
thefirst law can be written using a control mass formulation,U = TS pwVw pnVn +w ,iAw ,i + F Ap . 1
8/3/2019 Capillarity and Its Thermodynamics
11/31
Here work done by the system is taken to be positive and theprimes indicate the derivative with respect to time. The last
term on the right considers the work done per unit time tomove the fluid-solid contact line. Fig. 1. Here F is the forceexerted by/on the contact line per unit of its length per unitof its velocity, is the contact line length per unit of fluidprojected area, Ap , is the velocity of the fluid-solid contact
line with its magnitude, and the average is performed overthe REV; see the Appendix. The summed terms account forwork corresponding to changes in the interfacial area of thefluid e.g., 19, with w indicating the wetting phase and i the nonwetting phases i.e., solid matrix or nonwetting phase being displaced. We consider an isothermal, quasistaticprocess where fluid entropy and internal energy can beconsidered to be constant. Such a process would approximatea slow capillary rise experiment conducted at constant tem
8/3/2019 Capillarity and Its Thermodynamics
12/31
pn pwVw = w ,iAw ,i F Ap , 2where we have made use of the fact that Vw =Vn . At the endof that process when = 0 , this can be rearranged to givePc = w ,iAw ,i / VU ,S ,T . 3Equation 3 shows that under this idealized model, the equilibriumpressure within the fluid matrix is determined by thechange in interfacial areas with respect to fluid volume andcan be shown to be a function of the medias saturation. Note that the last term in Eq. 1 only plays a role when the fluidsolidcontact lines are in motion, a situation of obvious nonequilibrium
8/3/2019 Capillarity and Its Thermodynamics
13/31
The result in Eq. 3 differs from that derived by Morrow12 in that no term for the interfacial energy of the nonwettingfluid-solid interface is included. However, since we definethe system as the region occupied by the wetting fluidonly, this term is rightly omitted. Equation 3 can also bederived as a subset of the more general analysis of Hassanizadehand Gray 8. Note that the restriction of Eq. 3 to aconstant entropy process does not violate the condition thatspontaneous processes of which soil wetting is one proceedtoward higher entropy as this requirement pertains to system
and surroundings. Entropy within the system can be writtenasS = S trans + S prod,
8/3/2019 Capillarity and Its Thermodynamics
14/31
where S trans results from heat and mass transfer into and outof the system and S prod represents entropy production withinthe system. The requirement that S =0 stipulates a balancebetween these terms, which then requires that the entropy ofthe surroundings increase even if that of the system is constant.Such a situation could well exist for isothermal, quasistaticprocesses such as those found in a slow capillary rise
experiment.In order to apply Eq. 3, a relation is needed for determiningthe change in interfacial area with respect to fluidvolume, and this can be obtained for a medium whose porespace exhibits fractal properties. A recent study by Bird andPerrier 20 suggests a relation for the pore size in a fractalsystem after T iterations,r = TL, 5where r is the characteristic size, L is the original lengthscale, and is a scaling factor. Correspondingly, the numberof such pores within a given set isn = L / rD
8/3/2019 Capillarity and Its Thermodynamics
15/31
with D the soil fractal dimension 20.If we assume that r represents pore radius, then we can take the interfacial area at a given scale to be sanr 2 withthe corresponding pore volume svnr 3, where sa and sv areshape factors e.g., 4 and 4/3, respectively, for a sphericalgeometry. If we assume that the interfacial areas Aw ,i aredirectly proportional to sanr 2, then the derivatives in Eq. 3are easily obtained,
8/3/2019 Capillarity and Its Thermodynamics
16/31
8/3/2019 Capillarity and Its Thermodynamics
17/31
EXPERIMENTAL RESULTS
A capillary rise experiment provides an ideal mechanismfor testing Eq. 8. Here fluid is drawn upwards from a saturatedregion by the network of pores above it with the capillary pressure increasing linearly with
height above the saturatedlevel. At any given elevation, small pores fill first withlarger ones doing so if their capillarity can overcome the
tension in the fluid surrounding them. The size of the largestpores filled therefore decreases with elevation above thesaturated region, which leaves an increasing fraction of thepore space empty and results in a decrease in moisture content.For a fractal system, all the pores in a given set are theneither filled or empty at a specific capillary pressure. Correspondingly,the value of Pc given by Eq. 8 increases as the
size of the largest pores filled, given by TL, decreases. Thesaturation of the media is defined as V / Vp =svnr 3 / Vp . HereVp is the total pore volume within a REV, V is the volume ofwater within that pore space, and the sum runs from thelargest pore filled defined by TL to the smallest, with T therefore being directly related to saturation.
8/3/2019 Capillarity and Its Thermodynamics
18/31
An eight-day capillary rise experiment was performed usinga 20 30 sieve SiO2 sand to obtain data on capillarypressure-versus-media saturation. The sand was held in analuminum chamber with interior dimensions of1.27 cm/ 20.3 cm/ 25.4 cm and the interior faces weretreated with a DowCorning silicon release compound in orderto render them hydrophobic. The only preparation of themedia was a thorough rinse with distilled water. Degassed,distilled water was introduced from the bottom of the chamberuntil a saturated region 5 cm deep was formed. Thechamber was sealed and allowed to equilibrate for 8 days at
a constant temperature of 25 C. The fluid content within theporous medium was measured, at 0.045 cm incrementsabove the saturated region, using neutron radiography as describedin Deinert et al. 21.
8/3/2019 Capillarity and Its Thermodynamics
19/31
Because the coefficients fw ,i are difficult to determine fornonidealized media, the factor f water,airwater,air+ f water,solidwater,solid2 D / L 3 D , as well as were
determinedby fitting Eq. 8 to data from the capillary riseexperiment. Taking the natural log of both sides of Eq. 8,we getlnPc = lnf water,airwater,air + f water,solidwater,solid D 2/ L 3 D T ln. 9Here Pc is the data vector of capillary pressures obtained
8/3/2019 Capillarity and Its Thermodynamics
20/31
8/3/2019 Capillarity and Its Thermodynamics
21/31
FIG. 2. Eight day capillary rise data and model fit.Data from an
8-day capillary rise experiment are shown alongwith the best fit of
Eq. 8 to those data. The model predicts a higherpressure at low
saturation then is shown by the data, however the8-day capillary
rise may not have provided sufficient time for thesystem to come toa true equilibrium
8/3/2019 Capillarity and Its Thermodynamics
22/31
experimentally, with T the correspondingfractal iteration
vector whose integer components rangefrom 0 to the number
of data points. A least-squares method wasused in conjunction
with Eq. 9 to determine bothf water,airwater,air+ f water,solidwater,solid D 2 / L 3 D and .
8/3/2019 Capillarity and Its Thermodynamics
23/31
DISCUSSIONFigure 2 shows a comparison of the data from the capillary
rise experiment along with the fit of Eq. 8, which reproducesthe shape of the data curve with good fidelity. Themodel predicts a higher capillary pressure at low saturationthen is shown by the data, however the 8-day capillary risemay not have provided sufficient time for the system to cometo a true equilibrium. While the total interfacial area in the
fractal model goes to zero as V goes to zero, the derivativeA / V increases monotonically, which is consistent with theexperimental results obtained by Cheng et al. 14. The specificentropy and
internal energy of water change by 0.1%between 0.1 and 1 MPa at 25 C e.g., 25. The capillaryrise experiment was conducted at this temperature, and for
the pressure ranges shown in Fig. 2 there would have beenvirtually no change in these quantities, which is consistentwith the assumptions under which Eq. 8 was derived. Thestrong fit of Eq. 8 to data also suggests that the pore spaceof the media indeed exhibits fractal properties
8/3/2019 Capillarity and Its Thermodynamics
24/31
For steady moisture content, Eq. 1 indicates that fluidpressure within a porous medium will be a functionPcU ,S ,T ,A,V . However, as with previous studies e.g.,2 4,7,8, Eq. 1 also suggests that movement of the
fluidsolidcontact lines within a REV will effect the pressure
relationship.Indeed, if PcU ,S ,T ,A,V is taken to be the equilibrium
pressure function, then, under appropriate conditions,Eq. 1 can be rewritten as
8/3/2019 Capillarity and Its Thermodynamics
25/31
Pc = PcU ,S ,T ,A,V /t , 10where the second term on the right considers nonequilibriumeffects arising from forces acting on the fluid-solid contactline during its
movement see the Appendix, with beingthe moisture content of the media. As with Eq. 2, the second
term on the right side of Eq. 10 represents the workdone in moving the total length of the contact line within aREV, but it is now written in a form similar to those found byprevious authors e.g., 9. While departures from equilibriumpressure are commonly attributed to this term, they
would also be expected to occur if temperature gradientswere to exist within the system, or if fluid interfacial areavaried in time even while moisture content remained fixed
8/3/2019 Capillarity and Its Thermodynamics
26/31
APPENDIXForces that act on the fluid-solid contact line during its
movement will cause nonequilibrium effects. Since pressureis defined as a quantity averaged over a REV 16, it isappropriate in Eq. 1 to also consider the average of theenergy required per unit time to move the contact line containedwithin such a region. This quantity is equal to theforce exerted by/on the contact line times the velocity at
which it moves, averaged over the REV, and can be writtenas F Ap . Here is the contact line length per unit ofprojected fluid interfacial area within the REV, Ap , is thevelocity of the fluid-solid contact line, is the magnitude ofthat velocity, and F is the force exerted by/on the contact lineper unit of its length per unit of its velocity see Fig. 3.
When does not vary with position over the volume ofthe integral, the work done per unit time in moving the fluidsolidcontact line may be written as F Ap . If the curvatureof the menisci with the REV are constant, then byconservation of mass Vw =Ap and Eq. 2 can be rewritten asPcVw = w ,iAw ,i F Vw . A1
8/3/2019 Capillarity and Its Thermodynamics
27/31
If the magnitude of the interfacial velocity does not varywithin the REV, then Eq. A1 can also be rearranged to givePc = w ,iAw ,i / Vu ,s ,T i k / i k FVw / Ap .A2Here i and k are the directions of the force exerted by/on the
fluid-solid contact line and the projected area vectors, respectivelysee Fig. 3 and where we have again made use of thefact that Vw =Ap . Noting that the moisture content=Vw / V REV, Vw can be written as V REV / t . The second term on the
right-hand side in Eq. A2 can then be written
as V REV / t i
k /
ik
F / Ap , which is positive wheni k / i k is positive and negative when it is negative seeFig. 3. Hence, Eq. A2 can be written
8/3/2019 Capillarity and Its Thermodynamics
28/31
8/3/2019 Capillarity and Its Thermodynamics
29/31
FIG. 3. Projected flow area. When the curvature of the menisci within aREV is constant, the volume flow rate into or out of the
REV is given by Ap , where the magnitude of the contact line andinterfacial velocities is assumed to be the same. The unit vectorsi , j ,k denote the directions of the force on the fluid-solid contact
line, the resultant force on the solid matrix, FR, and the direction inwhich the projected area moves, respectively. For a wetting fluid,the force on the fluid-solid contact line is in the direction of movementfor / t 0 and resists the movement of the fluid-solid contactline when / t 0.
8/3/2019 Capillarity and Its Thermodynamics
30/31
Pc = w ,iAw ,i / VU ,S ,T /t , A3where = F / Ap REV i k / i k dV . More generally, if termsfor U and TS are not dropped from Eq. 1, thenPc = PcU ,S ,T ,A,V /t , A4
where PcU ,S ,T ,A,V is understood to be the equilibriumcapillary pressure. Equation A4 is also obtained by
Hassanizadehand Beliaev 9 through a more general analysis.The coefficient takes into consideration the movement ofthe fluid-solid contact line within a REV and would need tobe determined experimentally in all but the simplest of
geometries
8/3/2019 Capillarity and Its Thermodynamics
31/31
Importantly, the parameter indicates that will havesome dependence on both media structure and moisture content.While the dependence of F on contact line velocity wasnot specified in the above derivation, simulations haveshown that the force required to move a contact line through
a smooth capillary tube is 1,22 and other studies haveobtained predictions of x , with x ranging from 0.4 to 9/723,24. Experiments done to measure the change in forcerequired to move a capillary interface through glass beadscontained in a thin tube have shown a dependence 1/2 2.
As a result, it is likely that the second term on the right-handside of Eq. 2 will increase in magnitude for higher velocities,indicating that will be correlated to /t.