782
Energy & Fuels 1988,2, 782-786
trogen species. Although we cannot definitively separate the contributions of the PACN and soot to the yield of gaseous nitrogen, the data suggest that, in the PAC conversion regime of our measurements, the net effect of the conversion of the nitrogen in PAC is the production of gaseous nitrogen species, most probably HCN.
Acknowledgment. We gratefully acknowledge the National Institute of Environmental Health Sciences
(Grants NIH 5 P30 ES02109 and NIH 5 PO1 ES01640)for support of this research. We also express our appreciation to Anthony Modestino and Peter Monchamp for help with the experiments and analyses. Registry No. Quinoxaline, 91-19-0; 2,6-dimethylquinoline, 877-43-0;phenanthridine,229-87-8; carbazole, 86-74-8;9-ethylcarbazole, 86-28-2;2-phenylindole,948-65-2;9-phenylcarbazole, 1150-62-5; 13H-dibenzo[a,i]carbazole,239-64-5.
Wettability Measurements of Coal Using a Modified Washburn Technique Geatesh K. Tampy,?Wen-Jia Chen, Michael E. Prudich,” and Robert L. Savage Department of Chemical Engineering, Ohio University, Athens, Ohio 45701 Received August 17, 1987. Revised Manuscript Received July 21, 1988
The Washburn technique, which involves the wetting of a packed powder bed of solid by a liquid, was modified by using the fundamental equations of fluid transport through a packed bed. The modified version eliminates some of the inadequacies of the original technique: ambiguous calibration steps to estimate “orientation factors” or “equivalent radii” are not needed. Instead, the wetting of the powder by the liquid is quantified in terms of the bed porosity, a sphericity factor, and a tortuosity constant. The porosity is easily measured; the sphericity and tortuosity are obtained from the literature. Free energy changes during wetting and contact angles can be measured by this modified technique. The experimental procedure for making the measurements was also modified. The apparatus was altered to allow for the application of an external pressure, which acted in conjunction with the surface forces in causing the liquid to wet the solid. Thereby the particle mean diameters could also be measured. The wetting of three Ohio coals by three hydrocarbon oils was studied by using the modified Washburn technique. The interfacial free energy changes of wetting, contact angles, and particle mean diameters were measured. The values of the particle diameters were similar to those obtained by another technique. The interfacial free energy of wetting, which can be measured by this technique, is a property of fundamental importance in wettability studies. This technique is also simple to perform, is sensitive, and is reliable. It should therefore prove to be a useful technique in studies involving surface characterization of powdered coal.
Introduction The performance of several physical coal cleaning processes (e.g., oil agglomeration, flotation, etc.) is dependent on the surface properties of coal. Surface characterization is not easy because coal is heterogeneous; also, most of these processes utilize pulverized coal. This work addresses the need for a simple method that can properly characterize the interactions of liquids with polydisperse, heterogeneous solids like coal in powder form. One of the ways of characterizing solid-liquid interactions like those occurring in coal beneficiation is by the computation of the contact angle. Two of the older methods for measuring the contact angle in powder systems are the Bartell plug method’ and the Washburn capillary rise method? More recently, Heertjes and Kossen3 have suggested a different approach for contact angle measurement, which involves making a compressed pellet out of the powdered solid and measuring the height of a drop of liquid placed on the pellet. However, Neumann and Good4have indicated that pelletization changes the ‘Present address: Ames Laboratory, Iowa State University, Ames, Iowa 50011.
0887-0624/88/2502-0782$01.50/0
nature of the solid surface. The Bartell method is experimentally cumbersome. The Washburn technique, which also involves a plug of the solid powder, is much simpler to perform. However, the conventional application of this technique suffers from certain inadequacies that are explained in the next section. In this work, a modified version of the Washburn technique, which does not suffer from many of the limitations of the original technique, is presented. It is also shown that along with the contact angle, the free energy changes during wetting can be directly computed by this method. The free energy change being a fundamental thermodynamic property, its measurement is of greater significance than many of the empirical parameters that may be obtained by the several other popular surface characterization techniques in use today. Many of the implications of the free energy measurements that have been made by this technique, howBartell, F. E.; Merril, E. J. J. Phys. Chem. 1932,36,1178.Bartell, Whitney, C. E. Zbid. 1932,36,3115. Washburn, E. W. Phys. Rev. 1921,17, 273. Heertjes, P. H.;Kossen, N. W. F. Powder Technol. 1967,2, 33. Neumann, A. W.; Good, R. J. In Surface and Colloid Science; R. J., Stromberg, R. R., Eds.; Plenum: New York, 1979;Vol. 11.
0 1988 American Chemical Society
Energy & Fuels, Vol. 2, No. 6,1988 783
Wettability Measurements of Coal
vamr riv
\I
mid YL v
#
I
so d (a)
(b)
Figure 1. Contact angles shown (a) for a liquid drop on a flat solid surface and (b) for liquid rising up a capillary of radius r. In the latter case, the interfacial force acting along the perimeter of the capillary (i.e. 27rrnryLvcos 0) may be represented b a capillary pressure that acts over its cross section (i.e. APcrK.
ever, are discussed in a separate paper;s this paper is devoted to the theory and experimental methodology of the modified Washburn technique.
Theory The contact angle, which is a parameter involving three phases, is measured at any point along the line where an interface involving two of the phases meets the third phase (Figure 1). The three interfacial tensions act as shown in Figure la, and a balance of the horizontal components yields6 YSL - Ysv = -YLv cos e (1) Since each term on the right also represents the free energy of an interface, the term -yLv cos 8 is the change in free energy (AG)when a solid-vapor interface is replaced by a solid-liquid interface. That is, it represents the free energy change during a phenomenon like wetting, in which one interface is replaced by another. Interfacial forces can cause a liquid to rise up a cylindrical capillary. This occurs when the “wetting” of the solid (constituting the capillary) by the liquid yields an interfacial energy that is lower than the solid-air interfacial energy. That is, since all systems have a tendency to move toward a state of lower energy, given a “wetting liquid” (i.e., one which causes a decrease in free energy) and a capillary of small enough radius (for large radii, the weight of the liquid is no longer insignificant compared to the interfacial forces), the liquid will spontaneously rise up the capillary. The capillary pressure causing such a phenomenon may be related to the liquid-air interfacial tension (yLv) and the contact angle (e) by what is known as the Laplace equations (see Figure Ib):
In the Washburn technique for contact angle measurement, the Hagen-Poisuelle equation for flow of liquid through a circular pipe is applied to describe the capillary rise phenomenon. The Hagen-Poisuelle equation simply states that the rate of flow (dlldt) of liquid through a pipe varies directly with the driving pressure (AP)and the cross sectional area available for flow; it varies inversely with the liquid viscosity q, and with the length 1, through which the liquid had to flow. Or dl - r2 _ - -AP (3) dt 8q1 In the Washburn technique, the above equation is used with the capillary pressure (AP,)replacing AP,and the height, h, through which the liquid has risen replacing the (5) Tampy, G. K.; Prudich, M. E.; Savage, R. L.; Williams, R. R. Energy Fuels, following paper in this issue. (6) Davies, J. T.; Rideal, E. K.; Interfacial Phenomena, 2nd ed.; Academic: New York, 1963.
length, 1. By substitution of the expression for AP, from eq 2, if the rate at which the liquid rises up the capillary is measured, the contact angle can be easily computed if the liquid viscosity, liquid-air interfacial tension, and the capillary radius are known. It must be noted, from the expression yLvcos 19 that appears in eq 2, that the free energy change of wetting can be directly computed by this technique. The Washburn technique is usually applied to study the wetting of powdered solids by liquids. To do this, the powder is packed into a tube to form a porus plug, and the rate of rise of liquid up the porous plug of powder is measured. The porous plug is considered to be a bundle of cylindrical capillaries. In the conventional technique, the capillary radius is then represented by an equivalent or effective radius r’, and an orientation factor, c. The radius r in eq 2 and 3 is replaced by cr’, and this substitution is assumed to account for all the system nonidealities (e.g. variations in the bed porosity and the tortuous nature of the flow path) that exist because the bed is not really a collection of parallel cylindrical capillaries. With these changes, the Washburn equation in integrated form becomes h2 cr’y~vCOS 8
_ --
(4)
t 211 Before the contact angle can be determined from the above equation, cr’ must be estimated. This is done by a separate “calibration“ step. Two methods are generally used for this calibration. In the fiist method,78 the wetting experiment is initially conducted with a liquid that ”completely wets” the solid. For such a liquid, the value of cos 8 is unity, thus eq 4 directly yields a value for cr’; this value of cr’ is assumed to hold for all other liquids wetting the powder. Considering for a moment that such an assumption is valid, this would involve the onerous task of finding a liquid that completely wets all of the chemically different fractions (organic and mineral) that are bound to be present in a heterogenous substance like coal. In the second method? cr’ is determined by using mercury. The contact angle for mercury with most solids is known to lie between 130 and 140O. A value in this range is assumed, and cr ’is calculated from mercury porosimetry data. The main drawback to this approach is that mercury does not spontaneously wet the coal surface (as the contact angle greater than 90° suggests). Mercury will have to be forced in with an externally applied pressure, and the extent to which it penetrates the powder bed will depend on the applied pressure. Therefore, the assumption that porosity is constant, which is implicit in the assumption of constant cr’, is clearly not valid. The ambiguities involved with the calibration steps described above may be removed by a more thorough analysis of the flow of fluid though the packed powder bed. Flow of fluids through a packed bed of solids has been extensively studied and is well documented in standard chemical engineering texts.lOJ1 The approach is similar to that used in the conventional Washburn technique: the bed is assumed to consist of a bundle of capillaries, and the Hagen-Poisuelle equation is applied. However, the treatment differs in the way the equation is modified to (7) Bruil, H. G.; van Aartaen, J. J. Colloid Polym. Sci. 1974,252, 32. (8) Liang, W.; Jiang, L. J. Coal Qual. 1987, 6, 44. (9) Hiemenz, P. C. Principles of Colloid and Surface Chemistry, 2nd ed.; Marcel Dekker: New York, 1986. (IO) Bird, R. B.; Stewart, W. E.; Lightfoot, E. N. Transport Phenomena; Wiley: New York (1960). (11) McCabe, W. L.; Smith, J. C.; Unit Operations of Chemical Engineering, 3rd ed.; McGraw-Hill: New York, 1976.
Tampy et al.
784 Energy & Fuels, Vol. 2, No. 6, 1988
make it applicable to tangled tubes of complicated cross section. To account for the complicated cross section, the hydraulic radius is introduced. The hydraulic radius is defined as the ratio of the cross sectional area available for flow to the wetted perimeter. Thus for a circular tube, the hydraulic radius R is given by
The advantage of using the hydraulic radius is that it may be expressed in terms of the bed porosity (e) and the wetted surface per unit volume of bed (a), as shown by Bird et al.1° cross section available for flow R= (6) wetted perimeter Considering flow through unit height of the bed volume available for flow R= total wetted surface volume of voids volume of bed
=
[ [
I
wetted surface] = volume of bed
a t
Table I. Coal Properties (Proximate Ana1ysis)O coal % ash % VM % FC PSOC 276 13.5 37.2 49.3 PSOC 305 22.6 32.9 44.5 PSOC 751 6.0 41.5 52.5 "Key: VM, volatile matter; FC, fixed carbon. All values were reported on a dry basis.
oil No. 2 fuel oil Varsol pentane
Table 11. Oil Properties' surface tension, dyn/cm viscosity, cP density, g/cm8 27.9 27.5 18.1
0.878 0.788 0.622
3.40 1.04 0.24
" Properties determined at 25 OC. Pressure Re ulator
(7)
(8)
__t
-Nit
wen
The parameter "a" is related to the specific surface a, (the total particle surface area per unit volume of bed) by a = a,(l - t) (9)
For spherical particles, the mean diameter D is related to a,:
D = 6/a,
For nonspherical particles, a sphericity factor 9 is introduced, and the hydraulic radius is then given by R = Dt@/6(1- t) (11) The value of @ may be obtained from the literature,12and is 0.73 for fine coal particles. The particle mean diameter and bed porosity are measurable quantities. One must also account for the fact that the fluid penetrates the bed by a tortuous path that is actually greater than the measured height, h. A vast amount of experimental data that have been collected by various researchers suggest that the theoretical formula may be made to fit the experimentaldata if a tortuosity constant is introduced in the denominator of eq 4, to account for the increased flow pathaloThe value of the tortousity constant is found to lie between 2 and 2.25. Bird et al.1° suggest the value 25/12, which is the value adopted in this work. For two beds comprising the same powdered solid and being wetted by the same liquid, the tortousity factor may be assumed to be the same, provided the packing of the powder in the two beds are similar. Taking advantage of this fact, the experimental procedure for the Washburn technique may be improved by the application of an external pressure @, which is of the same order of magnitude of the capillary pressure, @, and acts along with AP, to drive the fluid through the bed. With all these modifications, eq 4 becomes
_ h2 -t
[
( 0 ~ 4 ) ~6(l;;jAG 75(1 - C ) ~ V
(12) Perry, R.H., Chilton, C . H.;E&. Chemical Engineers' Handbook, 5th ed.; McGraw-Hill: New York, 1973.
Manometer
Figure 2. Experimental apparatus for the determina on of the contact angle by the modified Washburn technique.
Thus, two rates of penetration of the powder bed by the liquid may be measured-one with only AP, providing the driving force and another with the driving force being (AP, + AP,).The two rates obtained with the two different driving forces yield two equations. These two equations may be solved for the interfacial free energy change during wetting (AG or yLvcos 0) and the particle mean diameter, D.
Experimental Section Three coals, PSOC 276 (Ohio No. 8, hvAb), PSOC 305 (Ohio No. 11, hvBb), and PSOC 751 (Ohio No. 6, hvBb) were chosen for study. These coals are among a group of Ohio coals currently being evaluated for their beneficiation potential a t Ohio University. Upon their receipt from the Penn State Coal Sample Bank, the coals were stored under nitrogen, wet ground, and vacuum dried. Properties of the coals are given in Table I. Three oils were also used. These included No. 2 fuel oil, Varsol (a commercial petroleum distillate), and pentane. Relevant properties of the oils are shown in Table 11. Mercury porosimetry data were obtained by a Micrometrics mercury porosimeter with which both mercury intrusion and extrusion could be performed over a range of 0-50000 psi of pressure. Particle mean diameters were obtained by using a centrifugal photosedimentometer that operates by means of a light extinction method (Hitachi Horiba Capa 300) and has a lower
Energy & Fuels, Vol. 2, No. 6,1988 785
Wettability Measurements of Coal 1 2 r
35 30 26
3
E0
20
%
15
L 00
1
100
10000
PRESSURE (PSI)
10
Figure 3. Mercury intrusion data for two Ohio coals.
-0-
Table 111. Specific Oil Draw up for Coal Samples draw up, mL/g coal sample No. 2 fuel oil Varsol pentane 0.92 0.88 0.90 PSOC 276 PSOC 305 0.88 0.87 0.82 0.83 0.82 PSOC 751 0.85
particle size limit of 0.04 pm. Mean diameters were also obtained using the modified Washburn technique. The apparatus shown in Figure 2 was used for the capillary rise experiments. The dried coal powder was packed into glass tubes (120mm long X 10 mm i.d.) having fritted glass bottoms. A standard tapping procedure waa used to insure that the packing, and hence the tortuosity, did not vary much from one powder bed to another. It was found that after 500 taps, dealt out 50 at a time, interspersed with the addition of more coal to the tube, the variation in the fiial height of the sample in the tube did not change more than 2 mm (i.e., by more than 1.7%). The tubes were weighed before and after filling, and the mass of coal per unit volume of the bed was computed. For the same coal, this mass was found to vary by less than 0.05 g for different experimenta. The rate at which the liquid penetrated the packed bed was measured after the sides of the tubes were marked with graduations and the tubes were placed vertically so that the bottom of the tube was just in contact with the wetting liquid. If the liquid were to wet the solid, the rise of liquid up the bed could clearly be observed, and the height could then be measured as a function of time. The application of an external pressure would cause the height to change at a faster rate. The magnitude of the external pressure, when applied, was measured from the manometer reading (see Figure 2). A t the conclusion of the experiments,the tubes were once again weighed. From this weight, the weight of liquid per unit volume of the bed, which is the void volume or bed porosity (c) filled by the liquid, was calculated.
Results The mercury intrusion data for PSOC 276 and PSOC 751 are shown in Figure 3. I t is seen that the volume of mercury penetrating 1 g of sample is a function of the applied external pressure, varying from 0 to 1.15 mL as the external pressure changes from 1to 50 O00 psi. Table I11 shows the specific draw up of oil for the three coals under study. These values correspond to the amount of oil that has penetrated the bed of powdered coal aided purely by the surface forces, with no external pressure applied. Thus it is clear that penetration of the beds by ~~
~
5
-A-A-
PSOC PSOC PSOC PSOC PSOC PSOC
176; 276; 305; 305; 751; 751;
RUN RUN RUN RUN RUN RUN
1 2 1 2 1
2
0
TIME (sec) Figure 4. Representative capillary rise data for the intake of No. 2 fuel oil by coals PSOC 276,PSOC 305,and PSOC 751,with no external pressure. z
30
AP-0 psig; Run 1 AP-0 psig; Run 2 AP-1.67 psig 3- dP-1.98 paig -A-0-
+
20 t,
b N
10
0
0
500
1000
TIME (ecc) Figure 5. Representative capillary rise data for the intake of No. 2 fuel oil by PSOC 276, with and without externally applied pressure.
liquids with vastly different wetting properties are by no means similar. Therefore, the assumption that the equivalent radius is constant for liquids with vastly different wetting properties is a major drawback of the conventional Washburn technique. In the modified Washburn equation, the variations that are seen in Table I11 are
Table IV. Free Enern Changes of Wetting and Contact Andesa Varsol pentane AG,dyn/cm cos 8 u AG,dyn/cm cos 8 u AG,dyn/cm cos 8 0.01 5.0 0.11 0.03 3.1 0.12 7.3 0.26 0.01 2.4 0.04 0.04 1.4 0.05 3.3 0.11 0.7 0.09 0.02 1.2 0.07 2.0 0.18 0.01
No. 2 fuel oil
coal P S O C 276 PSOC 305 PSOC 751 a
u
= standard deviation for cos 8, n = 6.
U
0.02 0.01 0.01
Tampy et al.
786 Energy & Fuels, Vol. 2, No. 6,1988 Table V. Oil Agglomeration Results for the Three Coals and Oils AG, % ash % Btu coal oil dvn/cm reduction recoverv 42.8 94.1 PSOC 276 No. 2 fuel oil 7.3 43.8 97.6 Varaol 3.3 47.1 62.2 pentane 2.0 45.5 97.0 PSOC 305 No. 2 fuel oil 3.1 23.2 84.4 Varsol 1.4 24.7 40.3 pentane 0.7 41.0 99.7 PSOC 751 No.2 fuel oil 5.0 38.6 98.8 Varsol 2.4 22.8 50.8 pentane 1.2 Table VI. Mean Particle Diameters (rm) Obtained by the Modified Washburn Technique and by the Use of a Horiba CAPA 300 Particle Size Analyzer, Which Are the Average of n Values, and the Corresponding Standard Deviations
(4 coal PSOC 276 PSOC 305 PSOC 751 U
n
by
mod Washburn techn 11.3 11.9 10.1 0.8 6
from CAPA 300 8.4 9.5 8.1 0.8 10
accounted for by the parameter e, which is included in the equation. In the capillary rise experiments, the square of the measured height was plotted as a function of time, as shown in Figure 4. A good straight line fit was observed in every case. Reproducibility was also good: repeated experiments with the same coal-oil systems yielded lines with similar slope. On the other hand, the slopes obtained for different coal-oil systems were different, showing that the technique is fairly sensitive to different interfaces. Figure 5 shows results with and without an externally applied pressure. The free energy change in wetting, the contact angles, and mean particle diameters were calculated from these experiments. The values for contact angles and free energy changes are shown in Table IV. The average standard deviation
for cos 6 values was 0.18 (based on 54 sets of experiments), which translates to a deviation of about loin the calculated contact angle. It is more meaningful, however, to look at the values obtained for the free energy changes. An analysis of the free energy results are discussed in a separate paper? Table V shows oil agglomeration results. The best ash reductions and Btu recoveries were obtained for coal/oil combinations that showed a greater decrease in free energy. It must be noted, however, that other factors like the degree of mineral matter liberation also affect ash reduction and Btu recovery. Table VI compares the mean diameter values obtained by the modified Washburn technique to those obtained by using a photosedimentation technique. Considering the fact that results from the latter technique depend on the use of an average value for particle density, the results are in fairly good agreement.
Conclusions A modified Washburn technique of contact angle measurement has been proposed that does not suffer from some of the inadequacies of the original technique. This modification removes the necessity of identifying an ideal wetting liquid or estimating equivalent radius through peripheral experimentation. The decrease in free energy associated with the replacement of a solid-air interface with a solid-liquid interface can be measured by using the modified method. Such measurements should be useful in predicting the efficiencies that may be expected in surface-based beneficiation processes such as oil agglomeration and flotation. Additional processing information, such as degree of mineral liberation, may be required in order to make quantitative predictions of beneficiation performance.
Acknowledgment. We thank the State of Ohio for its financial support and the Penn State Coal Sample Bank for supplying the coal samples used. Grateful thanks are also due to Dr.W. Kneller of the University of Toledo for his help in obtaining the mercury porosimetry data. Registry No. Pentane, 109-66-0.
