Ind. Eng. Chem. Res. 1992, 31, 574-580
574
Levy, J. M.; Ritchey, W. M. J . High Resolut. Chromatogr. Chromatogr. Commun. 1987, 10, 493. Ohgaki, K.; Tsukahara, I.; Semba, K.; Katayama, T. Kagaku Kogaku Ronbunshu 1987,13,298. Paulaitis, M. E.; Alexander, G. C. Pure Appl. Chem. 1987, 59, 61. Sanders, T. A. B.; Younder, K. M. Br. J. Nutr. 1981,45, 613. Shafer, K. H. Anal. Chem. 1983,55, 1939. Squires, T. G.;Venier, C. G.; Aida, T. Fluid Phuse Equilib. 1983,10, 261. Suzuku, Y. Doctor's Thesis, Tohoku University, Japan, 1990. Taft, R. W.; Kamlet, M. J. J . Am. Chem. SOC. 1976,98,2886. Taniguchi, M.; Nomura, R.; Kamihira, M.; Kijima, I.; Kobayashi, T. J . Ferment. Technol. 1988, 66, 341. Yamaguchi, K.; Murakami, M. J . Jpn. Oil Chem. SOC.1986,35,260. Yamamoto, N.; Saitoh, M.; Moriuchi, A.; Monura, M.; Okuyama, H. J . Lipid Res. 1987,28, 144. Yonker, C. R.; Smith, R. D. J . Phys. Chem. 1988,92,2374. Yonker, C. R.; Gale, R. W.; Smith, R. D. J. Chromatogr. 1986a,371,
pressure dependence of the separation efficiency.
Literature Cited Arai, K.; Saito, S. J. Jpn. Oil Chem. SOC.1986, 35, 267. Brennecke, J. F.; Eckert, C. A. AIChE J. 1989,35, 1409. Hirano, A,; Hamazaki, T.; Terano, T.; Nishikawa, T.; Tamura, A.; Kumagai, A.; Sajiki, J. Lancet 1980, 11, 1132. Ikushima, Y.; Goto, T.; Arai, M. Bull. Chem. SOC.Jpn. 1987, 60, 4145. Ikushima, Y.; Hatakeda, K.; Ito, S.;Saito, N.; Asano, T.; Goto, T. Ind. Eng. Chem. Res. 1988,27, 818. Ikushima, Y.; Saito, N.; Goto, T. Ind. Eng. Chem. Res. 1989a, 28, 1364. Ikushima, Y.; Hatakeda, K.; Saito, N.; Ito, S.; Goto, T. Kagaku Kogaku Ronbunshu 1989b,15, 511. Ikushima, Y.; Saito, N.; Hatakeda, K.; Ito, S.; Goto, T. Chem. Lett. 1989c, 1707. Ikushima, Y.; Saito, N.; Arai, M. Bull. Chem. SOC.Jpn. 1991a, 64, 282. Ikushima, Y.; Saito, N.; Arai, M.; Arai, K. Bull. Chem. SOC.Jpn. 1991b, 64, 2224. Karger, B. L.; Eon, C.; Synder, L. R. J . Chromatogr. 1978,125, 71. Kim, S.; Johnston, K. P. Ind. Eng. Chem. Res. 1987,26, 1206. Lawson, D. D. Proceeding of the DOE ChemicallHydrogen Energy Contractor Review Systems; National Technical Information Service: Springfield, VA, 1978.
83.
Yonker, C. R.; Frye, S. C.; Kalkuaf, P. R.; Smith, R. D. J . Phys. Chem. 1986b,90, 3022. Wu, R.-S.; Lee, L. L.; Cochran,H. D. Ind. Eng. Chem. Res. 1990,29, 977.
Receiued for review April 25, 1991 Revised manuscript received September 16, 1991 Accepted October 13, 1991
Removal of Volatile Organic Compounds from Groundwater Using a Rotary Air Stripper Surinder P. Singh,*l+James H. Wilson,' Robert M. Counce,*JJohn F. Villiers-Fisher,' and Harold L. Jenningst Robotics & Process Systems Division and Chemical Technology Division, Oak Ridge National Laboratory,ll Oak Ridge, Tennessee 37831
Andrew J. Lucero and Gregory D. Reed Chemical Engineering Department and Civil Engineering Department, The Uniuersity of Tennessee, Knoxville, Tennessee 37996
Richard A. Ashworth and Mike G. Elliott United States Air Force, Tyndall Air Force Base, Florida 32403
The performance of a centrifugal vapor-liquid contactor was evaluated for air stripping of jet fuel components from groundwater. Hydraulic test data indicated that the Sherwood flooding correlation, which has been proposed for use in designing centrifugal vapor-liquid contactors, overestimates the rotational speeds a t which flooding occurs. A concept of the area of a transfer unit (ATU) was introduced in the mass-transfer testa to account for the change in fluid loading with the radius of the packing torus. A new correlation based on the specific surface area of the packing for predicting ATU described the experimental data with a fair degree of accuracy. The power consumed in rotating the packing torus was found to depend mainly on the liquid flow, outer rotor radius, and rotational speed. Previous claims in the literature that the centrifugal vapor-liquid contactor is resistant to fouling because of high shear forces were not found to be valid for groundwater with high iron content. 1. Introduction Air stripping is a process by which groundwater contaminated with slightly soluble volatile organic compounds *Address correspondence to this author at the Oak Ridge National Laboratory, P.O. Box 2008, Building 7601, MS/6306, Oak Ridge, T N 37831-6306. Robotics & Process Systems Division, Oak Ridge National Laboratory. Chemical Technology Division, Oak Ridge National Laboratory. *Alsoin the Chemical Engineering Department, The University of Tennessee, Knoxville, TN. 11 Managed by Martin Marietta Energy Systems, Inc., for the U.S. Department of Energy under Contract No. DE-ACOS840R21400.
*
0888-5885/92/2631-0574$03.00/0
(VOCs) may be purified by transferring the contaminants from the water to air. For efficient transfer of the contaminants from the liquid to the gaseous phase, intimate contact between the two phases is required. The contact between the two phases may be accomplished by mechanical surface aeration, diffused aeration, spray or tray towers, open-channel cascades, spray fountains, and countercurrent packed towers. In 1986, the United States Air Force and Coast Guard, in a joint project, evaluated the performance of a countercurrent packed vapor-liquid contactor that uses centrifugal force to drive the liquid through the packing for air stripping of jet fuel components from groundwater (Dietrich et al., 1987). This evaluation showed that the contactor was very effective in removing VOCs from groundwater. In 1987, the Air Force contracted Oak Ridge National Laboratory (ORNL) to perform fur0 1992 American Chemical Society
Ind. Eng. Chem. Res., Vol. 31, No. 2, 1992 575 I
LIQUID INLET
por-liquid contactor (Singh, 1989) is
8 ATIONARY ,(' LIQUID DISTRIBUTOR
Figure 1. Schematic of the centrifugal vapor-liquid contactor.
ther evaluation of this type of contactor and to develop design concepts which could lead to wider application of this technology. This paper presents the results of the hydraulic and mass-tranfer performance of a centrifugal vapor-liquid contactor for air stripping of jet-fuel components from groundwater and proposes new design correlations. 2. Air Stripper Design Equations The centrifugal vapor-liquid contactor (Figure 1) is composed of two major components: (1) the rotating packing and (2) the stationary housing. The liquid phase is fed into the center of the rotating packing and flows outward due to the centrifugal force. After exiting the packing, the liquid phase impacts the housing wall and flows by gravity out of the unit. The vapor phase is introduced into the annular space between the packing and the housing and flows inward due to the pressure-driving force. The rotating packing of the centrifugal vapor-liquid contactor is shaped like a torus. The hydraulic capacity of a packing unit is determined by the area normal to the flow at the inner radius (2?rrlZ) as the highest fluid velocities are encountered at this location. The outer radius of the packing is limited by the strength of the packing material and the support basket used to contain the packing. The overall packing dimensions (outer radius and axial length) are constrained by mechanical considerations such as bearing loads and vibration moments. The concepta used to design conventionalpacked towers can be modified for the design of centrifugal vapor-liquid contactors. For a conventionalpacked tower, the diameter of the tower and the depth of packing are the two variables to be determined in design. Similarly, for the centrifugal vapopliquid contactor, the cross-sectional area at the inner radius and the packing outer radius are needed for the design of a unit. An additional complication arises in the design of the centrifugal vapor-liquid contactor in that the cross-sectional area at the inner radius is a function of both the radius and axial length. The results in an iterative design process to obtain the optimum solution. The cross-sectional area required at the inner radius is a function of the desired hydraulic capacity. Munjal(1986) and Fowler and Kahn (1987) have presented data indicating that the Sherwood flooding correlation may be used to determine the cross-sectional area at the inner radius. The data presented, however, are very limited, and more tests were desirable to further validate the application of the Sherwood flooding correlation. The equation needed to calculate the outer radius of a packing torus is derived using the transfer unit concept. The final form of this equation for the centrifugal va-
In
lKLa
[ [ x1x2
Y2/H - Y 2 / H ](1 - 1/S) + (1/S) (1- l/S)
1
r
a h 2- r12) (1)
This equation is similar to that used for the design of conventional packed towers except that use of cylindrical coordinates for the centrifugal vapor-liquid contactor results in the square of the radius. Since the right-hand side of the equation is in units of area, the first term on the left-hand side (Q/ZKLU)is defined as the ATU; the remainder of the left-hand side of this equation is the number of transfer units. The number of transfer units is independent of the coordinate system and is the same for both the conventional packed tower and the centrifugal vapor-liquid contactor. In order to use eq 1,it is necessary to know the values of the Henry's law constant and the mass-transfer coefficient or ATU. Experimentally determined Henry's law constants for many environmentally harmful VOCs are available in the literature (Leighton and Calo, 1981; Gosaett, 1987; Ashworth et al., 1988). Experimental values of mass-transfer coefficients for the centrifugal vaporliquid contactor were almost nonexistent prior to this work. Even the small quantity of available data was difficult to interpret due to incompleteness or scatter. Two empirical correlations that might be used to estimate the mass-transfer coefficient have been proposed in the literature. Both of these empirical correlations are based on the penetration theory. The first correlation is that given by Tung and Mah (1985):
k L d / D = 0.96(a,/~,)'/~Sc'/~Re~/~Gr~/~ (2) and the second one is that used by Vivian et al. (1965):
k L d / D = 0.023S~'/~Gr~,~~(dL/p~)~/~ X (1- 1.02 e~p(-0.15(dL/p~)~.~)) (3) Equation 2 requires an estimate of effective area for mass transfer. Tung and Mah assume that this is equal to the wetted area given by Onda et al. (1968): a,/a,
= 1 - e x p ( - 1 . 4 5 ( ~ , / a ) ~ ~ ~ ~ ~ e ~ ~(4) 'Fr~~~~W
The accuracy of eqs 2 and 3 for the design of a centrifugal vapor-liquid contactor is questionable because the data needed to establish their validity were previously lacking. 3. Experimental Section 3.1. Equipment. The centrifugal vapor-liquid contactor employed in this study was a part of a larger system (Figure 2) that was used to demonstrate innovative air stripping techniques and materials in concert with emissions control technologies. The project was sponsored by the Air Force Engineering and Services Center (AFESC) at Tyndall Air Force Base, Florida, and the tests were conducted at Eglin Air Force Base in Florida. The groundwater at Eglin Air Force Base is contaminated with JP-4 jet fuel. Two types of packings were evaluated in the centrifugal vapodiquid contactor. The fiit packing (Sumitomo) was a metal sponge-like material made of 85% nickel and 15% chromium. This material had a specific surface area of 2500 m2/m3and a void fraction of 0.95. The second type of packing was made from wire gauze and had a specific surface area of 2067 m2/m3and a voidage of 0.934.
576 Ind. Eng. Chem. Res., Vol. 31, No. 2, 1992 CLEAN , \ , A I R
-. BACUED BED STRIPPER
ACTIVATED CUMW AD8WmER
CATALVTIC INCINERATOR
CLEAN WATER TO S U f A C E T O RECHAROE TO ETC
Figure 2. Groundwater decontamination by air stripping.
Since it is extremely difficult to withdraw liquid samples from the inside of the packing, three different rotors utilizing the Sumitomo packing were initially used in an effort to estimate the end effects in the mass-transfer tests. Each rotor had an inner radius and axial length of 12.7 cm (5 in.). The outer radii of the three rotors, excluding the support plate, were 22.9 cm (9 in.), 30.5 cm (12 in.), and 38.1 cm (15 in.) giving packing depths of 10.2 cm (4 in.), 17.8 cm (7 in.), and 25.4 cm (10 in.), respectively. The wire gauge packing was only used in the rotor with outer radius of 22.9 cm (9 in.). A liquid sample tube was installed inside the housing -0.64 cm (0.25 in.) away from the outer radius of the packing torus. The radial location of the sample tube was varied to match the outer radius of each rotor. The purpose of this type of sampling system was to minimize the variability between the end effects that may be caused by rotors with different outer radii. 3.2. Procedure. A hydraulic run was started by setting the rotor speed at 1000 rpm and then establishing the liquid and gas flows at the desired values. After 3 min, both the pressure drop across the packing and power consumption were measured. The rotor speed was then decreased by 100 rpm and the two dependent variables were remeasured. This procedure was repeated until either the desired air flow rate could not be maintained because of high pressure drop or the inside eye of the rotor filled with water. In order to reduce the number of runs required to characterize the mass-transfer performance, a central composite design was chosen for the experiments (Anderson and McLean, 1974). The center point in this design was the first experiment performed and was repeated after every two runs. This was done to achieve the necessary degrees of freedom required for error estimate and to detect any changes that may be occurring in the packing. A mass-transfer run was begun by setting the desired rotor speed, liquid flow, and gas flow. The exit air stream from the stripper was then continuously monitored using a total hydrocarbon analyzer to ensure attainment of a steady state prior to taking samples. When the total hydrocarbon analyzer reading did not change for 30 min, liquid samples were collected into prelabeled 40-mL glass bottles that contained 0.5 mL of 50% sodium hydroxide (all sample taps were left running continuously at a rate of -250 mL/min in order to collect representative samples). In collecting the samples, the bottles were allowed to fill until overflowing and then sealed with a cap equipped with a Teflon septum. The bottles were then checked for absence of air bubbles, shaken for 30 s, and placed in a refrigerator until analyzed. The sample bottles
,
15
I
SUMITOMO PACKING G = 2.83 x 103 umin (la, sdm) G = 8.50 x 1 6 Umin (300d m )
!
-
% 10 8
0
200
400
600
800
loo0
1200
ROTOR SPEED (rpm)
Figure 3. Pressure drop with both liquid and gas phases flowing (liquid flow = 0.63 L/s; 76.2-cm-diameter rotor).
were used once and then discarded. The liquid samples were analyzed using the Environmental Protection Agency Method 602 for purgeable aromatics. All the liquid samples were analyzed as basic solutions because during initial checkout, it was observed that the glass sample tubes on the automatic laboratory sampler adsorbed the VOCs, leading to large errors for the low concentration samples. Consequently, sodium hydroxide, which tends to preferentially adsorb on the adsorption sites of the glass tubes, was used to make the samples basic. 4. Results and Discussion 4.1. Hydraulic Performance. A typical family of pressure-drop curves with both the liquid and gas phases flowing is shown in Figure 3. The pressure drop initially decreases with decreased rotational speed at constant fluid flow rates. After some critical rotor speed is reached, the pressure drop begins to increase very rapidly. This demarcation is very sharp, with changes in rotation speeds of
