Вы находитесь на странице: 1из 2

Bachelor Thesis, Department of International Development Engineering, Tokyo Institute of Technology, 2007 Extended Abstract

Mass Transfer Rate in Liquid-liquid Extraction of


Coal Tar Absorption Oil Using a Spray Column
MASUDA TOSHIYA
Student ID : 03-2252-5
Supervisor: EGASHIRA RYUICHI
Department of International Development Engineering, School of Engineering,
Tokyo Institute of Technology, Tokyo 152-8552, Japan

Introduction Table 1: Experimental conditions


No. feed: absorption oil solvent: aqueous methanol
-1 -2 -1 -2
Coal tar, one of the byproducts from coal carbonization, Rt [kg・h ・m ] yw,b [-] Eb [kg・h ・m ]
contains many useful compounds to chemical industry. 1 8×102∼3×103 0.5±0.001 2×103±500
Absorption oil (AO) which is one of the fractions 2 9×102∼4×103 0.7±0.004 2×103±400
(b.p.=470~550 K) contains nitrogen heterocyclic compounds 3 1×103∼5×103 0.5±0.005 3×103±600
(nitrogen compounds) such as quinoline (Q), isoquinoline 4 2
4×10 ∼5×10
3
0.7±0.007 3
3×10 ±200
(IQ), indole (I), etc., and the other compounds such as
naphthalene (N), 1-methylnaphthalene (1MN), 2-
methylnaphthalene (2MN), biphenyl (BP), dibenzofuran 2. Results and Discussion
(DBF) etc. Generally, these compounds in the absorption oil
are separated by reactive extraction with acid and/or base. 2.1 Fundamental relationship
However, this process is relatively costly because of difficulty The yield, Yi, and the separation selectivity relative to 2MN of
of recovering the acid and base and of corrosion of the component i, βi/2MN, were defined as follows:
equipment by the acid and base. For this separation, the Yi=Et·yi,t/(Rt·xi,t) (1)
ordinary solvent extraction with aqueous methanol (MeOH βi/2MN=yi,t/y2MN,t/(xi,b/x2MN,b) (2)
aq.) has been suggested as alternatives to the reactive The overall mass transfer coefficient of the continuous phase,
extraction and the various kinds of efforts are devoted to apply Kc,i was calculated by,
this ordinary technique to the absorption separation [1-9]. d(E·y)/dz=−Kc,i·a·(yi*−yi) (3)
In this study, the relationship of the mass transfer coefficient where yi* was estimated from xi with the equilibrium
in a spray column[7], one of the simplest practical extractors, relationship and a was estimated from dp.
is examined for designing the separation process. 2.2 Yield
The yields calculated by Eq.(1) are plotted versus Rt in Fig.1.
1. Experimental Yi with the solvent of yw,b=0.3 were higher than those with
yw,b=0.5. In this study, Yi of nitrogen compounds were higher
than those of other compounds and these groups of
The equipment used in this work was the same as in previous
compounds could be separated into each other group. The
work [7]. The column was made of Pyrex glass and behavior
maximum yield of nitrogen compounds came up to around
in the column could be observed. The distributor of dispersed
Rt=2×103 kg·h−1·m−2.
phase had 8 needle nozzles. The distributor of continuous
2.3 Separation selectivity
phase had outlets on the side of cylinder to pretend falling
The separation selectivity calculated by Eq.(2) are plotted
drops entering in it. First, AO and MeOH aq. were prepared in
versus Rt in Fig.2. βi/2MN did not change so much as Rt
the tanks. AO, dispersed phase, was fed to the top of the
changed. βi/2MN of nitrogen compounds with the solvent of
column, and MeOH aq., continuous phase, to the bottom and
yw,b=0.5 were higher than those with yw,b=0.3.
these two phases were contacted countercurrently. The flow
2.4 Mass transfer rate
rates were kept constant by checking the flow indicators. The
Fig. 3 gives the effect of the superficial mass flow rate of
level of interface between accumulated dispersed phase and
dispersed raffinate phase, Rt, on the overall volumetric mass
continuous phase at the bottom was maintained by adjusting
transfer coefficient, Kc,i·a, obtained by Eq.(3). Kc,i·a increased
the drain valve of dispersed raffinate phase. The inlet and
with increasing Rt, mainly because the specific interfacial area,
outlet flow rates were determined by weighing the mass
a, increased with Rt. However, at Run No.4 condition, Kc,i·a
collected in about 10 minutes. The holdup of dispersed phase
did not change with increasing Rt. This is because at these
in the column was determined by measuring the rise of the
conditions, the axial dispersion of the raffinate phase was
interface when the feed of AO and drain were stopped at the
significant.
same time. Each phase was analyzed by GC-17AAF Ver.3
In general, the relationship among the overall and local
(Shimadzu Corp.) to determine the composition. The density
height per transfer units can be represented as follows:
of each phase was measured by pycnometers. The principal
E
experimental conditions are summarized in Table 1. The H.T.U.OC,i=H.T.U.c,i+ mi ⋅ b ·H.T.U.d,i (4)
liquid-liquid equilibrium relationships were also measured for Rt
estimating the mass transfer coefficients, beforehand. The intercept and slope in the plot of H.T.U.OC,i against
mi ·Eb/Rt are, thus, H.T.U.c,i and H.T.U.d,i, respectively,
provided these local H.T.U.s are constant independently
Received on February 16, 2007. Correspondence concerning this of mi , Eb, and Rt. This plot with the results in this study is
article should be addressed to T. Masuda (E-mail address: presented in Fig.4. The result showed that in the range of this
tmasuda@ide.titech.ac.jp).

Tokyo Institute of Technology


102
study H .T .U .oc ,i increased as mi ⋅ E b / Rt increased.
However, some data did not show the tendency especially in

βi/2MN [-]
the range of higher mi ⋅ E b / Rt . From this result, in this
101
separation process, the continuous phase was more dominant
than the dispersed phase in the range of mi ⋅ E b / Rt lower
than 2. 100
102 103 104
Rt [kg・h-1・m-2]
Conclusion
Fig.2 change ofβi/2MN against Rt
In this study, the experiment was carried out favorably under
the conditions. The better relationship of the mass transfer (the keys are same as Fig.1)
coefficient was obtained by estimating H .T .U . which is one
of the factors to calculate the height of the columns. It was 106

Kc,i・a [kg・h-1・m-3]
known that the continuous phase was dominant in the
105
separation process and from this study since the dispersed
phase has a little resistance, the dominant of the continuous
phase had been comfirmed. 104
More consideration need to be done since some data were not
in the relationship obtained. 103

Acknowledgement 102
AO was provided by JFE Chemical Corporation. 102 103 104
Rt [kg・h-1・m-2]
Nomenclature Fig.3 change of Kc,i・a against R
a = specific interfacial area [m2·m−3]
dp = drop diameter of the dispersed phase [m2·m−3] (the keys are same as Fig.1)t
E = superficial mass flow rate of the continuous phase
[kg·h−1·m−2]
H = holdup [–]
Kc = overall mass transfer coefficient of the continuous phase 9
i Q IQ I N
[kg·h−1·m−2] 8 Key
R = superficial mass flow rate of the dispersed phase
[kg·h−1·m−2] 7 i 1MN 2MN BP DBF
V = volume in the column [m3] 6 Key
H.T.U.oc,I [m]

x = mass fraction in the dispersed phase [–]


y = mass fraction in the continuous phase [–] 5
Y = yield [–]
Z = effective contact height of column [m] 4
z = distance from top of the column [m] 3
βi/2MN = separation selectivity of component i relative to 2MN [–]
2
1
1 0
Key Run i 0 1 2 3 4 5 6 7 8 9 10
0.9 1 mi・Eb/Rt [-]
0.8 2 Q
3 Fig.4 change of H.T.U.oc,I aginst mi・Eb/Rt
0.7 4 Literature Cited
0.6 Key Run i [1] Ukegawa, K., Matsumura, A., Kodera, Y., Kondo, T., Nakayama,
T., Tanabe, H., Yoshida, S., Mito, Y., Sekiyu Gakkaishi (J. Jpn.
Yi [-]

1
0.5 Petrol. Inst.), 33, (4), 250 (1990)
2 IQ [2] Kodera, Y., Ukegawa, K., Mito, Y., Komoto, M., Ishikawa, E.,
0.4 3 Nagayama, T., Fuel, 70, (6), 765 (1991)
4 [3] Matsumura, A., Kodera, Y., Kondo, T., Ukegawa, K., Tanabe, H.,
0.3 Yoshida, S., Mito, Y., Shigen to Kankyo (Journal. of NIRE), 4,
Key Run i (2), 147 (1995)
0.2 1 [4] Egashira, R., Nagai, M., Sekiyu Gakkaishi (J. Jpn. Petrol. Inst.),
2 43, (5), 339 (2000)
0.1 I [5] Egashira, R., Salim, C., Sekiyu Gakkaishi (J. Jpn. Petrol. Inst.),
3 44, (3),178 (2001)
0 4 [6] Salim, C., Saito, J., Egashira, R., Journal of the Japan Petroleum
102 103 104 Institute, 48, (1), 60 (2005)
Rt [kg・h-1・m-2] [7] Saito, J., Dissertation for a master’s degree of engineering, Tokyo
Fig.1 change of yields against Rt Institute of Technology (2006)
[8] Salim, C., Egashira, R.; Journal of the Japan Petroleum Institute,
49, (6), 326 (2006)
[9] Egashira, R., Watanabe, K; Solvent Extraction Research and
Development, Japan, 12, to be printed (2007)