In our previous report10) we made it clear that water vapor contained in air greatly influenced the wet-bulb temperatures of organic solvents. Based on the idea that one might obtain higher rate of vaporization by using water-wet air rather than dry air, our Studies were coptinued with a view to throwing light on how the vaporization of organic solvents into water-wet air would be affected by water-vapor content thereof.
The present results obtained with toluene, ethyl-alcohol and methyl-alcohol clearly proved that we were right in supposing that water-wet air would bring about higher rate of vaporization than dry air, as shown in Figs. 3 & 4. Here the curves for methyl-alcohol are omitted in order to avoid complication.
I. Effects of Temperature of Air and Its Water-vapor Content upon ht:
(i) In Dry Air:
The surface temperatures of liquids in the vaporization pan were calculated by means of Eq. (5), derived from Eq. (1) and Mizushina & Nakajima's Eq. (4)5), which is said to be applicable over the region where Sc, Pr=0.5-2.5. The results tabulated in Table 1 show that these surface temperatures are 1-2°C lower than the wet bulb temperatures of the same organic solvents, which we measured in our previous experiments10)-Cf. Fig. 5-(A), (B) & (C). Of these two kinds, the lower temperatures seem to be the real surface temperatures, uninfluenced by radiation heat, etc. The values of hs, calculated by using Eq. (6), are plotted in Fig. 6.
(ii) In Water-wet Air:
In Table 3 are shown the surface temperatures (t_a1) calculated by using Eq. (8), where hs is the value obtained from Eq. (6) and where the absorption heat etc. are ignored because they are negligibly small as compared with condensing heat. Here again the surface temperatures (t_a1) are almost 1-3°C lower than the wet bulb temperatures. hr can be obtained from Eq. (9). These data are plotted in Fig. 6 as h_t=h_s+h_r vs. dry bulb temperatures of air. The values of hr are known to be considerably larger than those of hs.
II. Effects of Air Velocity upon ht:
The experimental data are plotted in Fig. 7 and correlated as shown in Figs. 8 & 9, by means of dimensional group in Eqs. (17) & (18); b2 and n appearing in these are shown in Table 4, too. ε in the equations is a factor to show the effect of water-vapor content of air upon the vaporization of organic solvents and is equal to 1 when the experiment is conducted in dry air. The powers of Reynolds number in these equations agree with the results obtained by other workers1, 2, 6, 8). The power of Prandtl number is what Mizushina and Nakajima5) proposed in their report. The valueS of n. which are the power of ε, seem to agree with values represented by the slopes of linear lines in the diagrams obtained formerly by us10). These lines show the relation between the wet-bulb depression of organic liquids (t-t_a1) and the wet-bulb depression of water (t-t_w1)-Cf. Fig. 5 (A), (B) & (C).

抄録全体を表示

It was recently found by Carner4) (with nitrobenzene-water system) and Terjesen6-8) (withCCl4-water system) that an addition of a very small amount of surface active agents decreased the extraction of acetic acid from droplets of organic solvents.
The authors carried out the experiment with 8 organic solvents to determine what surface active agents worked as poisonous substances for them. The results revealed that the range of poisonous substances for these organic solvents covered not only the surface active agents as shown by Carner and Terjesen but also other high molecular substances (cf. Tables 1 and 2). Moreover these poisonus substances were found to possess elective affinity, which was particularly obvious in polyethylenglycol (PEG mol wt.≅4000). Generally speaking, benzene, toluene, hexane, heptane and cyclohexane were easily contaminated with the poisonous substances, while butylacetate, isopropylether and methyl-isobutylketone were relatively not so.
Photographs were taken by schlieren method, with the apparatus shown in Fig. 2, to make clear the mechanism of solute transfer. Fig. 7-I-a, II-a, III-a and IV-a show acetic acid transfer to pure water and Fig. 7-I-c, II-b, III-b and IV-b, to additives-containing water, In the latter group, but not in the former one, vortex ring motions were clearly observed behind the drops. Generally such a vortex ring was observed when the rigid sphere moved in a fluid at Re>25, so it was naturally supposed that the contaminated droplets were of stagnant sphere. Values of C2/C1 observed of the most contaminated drops of several solvents were compared with those calculated of rigid sphere by means of Ep. (3) as shown in Fig. 8 and Table 3. They showed good agreement in butylic acid (where m is small) transfer as well as acetic acid transfer, revealing that k_c is very large in comparison with kd in the period of free rising.
It was also found that the mechanism of transfer was effected even by the most purified solute, as well Fig. 7-V is a schlieren photograph of heat transfer from the pure benzene drop (none solute) to pure water, which is different from solute transfer shown in Fig. 7-I-a. Consequently we may safely conclude that the mass transfer from droplets in liquid-1iquid system where solute is contained is different, in mechanism, from the heat transfer from droplets in liquid-liquid system where no solute is contained, and that the method of Colburn and Welsh, in which, in order to obtain the data on individual coefficients, two pure liquids of limited solubility are contacted in the abscnce of a third solute, cannot be applied to the study of mass transfer from droplets.
It is proposed that, in practice, when the contamination is unavoidable, PEG aqueous solution (0.2 -1%) should be added to water phase for the purpose of increasing the extraction rate, whichever side of the phases might be contaminated with poisonous substances.

抄録全体を表示

Studies were made on the distribution of flow in packed beds with the aid of a diffusion-controlled electrode reaction, that is oxidation of ferrocyanide ion in a potassium nitrate solution on a platinum anode.11)
The experimental apparatus employed was as shown in Fig. 1. The test column was packed with the glass beads of the same size to a height of about 160 mm over the screen which was placed a little above the cathode of platinum plate. The average diameters of the packings used were of 4 sizes: 0.945, 3.17, 5.38 and 7.05mm, each for one experiment, and the inside diameters of the columns were of 2 sizes: 26 and 50.5mm, respectively. An electrode assembly used for the anode consisted of five rings of 0.1mm-diameter platinum wire. The radii of these rings were determined by the equation, 13)
r/R=[(2n-1)/10]^1/2, n=1, 2, 3, 4 and 5
(1)
Cemented on a large-mesh Saran screen, these five platinum rings were held concentrically. The leading wire was of PVC-coated 0.2mm platinum wire.
The electrode assembly was inserted into the bed and the measurements of limiting current, 3), 11) were made at various liquid flow rates. The mass-transfer coefficients were calculated by the equation:
k_F=i/FAC
(2)
The flow-distribution was measured at the positions shown in Fig. 2 and some of the experimental results in Fig. 3 and Table 1, are expressed in Sherwood number. A survey of these results show that the distribution of flow is mostly independent of the total flow rate. The distribution was smooth and the reppoducibility good when the sizes of packings used were 0.945 or 3.17mm, but not so when they were larger.
These experimental results were correlated in the form (Fig. 4)
Sh/(Sc)^0.3=kRew^m
(3)
Therefore, by means of the relation:
the values of(Rew)/(Rew)av at different radii could be calculated. The calculated results are shown in Table 2 and Fig. 5. It may be concluded that the distribution of flow in packed beds is practically uniform across the diameter of the pipe.

抄録全体を表示