Document ev1QBJRNpZO807aKRRJxXQrrM

A SIMPLE METHOD OF VAPOR-PRESSURE MEASUREMENT FOR HIGH BOILING LIQUID SUBSTANCES Hiroshi ISHII and Katsuya NAKAYAMA [Electric Laboratories, Government Research Institute, , Tokyo, Japan] Although there are many studies on the vapor pressure measurement of high-boiling liquids, rela tively little work has been done on a simple way of measuring the vapor pressure at relatively high tem perature . We have used a U-shaped barometer to develop a new, simple vapor-pressure measuring appa ratus. This method is simple and it was used to measure the vapor pressure of esters, petroleum sys tems, silicone oil systems, diphenyl chloride system, alkylnaphthalene systems and polymonochlorotrifluoroethylene oil systems over a pressure range of from 0.5 to 20 mm Hg and at a temperature range of from 100C to 300C, with satisfactory result. INTRODUCTION case. In fact, changes in relation to a change in T. Therefore, it is not valid to make an extrapolation over a Studies on the vapor pressure mea wide range. In order to obtain an accu surement for high-boiling liquids have rate data for vapor pressure and heat of been made for a long time. In a lower evaporation at a certain temperature, one temperature region, the vapor pressure is has to determine such parameters at that low. In this range, dew point method1, particular temperature. spring balance method2 and others3 have The principle of the device constructed been used. In spite of its industrial by us is based on a direct measurement of importance, there is virtually no simple vapor pressure by an U-shaped barometer. method available for measurement of vapor One of its characteristics is to use the pressure in a range of from 0.5 to 20 mm Hg ' liquid sample itself as the column fluid of the barometer. If, under equilibrium condition, a vapor phase can be handled as an ideal gas, there is following relationship, according to the law of thermodynamics, between the pressure P and the temperature T. Traditionally, barometer(or manometer) of this type which employs a liquid column has the following advantages : (i) The construction is simple and therefore, method d In P _ A H dT "RT2- /\ '' of use and construction can be done easily. Moreover, since there is no movable part, the apparatus is relatively free of troubles, Now, if is independent of tempera (ii) Pressure is measured directly from ture, then the equation will become a first order function of log P and T-1. This, however, is not true in an actual K0GY0 KAGAKU ZASSHI, Volume 62, No. 3, pp385 - 388 (1959). i ; -,v.| DSW 297128 STLCOPCB4067041 * .mitt.. a difference of the heights of the heads of the liquid columns. Therefore, there is no need to make a sensitivity calibra tion. Measurement of an absolute value can be made, independently of molecular weight, as long as the specific gravity is known. On the other hand, they hav the following disadvantages: (i) the fluid itself may react with the sample vapor, U-Shaped barometer (int. diameter 4 mm }) (ii) the fluid may dissolve the vapor, or (iii) error may originate from the vapor pressure of the fluid itself,.or (iv) temperature correction for the fluid column may be needed. By using the same substance for the sample and the column fluid, shortcomings (i),(ii), - Chimney (30 mm ^) Insulating material ^Temperature probe (int, diameter 2.5 mm p) Viewing hole > Cooling water and (iii) were corrected. Boiler Asbestos The second characteristic is that a portion of the vapor is constantly released into vacuum which is usually higher than 10-3 mm Eg. By such mech anism, lower molecular weight impurities Diatomaceous earth brick FIG. 1 Apparatus V formed from thermal decomposition, even in extremely small quantities, are Through the cock, the apparatus is constantly removed, thus preventing them from being accumulated in the system. connected to a vacuum pump to maintain a pressure of 10"3 mm Hg or lower. Inert Since there is a hole, the initial gas is placed in the exhaust system, at "de-gassing" procedure can be carried a constant rate of flow, and its pressure out thoroughly. Even when the sample ' is read by another mannometer, and adjusted is made of a mixture, fractional disti accordingly. Thus, vapor pressure can lling effect can be avoided because be measured for a high vapor pressure range, almost no condensation process takes without extending the fluid column portion. place in the vapor chamber, and the rel eased vapors are totally condensed and Boiler and chimney portions are heated refluxed by the condensor. Thus, the by a suitable means to maintain an uniform composition of the sample within the boiler will remain unchanged. This temperature throughout. A glass tube (internal diameter 2.5 mm ^) is attached latter aspect eliminate the biggest to the middle and upper portion of the disadvantage of the distillation method. chimney, through which a thermocouple is Thus, the present method can be most inserted to measure the temperature. Tem advantageously used on industrial products perature at these two areas has to be the for which the measurement of the vapor same. The thermocouple and the tube pressure of a mixture is frequently imp attached to the boiler portion is used for ortant. 1 measuring the temperature of the liquid phase. The thermocouple and the tube attached at the middle portion of the APPARATUS chimney extends near the opening of the U-shaped barometer within the chimney, The vapor pressure-measuring apparatus and in this area the vapor pressure and constructed by us was made totally of the temperature are measured. hard glass(Toshiba Telex). Outline of the apparatus is illustrated in Fig. 1. There is a hole(diameter 1.2 mm ^) on the top of the chimney, through which the 2 0S\N 297129 STLCOPCB4067042 ( V. tM vapor is released Into vacuum. The condensed liquid passes through the side tube and returns to the boiler. tion of data. Therefore, in our opinion, such a beak-like device should be attached. The glass tube located at the center of the chimney has pointed tip, and its function is to supply liquid to the U-shaped barometer. With the^ rising temperature, condensed sample fluid will accumulate within the barometer. If one wants to resupply, one has only to blow an air from external sources into this center tube to forcibly cool down the system. It is, however, not adv isable to force the system to cool too rapidly, because by such doing the temperature of the entire vapor phase. tends to decline. mm FIG. 2 Opening on the U-tube within the chimney With this procedure, the temperature gradually elevates. After the height of the head of the liquid column protr uding out of the U-shaped barometer has reached a sufficiently high level, the temperature is lowered slowly and the temperature and the pressure are meas ured . In such case, the shorter side of the U-tube being inserted into the chimney has always its fluid in the overflow state. In other words, one of the liquid column of the U-tube is fixed. The internal diameter of the U-shaped barometer is 4 mm and its outside surface can be cooled with water. Therefore, the temperature can always be maintained constant. The end of theU-tUbe located inside is open, and its meniscus may be hori zontal, concave or convex. Accordingly, the reading of the head of the liquid column may change. In order to make the meniscus of the inner arm of U-tube to have the same shape of meniscus as that of the outer arm of the U-tube, a beak-shaped glass piece was attached. When the liquid is in overflow state, as illustrated in Fig. 2, one can not obtain a normal meniscus. Due to vari ation of the shape of the menisci of two aims of the U-tube, systematic errors always exist. In view of the magnitude of the data, however, this does not cause significant problem. But, rather, identical variations always reproduced, which reduces the distribir*- MEASUREMENTS AND RESULTS An oil pump and a three-stage glass oil diffusion pump were used as the vacuum pump to obtain a vacuum of 10-5 mm Hg. Boiler and chimney portions were heated separately, each using 0.5 mm $ nichrome wire and heated electrically. Heating is controlled by the use of sliducks which work independently. The heat lost by the vapor which is released from the small hole on the upper part of the chimney is only about 4 W, even when the vapor pressure within the chimney is as high as 20 mm Hg. Therefore, there will be no problem. The entire area may be enclosed in a copper block and heated externally, but in an actual practice, we have used the following mechanism. In the boiler portion, nichrome wire wound around as a coil is embedded in a trench which has been formed spirally around a piece of a brick of diatomaceous earth. The entire body is wrapped with asbesto sheets to effect uniform heating and to maintain insulation. In chimney portion, it was wrapped with one layer of glass fiber tape. Nichrome wire is then wound over the glass fiber at a pich of about 1 cm. The wire layer is then covered and wrapped with asbesto layer and glass fiber tape to effect its insulation. _ 3 DSW 297130 STLCOPCB4067043 ' 4' ' With this kind of insulation, it required about 100 W for the chimney portion and about 50 W for the boiler portion(room temperature 28 `te), in order to heat them to 200 C. ' To measure the temperature, a Cu-constantan thermocouple that has been temperature-corrected(measuring the melting point of lead and tin and boiling point of water) is inserted deeply into the side arm tube, and the voltage is read with a mV meter directly. The temperature of the center portion of the chimney portion was about 1C higher than that of the upper portion of the chimney portion. Also, the liquid phase showed higher temp erature (3 C) than the center portion of the chimney portion. Temperature of the vapor phase is the reading taken at the center portion when the temperature diff erence between the upper and center portions of the chimney portion is less than 1C. About 20 cc of the sample was charged in, and this amount was sufficient. The specific gravity of the sample was measured with a float. A care was taken to keep the temperature of the water for cooling near 25C, at which specific gra vity measurement was taken. Since the amount of liquid evaporated was extremely small, surface evaporation alone was sufficient. Even when the vapor pressure was high, there needs no concern for boiling. Uhless the chimney portion is well insulated and the temperature is kept well, evaporation will be quite abrupt and boiling might take place. It is timewise advantageous and also reliable to take the measurements, as the temperature is lowered. Although we did not see any delay in pressure reading due to reading of temperature, but we still paid attention in such a manner that the vapor pressure readings were taken after the temperature was lowered cautiously by 5C. Ordinarily, it took about 60 minutes to take ten measurements at an interval of 5C over a temperature range of 50C. Actual operation of the procedure is illustrated in Fig. 3. Results are illu strated in Fig. 4 and also in Table 1. DISCUSSION _ Possible causes of errors are indicated below. (1) Problem of overheating and problem of non-uniform temperature distribu tion within the gas phase. (2) Pressure gradient caused by the vapor flow which accompanies the release of the vapor from the hole. (3) Error due to temperature difference at both arms of the U-tube. As far as (l) is concerned, since the vapor flows at a certain proper velocity through a large chimney and therefore it is little affected by the wall of the vessel, one may expect the temperature difference at various locations within the gas phase to be extremely small. In fact, in the chimney portion, the temperature difference between the upper and the middle portions was only about 1C. And, in such case, the difference is not significant. There fore, this will cause no significant problem. With regard to (2), the amount of vapor released can be calculated from the equation which was originally derived in the fluid mechanics. And, the pressure gradient can be estimated from the resistance against ,,S1N 297131 STLCOPCB4067044 X \J *1 2auo O <H DSW 297132 STLCOPCB4067045 if ;\ - A Jk. the pressure gradient can be written as following. AP = 128??. L TE V D4 (3) In this equation, 7t is the vis cosity coefficient of the gas, but its exact value is not known. Therefore, by assuming the mole cular diameter as about 8 x 10"8 cm to calculate \ as = 130 x 10" poise, and substi tuting this value, along with V = 88 cc/sec and D = 3 cm, into the above equation, the pressure gradient will be 4 x 10" mm Hg/cm. This value is too small to be of any significance. However, since the volume the vapor flow within the chimney. of the chimney is about 200 cc, it may be estimated that the Let us assume a tank that is filled with a gas having a molecular weight M, having a pressure P0(dyne/cm2) at a vapor within the chimney is replenished anew every 3 seconds. There needs a particular caution for the temperature T0(K). Let us also assume term (3). Since the volume expansion that the gas is constantly spouting out from a hole(cross sectional area S(cm2)) coefficient of ordinary organic liquid is in the order of magnitude of 1 x 10-3 or into the vacuum. The amount of gas smaller, there will be 10 $ error if escaping into the vacuum during unit time the temperature difference between two (G (g/sec) can be expressed as follows2..... arms of the U-tube is 100C. Further, 2K *+l \ic+l/ (2) since the height of the liquid column in side the chimney portion is about 8 mm, the error will amount to about 0.8 mm. Where, R is gas constant, ft is the ratio But, the actual problem occurs when (Cp/Cv) of the specific heat at constant the difference of the liquid column heads pressure(Cp) to that at constant volume is extremely small. Fortunately, in such (Cv). In case of diethyl phthalate case, the temperature is also low, and and at the highest chimney pressure (P0 = 20 mm Bg), the pressure will corr espond to an energy of 2.7 x 104 dyne/cm2. therefore the problem will not be serious. When two kinds of petroleum series oils were used to measure the pressure at 192C If the temperature in this case is assumed and at 224C (difference of liquid heads= roughly as 443K and the specific heat 8 mm)' deviation from the vapor pressure ratio K is assumed as 1.02, M will be curve was only less than 0.5 mm, in terms 222 and the cross-sectional area of the of liquid column. This means that the spout will be 1.13 x 10"2 cm2. Therefore, reading error is very small. G will be 1.4 x 10"2 g/sec. This will correspond to a vapor flow of about Usually a cooling water is circulated 88 cc per second. around the portion of the barometer that protrude outside to maintain a constant In a cylinder having a length of L(cm) temperature. Even when the temperature and a diameter of D(cm), if the gas is differs by 5C, the change of specific flowing at a velocity of V(cc/sec) and gravity is only about 0.5 $. Therefore, the pressure difference between two ends the variation is not significant. of the cylinder is Ap(dyne/cm2), then dsw 297133 STLCOPCB4067046 ;5 --ntHiT'lifi'll rt iWhma Even for the samples that tend to solidify or increase its viscosity at normal temperature, measurements can be taken. In such case,. the U-tube portion of the apparatus has to be suitably heated. Measurements were taken for dibutyl phthalate, di-2-ethylhexyl phthalate, and di-2-ethylhexyl sebacate and the results are compared with the data reported in literature.7 The data seem to fall on a single curve, over a wide temperature range, as illustrated in Fig. 5. Based on these results, we conclude that this simple vapor pressure-measuring appaeatus has broad applicability. Acknowledgement The authors are grateful to Mr. J. Kojima, Nippon Glass Ryoki, K.K. for constructing the glass portion of the apparatus. * cp) in literature LITERATURE CITED1 2 3 4 1) S.F. Kapff, R.B. Jacobs : Rev. Sci. . Instr. 18, 581(1947); A.C. Werner : Ind. Eng. Chem., 44, 2736(1952). 2) for Example, F.M. Erusberger, H.W. Pitman : Rev. Sci. Instr., 26, 584 (1955). 3) G. Milazzo : Chem. Ing. Tech., 28, 646 (1956). 4) K.C.D. Hickman : J. Franklin Inst., 221, 383(1936); H.S. Myers, Fenske : Ind. Eng. Chem., 47, 1652(1955). 5) For example, I. Tani : Science of Flow, Iwanami Zensho, 1955, ppl89. 6) For example : S. Dushman : Scientific Foundation of Vacuum Technique, p.84 (1949); A. Guthrie, R.A. Wakerling: Vacuum Equipment and Techniques, p.27(1949). 7) Also, Gardner, Brewer : Ind. Eng. Chem. 29, 179(1937); E.S. Perry, W.H. Weber : J. Am. Chem. Soc.,^71, 3726(1949). 7 DSW 297134 STLCOPCB4067047