Vapor Liquid Equilibria Ethylene Oxide Acetaldehyde and Ethylene Oxide Water Systems PDF
Vapor Liquid Equilibria Ethylene Oxide Acetaldehyde and Ethylene Oxide Water Systems PDF
Vapor Liquid Equilibria Ethylene Oxide Acetaldehyde and Ethylene Oxide Water Systems PDF
T h e vapor-liquid equilibria for the systems ethylene Acetaldehyde was fractionated on the same column in order
oxide-acetaldehyde and ethylene oxide water were deter- to separate it from polymers and acetic acid. The distillate was
either used immediately after fractionation or was stored a t
mined at atmospheric pressure and also at higher pres- -10" C. in a closed bottle. Vapor pressure determinations in-
sures for some compositions. The vapor pressures of dicated the high purity of the samples.
ethylene oxide and acetaldehyde were determined be-
tween 0" and 35" C. The system ethylene oxide-water is VAPOR P R E S S U R E DETERMILTATIONS
discussed in terms of its activity coefficients.
If the vapor-liquid equilibria are to be expressed in terms of
activity coefficients, the vapor pressures of ethylene oxide and
PURIPICATION O F COMPONENTS
TABLE OF ETHYLEYE
11. VAPORPRESSURE OXIDE
The ethylene oxide used contained small amounts of acetal- (-273.2' C. 0' K.)
dehyde and water. It was purified by mixing with 5% tri- Temp., C. Mm. H g Temp., O C. Mm. Hg
ethanolamine and fractionating on a column, 2.5 cm. in diameter
and 60 cm. high, packed with glass Fenske helices. The column 0.3 506 16.05 939
0.4 508 16.6 958
was operated a t a 10 t o 1 reflux. The first 5% v a s rejected and 1.1 522 16.75 964
a residue corresponding to 20% of the charge was left in the re- 1.2 526 21.6 1156
boiler. The remaining distillate vias used for the experimental 10.5 700 24.9 1294
10.7 766 29.0 1518
work. 11.1 775 30.35 1574
12.55 823 31.8 1654
1 Present address, Coal Tar Research Association, Oxford Road, Gomer-
15.7 928
sal,Leeds, England.
1434
July 1950 INDUSTRIAL AND ENGINEERING CHEMISTRY 1435
Giauque and Gordon ( 4 ) have also reported data of ethylene The Othmer still is normally operated with liquids boiling
oxide vapor pressures. Over a range from -50' to +12' C. the above the ambient temperature. This results in a slight reflux
equation taking place initially in the outer jacketing space. Under these
log p = - (2045.70/T) - 0.021507 T +
2.3328 X 10-6 T a = 14.31363
conditions, and with liquids whose boiling points are not far
apart, the temperature recorded in the vapor stream will ap-
proach closely that recorded in the boiling liquid. With the
is in excellent agreement with their experimental results. liquids under consideration here the two observed temperatures,
Using these equations to find the boiling point at atmospheric in the vapor and liquid, respectively, differed appreciably. When
pressure (760 mm.) the following result: ethylene oxide-acetaldehyde mixtures were being distilled at
Moor et al. ( I S ) : 11.2' C. atmospheric pressure the vapor stream was always slightly su-
Maass and Boomer (11): 10.7 * 0.1' C. perheated unless the top of the still was cooled below the dew
Giauque et al. ( 4 ) : 10.50 * 0.05' C. point of the vapor. When this was done the same temperature
Present work: 10.4 * 0.2" C. was observed in the liquid and in the vapor. During the experi-
The boiling point of ethylene oxide has also been reported (19) ments with ethylene oxide-acetaldehyde mixtures some observa-
as 10.7"C. by Timmermans and Hennaut-Roland, but no further tions were made with and without cooling the top of the still.
vapor pressure data are recorded in that paper. No significant difference in the vapor and liquid compositions
The freezing point of the ethylene oxide used for these experi- was found for this change in conditions.
ments was -113 * 0.5' C. This may be compared with melt- With ethylene oxide-water mixtures of high water content in
ing pointa in the literature of -111.4' C. (II), -111.7' C. (19), the glass equilibrium still the observed temperature in the vapor
and -112.51 * 0.05' C. (a). was found t o be lower than that in the boiling liquid, as a result
of the very large difference between the boiling point of the liquid
and the dew point of the vapor.
111. VAPORPRESSURE
TABLE OF ACETALDEHYDE
(-273.2'' C. 0" K.)
Temp., C. Mm. Hg Temp., .a C. Mm. H g
-0.2 332 13.3 577
2.7 375 17.6 682
6.7 443 20.7 766
9.3 494 30.8 1120
11.6 531 34.4 1259
From these equations the following boiling points a t 760 mm. can
be calculated:
Gilmour (6): 20.2' C.
Present work: 20.4 * 0.2"C.
The agreement in the case of acetaldehyde is excellent. Agree-
ment on ethylene oxide is not so general. It should, however, be
noted that the present results are in good agreement with the re-
sults of Giauque and Gordon ( 4 ) of recent date. In any case, it
is considered that errors which might arise from the presence of
impurities of the order indicated by the boiling point differences
are small in relation t o those arising from the inaccuracy of the
analytical methods used in the vapor-liquid equilibrium deter-
minations,
caution, each day's work was started with freshly prepared liquid Above 760 11111. Pressure
mixtures. 34 35.4 1.41 0.27 96.56 99.34 0.97 8.4
34 35,4 1.36 0.30 9 6 . 6 8 99.27 0.97 9.3
34 ... 2.20 0.45 9 4 . 8 5 98.90 0.96 9 3
ANALYSES 34 37.5 4.70 0.67 8 9 . 3 98.36 0.99 5.6
65 56.3 7.9 0.99 8 2 . 7 97.58 1.13 a.7
Three methods of analysis were used in this work. For the
ethylene oxide-acetaldehyde system, amounts up to 10% acet-
aldehyde were determined by t'hr silver oxide met,hod ( 1 7 ) . For
the remainder of the range, ethylene oxide was determined by a Figure 1 shows the plot of activity coefficient on a logarithmic
modified Lubatti method ( 1 0 ) in which the ethylene oxide reacts scale against' the liquid mole composition of each component.
with hydrochloric acid to give ethylene chlorohydrin. This The activity coefficients for the ethylene oxide-acetaldehyde
method is subject to certain errors which tend to give low results. can be seen from Table IV to be, within the limits of experimental
The reasons for this have been discussed hy Lichtenstein and error, unity over the whole range except for a slight decrease in the
Twigg (9) and a more detailed study has been made in t,his labora- region of below 1% acetaldehyde. These data, therefore, have
tory and is being prepared for publication. In the ethylene not been represented graphically, and it is concluded that' tjhe
oxide-water system, the mixtures cont,aining less than 5% water system obeys Raoult's lam with close approximation. This is not
were analyzed by the Karl Fischer reagent. surprising in view of the many similarit,ies to which reference has
The accuracy of the analyses is estimated to be 170for ethylcne already been made by Maass and Boomer ( 1 1 ) .
oxide, 2y0for acetaldehyde, and 2% for water, the percentage in The system ethylene oxide-water is shown graphically iii
each case being on the result. The significance of the accuracy Figure 1. The equilibrium curve is shown only for the rango
of the analyses is discussed below. 95 to 100 mole % ethylene oxide in the vapor phase in order to
show this section clearly.
RESULTS The Gibbs-Duhem equation may be expressed as
While the curves relating liquid to vapor composition and to
boiling point are sufficient for distillation calculations, the ex-
pression of the results in terms of activity coefficients is much to
be preferred, as it admits an immediate appreciation of the de- Assuming the vapors to behave as ideal gases this may bc rc-
viation from the behavior of an ideal mixture. It also enables written for a binary mixture as
some evaluation of the consistency of the data and, if necessary,
extrapolation from a few experimental results over the a-hole
range of compositions. The significance of the activity coefficient
and the application of vapor-liquid equilibrium data is discussed
by Carlson and Colburn ( 2 ) . The activity coefficients have been This equation has been integrated in several ways, notably by
calculated from van Laar (8),Margules (IS),and Scatchard ( I t ? ) , and it is of in-
July 1950 INDUSTRIAL AND ENGINEERING CHEMISTRY 1437