December.

1933

INDUSTRIAL AND EXGINEERING CHEMISTRY

(14) International Critical Tables, Vol. I11 a n d V, McGraw-Hill,

1927.
(15) Kirkbride a n d hfoCabe, IKD.ESG. CHEM.,23, 625 (1931).
(16) Kraussold, Forsch. Gebiefe I n o e n i e u r w . , d3, 21 (1932).
(17) Landolt-Biirnstein, Phys. Chem. Tabellen, Springer, 1912.
(18) Lawrence and Sherwood, ISD. ESG. CHEM.,23, 301 (1931).
(19) RlcAdams a n d Frost, I h i d . , 14, 1 3 (1922).
(20) hlerkel, "Die Grundlagen der ~ ~ ~ i r m e ~ ~ b e r t r a g l ip.
n g 140,
,"
Theodor Steinkoyff, Dresden and Leipzig, 1927.
(21) h l o n r a d , IXD. ESG. CHEX, 24, 505 (1932).
(22) Monrad and Badger, I b i d . , 22, 1103 (1930).

(23)
(24)
(25)
(26)
(27)
(28)

1331

Morris a n d Whitman, Ibid., 20, 234 (1928).

Kusselt, 2. Ver. deut. Ing., 54, 1154 (1910).
I b i d . , 60, 541 (1916).
Othmer, IKD.ESQ. CHEY.,21, 576 (1929).
Sherwood, Kiley, a n d Mangsen, Ihid., 24, 273 (1932).
Webster, Trans. I n s t . Engr. ShipSuilders, Scot., 57, 58 (1913).

REC&VEDM a y 13,1933. Presented before the Division of Petroleum Chemiatry at the 85th Meeting of the American Chemical Society, Warrhington,
D. C., March 26 t o 31, 1933.

Studies in Distillation
11.

Liquid-Vapor Equilibria in t h e Systems Ethanol-Water, MethanolWat,er, and Acetic Acid-Water'

L. VALL LACE CORNELLAND RALPH.E.

fiIONTONNA,

University of Minnesota, h'hneapolis, Minn.

The nieihod of Rosanoff, Bacon, and White

the i n d i c a t e d level in a suithas been used for the determination, at atmosable l i g h t oil, held a t a temmental study of the plate
perature high enough t o
efficiencies Of a
pheric pressure, of the liquid-vapor equilibrium
p r e v e n t condensation of the
column for different binary mixOf
systems ethanol-water, mefkanol- v a p o r s . The oil b a t h was
tures, it was found that values
considerably Over 100 per writ
water, and acetic acid-water. This method is
heated electrically, the temwere obtained for ethanol-water
shown to be consistent and reliable. The equilibperature being controlled to
rium data obtained by it are
when the most reliable equigraphi- *0.5 " c. b y t h e m e r c u r y
bulb, M, through a relay. The
librium data were used* The
cally with all other data found in the literature.
t e m p e r a t u r e w a s varied b y
method of calculation was that
raising or lowering the contact
proposed hy McCabe and Thiele
(12) and gave the plate requirements for theoretically perfect wire in the capillary tube, C.
The heater, H , in the still, S, was of KO,24 nichrome wire
operation; hence, it was indicated that the equilibrium data
were not accurate. It was decided, therefore, to determine in the work on the alcohols, but for acetic acid-water it was
the eauilibrium curve for ethanol-n-ater by a reLable method, necessary to change to KO.30 platinum wire. This heater
was a small coil wound in the form
and later also those for methanolof a spiral as indicated in Figure 1,
.AY
water and acetic acid-mater.
and the l e n g t h s of wire were as
uT0 AC.
EXPCRIMEKTAL
PROCEDURE
f o l l o w s : KO.24 n i c h r o m e , 36
31L LEV
inches (91.4 cm.); No. 30 platinum,
There are a number of different
_ _ - - - 11-- - -38 inches (96.5 cm.). Contact was
methods for the determination of
made with the mercury in the leadequilibriuni curves of binary liquid
in tubes by means of loops of No.
m i x t u r e s . d careful review of
24 platinum wire sealed through
the literature indicated that the
the glass.
methods of Rosanoff and his coI n the work with the alcohols, the
workers are r e l i a b l e . They are
corks which were in contact with
r e c o m m e n d e d by Y o u n g (23).
hot vapors were covered with lead
Rosanoff, Lamb, and Breithut (19)
foil, but t h i s was r e m o v e d f o r
a n d R o s a n o f f arid Easley (18)
acetic acid-water.
have developed an a c c u r a t e but
The oil bath was heated to a temi n v o l v e d m e t h o d . Rosanoff,
p e r a t u r e 2" to 5" C. above the
Bacon, and White (17)have worked
initial boiling point of the liquid
out a less involved method based
mixture to be tested. The apparaon an entirely different principle.
tus was dried before each run by
Young (22) states that these two
drawing air through it for 10 to 15
methods have been found to give
m i n u t e s . Then the temperature
M
results in good a g r e e m e n t , a n d
inside the still was allowed to rise
for this reason the simpler method
to 4" to 5" C. below that of the oil
of Rosanoff, Bacon, and White was
bath, a n d a b o u t 130 ml. of the
selected for this work.
binary mixture were run into the
A diagram of the apparatus is
ALL DIMENSIONS
inner boiling vessel, S, through A .
shown in Figure 1. The still, of
IN CENTIMETERS
GLASS WORM
The composition of this liquid had
Pyrex g l a s s , w a s i m m e r s e d t o
been determined previously. The
CONDENSER
* The first paper in this s e r i e s was
addition
of the liquid lowered the
p u b l i e h e d by L. H. Shirk and R. E.
FIGURE
1. APPARATUS
FOR DETERMINATION
OF
Montonna, IND. ENQ.CHEM.,19, 907-11
COMPOSITION
OF VAPORS
FROM BOILING
BINARY temperature shown by the still ther(1927).
mometer, and, as soon as this had
SOLUTIONS

S THE course of an experi-

1332

I N D U S T R I A L A N D E I\; G I I\; E E R I N G C H E M I S T R Y

risen to a value 2" to 3" C. below the initial boiling point of

the binary mixture, the electric heater, H , was turned on, the
distillation commencing in about 5 minutes. Enough heat was
supplied to distill over 60 to 70 grams in 30 t o 45 minutes.
During the distillation
the oil b i t h temperature
w a s maintained 2" t o
4" C. above that in the
s t i l l . T h i s difference
was so small that there
could hardly have been
any superheating of the
inner chamber, and yet,
as will be shown later,
there was evidently no
condensation. For
a c e t i c acid-water and
for the runs at t h e
higher concentrations of
the alcohols, four fracFIGURE2. DATA
FROM APPARATUS
OF ROSANOFF,
BACON,AND WHITE t i o n s Of l 3 to l 5 ml*
were collected, but for
ON ETHANOL-WATER
the o t h e r alcohol runs
five were taken. The fractions were collected in small vials
(20 t o 25 ml.) connected to the bottom of the condenser by a
cork with only a small opening to the air. Since the condensed distillate was quite cold (about 15" C.), no other precaution was needed t o prevent loss by evaporation.
At the end of the run the waste liquor was siphoned from
the still, and a sample was bottled and analyzed if desired.
The fractions of distillate were weighed and their compositions
determined. From these data were calculated the percentage
compositions for different total weights of distillate, assuming
that the fractions were combined successively. By plotting
these values, it was possible t o extrapolate back t o zero
weight of distillate and obtain the composition of the first
infinitesimal fraction of vapor evolved from the liquid, which
was the desired vapor composition (curve A , Figure 2 ) . By
similar calculation the composition of the last infinitesimal
fraction of vapor evolved from the residual Iiquid in the still
could be found (curve B, Figure 2). For a more detailed description of the method of calculation, the reader is referred
to the original article of Rosanoff, Bacon, and White (17).

FIGURE3. ETHANOL-WATER
EQUILIBRIUM
DIAGRAM

Vol. 25, No. 12

MATERIALS.Redistilled water was used in all runs.

Absolute alcohol was prepared from c. P. 95 per cent undenatured ethanol by treatment with lime followed by careful distillation. The finished product was a fraction distilling within
0.1" C., and contained 99.8 to 99.9 per cent ethanol by weight
(specific gravity, 0.7898 a t 20" C.).
General Chemical Company c. P. methanol was fractionated in
a 12-inch column sled with short glass tubes, the middle fraction (distilling over within 0.1" C.) being used in this work.
This fraction gave negative tests for acetone, ethanol, aldehydes,
and reducing substances when tested by the methods given by
Murray (f3), and contained about 99.8 per cent methanol by
weight (specific gravity, 0.7923 a t 20" C.),
Grasselli's reagent grade acetic acid was distilled through a
short column from chromic oxide to remove any formic acid, the
first and last fractions being discarded. The middle fraction
analyzed 99.5 per cent acetic acid according to the freezing point
(data of Worden, 8 f ) and by titration with sodium hydroxide
(freezing point, 15.64" C.).
METHODS
OF ANALYSIS.For both ethanol and methanol the
compositions of the samples were determined at 20' C. by means
of a 10-ml. pycnometer, using the specific gravity data of the
U.S. Bureau of Standards as given in the Handbook of Chemistry
and Physics ( 4 ) for ethanol, and those from International Critical
Tables (IO) for methanol. The thermometer in the pycnometer
stopper was checked against a Bureau of Standards thermometer.
The compositions of the original and residual liquids were determined in duplicate, while those of the various fractions were obtained from one carefully made specific gravity determination, a
check being run only when there was indication that this was
necessary.
In several runs with ethanol and with methanol, determinations
were made with the AbbB refractometer, for reasons which will
be given later. Immersion refractometer scale readings were
taken from the Handbook of Chemistry and Physics (5),and these
readings were converted into refractive indices by means of a
table published by Zeiss ( 2 4 ) .
The compositions of the acetic acid samples were determined
by titration with 0.5 N sodium hydroxide solution with phenolphthalein as indicator, the solution being boiled to remove
carbon dioxide. Duplicate determinations were made on the
original liquid and the first two fractions, the others being checked
only when it seemed necessary to do so. The samples for analysis
were weighed out in small stoppered vials.

PRELIMINARY
TESTSON THE APPARBTCS. Eight preliminary runs were made with ethanol-water to test the reliability
of the apparatus. The data are given in Table I. With the
exception of runs E and F, in which the conditions were the
most abnormal, the apparatus gave consistent results over
a wide range of operating conditions.

FIGURE4. METHANOL-WATER
EQUILIBRIUM
DIAGRAM

I N D U S T R I A L A N D E N G I N E E R I NG C H E M I ST R Y

December, 1933
TABLEI.

PRELIMINARY TESTSON APPARATUSO F

B WON, AXD WHITE (17 )

ROSANOFF, TABLE
111. EXPERIMENTAL
LIQUID-VAPOREQUILIBRIUM
DATA
FOR ETHbNOL-WATER, BASEDON RESIDUAL
LIQCID1 3 STILL

(Ethanol-water)
TEMP.OF
VAPOR
OF
or
FRICOIL
VERUN
RUN
CHARQE"T I O N S
BATH
LOCITVb
Mzn.
M1.
C.
Cm./soc.
92-101
2.3
B
1120
130
100-101
2.8
130
E
90
92-101
4.6
130
60
C
100-101
7.0
100
30
G
92-101
7.5
130
35
A
100-101
8.0
130
35
D
94-101
8.7
130
35
H
100-101
9.5
21
100
F
5 Composition of liquid charged to still
= 10.75 per vent
weight; the initial boiling point was about 90' C.
b Average velocity, for the entire run, u p the inside of the
(Figure 1).
LENQTH VOL.

EtOH

RUN

IN
VAPOR

Wt. %
54.0
53.7
54.3
54.2
54.5
54.3
54.5
54.9
ethanol by
3-cm. tube

24
25
26
27
34
28
29
30
31
32
35
36
37
39

LIQUID-VAPOR
EQUILIBRICX DATA TABLE
TABLEII. EXPERIMENTAL
IV.
IFOR ETHANOL-WATER
RCN

53
50
54
55
48
3
46
4
44
5
20
6
7
8
11
21
12
43
14
15
42
17
33
18
52
19
22
23
24
25
26
27
34
28
29
30
31
32
35
36
37
38
39
40
41

1333

(730 to 750 mm. of mercury)

EtOH I N VAPOR
EtOH I N LIQUID
Wt.
Mole
Wt.
Mole
%
fraction
%
fractzon
6.2
0.025
0.6
0.003
0.046
1.1
0.005
10.9
0.089
2.0
0,008
20.0
0.088
2.1
0,009
19.9
26.2
0.122
3.0
0.012
0.138
0.014
29.0
3.44
0.190
0,020
37.4
4.9
42.3
0.223
0.025
6.15
44.8
0.241
0.028
6.8
49.3
0.275
0.035
8.5
0.307
0,042
53.1
10.14
0.359
0,058
58.9
13.6
0.380
0.065
61.1
15.14
0.400
0.075
63.0
17.14
0.417
0.084
64.7
19.04
65.1
0.422
0.086
19.4
66.3
0.435
0.094
20.9
0.451
0.103
67.8
22.7
69.1
0,466
0.117
25.35
0.473
0.126
69.7
26.85
0.493
71.3
0.143
29.9
72.3
0,505
32.95
0.161
72.9
0.513
0.163
33.3
73.7
0.523
0.184
36.6
0.527
74.0
0.194
38.0
0.537
74.8
41.24
0.215
0.562
75.9
45.1
0.243
0.564
76.8
48.9
0.272
0.576
77.6
0.305
53.0
0,590
78.7
0.341
57.0
79.9
0.608
0.382
61.3
80.9
0.623
0.427
65.6
81.7
0.636
0.460
68.6
0.650
0.488
82.6
70.9
0.671
0.537
83.9
74.8
85.0
0.690
0.581
78.0
86.5
0.715
0.633
81.5
87.7
0.735
0.672
83.9
88.8
0.756
0.706
86.0
90.0
0.778
88.0
0.741
0.803
0.780
91.3
90.1
93.3
0.844
0.834
92.8
94.31
0.866
0.864
94.20
96.16
0.907
0.910
96.28
97.88
0,948
0.952
98.07

When the weight distilled-per cent composition data for

any of the runs in Table I were plotted, the curve produced
had an S-shape, as shown by curve C in Figure 2. For liquids
of lower alcohol content the curves were similar to curve D in
Figure 2 . As may be seen, it was difficult to determine just
where the curve should cross the y axis with only five points
as a guide. I n order to overcome this difficulty, for ethanolwater and methanol-water, the Abbe refractometer was used
to determine the compositions of the distillates from liquids
below 10 per cent by weight. Since the refractometer was
not as accurate as the pycnometer, two runs were made on
liquids of identical composition, one in the usual way using
the pycnometer, and the other with fractions of 3 t o 5 grams
for the first 20 to 25 grams distilled, using the refractometer.
The refractometer points gave a reasonably accurate indication of the exact shape of the curve near the y axis, and the
pycnometer curve was drawn with as nearly the same shape
as possible, as shown by curves D and E in Figure 2. The
y intercept of the pycnometer curve was taken as the vapor

RUN
22
20
18
16
15
24
14
13
25
12
11
26
10
2
30
9
27
8
7
3
6
28
5
4
29

(730 to 750 mm. of mercury)

EtOH IN LIQUID
EtOH
Wt.
Mole
W1.
%
fractaon
%
0.118
69.8
25.5
72.2
0.166
33.7
75.2
0,218
41.7
77.1
0,286
50.6
77.7
0.306
53.1
79.0
0.356
58.6
80.6
0.417
G4.7
82.3
0.477
(0.0
84.2
0.553
76.0
85.7
79.3
0.600
87.1
82.6
0.651
89.2
86.7
0.718
90.7
89.3
0.765
94.22
94.05
0.861

IN

VAPOR
Mole
fractzon
0.475
0.504
0.543
0.568
0.577
0.595
0.619
0.646
0.677
0.701
0.725
0.763
0,792
0.865

EXPERIMENTAL
LIQUID-VAPOR
EQUILIBRIUM
DATA
FOR METHANOL-WATER
(730 t o 750 mm. of mercury)
MeOH I N LIQEID
MeOH
Wt.
Mole
Wt.
%
fractaon
%
9.5
1.3
0.007
18.5
3.0
0.017
27.8
5.06
0.029
35.0
7.0
0.041
0.060
44.0
10.2
0,090
53.5
15.0
0.091
53.1
15.1
0.124
61.5
20.2
0.140
63.8
22.4
25.1
0.159
66.8
70.5
30.2
0.196
0.233
73.8
35.1
76.3
39.6
0,269
0.354
81.0
49.3
49.55
0.356
81.0
0.357
81.1
49.7
84.6
59.5
0,452
84.7
60.0
0,457
88.4
69.9
0.566
91.6
78.9
0.677
92.5
81.3
0.710
0.825
95.7
89.4
0.841
96.4
90.4
0.913
98.1
94.9
95.1
0.916
98.1

I N \-.APOR

Mole
fractzon
0.056
0.113
0.178
0.233
0.307
0.393
0.389
0.473
0.498
0.530
0.573
0.613
0.644
0.705
0.705
0.706
0.755
0.756
0.810
0.860
0.874
0.925
0.938
0.966
0.966

LIQUID-VAPOR
EQUILIBRIUM
DATA
TABLEV. EXPERIMENTAL
FOR ACETICACID-m'ATER

14
7
15
6

(730 t o 750 mm. of mercury)

Hz0
HzO IN LIQUID
wt.
Mole
Wt.
%
fraction
%
16.0
9.9
0,268
16.9
10.4
0.278
23.1
15.0
0,370
29.1
19.6
0.449
24.2
0.515
34.8
0.592
42.1
30.3
0.636
46.8
34.3
0.686
52.4
39.6
56.6
44.1
0.724
61.6
49.5
0.765

20
18
3
19
5
21

74.9
75.4
80.0
84.7
89.9
90.3

RUN
11
10
12
9
13

0.909
0,911
0.930
0.949
0.967
0.969

81.2
81.4
85.0
88.5
93.6
92.8

I N VAPOR

Mole
fraction
0.388
0.404
0.500
0.578
0.640
0.708
0.745
0.785
0.813
0.842

0.935
0.936
0,950
0.962
0.976
0.977

composition in equilibrium with liquid of the composition

of the initial charge.
EXPERTMEUTAL
DATA. The experimental results given in
Tables 11, IV, and V are based on the liquid originally
charged to the still, while those in Table I11 are based on
the residual liquid. The results in Table 111,with the exception of runs 24 and 25, are in excellent agreement with those
in Table 11. The writers feel that this is a further indication
of the reliability of the method.
The system ethanol-water was investigated more thoroughly than the other two, for two reasons-the wide disagreement between the previously published results for this system,

1334

I hTD U S T R I A L A N D E N G I N E E R I D; G C H E %I I S T R Y

Vol. 2 5 , No. 12

The experimental results were carefully plotted, both in

mole fraction and in weight per cent, and the coordinates of
points read from these curves are given in Table TI.
TABLEVI.

COORDINATES OF CCRVE DRAWN THROUGH

PERIMESTAL POISTS FOR ETHANOL-KATER

EX-

(730 t o 750 mm. of mercury)

IN

IN

LIQUID VAPOR

IN

IN

LIQCID VAPOR

IN

IN

LIQUIDVAPOR

IN

IN

0.740
0.760
0.780
0.800
0.820
0.840
0.860
0.880
0.900
0.920
0.940
0.960
0.980

0.777
0.790
0.804
0.818
0.833
0.848
0.864
0.881
0,898
0.917
0.938
0.956
0.978

14.0

83.6
84.3
85.0
85.8
86.8
87.7
88.8
90.0
91.2
92.6
94.2
95.9
97.8

LIQUIDVAPOR

MOLE FRACTION OF ETHANOL

0.010
0.020
0.030
0.040
0.050
0.060
0.080
0.100
0.120
0.140
0.160
0.180
0.200

0.104
0,190
0.250
0.297
0.332
0.364
0.410
0.442
0.468
0.488
0.505
0.51Q
0.531

0.220
0.240
0.260
0.280
0.300
0.320
0.340
0.360
0.380
0.400
0.420
0.440
0.460

1.0
2.0
3.0
4.0
5.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
20.0

10.3
19.2
26.3
32.5
37.7
41.9
48.1
52.7
56.5
59.5
61.9
63.9
65.6

22.0
24.0
26.0
28.0
30.0
32.0
34.0
36.0
38.0
40.0
42.0
44.0
46.0

0.541
0.551
0,560
0.568
0.576
0,584
0.591
0.599
0.607
0.614
0.621
0.629
0.637

0.480
0.500
0.520
0.540

0,560
0.580
0.600
0.620

0.640
0.660
0.680

0.700
0.720

0.646
0.654
0.663
0.672
0.681
0.690
0.699
0.709
0.719
0.730
0.741
0.753
0.765

WEIQHT PER CENT ETHANOL

FIGURE
5. ACETIC ACID-WATEREQUILIBRIUM
DIAGRAM

and a desire to test the consistency of the apparatus. No

special attempt was made to determine the composition of
the constant-boiling mixture of ethanol and water.

DISCUSSION
OF RESULTS
ETEL~soL-WATER.Equilibrium curves for this system a t
atmospheric pressure have been determined by Evans (S),
Blacher, as given by Hausbrand (9), Lewis and Carey ( I I ) ,
Rayleigh ( I @ , Bergstrom as quoted by Hausbrand ( 6 ) , and
Sorel (60). Blachers data were obtained by interpolation
from a graph reported by Hausbrand, and are not therefore
extremely accurate. The data of Sorel were taken from
Elliott (6). A comparison of the various sets of equilibrium
data is given in Figure 3. The curve is that drawn through
the points determined experimentally in the present work,
only the data from Table I1 being plotted.
The data of Len+ and Carey did not appear in the literature until after the completion of the present work, and their
results are in excellent agreement with the latter, being only
a little low in the middle portion of the curve. It is obvious
that the data of Evans and of Sorel are seriously in error,
while those of Rayleigh are evidently low. Blacher and
Bergstrom agree fairly well with the results of the present
work.

67.2
68.4
69.4
70.4
71.3
72.0
72.8
73.4
74.0
74.6
75.1
75.6
76.1

48.0
50.0
52.0
54.0
56.0
58.0
60.0
62.0
64.0
66.0
68.0
70.0
72.0

76.6
77.1
77.5
78.0
78.4
78.9
79.4
79.9
80.5
81.1
81.7
82.2
82.9

(6.0

78.0
80.0

82.0
84.0
86.0
88.0
90.0
92.0
94.0
96.0
98.0

METHANOL-WATER.
Only three equilibrium curves for
this system a t atmospheric pressure could be found in the
literature. These were determined by Bergstrom, as quoted
by Hausbrand ( 7 ) , Bredig and Bayer ( I ) , and Blacher as
given by Hausbrand (9). A graphic comparison of the various sets of data is given in Figure 4. Bergstroms curve is
quite a little higher than that found in the present work. The
data of Bredig and Bayer are high also, but their data are
badly scattered. The data of Blacher were obtained in the
same way as for ethanol-water, and, u hile the reading of the
points from the graph may not have been extremely accurate,
every one of them checks the present curve almost exactly.
The coordinates of points read from smooth plots of the experimental data are given in Table VII.
TABLEVII.

CURVEDRANNTHROUGH ExPOINTS
FOR METHANOL-WATER

COORDINATES OF

PERIYESTAL

(730 t o 750 mm. of mercury)

IN

IN

LIQUID VAPOR

IN

IN

LIQUID VAPOR

IN

IN

IN

IN

LIQUID VAPOR LIQUID VAPOR

MOLE FRACTION O F METHANOL

0.140
0.160
0.180
0.200

0.420 0.739
0.440 0.749
0.460 0.759

0.480
0.500
0.520
0.540
0.560
0.580
0.600
0.620
0.640
0.660
0.680
0.700
0.720

0.768
0.778
0.787
0.797
0.806
0.815
0.825
0.834
0.843
0.852
0.862
0.871
0.880

WEIQHT PER CENT METHANOL

1.0
2.0
3.0
4.0
5.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
20.0

WEIGHT PERCENT OF MORE VOLATILE C O U P O M h T N LIQUID

FIGURE
6 . EQUILIBRIUM
DIAGRAMS

22.0
24.0
26.0
28.0
30.0
32.0
34.0
36.0
38.0
40.0
42.0
44.0
46.0

48.0
50.0

52.0

54.0
56.0
58.0
60.0
62.0
64.0
66.0
68.0
70.0
72.0

87.6
88.3
89.1

74.0
76.0
78.0
80.0
82.0
84.0
86.0
88.0
90.0
92.0
94.0
96.0
98.0

ACETICACID-WATER. Equilibrium curves for this system

a t atmospheric pressure have been determined by Rayleigh
(If?), Blacher as quoted by Hausbrand (8), Bergstrom as

Effect of Addition Agents upon the Corrosion

Rate of Aluminum by Alkalies
Effect of Various Substances
F. H. RHODESAND F. W. BERKER,Cornel1 University, Ithaca, N. Y.
N THE pickling of steel with acid it is well known that
the action of the acid upon the metal may be minimized
b y adding t o the acid certain inhibitors that retard the
corrosion of the metal without at the same time seriously
interfering with the solution of the scale. I n the present
investigation the effects of various inorganic salts and organic
compounds upon the rate of corrosion of aluminum by alkalies
has been determined, with a view to the possibility of finding
some substance that would minimize or prevent the corrosion
of the metal by alkaline solutions.
Some work on the effects of addition agents on the rate of
corrosion of aluminum by alkalies has already been published,
Rohrig (5)found that the addition of sodium silicate to a solution of sodium carbonate decreases the rate of action of the
solution on aluminum, although the silicate increases the rate
of attack of the metal by sodium hydroxide in moderately

Into the reaction tube, M , was placed sufficient of the soluton to be tested to cover the test strip of aluminum. Gaseous
hydrogen was passed through the solution for about 3 minutes
in order to saturate it with hydrogen and thus eliminate any error
due t o the solubility of the gas in the solution. The reaction tube
wvas then immersed in a bath of water maintained at a constant
temperature of 30" C. and allowed t o stand until the solution had
attained the temperature of the bath. A cleaned and weighed
strip of aluminum, 5 cm. long, 1 cm. wide, and 0.0635 cm. thick
was introduced, and the reaction vessel was connected immediately with a Hempel buret in which the volume of evolved hydrogen was measured. Readings were taken every minute for 20
minutes. At the end of this period, the aluminum was removed,
rinsed, and again weighed. The loss in weight of the metal
served as a check upon the amount of hydrogen evolved as a
measure of the total extent of the corrosion. The aluminum
used was commercial sheet aluminum, 99.3 per cent pure.

EFFECTS
OF ADDITIONAGENTSON CORROSION
ALKALIESALONE. I n the first series of experiments the
rate of corrosion of aluminum in sodium hydroxide solutions
of various concentrations was measured. The p H of each
solution was determined, using a standard hydrogen electrode
balanced against a saturated calomel electrode. The results are shown in Figure 2, in which the values for p H and
for rates of corrosion are plotted against the concentrations,
and in Figure 3, in which is shown the variation in the rate
of corrosion with the change in pH. The rates of action of
molar and 0.5 molar solutions of potassium hydroxide were
also measured :

ALKALI
Sodium hydroxide
Potassium hydroxide
Sodium hydroxide
Potassium hydroxide

FIGURE
1. APPARATUS
FOR DETERMINATION
OF CORROSION
RATE
E, F.
M,

Control stopcocks
Reaction bulb

S. Capillary spiral
2'. Inlet tube for hydrogen
X. Hydrogen inlet control

Y . Inlet for alkaline solution

CONCENTRATION RATIO
Molar
Cc. Hn/min.
1
2.50
1
2.13
0.5
1.60
0.5
1.42

Although potassium hydroxide is more nearly completely dissociated than is sodium hydroxide in solutions of these concentrat,ions, there is a distinct and consistent difference in
the rate a t which they attack aluminum.
SALTS. I n the next series of experiments the effects of
salts were investigated. I n each experiment the salt, in the
concentration indicated, was dissolved in a 0.5 molar solution
of sodium hydroxide.
SALT

None
NaCl
NaCl
NaCl

CONCN. RATE
Molar Cc. Hl/min.
1.6
0.5
1.96
1.0
1.93
1.5
1.89

...

SALT
KCl
NazSOo
KnSO,
NrtsPO4

COXCN. RATE
Molar Cc. Hdmin
0.5
1.66
0.5
1.98
0.5
1.83
0.5
3.98

concentrated solutions. Seligman and Williams (6) also

investigated the effects of sodium silicate as a n addition
The addition of sodium chloride increases the activity of
agent in the cleaning of aluminum by alkalies. The use of
silicate as a retarder in this process was patented by Lea ( 3 ) . the hydroxyl ions from the sodium hydroxide and markedly
Rohrig (5) found that certain colloidal substances, such as accelerates the corrosion. The acceleration is most marked
glue, agar agar, and gelatin, exhibited inhibiting action. in solutions of 0.5 molar concentration and decreases as the
Other substances for which similar action has been claimed concentration rises above this value. -4 similar effect is
include ammonium compounds, dichromates, sodium molyb- observed with solutions of sodium chloride in molar alkali.
date, permanganates, and salts of the noble met.als (1, 2, 4). Sodium sulfate shows a n effect similar to that of sodium chloIn the present work the rate of action of the alkaline solu- ride but even more pronounced. The addition of sodium phostions on the aluminum was determined by measuring the phate increases the rate of corrosion very greatly, the increase
rate of evolution of the hydrogen formed by the reaction. being due in part at least t o the hydrolysis of the salt and the
The construction of the apparatus used is shown in Figure 1. resulting increase in concentration of hydroxyl ions.
1336