WEIR-A MMONIA-CA RBON DIOXIDE-WA T E R EQUILIBRIUM w

close contact for a sufficient time can coalesce if the interfacial film does not completely inhibit
diffusion of the dispersed phase; the rate of coarsening will depend both on the emulsifying agent
and on the particular liquid concerned.

Conclusion
This suggested mechanism can account for many observations on emulsion stability and is
independent of the particular nature of the interfacial film, i.e., whether it is ionic or non-ionic,
unimolecular or multimolecular in thickness, viscous or rigid, soluble or insoluble. It is interesting
that, although the diffusion of molecules of an emulsified liquid through the interfacial film does
not appear to have received previous attention in relation to emulsion stability, it is an integral
feature of Harkins theory of emulsion p~lymerization.~

15 Goddington Lane
Orpington, Kent
References
See Stamm, A. J., & Kraemer, E. 0..J . phys. Chem.,
1926, 30, 992; Hildebrand, J. H., ibid., 1941, 45,
1303
PSee Clayton, 1954 The Theory of Emulsions
and their Technical Treatment, 5th edn. (by
C. G. Sumner), Chapter VI (London : Churchill)
Received 12 October, 1956
Cockbain, E. G., 62 McRoberts, T. S., J . Colloid Sci..
1953, 8, 440; see also J. T. Davies, Perf.
essent. Oil Record, 1952, 43, 338, 353, 365
Harkins, W. D., J. chem. Phys., 1945, 13, 381;
J . Amer. chem. Soc., 1947, 69, 1428

THE AMMONIA-CARBON DIOXIDE-WATER EQUILIBRIUM IN

BOILER FEEDWATER CONDITIONJNG
By C. D. WEIR
Charts are presented showing the equilibrium relations between ammonia and carbon dioxide in
aqueous solutions a t temperatures to 200 over the range of p H values of interest in feedwater con-
ditioning, A criterion-equivalent pH-is defined and tentatively advanced as a basis for the
design of high-temperature feedwater conditioning. The equilibria between gaseous and dissolved
CO, and NH, are discussed and applied, as an example, to the estimation of the behaviour of a power
plant de-aerator.

Introduction
As originally discussed by Wesly, the conditioning of boiler feedwater by ammonia was of
two distinct types. I n one form, suitable for application to drum-type boilers, the treatment
consists of the addition of ammonium salts to the treated make-up which is then fed either direct
to the boiler or to an evaporator. Alkalinity must be present in the make-up since the ammonium
salts employed react acid and, for this reason, base-exchange softening is recommended.
Ammonium sulphate or, preferably, ammonium monohydrogen phosphate, may be used and
reacts in the boiler drum or in the evaporator with the sodium alkalinity present to form the
corresponding sodium salt and free ammonia:
3NaOH + (NH,),HPO, = Na,PO, +
2NH, +
3H,O.
The ammonia thus released cycles in the system and so provides the necessary alkalinity
in the condensate and feed lines.
The method of designing the treatment recommended by Wesly is as follows: the ammonia,
as NH,, equivalent to the carbon dioxide content of the condensate is determined and an excess
of 6 p.p.m. NH, allowed: this gives the total equivalent of ammonium salts to be added to the
feed to the boiler or evaporater. The phosphate equivalent, as Po,3-, of the residual hardness in
the feed is then determined and an excess of 1 to 1 . 5 p.p.m. PO, allowed. This gives the

J. appl. Chem., 7, September, 1957

508 WEIR-AMMONIA-CA RBON DIOXIDE-WA T E R EQUILIBRIUM

amount of ammonia to be added as the mono- or di-hydrogen phosphate and this, as NH,, is
subtracted from the total ammonia required. The difference is added as the sulphate or
bisulphate.
The second type of treatment devised by Wesly, intended specifically for forced-flow boilers,
is the direct injection of free ammonia, as a gas or in water solution, into the condensate line.
As before, the treatment is designed by estimating the CO, content of the condensate and adding
the stoicheiometric equivalent of ammonia plus an excess of 6 p.p.m. as NH,. In both cases
pH 9 . 5 a t 25" is desired in the condensate.
In common with all similar methods of designing water-conditioning treatment, that out-
lined above involves procedural approximations which do not always give the optimum treat-
ment conditions, and a more exact method based on a consideration of equilibrium conditions is
now presented. The equilibria have been estimated over a range of temperatures, and these
data, in conjunction with concept of ' equivalent pH ' also developed, provide a basis whereby,
it is tentatively suggested, the behaviour at different temperature levels of a treated condensate
or feed may be predicted.
The primary advantage of ammonia treatment as applied to forced-flow boilers is that the
concentration of treatment chemicals to the extent of precipitation on the heating surfaces
is impossible. This does not imply that the ammonia in the system is equally distributed
between the steam and water phases; this matter may assume importance in connexion, for
example, with feed heaters, ejector condensers and the de-aerator. Within the limits of the
available data it is possible to estimate this distribution. No great precision is to be expected,
but the estimates given may be presumed to provide a t least a guide to the behaviour of the
system.

Equivalent pH
In oxygen-free solutions certain common constmctional materials, in particular the mild
steels and iron, tend to exhibit a minimum corrosion rate at a certain pH value of the solution.
At lower temperatures the corrosion rate may vary little over a rather wide range of p H values
round that corresponding to the minimum but, as the temperature is raised, this range becomes
narrower.
If it is asserted that, at a given standard temperature, say 25", a certain pH, pH", will result
in minimum corrosion and if it is assumed as a tentative working hypothesis that at any other
temperature minimum corrosion will occur at the same ratio of hydrion to hydroxyl ion and that
the position of this minimum on the pH scale is independent of the actual concentrations (or
activities) then we can define an ' equivalent pH ', pH?, which relates the ratio [H-]/[OH'] a t
any temperature t to the pH corresponding to the same value of the ratio a t the standard tempera-
ture. It is to be noted that nothing is said as to the numerical value of the minimum corrosion
rate which will normally vary with the temperature.
We have, by hypothesis,

or [H']t e - Kwt
- -o[H.]03 ..................................................... .(I)
KW

whence pH: = pHt - 3p .................................................

where Kwo, KWtare the dissociation constants for water a t the standard temperature and t
respectively. The concept is thus a very simple one, and provides a means of relating the hydrion
concentration at high temperatures to that a t the standard temperature which is independent of
any variation in dissociation constants other than that of water.

The ammonia+xubon dioxide-water equilibrium

In order to utilize this concept and in general to predict the behaviour of a condensate con-
taining CO, and conditioned by ammonia, it is necessary to know the pHt values corresponding

J. appl. Chem., 7, September, 1957

 WEIR-AMMONIA-CA RBON DIOXIDE-WA T E R EQUILIBRIUM 507
to a range of ammonia and carbon dioxide concentrations. For the determination of the con-
dyctivity of condensates, curves for 25" only have been prepared by Meyer & Werner.= Unfor-
tunately this data does not cover the somewhat troublesome region near pHo = 7.8, nor is the
mode of plotting convenient for present purposes. For these reasons the 25"-data have been
recalculated for the present paper.
Because of the low concentrations involved, it has been considered permissible to assume
the activities of the ionic constituents to be equal to their concentrations, but in the region of
pHo = 7 . 8 , none of the simplifications usually made in the estimation of acid-base equilibria can
be utilized. The relevant equilibria are as follows:

........................................ . (3)
*

......................................

[CO3''I [H'I .......................................................

= K
: (5)
[HCO3' 1
and [H'] [OH'] = Ktw ........................................................ (6)
Values of Kwoare available over a wide range of temperatures3 and for K: the data of Noyes4
have been utilized.* Values of K: are available up to 50" and K,t a t 25" and 38.6 There are
several possible assumptions on which KZtand Kat may be extrapolated, but Harned's empirical
equationa
p K t = p K e - 5 * 0 x 10-6(t--B) ............................................. * (7)

in which KO is the maximum value of Kt and 0 the corresponding temperature, seems the most
reliable.** In the particular cases involved tne error can, in any event, scarcely be large.
In addition to the equilibrium relations (3) to (6)we have also the stoicheiometric equations
CNH, = [NH,'] + [NH,OH] ............................................... * (8)
and CCO, = [H,CO,] + [HCO,'] + [CO,"] ..................................... .(9)
From equations (3) to (6) and (8) and (9) the ratios

........................... .(10)

uco3" = [CO,l7
- ....................................................... * (12)
ccoz
may be calculated for any pHt in the usual manner. Finally from the ionic balance
[H'] + ~ [OH']
u N H ~ * C N H= + +
( U H C O ~ 2 ~ 0 ,)Ccoz.. .......................
II
* (13)

the value of, say, CNH, may be obtained for any given pHt and CCO,.
The results of these calculations are presented in Figs. 1-5, which are plotted to a base of
equivalent pH, i.e. pH,", a scale of pHt is however also given in each figure. The main feature
of these curves which requires comment is that, for a given CO, content, the ammonia
concentration required to maintain a predetermined value of pH,O increases markedly with the

*The precision of both these sets of data may be questionable, particularly a t the higher temperatures.
Pending further work they have been adopted as the best available.
**The alternative assumptions that either the differential specific heat or the enthalpy of reaction are
constant in the exact relation
dlnK
-=--
AH
dT RTZ
do not generally fit the experimental results for electrolytes.

J. appl. Chem., 7, September, 1957

508 WEIR-AMMONIA-CARBON DIOXIDE-WATER EQUILIBRIUM

50 55 60 65 7.0 7 5 8.0 8.5 9.0 9%

PH

FIG. 1 . Eouilibriunz between N H . 1ao.l and CO.

I \ I ,
FIG.2. Equilibrium between N H _, (aq.)
. - . and CO,
(q.) at 25" (aq.) at 50'

50 55 60 65 70 7.5 8.0 8.5 90

PH
FIG. 3. Equilibrium between N H , (ag.) FIG.5. Equilibrium between NH, (ag.) and CO, (aq.) at
and CO, (q.) at 100" 200"

temperature. This is almost entirely due to the smaller dissaiation constant of ammonia at the
higher temperatures. As far as the writer is aware the importance of this effect in relation to the
efficacy of ammonia treatment has not previously been remarked upon in the literature.

Vapour pressures of ammonia and carbon dioxide above their solutions

I t has been pointed out in the Introduction that it is of some importance to be able to
estimate the vapour pressures of NH, and CO, above a condensate. For ammonia, the
equilibrium condition is
NH,(soln.) + NH,(gas)
whence Ft,h. + + .
RTln [NH,] = Ftgas RTln P N H ~ . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . .(I41
in which Ft,h., Ftgasare the standard free energies' of ammonia in aqueous solution and in the
gaseous state respectively a t temperature t and PNH, the vapour pressure.

J. appl. Chem., 7, September, 1957

 FVEI R-AMMONIA-CA RBON DIOXIDE-WA TER EQUILIBRIUM 509
Since AFt = Ftsola -Ftgas = RTlnK,t
where K: is the corresponding equilibrium constant, it follows that
p (PNH3
, ~ ) =
pK2
.............. .................. ..................... ...(15)
Now
"H,] = CNH3 -"H4']
and it is readily shown from the relations of the preceding section that
............................ .(16)

whence it follows that

K,t C N H ~............................................ (17*)
PNH3 =

1 + Klt
j p . 1
In the case of CO, we obtain in a similar fashion

WI2 Cc02 .................................. (18)

pcoz = KSt
+
[H'], K,t[H'] +
K,tK2t
in which K,t is the equilibrium constant for the reaction
CO,(soln.) + CO,(gas).
I t remains to evaluate the constants K4t and I<:. The necessary experimental data have not
been found in the literature for the concentrations of interest, but by integrating the van't
Hoff equation
In K - - A
d-- -RHT"2 .......................................................
dT . (19)
on the reasonable assumption that A H o is independent of temperature for the physical process**
of solution, values K: and K: have been obtained and used in plotting the curves of Figs. 6and7.
It will be observed that uncertainties in the numerical data render the effect of the
procedural assumptions insignificant.

Applications
The two following simple and quite arbitrary examples may serve to illustrate the appli-
cation t o practical use of the results of the previous sections.
(1) It is required to treat with ammonia a condensate containing 1 p.p.m. CO, so that the
equivalent pH at 150" shall be 8.
Reference to Fig. 4 shows that 1.7 p.p.m. of NH, are required. The actual pHt at 150" is
6.86 and the pH" (25') of the treated condensate is found to be 9.3 from Fig. 1.
The ammonia addition required when amved a t in this way corresponds to the stoicheio-
metric equivalent plus an excess of about 100% and is about four times the ammonia con-
centration required to give pH 8 at 25".
(2) A condensate containing 2 p.p.m. of CO, and 2 p.p.m. of NH, (pHo= 9.15) enters a simple
de-aerator in which 1% is evaporated a t 100". The NH, and CO, contents and the pHO a t the
outlet are required.
Let C,, C, be the concentrations of CO, and NH, respectively at the inlet, Ctl, C', the corres-
ponding concentrations at the outlet and C1",C i t the concentrations in the evaporate.

*It is preferable to use this relation rather than the equivalent expression
PNH3 = (1 - - d K 4 C N H 3
in view of the small values of the bracketed expression at low values of the pH.
**i.e. A H " relates, e.g., to the equilibrium
NH,(soln.) + NHJgas)
not NH,' + +
OH' + NH3(gas) H,O

J. appl. Chem., 7, September, 1957

510 WEIR-AMMONIA-CA RBON DIOXIDE-WA T ER EQUILIBRIUM

5
PH'
FIG.6. Equilibrium between NH, ( s o h . ) and FIG.7. Equilibrium between CO, ( s o h . ) arid
NH, (vapour) f o r 1 p . p . m . in solution CO, (vapour) for 1 p.9.m. CO, i n solution
We have
c, = (1 - X ) C , ' + XC,"
c, = (1 --X)C,' + XC,)'
where x is the fraction evaporated.
Also, on the assumption of equilibrium,
C," = f,(pH)*C,' and C i ' = f,(pH)*C,'
in which the functions f, and f, of the pH are those expressed graphically in Figs. 6 and 7 res-
pectively.
From these equations we obtain immediately

C'= c2
2
1 -x(l -f2)
'
which have to be solved simultaneously.
The procedure is to plot, as functions of pH, the
right-hand sides of these equations in the pH diagram
appropriate to the temperature in question; the point of
intersection is the required solution.
This is shown for the data of the present example
in Fig. 8 which is an extract from Fig. 3. The line
AB corresponds to C,/{l -x(l -fl)} and CD to
C,/{l -x(l -fa)}. The point E represents the con-
ditions a t the de-aerator outlet. I t will be noted that
i in relation to the CO, only a small proportion (in this
.o 4 case about 10%) of the ammonia is removed. The
0
.5 I" proportion may be increased by increasing the pH (for
example by adding sodium hydroxide) or by raising the
,O temperature and so increasing the proportion of un-
dissociated ammonia in the condensate.
From the diagram of Fig. 8:
C, = 0.55 p.p.m.
C, = 1.8 p.p.m.
7.5 8.0
pHt = 8.55
PH' and from Fig. 1
FIG.8. pHo = 9.4
J. appl. Chem., 7,September, 1957
 WEIR-AMMONIA-CA RBON DIOXIDE-WA T E R EQUILIBRIUM 511
Acknowledgments
Acknowledgments are due to the Authorities of the University of Glasgow, who provided
facilities during the preparation of the paper and to Messrs. Merz and McLellan, for permission
to utilize material from a document prepared by the writer on their behalf in May 1955.
Grateful thanks are also due to Dr. M. Werner, Opladen, Germany, for much helpful
criticism of the original document.

University of Glasgow
Department of Mechanical Engineering
References
Wesly W., Kovros. Metallsch.. 1942,18, 158
a Aleyer, A., &Werner, M., Mitt. Ver. Grosskesselbesitz.,
1950, 10, 132
See e.g., Harned, H. S., & Owen, B. B., ' The
Physical Chemistry of Electrolytic Solutions ',
A.C.S. Monograph No. 95. 1943 (New York:
Reinhold Publishing Corp.); also Uhlig, H. H.,
ed., 'Corrosion Handbook', 1948, (New York:
John Wiley & Sons, Inc.)
Received 5 November, 1956
Noyes, A., & Kato, Y.. Publ. Carneg. Inslw,
1907, 68, 178; also Uhlig, H. H., loc cit.
6Harned, H. S., & Davis, R., J . A m e r . chem. Soc.,
1943.65, 143; also Uhlig, H. H., Zoc. cit.
Harned, H. S., & Embree, N. D., J . Arnev. chem. Soc.,
1934, 65, 1050
'Latimer, W. M., Oxidation Potentials ', 1952,
(New York: Prentice-Hall Inc.)

Appendix
The values of the dissociation constants used in the preparation of the equilibrium curves
are given in Table I, in which the extrapolated values were obtained from Harned's equation
as discussed in the text. I t may again be emphasized that the values at the higher temperatures,
whether experimentally determined or extrapolated are open to considerable doubt: they are,
however, probably the best available at the present time.

Table I
Dissociation coitstarzts for ammonia, carbon, dioxide and water

Temp., NH3 COZ Ha0

OC -log K, -log I<,
--log Kz -log K,
25 4.74 6.35 10.33 13.98
50 4.74 6.29 10.17 13.26
100 4.87 6.24* 9.97* 12.26
150 5.17 6.44* 10.14* 11.84
200 5.59 6.90* 10.31* 11.38
*Extrapolated.

Table I1 gives values of the ratios aNH3, t1Hco3 and ~ 0 over 3 the range pHt 5 . 5 to 10
units, and is included here since the values of these ratios are available in the literature at 25"
only, and it is thought that the estimated values a t higher temperatures may prove of service in
equilibrium calculations other than those considered in the present paper.

Table II
(a) Carbon dioxide aHco3= [HCOa']
- (b) Carbon dioxide aco3 =
Ccoa
-
PHt 50 100" 150" 200 50" 100" 150" zoo0
5.5 0.140 0.152 0.103 0,034 0 * 000 0.000 0.000 0.000
6.0 0.339 0.362 0.266 0.112 0.000 0.000 0.000 0.000
6.5 0.619 0.643 0.535 0.285 0 * 000 0.000 0 * 000 0.000
7.0 0.837 0.850 0.784 0.557 0.001 0.001 0.001 0.001
7.5 0.942 0.945 0.919 0.798 0.002 0.003 0.002 0.001
8-0 0.981 0.972 0.957 0.923 0.007 0.01 1 0.007 0.005
8.5 0.974 0.961 0.960 0.961 0.021 0.033 0.022 0.015
9.0 0.936 0.902 0.930 0.947 0.063 0.097 0-068 0.046
9.5 0.824 0.745 0.813 0.865 0.176 0.254 0.187 0.134
10.0 0.598 0.481 0.578 0.671 0.402 0.519 0.422 0.329

J. appl. Chem., 7, September, 1957

512 PENKETH-OXIDATION POTENTIALS OF A N T I O X I D A N T S

Table 11-continued.
"Hi 1
(c) Ammonia aNHg= -
CNHB

:Hi 50"
0.999
100"
0.987
150"
0.915
200"
0.661
6.0 0.997 0.960 0.773 0.382
6.5 0.992 0 * 886 0.517 0.163
7.0 0.970 0.721 0.253 0.058
7.5 0.913 0.437 0 * 097 0.019
8.0 0-768 0.197 0-033 0.006
8.5 0.512 0.072 0.016 0.002
9.0 0.249 0-024 0.003 0.001
9.5 0,095 0.008 0.002 0.000
10.0 0.032 0.002 o*ooo 0.000

TIE OXIDATION POTENTIALS OF PHENOLIC AND AMINO

ANTIOXIDANTS
By G. E. PENKETH

A simple and rapid polarographic method of measuring the relative oxidation potentials of organic
compounds has been developed, and values for a large number of phenolic and amino compounds have
been determined to find whether any relationship exists between Oxidation potential and antioxidant
activity. Buffered electrolytes a t various pH levels were used and when the determined potentials are
plotted against the pH of the electrolyte, most monohydric phenols give straight-line graphs which
correspond closely to the theoretical equation for oxidations in which hydrogen ions are involved
in the electrode process
E = E , - 0.058 p H at 20"
where E, is the standard oxidation potential. Other phenolic compounds give graphs showing a
' break point ' a t about pH 6 , and have different slopes above and below this value.
Most of the hydroxyanisoles and amines exhibit slight curves when potential is plotted against pH.
Sufficient alkylated phenols have been examined to establish definite trends in the effects of substi-
tution; thus ortho and para substituents lower the oxidation potential by about 0.08 v while meta
substituents have a much smaller effect. These effects are largely additive and the oxidation potentials
of polyalkylated phenols can be predicted from a knowledge of the effects of the separate substitutents.
With certain 2 : 4 : 6-compounds containing large groups, the measured oxidation potential is higher
than that predicted and this is attributed to stenc hindrance.
In general, as the oxidation potential decreases, the antioxidant activity increases. The relation-
ship is not invariably close and, with certain substituted hydroxyanisoles, the antioxidant activities
measured are even higher than would be anticipated from their low potentials. No compound of
high antioxidant activity has been found to possess a potential higher than 0.70 v and hence the
polarographic method should prove to be valuable for a rapid preliminary assessment of possible anti-
oxidants.

Introduction
A simple method of measuring the relative ease of oxidation of organic compounds would
be very useful for many purposes. In the antioxidant field, Moureu & Dufraissel observed that
many compounds which are oxidation inhibitors are themselves relatively easily oxidized, and
hence the measurement of their oxidation potentials should be a guide t o their inhibitor efficiencies
and could prove to be a rapid method of evaluation. Whereas the oxidation/reduction potentials
of such reversible systems as hydroquinone/quinone can be readily determined, the usual methods
cannot be applied to compounds that exhibit irreversible oxidation (or reduction) and it is in
this class of compound that the majority of antioxidants are to be found.
Conanta devised a method of measuring ' apparent ' oxidationlreduction potentials of such
irreversible systems. He used a series of reference systems of various known potentials and added
a quantity of the phenol under investigation to each. Below a certain potential no reaction
occurred and above it the potential dropped at a rate which increased as the potential of the

J. appl. Chem., 7, September, 1957