A STUDY OF T H E D I S T R I B U T I O N OF A L U M I N U M

HYDROXIDE BETWEEN WATER AND SATURATED STEAM

M. A. S t y r i k o v i c h , O. I. M a r t y n o v a , K. Y a . Katkovskaya,
I. Y a . D u b r o v s k i i , and ~.I. Mingulina

Translated from Atomnaya ~nergiya, Vol. 15, No. 2,

pp. 161-168, August, 1968
Original article submitted July 6, 1962

In atomic power stations operating with a single-loop arrangement the steam must be purer than that in thermal
power stations. Deposits in the flow part of the turbines receiving superheated steam* from the reactor are extremely
undesirable; if their removal requires the turbine to be opened and m e c h a n i c a l l y cleaned; the presence of even small
amounts of compounds with long-lived radioactive isotopes in these deposits causes additional difficulties.

Contamination of the s t e a m - w a t e r line of steam-generating equipment and the flow part of the turbine is due
to the mechanical entrainment of moisture drops by the steam and the solubility of inorganic compounds in the steann;
mechanical entrainment of the moisture drops only plays an important part in the contamination at low pressures.
The solubility of various compounds in the steam increases with the pressure and becomes the main factor in con-
t a m i n a t i o n at high and superhigh pressures. A large part of the inorganic compounds contained in the water of the
first loop, except for silicic acid and products of the corrosion of structural materials include salts which are prac-
tically completely dissociated in aqueous solution. A characteristic feature of almost all corrosion products is the
fact that in aqueous solutions they form weak acids or weak bases, i,e., they are present in water in the form of prac-
tically undissociated molecules. These compounds include hydroxides of iron, chromium, cobalt, a l u m i n u m , zir-
conium, etc. Some of them have amphoteric properties.

The transfer of nonvolatile compounds to saturated steam due to solubility in it is determined by the distribu-
tion coefficient K of the given compound between steam and water. Generalizations of the existing experimental
data on the solubility of various substances in saturated steanq showed that in the case of strong electrolytes,K (at a
given pressure) is at a m i n i m u m for compounds with trivalent ions; it increases for compounds with divalent ions and
still more for those with monovalent ions, i.e., it increases with reduction in the degree of hydration of the ions,
Maximum values of K should characterize compounds which are almost completely in the form of molecules in the
aqueous solutions [1]. This indicated that products of the corrosion of structural materials would dissolve in saturated
steam in large amounts.

A check of the solubility of corrosion products in saturated steam was started with experimental investigations
into the transfer of aluminum hydroxide to saturated steam. The corrosion products of a l u m i n u m can exist in solu-
tion depending on the pH of the system or in the form ofdifficultlysoluble hydroxide and a number of basic salts (in
weakly acid, neutral, and weakly alkaline media), or in the form of readily solubIe aluminate ions AtO z- or AlOe-
(in alkaline medium), or, finally, in the form of readily soluble A1s+ ions (in acid medium).

There are no analytical methods for the separate quantitative determination of the acid and base Forms of
AIOOH. The existing method of colorimetric determination of small amounts of a l u m i n u m is based on the capacity
of the A13+ ion to form a colored complex compound with the organic reagent stilbazo. During the preparation of the
solution for the colorimetric determination an acid medium was set up, corresponding to the transfer of all a l u m i n u m
to the A1s+ ion [2] .

In the experimental investigation of the distribution of aluminum oxide hydrate between water and saturated
steam (with the condition of negligibly low moisture content of the steam) it is therefore only possible to determine
the so-called visible distribution coefficient Kvis, the ratio of the total A13+ concentration in the steam to the total
A13+ concentration in water (neglecting the form of existence).

* If the turbine operates with saturated steam, then the impurities contained in it will be transferred to the moisture
which forms when the steam expands in the turbine.

855
 The main problem of the present investigation was to establish the dependence of Kvi s of A1OOH on the pressure
and pH of the solution. The investigation was conducted with a high-pressure bubbling bench. The steam developed
in the eleclzical steam generator from high-purity water passed into the bubbling column th~'ough a layer of solution
of the compound under investigation. Each experiment was conducted at a constant steam flow rate, corresponding
t o very small load of "evaporation surface" (from 0,2 to 0.4 t / m z 9 h). In combination with a considerable height
of the steam space in the column (800-1200 mm), this ensured good natural separation of the moisture, i.e., negligibly
low moisture content of the steam and the transfer of impurities to the steam only due to solubility in it.

Ken, %
35
i ~--~ o Ken' %

o/ C)-
i
3O 15 w

ZSI
|

/
2O

15 f
I
10

06 A

7 8 9 10 11 12 ~s 7 8 9 10 11 12 13
a pHo b pXe

Fig. 1. Dependence of Ken of A1OOH on the pH 0 of the solution: a) at a pressureof185

arm; b) at a pressure of 100 arm.

The achievement of practically complete equilibrium between the steam and solution was ensured by prolonged
time of contact of the steam with a solution of the investigated compound (approximately 5-8 see at 185 atm and
1.5-2.5 sec at 100 atm).

To prevent overheating or condensation of the steam in the bubbling column the internal working vessel was
surrounded by an isothermal steam jacket. The selection of samples of saturated steam was made representative by
condensation under the total pressure of the whole current of steam passing through the solution of investigated com-
pound into the bubbler.

KAIOOH
vis Samples of solution from the bubbling column were cooled in a cooler under

:1 \! total pressure and wer6 taken periodically. To prevent suspended particles entering
the sample, there was a silver filter at the inlet to the sampling tube.

t'
/gSatm During adjusting experiments it was noticed that the A1OOH had a tendency
to form colloidal solutions which coagulated on prolonged standing at room tempera-
/ ture; it also tended to be sorbed on the walls of the glass flasks. To prevent these
, / \ phenomena a solution of hydrochloric acid (1 : 1) in amounts of 1 ml per 100 ml of
sample was added to the flasks used for samples of the steam condensate. Acid(5 ml
2 per 100 ml) was added immediately after the pH 0 was measured.*

In work with alkaline solutions of A1OOH contact with atmospheric carbon di-

'L oxide reduced the pH 0 of the solution and difficulty soluble basic salts were formed.
Samples were therefore taken and kept in bottles with alkali seals. Carbonate ions
were removed from the feed water of the steam generator by passing it through a
2 strongly basic anion-exchange filter.
:10-3 The initial solution for each experiment with a certain concentration of alumi-
5 C 7 8 9101112
pHo num and pH 0 value was mainly prepared from sodium aluminate (NaA10 z) and partial-
ly from A1OOH, The concentration of aluminum in the initial solution, samples of
Fig. 2. Value ofKvi s of A1OOH
atpressures of 100 and 185 atm. * The quantity pH 0 is the hydrogen exponent at room temperature.

856
Ktr
solution, and steam condensate were determined by a photocolori-
I l Ill I
,[ ' ] 1II I metric method with stilbazo reagent; the pH 0 was determined by
I a potentiometrie method with a glass electrode. To prevent the
effect of carbon dioxide on the pH0 of the solutions the measure-
ments were conducted in a special box through which COz-free air
was passed.
! \t i
The transfer of A1OOH to saturated steam was initially stud-
led at 185 aim. This pressure was chosen so that the a l u m i n u m

,o!
2 concentrations in the saturated steam could readily be determined
by a photoeolorimetric method.

An experimental investigation at 185 atm was conducted

2 r G810' 2 * 6'8102 2 r gl~ t over a wide range of pH 0 values (from 7.0 to 11,8). The A1OOH
concentration in the solution was 0.1-430 mg/kg* ; the concen-
trations of solution in most of the experiments were small - about
Fig. 3. Dependence of Ktr on the pressure,
0.5-10 mg/kg. The A1OOH concentration in the steam in various
experiments was 0.03-15,35 mg/kg.

A second series of experiments was conducted at 100 atm. The results obtained at 185 atm indicated that at
100 aim the concentration of A1OOH in the steam would still readily be determined by a photocolorimetric method.

In experiments at 100 atm the pH 0 range was somewhat wider - from 6.8 to 12.25. In these investigations the
A1OOH concentration in the solution was 0.3-147.0 mg/kg, and in steam 0,014-0.13 mg/kg,

The results of an experimental investigation of the transfer of A1OOH to saturated steam at 185 and 100 atm are
shown in Fig. 1. The scatter in the obtained data may be due to the very sharp dependence of the e n t r a i n m e n t coef-
ficient Ken of A1OOH (i. e., the ratio of its content in the steam to the content in solution) on the pH 0 of the solution.

For the same pH 0 values the quantity Ken for a pressure of 185 atm is greater than for 100 atm, This agrees
with the known fact that the solubility of substances in saturated steam increases with the pressure.

Over the whole investigated range of pressures and values of pH 0 the values of Ken of A1OOH are several orders
higher than the possible drop entrainment in the bubbling column during the experiments. This indicates that the
obtained experimental data on Ken are equal to the values of g v i s.

At room temperature the first dissociation constant of A1OOH with respect to the acid type (KI ~ 10 -Iz) is several
orders less than with respect to the basic type (only the product K1 9 K2 , K3 ~-. 10 -33 is known). Consequently, the dis-
solution of A1OOH at room temperature produces a weakly alkaline medium (pH 0 ~ 8); i.e., the greatest fraction of
molecules in the total content of a l u m i n u m compounds in the solution is found at this value of pH 0, The m a x i m u m
values of Kvis of A1OOH at 100 and 185 aim should also probably be found in a weakly alkaline m e d i u m (Fig. 2) and
correspond to the m a x i m u m possible quantities of undissociated A1OOH molecules. The m a x i m u m values of Kvi s,
close to the true distribution coefficients Ktr of the molecular form of A1OOH, were therefore used to plot the depend-
ence of Ktr on the pressure (Fig. 3). For comparison, the pressure dependences of Ktr are plotted for silieie and boric
acids.

The analytical expression for the dependence of Ktr of A1OOH on the pressure has the following form:

KAIOOH [AIOOH] s /Y w']-~

tr - - [AIOOH] w-- ~.~ss .' '

where )'w and Ys are the densities of water and steam at the given pressure, respectiveIy. The values of Ktr, deter-
mined using a graph or analytical relationship, will correspond to the m a x i m u m possible transfer of A1OOH to satu-
rated steam due to solubility.

We have made an experimental investigation of the distribution of A1OOH between water and saturated steam
at 100 and 185 atm over a wide range of pH 0 values. The investigation confirmed the assumption of considerable
transfer of A1OOH from water to saturated steam due to solubility. We used the results to establish the dependence

* The A1OOH concentration is expressed in terms of the A1a+ ion.

857
of the true disl~ibution coefficient of A1OOH on the pressure. It corresponds to pH 0 values of about 8.0-8.7, With in-
crease and reduction in the p H 0 from the value at the point of the m a x i m u m the visible coefficient of distribution
decreases sharply.

LITERATURE CITED

1. M . A . Styrikovich, O. t. Martynova, I. Kh. Khaibullin, and E . I . Mingulina, Teplo~nergetika, No. 9,50(1959).

2. S. Yu. Fainberg and A. A. Blyakhman, Collection of Transactions of the State Scientific Research Institute for
Nonferrous Metals [in Russian], No. 12, Moscow, Metallurgizdat (1956),

858