THE PRUSSIAN BLUE PARADOX

DAVID DAVIDSON
Brooklyn College, Brooklyn, New York

The reactions between ferric and ferrocyanide ions to occurrence of the ionic redox equilibrium: Fe++++
form Prussian blue and of ferrous and ferricyanide ions
to form Turnbull's blue are profoundly influenced by the
[Fe(CN)6IS =Fe* +[Fe(CN)%]=, which is largely
displaced towrd the right.

INTRODUCTION 12H2O (in tenth-formal KHSOd* the orieinallv color-

less liquids form a yellow soluti& which quickly alters
" .
T HE action of ferric ion on ferrocyanide ion re-
sults in the production of blue substances bearing
the generic name Prussian When an excess
to green and finally to blue within one minute. If the
taken are t h e times as strong (0.0006
formal) the succession of colors is
~~- cornnlete
~-- a~ ~ within
- ~ one
.---
~ ~
of ferric ion is employed, the product has the formula second while, if the solutions are only one-third as
FQ[WCN)EIJand is known as insoluble P ~ s s i a nblue. strong as the original solutions (13.00007 formal), the
From equimolecular quantities of ferric and ferro- ,,,tion requires the better part of an hour.
cyanide ions, a product is formed which contains the
metallic ion of the soluble ferrocyanide employed. THE REDOX EQUILIBRIUM
Thus, with potassium ferrocyanide, for example, the A finite time is thus seen to elapse between the
blue product has the formula KFe[Fe(CN)E] and is mixing of ferric and ferrocyanide ions and the appear-
known as soluble Prussian blue. The term soluble is a ance of Pmssian blue. It may now be worth while to
misnomer, however, and refers not to the true solu- consider the possibility of an oxidation-reduction
bility of the product (it is actually insoluble in water) reaction between these ions in solution preceding the
but to the ease with which it forms colloidal solutions. formation of Prussian blue.
Although it would appear from the formulas given A mixture of ferric and ferrous ions forms a reversible
above that the formation of Prussian blue involves redox system which imparts a definite potential to an
simply the mutual precipitation of femc and ferrocya- inert electrode such as platinum, as does, likewise, a
nide ions, and should, therefore, be an instantaneous mixture of ferro- and ferricyanide ions. From the
reaction, it has actually been found to be a relatively single electrode potentials of these systems one may
slow reaction.' It might he anticipated, however, predict the following equilibrium:
that excess of ferric ion would accelerate the formation
of Prussian blue. Actually, the opposite is found to Fet++ + [Fe(CN)al-= Fe++ + [Fe(CN).Ie
be the case. Excess of ferric ion hinders the formation and calculate the equilibrium constant:
of Prussian blue.
To account for the slowness of Prussian blue forma-
tion, Vorlander took refuge in the labyrinths of colloid
chemistry. He corrtended that ferric salt solutions The size of the equilibrium constant indicates that the
were to be considered incomplete colloids whose reversible reaction is largely displaced toward the
particles had a size between that of true solutions right. In other words, when ferric ion is added to an
and those visible in the ultramicroscope. Hence, the equimolecular quantity of ferrocyanide ion, an oxida-
slow reaction with ferrocyanide ion. The results ob- tion-reduction occurs in which the ferric ion is almost
tained in the present paper make it appear that this ex- completely reduced to ferrow ion by the ferrocyanide ion
planation is invalid and that the elucidation of the which is simultaneously ozidieed to ferricyanide ion.
-
-
Prussian blue paradox lies in the redox reaction: These dilute solutions are conveniently prepared by diluting
1 cc. of 0.01 formal solutions to 50 cc. For ferrocyanide ion
Re+++ + [Fe(CN).IE e Fe+++ [Fe(CN)dE dissolve 0.422 g. of GFe(CN)s.3H20in 100 cc. of water. For
ferric ion dissolve 0.482 g. of ferric alum, NH,Fe(SO&12H20
in 100 cc. of tenth-formal KHSO,. The bisulfate serves to pre-
A DEMONSTRATION vent hydrolysis of the ferric salt. (A perfectly colorless solution
is obtained.) Dilutions of the ferric alum solution are made with
~tis a simple matterto show that the formation of tenth-formal KHSOa. The ferrocyanide solution should be
freshly prepared.
Prussian blue is a slow reaction. On mixing equal
volumes, say 50 cc. of 0.0002 formal K4Fe(CN)s.3H20
r G, N, L,,, ,
, M, w D A L L .'T,,ermOdynamicS,..
,
McGraw-Hill Book Co., New York City. 1923, p. 396.
water) and 0'0002 NH4Fe(S04)a a E . M ~ L L E AND
R T . STANISCA, I. Prakt. Chon.,(2) 79, 81
(1909). These authors employed 0.71 v. and 0.41 v. for the
- respective potentials. The current values of 0.75 v. and 0.49 v.
1 D. VORI.IWDEB,
Bw., 4 4 181 (1913); Ko2bid-Z.. 22, 103 yield a value closer to 20,000 for Klw. The actual value is not
(1918). significant for the present discussion as long as it is large.
238
 Thus, for example, if 1.0 mole of ferric ion is treated beakers. The first is kept as a control. A few cc. of
with 1.0 mole of ferrocyanide ion, it may be calculated 0.01 M ferric ion is added to the second beaker, while
that, a t equilibrium, 0.9968 mole of ferrous and ferri- a few cc. of 0.01 M ferrocyanide ion is added to the
cyanide ions will have been formed and only 0.0032 thud. A blue color appears immediately in the thud
mole of each of the original ious will remain. beaker, slowly in the control, and still more slowly in
the second beaker. That is to say, while excess of
Start with
+
Fe+++ [ F e ( C N ) a l ' ~ F e + +
1.0 1.0
+
0
[Fe(CN)el'
0 ferrocyanide ion has the expected accelerating effect
At equilibrium (1 -3) (1-1) x x on the formation of Prussian blue, excess of femc ion
x'
-
- 100.000 has just the opposite effect; namely, it actually hinders
-(
x - 0.9968 the formation of the blue.

It is now conceivable that any or all of the possible FURTHER THEORY

combinations of the ions that participate in the redox In considering the explanation of this paradox, it
equilibrium may precipitate. If it is assumed, as should be kept in mind that Prussian blue is not formed
seems likely, that it is the ferric and ferrocyanide ions instantaneously, but only after the establishment of
which are actually involved in the formation of Prus- the redox equilibrium a t the expense of the ferric and
sian blue, then it becomes clear that the formetion of ferrocyanide ions. What effect will excess of femc
Prussian blue is slow because the ions required for it ion have upon this equilibrium? It has been calculated
@ractically disafipear when mixed, owing to the redox previously that the equimolar quantities of ferric and
reaction which they undergo instantaneously. Of ferrocyanide ious are reduced to 0.0032 of their original
course, as the small amounts of ferric and ferrocyanide values by the oxidation-reduction reaction. If a
ions remaining combine to form Prussian blue, the similar calculation is made for the action of an excess
redox equilibrium is continuously disturbed and the (say, tenfold) of ferric ion on one mole of ferrocyanide
originally formed ferrous and ferricyanide ions gradu- ion :
ally regenerate the reciprocal pair, thus: Fe+++
10
+ [Fe(CNZIZ stFe++ + [Fe(CN)JE
0 0
Pmsian Blue Start with 1
fast At equilibrium (10-x) (1 - z ) x z
Fe+++ f [Fe(CN).IC s Fe++ + [Fe(CN)alW
Jb slow

FURTHER DEMONSTRATIONS
In the previous demonstration it was shown that
the time required for the appearance of Prussian blue it is found that the excess of ferric ion simply shifts
depends upon the concentration of the reacting ious. the redox equilibrium toward the right and reduces the
It is possible, therefore, by choosing a suitable con- ferrocyanide ion to 0.0000011 of its original value or
centration, to study the redox reaction experimentally to 0.00034 of its value in the presence of an equimolar
before the formation of Prussian blue interferes. For quantity of ferric ion. Actually, the ion product,
this purpose, 0.0001 M solutions, which give a blue [Fe+++] [[Fe(CN)s]'], remains practically constant,
color in a period of time varying from fifteen to thirty and the expected effect of excess of ferric ion on the
minutes, are convenient. Fifty cc. of 0.0001 M ferric formation of Prussian blue is nullfied by the enormous
ion is placed in each of two 150-cc. beakers. To the reduction in ferrocyanide ion which it occasions.
first beaker is now added 50 cc. of distilled water, while Now, excess of ferrocyanide ion should have exactly
to the second is added 50 cc. of 0.0001 M ferrocyanide the same effect upon ferric ion that has just been found
ion. Ferric ion is now tested for in both beakers by for excess of ferric ion on ferrocyanide. In the demou-
adding 5 cc. of M KSCN. A positive test is obtained stration, however, excess of ferrocyanide was found'to
in the first, while the second gives a negative result. be effective in hastening the formation of Prussian
This indicates that ferrocyanide ion reduces ferric blue. How is this fact to be reconciled with what has
ion, thereby causing the practical disappearance of gone before? Calculation leads to the following re-
both of these ions according to the equation given sults.
above. That ferricyanide .is formed is indicated by
Start with
Fet++ +1
[Fe(CN)aSl e Fe++ + [Fe(CN)s]=
10 0 0
the yellow color obtained immediately on mixing the At equilibrium 0.000Wll 0.0000011 0.99999O9 0.9999909
colorless ferric and ferrocyanide ions. The presence
of ferricyanide may be confirmed by testing for it with That is, here the ferric ion is reduced to 0.0000011 of
ferrous ion. Thus, a few cc. of 0.01 M ferrous ion its original value. The ion product [Fe+++] [[Fe
added to the second beaker cause an immediate blue CN)sE]] again remains practically constant. It must
color (Turnbull's blue). now he recalled that any of the four possible oppo-
To test the action of excess of either of the reagents sitely-charged pairs of the four ions involved in the re-
on the formation of Prussian blue, 150 cc. of 0.0001 dox equilibrium may be precipitated. Of these, ferrous
M ferric ion is mixed with 150 cc. of 0.0001 M ferro- ferrocyanide is known to be insoluble. It is well,
cyanide solution and the mixture divided among three therefore, to consider the ion-product [Fe++] [[Fe
(CN)B]'] which prevails after the establishment of turning green and finally blne in about a second, a min-
the redox equilibrium in the various mixtures of ferric ute, and an hour, respectively.
and ferrocyanide ions for wbich calculations have been To study the effect of excess of ferrous on ferricya-
made. nide ions, 150 cc. of 0.0001 M ferrous ion is mixed with
150cc. of 0.0001 Mferricyanide and the mixture divided
among three beakers. The first is kept as a control,
while a few cc. of 0.01 M ferrous ion is added to the
second beaker, and a few cc. of 0.01 N ferricyanide ion
to the third. A blne color appears a t once in the
From this table it is seen that in the presence of ex- second beaker, a green almost as quickly in the third,
cess of ferrocyanide ion the ion product [Fe++] [[Fe while the control requires about fifteen to thirty min-
CN)s]l] is greatly increased, and hence the formation utes to produce a blue color.
of ferrous ferrocyanide is favored. Once formed, the It thus appears that unlike the inhibiting action of
colorless ferrous ferrocyanide may be oxidized to Prus- excess ferric ion on ferrocyanide, excess of ferrous ion
sian blue by the air or by the ferric ferricyanide re- really favors the formation of a blue product from
maining in solution. ferricyanide ion.*
A FINAL EXPLANATION
To account for the foregoing demonstrations it is
Thus far it has been demonstrated that an imme- necessary to return once more to the calculation of
diate redox equilibrium is established when ferric and equilibrium conditions. If an excess, say, tenfold, of
ferrocyanide ions are mixed, resulting in their nearly ferrous ion to one mole of ferricyanide ion is employed,
complete conversion to ferrous and ferricyanide ions, the following result is obtained.
as a consequence of which the assumed formation of
Prussian blue from the original pair of ions is slow. Start with
Fe+++
0
+ [Fe(CN)slS
0
=== Fe++ + IFe(CN)sl"
10 1
Now i t is a criterion of true equilibrium that the same At equilibrium 0.010 0.010 9.990 0.990
final state may be approached from both sides of the If these figures are compared with the previous cal-
reversible reaction. It may, therefore, be anticipated culations, it will be seen that excess of ferrous ion, while
that if equimolar quantities of ferrous and ferricyanide having only a slight effect on the ferricyanide ion,
ions are mixed they will react to a slight extent to form causes a tripling of both the ferric ion and the ferro-
ferric and ferrocyanide ions in order to satisfy the cyanide ion. (Contrast this with the effect of excess
equilibrium constant, and the same equilibrium con- of ferric ion on ferrocyanide ion where the gain in
centrations of the four ions involved will result as ferric ion is compensated by a reciprocal loss in ferro-
when equimolar quantities of ferric and ferrocyanide cyanide.) Hence, the accelerating action of excess
ions are employed as reagents. Hence, whatever of ferrous ion on ferricyanide.
blue product is formed in the first case should be formed Similar figures are obtained for the effect of excess
here also and a t the same rate. That the product of ferricyanide ion on ferrous ion; for example:
obtained from equimolar quantities of ferrous and ferri-
cyanide ions is the same as that from equimolar quan- +
~ e + + + ~[Fe(CNs]'=Fef+ + [Fe(CN.I)'
Start with 0 0 1 10
tities of ferric and ferrocyanide ions was considered by At equilibrium 0.010 0.010 0.990 9.990
Skraup as early as 18774 and was established by po-
tentiometric methods by Miiller and Lauterbach in The green color obtained in this case is probably the
1922.6 Vorlander, on the other hand, claimed that mixed color of the soluble Prussian blue and the yellow
unlike the formation of a blue from ferric and ferro- ferricyanide which is present in excess.
cyanide salts, wbich is slow, the production of a blue AN APPLICATION TO QUALITATIVE ANALYSIS
from ferrous and ferricyanide ion is instantaneous.
The qualitative test for nitrogen in organic com-
pounds frequently involves the detection of small
SOME FINAL DEMONSTRATIONS amounts of ferrocyanide which is produced by the
Paralleling the previous experiments with mixtures action of ferrous ion on the cyanide resulting from a
of femc and ferrocyanide ions, 50-cc. portions of sodium fusion. It is, therefore, desirable to take
0.0006, 0.0002, and 0.00007 formal solutions of K4Fe account of the ineffectiveness or even the hindrance
(CN)B (in water) are mixed with like concentrations of excess of ferric ion in producing Prussian blue.
of Mohr's salt, Fe(NHa)2(S04)2.6Hz0 (in tenth- During the "nitrogen test" considerable ferric ion is
formal KHSO*). The results are indistinguishable produced by the aerial oxidation of ferrous hydroxide.
from those obtained with the reciprocal pair of ions. Two methods of obviating.this disadvantaxe - have been
the originally yellow (due to ferricyanide) mixtures * I t is strictly this difference in the effect of excess of ferrous
- and ferric ions on the corresponding complex ions which Vor-
Z. A. SKRAW,
Ann., 186, 371 (1877). IXnder (loc. cit.) observed when he wrote that one reaction (ferric
WE.MULLERAND H. LATJTERBACH, J. prakt. Chem., (2) 104, ion and ferrccvanide) was slow and the other (ferrous ion and
241 (1922). ferricyanide) ihstantaneous.
recommended, both on empirical bases. Vorlinder to prevent the unfavorable shifting of the redox equilib-
suggests filtering the alkaline test mixture before rium (conversion of ferrocyanide to ferricyanide)
acidification, thus separating the iron hydroxides by the ferric ion. It is curious that the lack of an
from the ferrocyanide ion which appears in the filtrate. explanation for the use of fluoride has led one author
The filtrate is then acidified and treated with ferrous to recommend adding both ferric chloride and po-
sulfate, the requisite ferric ion for Prussian blue forma- tassium fluoride to the nitrogen test. Certainly, the
tion being produced by aerial oxidation. Kamm6 rec- common practice of employing ferric chloride in this
ommends the procedure of Viehoever and Johns' who test is both unnecessary and undesirable.
add K F to the test mixture, although no explanation
is offered for its use by these authors. It now appears ACmiOWLEDGMENT
that the function of the fluoride is to "tie up" the The author is indebted to his former students,
excess of ferric ion (produced by aerial oxidation) and Messrs. A. Schwebel and H. Rosenwasser, for checking
the experiments reported in this paper. They are
0. W, "Qualitative organic analysis;' 2nd ed.. John now engaged in studying the behavior of other revers-
Wiley & Sons,New York City, 1932, p. 135. ible redox systems of metallic ions, such as Cu++-
'A. VIEHOEVER AND C . 0. JOHNS, J. Am. Chem. Soc., 37, 601
(1915). Cu+, toward the ferrocyanide-ferricyanide system.