Kov.

, 1929 DICHLOROFLUORESCEIN AS AN ADSORPTION INDICATOR 3273

[CONTRIBUTION FROM T H E SCHOOL OF C H E M I S T R Y O F T H E UWIVERSITY OF M I N N E S O T A ]

THE USE OF DICHLOROFLUORESCEIN AS AN ADSORPTION

INDICATOR FOR THE ARGENTOMETRIC TITRATION OF
CHLORIDES
W. M. LAUER
BY I. M. KOLTHOFF, AND C. J. SUNDE
RECEIVED
JUNE 10, 1929 PUBLISHED
XOVEMBER
8, 1929

Introduction
K. Fajans and his collaborators’ have recently described a new type
of indicator for the argentometric titration of halides to which the name
of adsorption indicator was given. ’She substances used as indicators
are not sensitive to dissolved silver or halide ions, but change their color
a t the equivalence point where colored compounds are formed on the
precipitate due to adsorption phenomena. Other investigators* have
confirmed the results of Fajans and his collaborators and obtained accurate
results with the new type of indicator. Fluorescein has been used as the
indicator in the titration of chlorides but its application is limited. Good
results are obtained only if the chloride concentration is greater than
0.005 N and if the solution has a neutral or weakly alkaline reaction.
Other indicators which were applicable in weakly acid solutions have
been found, but the color changes are less distinct than is the case with
fluorescein.
When chloride ions are titrated with silver nitrate, the silver chloride
adsorbs silver ions a t the equivalence point or better slightly beyond.
The “silver body” then adsorbs fluorescein ions from the solution and
dark red silver fluoresceinate is formed on the surface of the precipitate.
Fluorescein possesses a weak acidic character; in weak acid solution the
concentration of fluorescein ions is so small that the color change no longer
takes place. The introduction of halogen atoms into the fluorescein mole-
cule has two effects: the dissociation constant is increased and the ad-
sorbability of the dye anions by the silver halide increases. It would
therefore be expected that substances with a stronger acid character than
fluorescein could be used as indicators in weak acid solutions. On the
other hand, the stronger adsorbability might either interfere or be favor-
able. During the titration of chloride ions with silver nitrate the silver
chloride formed a t the beginning of the titration adsorbs chloride ions. I n
the presence of the anions of the indicator, the latter also have a tendency
to be adsorbed and if their adsorption is larger than that of the chloride
ions, the color change will take place during the early stages of the titra-
(a) Fajans and Hassel, Z. Elektrochem., 29, 495 (1923); (b) Fajans and Wolff,
Z . anorg. allgem. Chem., 137,221 (1924).
* W. Bottger and K. 0. Schmidt, ibid., 137, 246 (1924); I. M. Kolthoff and I,. H.
van Berk, Z . anal. Chem., 70,369 (1927).
a Kolthoff, ibid.,71, 235 (1927).
3274 I. M. KOLTHOFF, W. M. LAUZR AND C. J. SUNDE: Vol. 51
tion. Such is the case when eosin is used. Therefore this indicator can-
not be used in the argentometric determination of chlorides, even though
i t gives highly accurate results when bromides, iodides and thiocyanates
are titrated even in weak acid solutions. On the other hand, stronger
adsorbability may exert a favorable influence in that it might make pos-
sible the titration of very dilute chloride solutions. I n this case, the ab-
sorption of fluorescein ions is too small and no color change is observed.
The indicator in order to be better than fluorescein must have a larger
dissociation constant and greater adsorbability than fluorescein. The
adsorbability, however, must not be so great that it replaces the chloride
ions from silver chloride. It was expected that dichlorofluorescein would
serve the purpose, and as is shown in the following part of this paper it
proved to be an excellent indicator for the titration of chlorides under
different conditions. That dichlorofluorescein is a stronger acid than
fluorescein can easily be shown by the fact that its color change interval
from green-yellow (without fluorescence) to a strongly fluorescent yellow-
green lies between I“ 3 to 5; so that the dissociation constant is of the
order of
Dich1orofiuorescein.-Dichlorofluorescein was prepared by condensing
monochlororesorcinol with phthalic anhydride. It was purified by con-
verting it into the diacetate and then saponifying the diacetate with alco-

Ho-o\:,o
c1-
/’ \
--cl
holic potassium hydroxide, in accordance with the
-OH procedure used by Orndorfl and his collaborator^.^
Dichlorofluorescein prepared from dichlorophthalic
anhydride and resorcinol has been known for some
time, but dichlorofluorescein containing two chlo-

‘p C-0 rine atoms in the “resorcinol part” of the molecule

has not been described hitherto. The most prob-
able formula for the dichlorofluorescein here de-
scribed is that indicated. All attempts to prepare a dichlorofluorescein by
the direct chlorination of fluorescein failed.
Experimental
The Preparation of Dic3lorofluorescein
Monochlororesorcinol was first prepared by the action of sulfuryl chloride on resor-
cinol: 100 g. of resorcinol was dissolved in 360 g. of absolute ether. This solution, placed
in a 2-liter round-bottomed flask fitted with an efficient reflux condenser, was then treated
with 120 g. of sulfuryl chloride (added dropwise). This operation was carried out in a
well-ventilated hood. After the addition of the sulfuryl chloride the ether was distilled
off on a steam-bath. After the low-boiling material had been removed completely, the
residue was subjected to vacuum distillation. Since monochlororesorcinol has a lower
boiling point than resorcinol, the unchanged resorcinol was separated by vacuum frac-
4 Orndorff and Hemmer, THIS
JOURNAL, 49, 1275 (1927), and various other studies
by Orndorff and co-workers.
Nov., 1929 DICHLOROFLUORESCEIN AS AN ADSORPTION INDICATOR 3275
tional distillation. The monochlororesorcinol distilled at 131O under a pressure of 6-7
mm. It was analyzed for chlorine by the method of Carius.
Anal. Calcd.: C1,24.54. Found: C1,24.50, 24.72.
Dichlorofluorescein.-Ten grams of phthalic anhydride mixed intimately with
19.7 g. of monochlororesorcinol was heated to 160' in an iron crucible by means of an
oil-bath. Five grams of anhydrous zinc chloride, finely ground, was then added with
stirring during a period of about ten minutes. The temperature was then raised to
175-180O and maintained for about two and one-half hours. After cooling, the reaction
product was ground and then boiled for fifteen to twenty minutes in dilute hydrochloric
acid (10 cc. of concd. acid to 200 cc. of water) and filtered.
The precipitate was dissolved a t room temperature in a saturated aqueous solution
of sodium bicarbonate, filtered, heated t o boiling and then acidified with hydrochloric
acid (approx. 4 N). After cooling and filtering, the precipitate was refluxed with acetic
anhydride for several hours (2 g. of crude dichlorofluorescein t o 10-12 cc. of acetic an-
hydride). The diacetate crystallized out upon cooling. It was filtered off and the above
process repeated four or five times. The diacetate (almost colorless) was then digested
with a small amount of alcohol, cooled and filtered. After drying for several hours
at 100' it was analyzed for chlorine.
Anal. (Carius). Calcd.: C1, 14.64. Found: C1, 14.49, 14.98.
The diacetate was then saponified by refluxing 2-3 g. with alcoholic potassium
hydroxide (6-7 g. of base in 40 cc. of alcohol) for two hours. Water was then added,
the solution transferred t o a distilling flask and the alcohol removed by distillation.
The solution was then filtered (if necessary), heated to boiling and acidified with hydro-
chloric acid. An ochre-colored product was obtained.
Application to Halide Titrations
Indicator Solution.-One-tenth gram of the indicator was dissolved in 60 cc. of
alcohol, 2.5 cc. of 0.1 N sodium hydroxide was added and the volume of the solution
was then made up t o 100 cc. with distilled water. An aqueous alcoholic solution (60-
70%) of approximately the same concentration of unneutralized indicator may also be
used.
Procedure.-For every 10 cc. of chloride solution 2 drops of indicator are added
and the titration is started with silver nitrate. The silver chloride remains in colloidal
solution which, near the end-point, has a yellowish-orange appearance. About 0.5%
before the equivalence point the silver chloride flocculates in voluminous flocks; the
titration is then conducted carefully. At the end-point the silver chloride suddenly
turns dark red.
Note 1 .-Concentrated chloride solutions should be diluted to a salt-concentration
of about 0.025 N . In this case the color change is extremely sharp. In six titrations
100 cc. of 0.025 N sodium chloride required 25.00 t o 25.03 cc. of 0.1 N silver nitrate.
Note 2.-Even very dilute chloride solutions can be determined satisfactorily.
I n these cases it is recommended not to add more than 2 t o 4 drops of indicator to 50
cc. of solution.
A 0.01 N chloride solution can be titrated with an accuracy of 0.5%. Near the
end-point the liquid appears more brownish; the titration is finished when the color
changes to distinct orange. One or two drops of 0.01 N silver nitrate in excess deepens
the color t o pink or reddish.
If the end-point is passed the excess of silver nitrate can be titrated back with a
chloride solution of known constant. This reverse titration should be made immediately
after the addition of silver nitrate. After longer standing the adsorbed indicator reacts
slowly.
3276 I. M. KOLTHOFF, W. M. LAUER AND C. J. SUNDE Vol. 51

The following data show that the method gives satisfactory results.
TABLEI
TITRATION OF DILUTE CHLORIDE S01,UTIONS
Sodium chloride, cc. 10 of 0.01, N 50 of 0.002 N 100 of 0.001 N
0.01 N AgN03, CC. 10.02 t o 10.05" 10.05 t o 10.15 10.10 t o 10.15*
0.01 N AgN03 calcd., cc. 10.00 10.00 10.00
a End-point of an orange color which turned pink after addition of 10.07 t o 10.09
cc. of 0.01 N silver nitrate solution.
With a larger excess of silver nitrate the color became more intense.
Especially in these very dilute solutions the back titration with chloride
gives a sharp end-point.
Solutions containing as little as 15 to 20 mg. of chloride in a liter can
be titrated with an accuracy of 1 to 2%. Therefore the method is superior
t o the ordinary Mohr procedure, as the latter has an appreciable titration
error (at those dilutions). It might be expected that the indicator can
be used advantageously in the titration of ,chloride in tap water. Special
tests were run in order t o see if calcium bicarbonate, magnesium bicar-
bonate and other salts had an interfering action. Though larger amounts
of salts, especially those containing higher valence cations affect the color
change a little, the results were still accurate to about 2% in 0.002 N chlo-
ride solution.
The tap water of Minneapolis contains only 2.6 mg. of chloride in a liter,
which is too small to be titrated. Moreover, i t seems that the tap water
of this city contains some interfering substance. If a 0.001 N chloride
solution prepared in tap water was titrated, no color change a t all was
observed. With artificial water of the same chemical composition as the
tap water of Minneapolis and containing 1 milliequivalent of chloride in
a liter a good end-point was observed. Even in the titration of 0.01 N
chloride solution the tap water had a distinct retarding action on the for-
mation of the color change. I n all these cases it was observed that in
the presence of tap water the silver chloride did not flocculate a t all or
very slowly with an excess of silver nitrate. Therefore i t was thought
that protecting colloids in the tap water might interfere with the obser-
vation of the end-point. The influence of different hydrophyllic colloids
was tested. Some of them, especially gelatin even in small concentra-
tion, make the color change much less distinct and the end-point appears
late. The presence of traces of colloidal material may explain the pe-
culiar behavior of the Minneapolis tap water. On the other hand, i t was
shown that a dilute solution of silicon dioxide (from sodium silicate and
acid) did not affect the color change of the indicator.
Another advantage of the new indicator over potassium chromate is
that the titration can be carried out in weakly acid solution. If the acidity
is regulated by some buffer in such a way that the PHis larger than about
NOV., 1929 DICHLOROFLUORESCEIN AS AN ADSORPTION INDICATOR 3277
4.0, the color change is still very distinct. In the titration of 0.025 N
sodium chloride solution in the presence of an acetate buffer with a PH
equal to or greater than about 4.4,the first color change to orange occurred
after overstepping the end-point 0.1 to 0.2%. An advantageous applica-
tion of the method is a direct chloride titration of those salts of heavy
metals which are slightly hydrolyzed. Mohr's method gives unsatisfac-
tory results in the above cases. In the titration of 0.025 N (to chloride)
solutions of copper, zinc, manganese, nickel, aluminum, calcium, strontium,
barium and magnesium chloride, the end-point was very sharp and ap-
peared a t the right place. In most cases the silver chloride flocculated
50 to 20% before the end-point was reached. Near the equivalence point
the precipitate was slightly colored. At the equivalence point the sudden
increase of the color to intense red could be observed with. an accuracy
of 0.05 cc. of 0.1 N silver nitrate. In the titration of very dilute solutions
it is better to work a t neutral reaction.
Bromides and iodides can be titrated in the same way as chlorides, even
in very dilute solutions. In the application of fluorescein as an indicator,
K. Fajans and H. WOE' found that a mixture of iodide and chloride re-
quired an excess of silver nitrate, the error being +1 to 1.5%. There-
fore, the use of the dichlorofluorescein for this purpose was tested. Dur-
ing the titration the same phenomena are observed that Fajans and W O E
described; about 2% before the end-point the silver iodide and chloride
flocculate, the precipitate has a yellowish-brown color which turns brown-
orange at the equivalence point and changes to orange-pink with an ex-
cess of about 0.5% silver nitrate. Though the color change is not as sharp
as in the titration of both halides separately, the end-point still can be
detected with an accuracy of 0.5%, if the mixture is so far diluted that its
concentration is no more than 0.01 N to halide. The change in color in
this case is much sharper than with fluorescein as an indicator.
Summary
Dichlorofluorescein is a very suitable indicator for the argentometric
titration of chlorides in very dilute as well as in weakly acid solutions.
MINNEAPOLIS,
MINNESOTA