Research Laboratories of the

Portland Cement Association

33 West Grand Ave., Chicago

A STUDY OF VARIOUS VOLUMETRIC METHODS FOR

DETERMINING THE FERRIC OXIDE CONTENT OF PORTLAND CEMENT

By Lorrayne LeMar

LIBRARY
RESEARCH AND DEVELOPMENT LABORATORIES
PORTLAND CEMENT ASSOCIATION
********

June, 1947
 TABLE OF CONTEMS 1

Page
ABSTRA.CT •••••••••••••••••••••••••••••• " • ., •.••••••". f ,"• • • • • • • • ' . 1

INTRODUCTION ••• "•••••••••••••••• ,•••••••.•••••• ,•••••••••••• ' . 2

LABORATORY PROCEDUBES:
1. Standardization of the Oxidimetric Solutions ••••••••••• ,••••• 2
2. Determination of the Ferric Oxide Content of Portland Cement. 3
DISCUSSION:
1. Standardization of the Potassium Dichromate Solution••••••••• 3
(a) With Ferrous Ammonium Sulfate ••••••••••••••••••••••••• 3
(b) With Iron Ore '...••..••• ' . 4
(c) Wi th Iron Wire ••••.••..••.••.••·••.•..••••• i •••••••• 6
(d) Sl1IIIl'l1ary of Beaul ts .••.••.••.•••.•...•. 'f • • • • • • • • • ' , • • • • •
7
(e) Methods of Determining the Endpoints of the Titrations. 8
2. Standardiza.tion of the Ceric Sulfate Solution . 9
(a) With Ferrous Ammonium Sulfate •••••••••••.•••••••••••• 9
(b) With Iron Ore , ••••..•••••••• 9
(9) With Iron Wire ••••••••••••••.••••.•••••••••••••• ·... It • • •
9
(d) With Sodium Oxalate .••..•••••••••••.•••••••••••.•• '• •• 10
(1) Potentiometric Method Using Iodine Monochloride. 10
(2) Indicator Method Using Ferrous Ammonium Sulfate-. 10
(e) Summa.ry of Beaults •.•.•••.••••..•.•.••.• , .••.•..•.••• 11
(f) Methods of Determining the Endpoints of the Titrations 11
3. Determination of the Ferric Oxide Content of Portland Cement. 12
(a) With Potassium Dichromate .••••.•••••••••••••••.•••••• 12
(b) With Ceric Sulfate •••.••••••••••••••••••••••••••••••• 14
4. Comparison of the Two Oxidimetric Solutions •.••••••• ~ •••.••• 14
(a ) Preparation " to • • • • 14
(b) .Stabili ty . 14
(c) Use with Hydrochloric Acid ••.••••••••••••••••••.••••• 15
(d) Standardization . 15
(e ) Indicators 0............."..... _ ....... ... 16
( f ) SlllDlD.ary ....................................."................ 17
5. Results in Tables 1 to 6 ......•... '. .•................•...•.. 18
SUMMARY: Advantages and Disadvantages of:
1'. Ceric 'Sulfate •••••••••••••••••.••••••••.•••••••••••••••••••• 20-21
2. Potassium DichrolJUlte it
21 .

3. Iron Ore ,. 0' . 22

4. Iron Wire - . 23
5. Diphenylamine Barium Sulfonate •••.•••••••••••.•••••.•..•.••• 23
6. N-methyldipheny1amine Barium Sulfonate •••••••••••••••••••••• 24
RECOMMENDATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .0 •• '.' , .
25
ACKNOWLEDGEMEl\1T ••• 0 •••.••••••••••••••••••••••••••••• " '.· • • • •
25
........................ " ". " " -" " . 26-27
 APPENDIX

REAGENTS:
1. Standard Oxidimetric Solutions ••.•.•••••••••.•••••••••••••••• 1
(a) Potassium Dichromate ••••..•••••.••...•••••.•••••••••••• 1
(b) Cerie Sulfate , II' • 'I' •••••••••••••• 1
2 • Primary Stan.dards -. . . . . . . . . . . . . ". . . . . II 1
(a) Sodium Oxa.late ., . 1
(b) Potassium Dichromate .... ........ .- . 1
( c) Iron Wire ' ~ , 1
(d) Iron Ore ,', '. 1
(e) Ferrous Ammonium Sulfate Solution •••.••..••.••••••••••• 2
;. Indicators ,' .. I •••••••••••••••••• 2
(a) Ferroin ~ II • " ••••••• 2
(b) Diphenylamine Barium Sulfonate ••..•.•..••••..•••••••.•. 2
(c) N-methyldiphenylamine Barium Sulfonate .•••.•••••.•••••• 2
4. MiscellaneouB Reagents '" ' ,. , . 3
(.a) Normal Sulfuric Acid It .
3
(b) Normal Hydrochlorio Acid .•••••.•..•••••.••••......••..• 3
(c) Phosphoric Acid Solution ••••••••.••..•..•••.•••..•..••• 3
(d) Stannous Chloride Solution ••••.•.•.•••..••••.•••..•.... 4
(e) Iodine Monochloride Solution ••••.•.•••••••••••.•..•..•• 4

APPARATUS:
1. Potentiometer ". It . 4
2.. Electrodes " " It . 4
3. Silver Reductor e II ••••
5

METHODS:
1. Standardization of the Ferrous Ammonium Sulfate Solution 5
2. Standardization of the Potassium Dichromate Solution •••••••• 6
(a) With Ferrous Ammonium Sulfate •••••••.••••.•••••••••••• 6
(b) With Iron Ore ., I • • • • • • • • • • 7
(0) With Iron Wire "e.• ,-. e" e.... It , lit 8
3. Standardization of the Ceric Sulfate Solution ••••••••••••••• 9
(a) Wi~h Ferrous Ammonium Sulfate •••••••••••••••••••••.••• 9
(b) With Iron Ore eo.......... 10
(e) With Iron Wire It • • '.......... •••••••••••••••••••••• 10
(d) With Sodium Oxalate Plus Iodine Monochloride •••••••••• 10
(e) With Sodium Oxalate Plus Ferrous Ammonium S.ulfate ••••• 10
4. Determination of the Ferric Oxide Content of Portland Cement. 11
(a) Titration of Cements with Potassium Dichromate ••••••• , 11
(b) Titration of Cements with Ceric Sulfa~e •••••••••••• ~.. 12
 Research laboratories of the
Portland Cement Association
33 West Grand Ave., Chicago

A STUDY OF VARIOUS VOLUMETRIC MEYrHODS FOR

DETElMINING THE FERRIC OXIDE CONTENT OF PORTLAND CEMENT

By Lorrayne LeMar*

-000-

ABSTRACT
Determinations were made on the ferric oxide content of port-
land cements with potassium dichromate and the indicators diphenylamine
barium sulfonate and N-methyldiphenylamine barium sulfonate as com-
pared with similar determinations made with ceric sulfate.and the
indicator ferroin. Results obtained with the indicators were checked
against potentiometric titrations • A s ilver reductor and the stannous
chloride method were used for reduc ing the iron. Also various primary
standards for determining the Fe20 s factor of the OXidizing solution
were compared. .
It was c9ncluded that aolution of potaseium dichromate is the
most suitable oxidizing agent for routine analYsis of the Fe20s con-
tent of portland cement because of' its stability, low equivalent
weight and low cost of' the salt, and ease of preparation of the
solution. Moreover, it can be used in a hydrochloric acid medium
without adverse side reactions.
. .
Iron wire was found to be the most suitable primary standard
for determining the Fe2 0 s factor of the oxidizing solution. Iron ore,
because of its slow solubility, its lack of homogeneity, and the
difficulty in handling the weighed sample of ore quantitatively, was
found to be less desirable than iron wire as a primary standard for
routine analyses.
A more reliable factor for the standard solution was obtained
by averaging the results of four or five determinations of the factor
(against 0.030 to 0.035g. Fe) than by using various factors determined
by One or two titrations whose volume approximated the volume required
by the cement. The reason is that individ\1al determinations of the
factor of the standard solution may vary considerably even thOUgh
titrated against the same weight of Fe, and theref'ore, single or even
duplicate determinations of the factor may not be accurate.
Certain changes are recommend,ed in the current A.S.T.M.
procedure (C 114-46) for determining the ferric oxide content of
portland cement.
* Associate Chemist, Research Laboratories, Portland Cement Assn., Chicago.
 INTRODUCTION

!ho voluaetr1c determination of iron involves the solution ot

\he 8aJ1ple, the conversion of the iron to the ferrous state, and the
et1Wltlt.at1ve addition of a standard oxid1metr1e solution until the
fenoWi iron is converted to the ferric state. The endpoint ot the
reactloft 1Da7 be determined with various indioators or it 1IB.1 be deter-
ained electrometrically. The percent of iron is calculated trOJll the
1e203 equivalent ot the oxidizing solutlon.The current A.S.T.M.
method for portland cement (0 114-46) uses potassium dichromate as
oxidant and d1pl:1enrlam:l.ne as indicator.
A number of factors may influence the accuracy ot the
volumetric determinat10n of iron, namely, the OXidizing solution
used, the method of standardizing the solution, the method of reducing
the iron, and the method of determining the endpoint. The purpose
of the investigation described in this Paper was to study the various
factors which influence the determination of the ferric oxide content
of portland cement and to compare results obtained by various methods.

IA:BORATORY PROCEDURES
1. Standardization of the Oxid1metric Solutions. In this
investigation two oxidtmetric solutions, potassium dichromate and
eerie sUlfate were used for determining the ferric oxide content of
portland cement. Potassium permanganate was not included in the study
inasmuch as it is a well-established fact that permangenate no~ly
gives high values in iron determinations tl ).

The solution of potassium dicbromate was standardized

against three primary standards:
(1) Solution of ferrous ammonium sUlfate,
(2) Iron ore,
(3) Iron wire.

The iron ore was reduced by a modified stannous chloride procedurej

the iron wire was also reduced by this method, end by passing the
solution ot the iron wire throUSl1 a silver reductor column. The end-
point of the titrations was determined with the 1ndicat~rs diphenylamine
barium sUlfonate and N-methyldiphenylam1ne barium sulfonate, and elec"
tl'ometrica.l1y with a Beckman Model G potentiometer..
The solution of cerie sulfate was standardized against four
primary standards:
(1) Solution of ferrous ammonium sulfate,
(2) Iron ore,
(3) Iron Wire,
(4) Sodium oxalate.
The iron ore was reduced by a modified stannous ehlo:ride procedure; tlle

*2*
iron wire was reduced by passing the solution through a silver reductor
column. The endpoint of the ceric sulfate titrations wa.s determined
with the indicator ferroin and electrometrically with the potentiometer.
The reagents, apparatus, and details of the preparation and"standardiza-
tion of the oxidimetric solutions used in the investigation are des-
cribed in the appendix of this report.
2. Determination of the Ferric Oxide Content of.Portland
C·ement.. The standard solutions of potassium dichromate and ceric
sulfate were used for determining the ferric oxide content of eight
cements ,ranging from 0.4% to 5.4% Fe2 0 3 • The same afore"mentioned
indicators and the potentiometer were used for determining the endpoint
of the titrations. In the potentiometric titration, the cements were
reduced with stannous chloride and by passing the solution through a
silver reductor. In the indicator method the cements were reduced by
the stannous chloride procedure only. Details of the procedures
followed are given in the appendix of this report.

DISCUSSION
1. Standardization of the Potassium Dichromate Solution.
(Results reported in Table 2). A change of primary standard for
determining the Fe2 0 3 factor for the dichromate solution is being
contemplated by Committee C-l of the A.S.T.M. For this reason, the
results obtained using the various primary standards are of particular
interest. The following three primary standards for the dichromate
solution were studied:
(a) Ferrous annnonium sulfate (MOhr's salt) which
is frequently employed in research work in this
country and as a comon primary standard in
Europe (particularly in Germany) for the
standardization of oxidimetric solutions;
(b) Sibley iron ore, standard sample No. 270 of the
National Bureau of Standards, which is the new
standard proposed by Committee C-l of the
A.S.T.M., and
(c) Iron wire which is the present standard specified
by the A.S.T .M.
. (a) Standardization.of Potassium Dichromate with Ferrous
Ammonium Sulfate. Mohr's salt is subject to variations in percent
of Fe due to changes in water content and to oxidation of the salt by
the air. ConsequentlY, a solution of the salt was prepared with a high
percentage of sulfuric acid and stored in a tightly stoppered brown
bottle to mitigate OXidation of the ferrous iron. Details of the pre-
paration and standardization of the ferrous ammonium sulfate solution
are given in the appendix; results are reported in Table 1.

* 3*
 Very sharp, clear-cut titrations were obtained with
either of the indicators. The diphenylamine barium. sulfonate changed
to a brilliant purple color and the N-methyldiphenylamine barium. su!..
fonate gave a rose-red color at the endpoint. The unusual brilliance
of the indicator color changes was a,ttribu..ted to the fact that no pre-
cipitate was present in this titration*.
Results using the potentiometer were also satisfactory
and checked the ~esults obtained by t~e indicator method. Here again
the absence of a precipitate in the solution was an advantage in that
practically no "poisoning" of the platinum. electrodes was noted.
However, the ferrous ammonium. sulfate solution was found
to have one decided disadvantage which limits its value as a practical
primary standard. Repeated standardization of the solution showed that
the ferrous salt was oxidized during storage in spite of the afore-
mentioned precautions taken to prevent such oxidation. The follOWing
data give the magnitude of the deterioration of the solution: The
initial normality found (potentiometrically) for the ferrous ammonium
sulfate solution was 0.05047. Two months later when the solution was
again standardized by the same procedure, the average normality was
found to be 0.04605. However, this deterioration of the solution in
no way affected the results reported in Tables 2 and 3 inasmuch as
little time elapsed between the initial standardization of the solution
and final use of it as a primary standard**.
(b) Standardization of Potassium. Dichromate with Iron Ore.
The potassium dichromate solution was also standardized against the
National Bureau of Standards' Sibley iron ore, No. 27b. The details
of the procedure are given in the appendix of this report.
Several difficulties were encountered in using iron ore
as a primary standard. When the sample was weighed in an ordinary
aluminum weighing pan,it was found to be extremely difficult to com-
pletely transfer the sample ·of ore to a beaker. Even after prolonged
brushing of the sample from the pan, some of the ore remained as evidenced
by the discoloration of a white towel used to Wipe the pan. For this
reason a glass weighing dish counterpoised by a similar glass dish was
used for weighing the samples of iron ore. Even using the glass dish,
great care had to be exercised in brushing the sample into the beaker to
prevent the ore from clinging to the edge of the weighing dish. The
following data illustrate that small differences in weight of the primary
standard or of the volume of the solution used to titrate the standard
greatly affect the numerical value for the Fe::ps factor of the solution:

* Theprec1pitate which is usually present in iron titrations is formed

by the reaction between stannous chloride used to reduce the iron and
mercuric chloride used to eliminate the excess stannous chloride.
Since the iron of Mohr's salt is already in the ferrous state" re-
duction with stannous chloride is not necessary.
** The ferrous ammonium. sulfate solution was prepared March 27, 1946" :first
standardiz'8d April 5, 1946, and "all subs"equent tests with this "
solution as a primary standard completed by April 11, 1946.

*4 *
 Primary Standard Vol. of Oxidizing Calculated
as g. Fe Solution Fe20s Factor
0.0100 ;.62 0.003962
0.0100 ;.57 0.004017
0.0393 13·90 0.004037
0.0392 13.90 0.004027
0.0400 14.15 0.004040
0.0400 14.10 0.004054

Complete solution of the ore presented another problem.

Overnight digestion of the ore in 25 IDl. of concentrated hydrochloric
acid at a temperature of about 60°0. proved to be the most satisfactory
method. The ore was digested in a covered beaker of 250 ml. capacity.
Other volumes of concentrated hydrochloric acid (10, 15 and 30 IDl.)
were also tried, but the use of 25 ml. led to the most satisfactory
results inasmuch as 10 tc) 12 ml. of hydrochloric acid were lost
during the digestion period.
Since it is difficult to completely dissolve iron ore,
some investigators have suggested the use of stannous chloride as an aid
in the dissolution of the ore. One such method involves treating the
sample of ore with hydrochloric acid, boiling for approximately thirty
minutes and adding a small excess of stannous chloride. Boiling is
continued and more stannous chloride is added if necessary until the
ore is completely dissolved. The excess stannous chloride is oxidized
with bromine water. After the bromine water is .expelled, the sample
is reduced and titrated in the usual manner. Another method utilizes
the same principle, except. that insufficient stannous chloride is
added to completely decolorize the iron solution. Since there is no
excess stannous chloride at this stage, there is no need to use bromine
water. In this method as in the previous one, after the ore is in
solution, it is reduced and titrated in the usual way.
The method of using stannous chloride to dissolve the ore
and bromine water to destroy the excess stannous chloride was tried in
this investigation. However, the resultant mixture so "poisonedll the
platinum electrodes that a satisfactory potentiometric titration was
almost impossible. The indicator method was also used after the
stannous chloride dissolution of the ore (Without bromine added) but
here again results were erratio.
After the ore was completely dissolved by the afore-
mentioned overnight digestion procedure, the iron was reduced from.
the ferric to the ferrous state. This was accomplished by the drop-
wise addition of a 5% stannous chloride solution until the iron
solution was decolorized*. Usually one drop in excess is added but

* As will be discussed later, a modified (not A.S.T.M.) stannous

chloride procedure for reducing the iron was followed in this investi-
gation.
*5*
with the ore two drops of stannous chloride solution were required
after the iron solution was deco10rized because occasionally no pre-
cipitate formed upon the later addition of· the mercuric chloride when
only one drop in excess was added. After the excess stannous chloride
was oxidized by the addition of mercuric chloride, the mixture was
usual1y* stirred vigorously for one minute with a mechanical stirrer to
allow the reaction to go to completion. Tests of stirring the mixture
'longer than one minute gave results of no greater reproducibility.
After the samples of ore were dissolved and reduced with
stannous chloride (the excess being oxidized with mercuric chloride),
the solutions were titrated with potassium dichromate. Erratic results
were sometimes obtained regardless of whether an indicator or the
potentiometer was used to determine the endpoint. These erratic re-
sultswere attributed chiefly to failure to get the ore completely in
solution, although other factors Buch as the nature of the ore and the
large quantity of stannous chloride r~quired to reduce the ore no doubt .
played a role. As shown previOusly(2), stannic chloride which arises from·
the OXidation of stannous chloride exerts a deleterious effect on the
titration of iron with an OXidizing solution. Since this effect is
exhibited in the potentiometric method as well as the indicator method
of titration, it is believed that the stannic chloride represses the
oxidative power of the dichromate. solution~
Because of the' analytical difficulties involved in the
use of iron ore, it is less desirable as a practical prima~Y standard
for oxidimetrie solutions than ferrous ammonium sulfate • Neither
proved to be as satisfactory as iron wire, which is discussed next.
(c) Standardization of Potassium Dichromate with Iron Wire.
Iron wire (Central Scientific Co·. Assay 99.89% Fe) was the third
primary standard used to determine the OXidizing power of the potassium
dichromate solution. The procedure followed with the wire was similar
to that used with the ore, except that in addition to the stannous
chloride method of reduction, samples of the wire were also reduced in
a silver reductor column. Specific details are given in the appendix.
In contrast to the iron ore, the quantitative transference
of the weighed sample of wire to a beaker was an extremeiy simple
process requiring only the use ofa pair of balance forceps.
Solution of the wire took place rapidly in dilute hydro-
chloric acid. If desired, solution of the sample could be effected
in approximately five minutes by warming the hydrochloric acid mixture
on a hot plate.
After the wire was dissolved, it was reduced by One of two
methods:
(1) :By passing the solution through a silver
reductor column,
(2) :By using a modified stannous chloride procedure.
* See appendix for the exact procedure followed.
*6*
 Reduction of the Wire in eo Silver Reductor Column. The
silver reductor was used in the hope of eliminating the disadvantages
of the stannous chloride method of reduction. As previously reported(2)
both stannic chloride and mercuric chloride have a repressing effect on
the reaction between iron and the OXidizing solution. Moreover, it has
been shown by other investigators(3)(4) that these salts have an in-
hibiting effect on the color formation of the indicators, some being
much more sensitive than others.
" When the samples of wire were titrated with dichromate
after reduction in the silver column, the titration resembled that of
the ferrous ammonium sulfate. The indicator endpoints were sha'l;P and
clear-cut. In the potentiometric titration there was little evidence
of contamination of the electrodes. Results of duplicate determinations
were markedly consistent.
Reduction of the Wire with Stannous Chloride. In the usual
stannous chloride method of reducing the iron, much more mercuric
chloride is added than is necessary to oxidize the excess stannous
chloride; consequently, it is possible for the remaining mercuric
chloride to compete with the oxidizing solution in reacting with the
iron. This interference by the mercuric chloride may be explained by
the equation:
Hg012 + FeCl2 -.... HgCl + FeCls

For this reason, two alterations in the current A.S.T.M. method (C 114-46)
of reducing the iron with stannous chloride were made. First, a 5~
instead of a lCf/o solution of stannous chloride was used, thus limiting
the amount of excess stannous chloride that would be added to the iron
solution. Second, the amount of saturated solution of mercuric chloride
added was changed from 10 to 5 ml. The mercuric chloride was added
after cooling the solution" of the sample and before diluting the sample
to the reqUired volume. Under these conditions,· the results obtained
after the stannous chloride reduction compared very favorably with those
obtained with the silver reductor. (Table 2). This eliminates the
necessity of setting up and maintaining a silver reductor column which
is less convenient to use than the stannous chloride method of reduction.
(d) Summary of Resu!ts • Iron wire proved to be the most
practical primary standard for the dichromate solution. One reason why
results with the iron wire were less erratic than those" obtained with
the ore may be the fact that the dissolved wire is largely in the ferrous
state whereas the ore is largely in the ferric state as indicated by
the following equations:
Iron wire: Fe + 2HCl --'liio- FeCl2 + H:2
Iron ore: .o3 + 6XCJ
Fee ->£- ~eCls + ;H20

Consequently, much less stannous chloride is reqUired to reduce the

wire solution. As mentioned previously, stannic chloride, which
arises from the oxidation of stannous chloride, exerts a deleterious
effect on the SUbsequent titration.

*7*
 However, the ready and complete sol.ubility of the wire
is its outstanding advantage since the most common cause of poor re-
sults with the ore was believed to be the failure to get the ore com-
pletely into solution. The use of stannous chloride to aid the solution
of the ore alleviated one trouble only to cause others of more serious
consequences.
The objection to ferrous ammonium sulfate as a practical
pril)18.ry standard is its instability. .
One other point, not preViously discussed but indicated
by the tables, is that with any of the three pril)18.ry standards the size
of the sample taken had an effect on the calculated FeaOs factor of the
dichrol)18.te solution. This was evident regardless of how the titration
was conducted. The smaller the weight or volume of the primary standard,
the lower the value for the FeaOs factor. However, with weights of
0.04 to 0.05 g. Fe, the factors became nearly constant. It will be
shown later that the difference in factor made no significant difference
in the values for the percent of FeaOs in the cements whether the
highest, lowest or average factor was used in the calculations.
(e) Methods of Determinipg the Endpoints of the Titrations.
It has been mentioned that the endpoints of the previously discussed
titrations were determined by three agents: .
(1) A Beckman Model G potentiometer,
(2) Diphenylamine barium SUlfonate,
(3) N-methyldiphenylamine barium sulfonate.

The principal reason for using the potentiometric method

in this investigation was that results obtained by this method would
serve as a basis for comparing results obtained with the indicators.
Moreover, because of the manner in which potentiometric titrations are
accomplished, it is possible for the analyst to discern what is happen-
ing in the solution (to follow the course and rate of reaction) at any
point along the way. More accurate and reproducib;J..e results were obtained
with the potentiometer, but this method takes approximately five minutes
longer for each titration than the indicator method. For this reason
the indicator method is considered the most practical.
This raises the question of what indicator to use. It has
been shown in an earlier report(2) that diphenylamine barium sulfonate
is superior to the present diphenylamine indicator specified by the
A.S.T.M. In this study a new oxidation..reduction indicator(5) was
investigated. The new indicator is N-methyldiphenylamine barium sul-
fonate. The synthesis of this compound and the preparation of the
indicator solution is given in the appendix of this report.
The new indicator was,found to be more stable, i.e., less
susceptible to oxidative cleavage, than diphenylamine barium sulfonate.
This stability of the new indicator is the principal source of its
advantages over the other indicators of the diphenylamine series. More
will be said later about the various indicators used in this investi-
gation. It will suffice at this point to say that the N..methyldiphenylamine
*8*
barium sulfonate proved to be the better indicator with potassium
dichromate. Results with this indicator closely agreed with results
obtained by electrolytic titrations, thus showing that the indicator
correction for the N..methyldiphenylamine barium sulfonate is negligible.
2. Standardization of the Ceric Sulfate Solution (Table 3).
A number of invest1gators(6)(7)(8)(9) have advocated eerie sulfate as a
superior reagent for oxidation-reduction titrations. Consequently, a
solution of eerie sulfate (approximately 0.05 normal) was used to deter..
mine if more reproducible results could be obtained with this oxidant
then are obtained With dichromate. It was hoped that with eerie sulfate,
the anomaly of the change in factor with varying weights of primary
standard would be eliminated. It was eliminated when sodium oxalate
was used as the primary standard.
The eerie sulfate solution was standardized against the
following four primary standar.ds:
(a) Solution of ferrous ammonium sulfate,
(b) National Bureau of Standards' iron ore, No. 27b,
(c) Iron wire supplied by Central Scientific Co.,
(d) National Bureau of Standards' sodium oxalate.
The indicator, ortho-phenanthroline ferrous sulfate complex
("ferroin"), and a Beckman Model G potentiometer were used to determine
the endpoint of the titrationa. Details of the procedure for standardizing
the eerie sulfate are given in the appendiX.·
(a) Standardization of Ceric Sulfate with Ferrous Ammonium
Sulfate. In standardiZing the eerie sulfate against the solution of
fel"rousammonium. sulfate, ,much the same prpcedure was followed as was
used with this reagent in standardizing the potassium dichromate solution.
As with the dichromate, the instability of the ferrous ammonium sulfate
renders it less desirable as a primary standard. Moreover, as with the
dichromate, the smaller the volume of the primary standard used, the
lower the factor for the eerie sulfate solution.
(b) Standardization of Ceric Sulfate with Iron~ Again
the difficulties of weighing and dissolVing the iron ore maae this
material objectionable to use as a primary standard. Also variations
in the size of the samples of ore produced a corresponding variation in
the factor for the solution. Ho,,,ever, with weights of iron ore corres"
ponding to 0.04 to 0.05 g. of Fe, the factors for the eerie sulfate
eolution became remarkably constant.
(c) Standardization of the Ceric Sulfate with Iron Wire.
The standardization of the eerie sulfate solution against iron wire was
conducted by the same procedure as was used in standardizing the
potassium dichromate, except that only the silver reductor was used for
reducing the solution of the iron Wire. The results using the iron wire
and the silver reductor were more reproducible than those using the iron
ore with stannous chloride reduction.
Variation in the size of the sample of wire again influenced
the factor for the solution. The variation in factor decreased as the
weight of w:l.rewas increased. With weights of 0.04 to 0.05 g. the factors
became nearly constant.
*9*
 (d) Standardization ofCeric Sulfate with Sodium Oxalate.
(1) Potentiometric Method Using Iodine Monochlor1de. The potentiometric
titration of sodium oxalate with cericsulfate solution was accomplished
by using an iodine monochloride solution as catalyst(lO)(ll). With
iodine monochloride, it is possible to carry out the titration at room
temperature. The preparation of th.e catalyst and the details of the
procedure are given in the appendix.
The titration of the sodium oxalate with ceric sulfate was
the most satisfactory of all from en analytical sta~dp()int. In this
titration, the factors were markedly consistent regardless of the weight
of oxalate taken. This agrees with the results reported by Willard and
Young(8). .

The one disadvantage of this titration is that the hydro-

chloric acid concentration in the solution of the oxalate had to be
carefully controlled. In the early titrationa it was eVi.dent by the
drifting of the potentiometric needle after the endpoint had bee~ reached
that the hydrochloric acid present in the solution was being oxidized.
This was corrected by reducing the quantity of hydrochloric acid and
increasing the quantitl of iodine monochloride solution specified by the
original directions (ll) • Willard and Young(8) have also shown that under
certain conditions, hydrochloric acid has a reducing effect on ceric
sulfate. '
(2) Indicator Method Using Ferrous Ammonium Sulfate. At ordinary tem-
peratures, the oxidation of sodiUm oxalate by ceric sulfate is not
sufficiently rapid to serve as a method of standardizingceric sulfate
solutions. In the p~tentiometric.titrationdiscussed above, the reaction
was catalyzed by the addition of iodine monochloride; but even under these
conditions the reaction is not SUfficiently rapid at room temperature to
use ferroin. At temperatures greater than 50°C., ferroin is destroyed(6).
For this reason the indicator method of standardizing the ceric sulfate
solution against sodium oxalate was performed with the aid of a ferrous
a.mmoniumsulfate solution. In this method, an'excess of ceric sulfate is
added, to a hot (90 to 95°C.) solution of the oxalate. The solution is
then cooled and a measured quantity of ferrous ammonium. sulfa.te solution
1s added. This solution need not be standardized., but its equivalent 1n
terms of the ceric sulfate solution must be known. The indicator is
a,dded at this point, and the titration with the ceric sulfate is con-
~ua to the indicator endpoint. The eqUivalent amount of ceric sulfate
solution reqUired by the ferrous ammonium sulfate is subtracted from the
total volume of ceric sulfate used for the titration and the remainder
represents the amount required by the sodium oxalate.
This appears to be a round-about procedure for titrating
eerie sulfate against ferrous ammonium sulfate. The chief advantage is
that the solution of ferrous ammonium sulfate need not be standardized.
Until the recent issuance of a standard sample of potassium dichromate
by the National Bureau of Standards, the standardization of a ferrous
a.mmonium sulfate solution presented a minor analytical problem.

* 10 *
 (e) SHDJ1I!B!l of Results. The most satisfactory method of
standardizing the eerie sulfate solution was with sodium oXalate. The
potentiometric method using iodii'le monochloride as catalyst was more
satisfactory than the indicator method. With sodium oxalate the size
of the sample had no effect on the factor of the solution. With iron ore,
iron wire or the ferrous ammonium sulfate solution, the size of the sample
had a demonstrable effect on the calculated factor of th~ solution.
In comparing the results of these various methods of
standardizing the eerie sulfate solution, the following significant
observation was made. Good agreement between average values obtained
with the various primary standards and between the indicator and po·
tentiometric methods were obtained when these determinations were made
during the same day or week (Tabie 3). However, the results clearly
indicate that the eerie sulfate solution deteriorated on standing.
This point is illustrated by the following data which show the average
factor obtained for the eerie sulfate solution and the date at which
it was standardized together with the method of standardization:

Date Fe 2 0 s Factor Method of Determining Factor

(1946) Found
4/10 0.003936 Potentiometric with ferrous ammonium SUlfate
4/11 0.003934 " 'I sodium oxalate
4/11 0.003933 Indicator With ferrous ammonium sulfate
6/14 0.003920 Potentiometric with iron ore
6/16 0..003915 1/ Ii Ii wire
6/18 0.003911 Indicator with iron ore
6/20 0.003913 " II "wire

7/30 0.003897 Indicator with sodium oXalate

7/30 0.003897 " II iron wire

In the titration of a cement sample, the use of any of the

above factors would make no major dif~erence in the calculated percent
of Fe2 0 3 , i. e •:
8 ml. x 0.003936 (highest factor) = 3.15~ Fe20S
8 ml. x 0.003897 (lowest factor) = 3.12% Fe20s
(f) Methods of Determining the Endpoint of the Tittations.
The diphenylamine barium sulfonate and the N-methyldiphenylamine barium
sulfonate indicators which were used With the dichromate were also used
in a few trial tests with eerie sulfate. The N-methyldiphenylamine barium
sulfonate. gave better results than the diphenylamine barium sulfonate,
but neither is as suitable as ferroin with eerie sulfate.
The ferroin endpoint was fairly sharp. The red on fading
left a yellow-green colored solution. kt no time could the blue
oxidized form of ferroin be observed. The ferroin endpoint was less
striking than either of' the indicators of the diphenylamine series with
potassium dichromate. Average values for the solution factor with the
ferroin indicator closely checked the average values obtained with the
potentiometer (Table 3).

* 11 *
 In the potentiometric titrations, the endpoint with oerie
sulfate was exceedingly sharp. One drop (0.04 to 0.05 rol.) of this
solution was sufficient to cause a break in potential averaging 300
to 400 millivolts. With potassium dichromate, the break in potential
averaged only 75 to 125 millivolts per drop of solution.
(3) Determination of the Ferric Oxide Content of Portland
Cement. The standardized solutions of potassium dichromate and ceric sul-
fate were used for determining the ferric oxide content of eight cements
ranging from 0.4% to 5.4% Fe 20 3 • The same methods previously described
were used to determine the endpoints of the titrations. 'In the indicator
methods the cement samples were reduced with stannous chloride. In the
potentiometric titrations, the cement samples were reduced by passing
the solutions through a silver reductor or by treatment with stannous
chloride. When the cement solutions were to be reduced in the s ilver
reductor, filtration of the sample was necessary to remove the undis-
solved silica which might otherwise clog the pores of the reductor.
The exact details of the procedures followed in the titration of the
cements are given in the appendix of this report., The use of the two
oxidimetric solutions with the various indicators and the potentiometer
for determining the ferric oxide content of the cements is discussed
below.
. (a) Titration of the Cements with Potassium Dichromate. As
pointed out in the discussion concerning the standardization of the
dichromate solution, the size of the sample of the primary standard
affected the calculated Fe203 factor for the solution. It is therefore
logical to assume that this same phenomenon would be present when
titrating cementa of varying Fe203 content. For this reason the
A.S.T.M. directs that in calculating the percent of ferric oxide in
cements, a solution factor should be used which has been determined
against a weight of primary standard that yielded a titration approxi-
mately equal to that reqUired by the cement sample in question(12) •
This procedure necessitates the individual standardization of the
oxidizing solution for each different Fe203 content of celllents. While
this procedure is sound in theory, the results reported in this study
show that the use of a specific factor rather than an average factor
for the oxidant is of no practical significance. The volumes of
dichromate solution usedwitli ea.ch indicator and the potentiometer
in titrating the cements were converted to percent of Fe203 by the
, following two methods (Table 4):
(A) The dichromate factor used to calculate the percent
of ferric oxide was determined with a weight of
primary standard that yielded a titration
apprOXimately equal to that reqUired by'the
cement sample. Separate factors for each
indicator,potentiometer and for the iron wire
and iron ore were used and are tabulated in
separate columns.
(B) An average dichromate factor was used to calculate
the percent of ferric oxide in the cements. The
average factor (from Table 2) for each indicator,
potentiometer, etc. was used.

* 12 *
 A comparison b&tween the A columns with the corresponding
B columns in Table 4 shows that in the majority of cases the results
were identical whether an average factor or a specific factor for the
dichromate solution was used. The maximum difference between any value
shown in oolumn A and the corresponding value shown in column B was 0.0;
percentage points (one case only).
However, when the values obtained with the indicators
are compared with those obtained with the potentiometer, a difference
(still insignificant) in values is noted. A comparison on this basis
(Table 5) shows that somewhat higher values were obtained with the
indicators than with the potentiometer. :Results with diphenylamine
barium sulfonate are slightly higher than those with N-methyldiphenylamine
barium sulfonate but neither gave results as high as the diphenylamine
indicator specified inA.S.T.M. C 114-46.
Whether the dichromate factor was determined with iron
ore or with iron wire also made a slight difference in the calculated
Fe20 S contents of the cements •. The reason for this is that iron
ore as a primar,y standard gave a higher factor for the dichromate
solution than iron wire. The higher factor with the ore as the primary
standard might be explained on the bas'!s that the ore may not have been
transferred entirely quantitatively .from the weighing dish to the
beaker. A loss of a portion of the ore would result in less of the
OXidizing solution being required to titrate the ore which would give
a falsely high value for the OXidizing power of the solution. Likewise,
failure to obtain complete solution of the ore would produce the same
effect. A high factor for the solution would give a correspondinglY high
value for the Fe20Scontent of the cements.
The method of reducing the cement samples had the most
noticeable effect on the calculated values of Fe20 S content. The
results obtained when reduction was accomplished with the silver
reductor averaged 0.02 percentage points lower than results obtained
after the stannous chloride method of reduction (Table 4). The low
results with the silver reductor may be due in part to the fact that
the silver reductor column reduces iron only, wbile other common inter-
fering elements remain in the oxidized state\l;). Another explanation
may be that because of the presence of undissolved silica, which
would clog the passages through the silver particles in the reductor
. column, the cemeItt samples had to be filtered prior to reduction. Care
was taken to wash the filter papers and residues thoroughly but it
is well-known that it is exceedingly difficult to wash the last traces
of iron from silica, Low results would be obtained if part of the iron
remained in the filter paper.
SummariZing the foregoing statements, it is evident that'
the primary standard used to determine the solution factor, the method
of determining the endpoint of the titrations, and the method of
reducing the samples have a greater effect on the percent of Fe20s
found than does the use of an average factor or a specific factor in
the calculations.

* 1; *
 (b) Titration of the Cements with Ceric SUlfate. As with
dichromate, both an average factor end a speoifio factor for oalou-
latins the peroel;'lt of Fe203 in the oew-ents were used. Again the
results were remarkably oonsistent regardless of the faotor used in
the oaloulations (Table 4).
On the other hand, there were slisht differenoes between
results obtained withtheindioator and those obtained with the poten-
tiometer. The greatest difference was noted when the cements were
reduoed. in the silver reduotor column. As previously observed with
the diohromate solution, res~ts after the cements were reduced in
the silver reduotor oolumn averaged 0.02 percentage points lower than
results obtained after stannous ohloride reduotion. Practicall1no
differenoe between the use of faotors determined with iron ore and with
iron wire could be seen in the calculated percentages of Fe20 S '
The foregoing discussion may be summarized as follows:
The ceric sulfate titrationa further supported the conclusion that
the use of a specif10factor for each cement rather than an average
faotor is of no practical signifio~oe. The method of determining
the endpoint of the titrations and the method of reducing the samples
had some effect on the oalculated peroentages of ferric oxide, but the
maximum difference between results obtained by any method using cerio
sulfate was only 0.05 percentage points. The maximum difference with
potassium. dichromate was 0.07 percentage-points (Table 4).

4. COIllPariEiopof the ~'W'O Oxid1metr1c Solutions. As atatedpre-

"ti10u.tf17 the purpose of using oerie sulfate in this investigation was to
de~ermine whether this oxidant possesses, suffioient advantages to
warrant using it in plaoe of potassium dichromate for determining the
ferric oxide content of oements. A few oomparisons between the two
solutions have been indioated, but a more complete disoussion will
now be given to show why potassium dichromate is the most practical
oxidant for ferric oxide determinations.
(a) Pre'Paration. One practical advantage of potassium
dichromate is that it-is a simple and inexpensive solution to prepare.
The preparation merely requires dissolving the salt in water, whereas
the preparation of the oerio sulfate solution is an involved process
neccessitating the use of a high percentage of sulfUrio acid. More-
over, the ammoniumnitrato-cerate from whioh the ceric sulfate solution
is prepared is expensive and ten times more of the cerate salt must
be used in preparing the solution than the amo'lmt of diohromate
required to prepare the same quantity of solution of the same normality.
(b) Stability. The remarkable stability of the dichromate
solution is another important advantage of this' reagent. No evidence
of deterioration of the potassium diohromate solution used in this
study was noted. The absolute oonstanoy of this oxidant was shown
by Carey(14). He reported that a 0.1 N solution of potassium diohromate,
stored in an eight-liter flint glass bottle, had remained unohanged in
titer during a period of 24 years.

* 14 *
 Such is not the ease with eerie sulfate. Solutions of this
oxidant are regarded as b@ing reasonably stable when compared with
permanganate solutions. Nevertheless, the fact that definite and eon·
tinuous deterioration does take place is evident not only from this
study but has also been shown by Smith and Getz(15) who investigated
the stability of eerie oxidim@tricsolutions of various normalities
and containing various acids. All the ceric solutions used in their
studies showed deterioration during storage. According to their .
results, some of the 0.05 N solutions changed as much as 4.72% in
six months. Their O.lB solution changed 2.04% in· 7~ mon'tjhs. This
Was a perchloric acid solution of the ceric salt which waS shown to
be the most stable of all. The authors stated that this perchloric
acid solution "was SUfficiently stable in ordinary light not to
require restandardization for a period of 20 days".
(0) Us~w.•tth H,ydrochloricAcid. Solution of portland cement
is largely dependellt upon the use of hydrochloric acid and in selecting
an OXidant for titrating such a solution, it is well to keep in mind the
possibility of side reactions between the oxidant and the acid. This
is another reason for selecting potassium dichromate for routine ferric
oxide determinations. Ceric sulfate may be used with hydrochloric
acid, but because of its greater oxidative power (resembling per-
manganate) it will react with hydrochloric acid under certain conditions.
On the other hand, potassium dichromate because of its lower oxidation
potential can safely be used in a hydrochloric acid medium.
(d) Standard.ization. However, in comparing the two oxidi-
metric solutions, the advantages of cerie sulfate should not be over·
looked. From an analytical standpoint, none of the methods of standariz-
ing the dichr0IM-te are entirely satisfactory. It was shown.in an
earlier report\2J that numerous variables such as acidity, temperature,
dilution, rate at which solution is added particularly as the endpOint
is approached, grams of Fe being titrated, and the presence of extraneous
salts markedly affect the reaction between the ferrous iron and the
oxidant. Thus erratic factors may be obtained with any ferrous
standard available for use with potassium dichromate. Moreover, there
exists a peculiar anomaly in the standardization of dichromate in that
with low weights of iron (below 0.04 to 0.05 grams of Fe) there is
a noticeable variation in the solution factor. The above points also
apply to ceric sulfate when standardized against ferrous iron. However,
because of the greater oxidative power of cericsulfate, it can be
standardized against sodium oxalate. With this primary standard .
results are much more reproducible and the size of the sample has no
effect on the factor of the solution. Thus in the comparison of the
standardization of the two solutions, the advantage lies withceric
sulfate. However, it has been shoVlIl that the anomalous variation in
factor produced practically no difference in the calculated percent
of ferric oxide in the cementa analyzedjconseguently, this advantage
is ofno gree.t significance.
Ceric sulfate as a volumetric reagent has one advantage which
cannot be refuted. Ceric sulfate can be used for a wide variety of
oxidimetric titrations whereas potassium dichromate because of its
lower oxidative power is of limited application. From the standpoint
of cement analyses, the most important constituent besides ferric oxide
which can be determined with eerie sulfate is calcium oxide.
* 15 *
 (e) Indicators. Those who ardently advocate the U$e of eerie
sulfate will bring up another point, namely; the availability of suitable
indicators for use with each oxidant. They will have in mind that no
indicator can match the numerous advantages of ferroin; and it is true
that regardless of its other advantages, any oXidant would be of limited
value to an analytical laboratory unless suitable indicators for use·
''lith that oXidant were available. Consequently, this subject vli11 be'
discussed somewhat in detail.
It has been shown by this work and by previous workers that
indicators of the diphenylamine series are not very satisfactory for
use with eerie sulfate. Likewise rerroin is not satisfactory for use
with diohromate because the color change is exceedingly slow to appear
with this oxidant. Thus, at the present time, the analyst has ceric
sulfate with ferroin on the one hand, and potassium dichromate with
diphenylamine indicators on the other from which to choose.
It is admitted that ever since Knop's initial introductton
of diphenylamine, numerous investigators have been seeking a less
capricious indicator for use with dichromate. Most of the new compounds
suggested have been di];lheny1amine derivatives. Di];lhenylaminesu1fonic
acid (or the barium salt) has been generally accepted as the best of
these and has been ado];lted in numerous industrial laboratortes(2) and
in the Federal S];lecification for cement. The principal advantages of
this indicator are threefold:
(1) It is soluble in water,
(2) 'Fhe color change is very sharp, being recognizable
eveil when the solutions themselves have
considerable color,
(3) The indicator is not affected by tungsten or
various other salts which interfere with
diphenylamine.
However, diphenylamine barium sulfonate is not perfect. It
has a high indicator correction. Moreover, this correction is not
constl]nji but varies with different conditions. Thlje Whitehead and
Wil1s\lo) in their review article concerning redox indicators report
that all indicators for dichromate-iron titrations have one or more
disadvantages. These authors, however, were not aware of the attri-
butes of the indicator N-methyldipheny1amine barium sulfonate. As a
matter of fact, the only mention they make of this indicator is at
the end of their paper in a list of compounds reported as being
unsatisfactorl for use as redox indicators. The authors state that
this compound is reported as giving a long range of intermediate
colors as the endpoint is approached which renders it unsatisfactory
as an indicator.
This objection originates from the work of straka and Oesper(17)
vTho made the only previous investigation in this country of the indi-
cator properties of methyldiphenylamine and its sulfonic acid derivatives.
They state that these compounds are not suitable as indicators "since
a long range of intermediate colors precedes the development of the
full red, and the establishment of the exact endpoint is consequently

* 16 *'
obscured"*.
In the present study no long range of intermediate colors
but a sharp and clear-cut endpoint was observed with the N-methyldiphenyl-
amine barium sulf9~te indicator. This supports the work of Knop and
Kubelkova-Knopova\5J who found that this compound is not only satisfactory'
as an indicator but actually excels all others for dichromate-iron
titrations. It has all the advantages of diphenylamine barium sulfonate
without the disadvantage of a high indicator correction. Knop and his
co~worker report that the indicator correction for N-methyldiphenylamine
barium sulfonate is practically nil and can be ignored except in special
micro-titration. The results of this investigation support that con-
clusion since titrations with this indicator closely agreed with poten-
tiometric titrations.
Moreover, according to Knop and his co-worker, N-methyldiphenyl-
amine barium sulfonate is the most sensitive of all indicators. It
requires only 0.002 to 0.003 ml. of a 0.1 N solution of potassium di-
chromate to produce a color change. Diphenylamine barium sulfonate
requires 0.005 ml. of a O.lN solution, and ferro in requires 0.01 ml
of 0.1 N ceric sulfate to produce a color change(ll). Consequently,
because of its low indicator correction, sensitivity, and other quali-
ties of solubility and stability, N-methyldiphenylamine barium sulfonate
is as good, if not better than,ferroin as an indicator for iron titra-
tions. This advancement places dichromate in a new strategic position
as a volumetric reagent.
The use of N-methyldiphenylamine barium sulfonate is handi-
capped by the fact that this indicator is not yet manufactured com-
mercially. Consequently, in order to take advantage of the many de-
sirable properties of a dichromate exidimetric solution, the analyst
must synthesize the indicator at present or be content with the use
of diphenylamine barium sulfonate in spite of its larger indicator
correction. It should be pointed out, however, that a comparison be-
tween results obtaineq with any of the three indicators, ferroin and
the above two diphenylamine derivatives, agreed within 0.02 percentage
points (Table 5). It is believed that such results are entirely sat-
isfactory from a practical standpoint.
(f) Summar;y of the Comparison of the Two OXidimetric Solutions.
Potassium dichromate is considered the most practical OXidant for de-
termining the Fe2 0 S content. of cements because of the low cost and
low equivalent weight of the salt, ease of preparing the solution,
unquestionable stability of the solution,a;vailability 'of· sili table
f.i.~d.ioaJt;~i.<f04'c:uee W14/lti d;~,ll6.'l'b.tiOO, a!I'l1i i ts n~.•'!'eactivity "With .
hcrdJ:"'ochloric ac:id. Cerle sadfats ha13the, adv.ante:ge cx:f'C)bB1hg'e-a.a:lier
ito atanda.rn.ize and can be used for determining other cement oxides
such as CaO.
* Straka and Oesper used a different method of synthesizing their com-
pound from that of Knop and Kubelkova-Knopova (German investigators who
were apparently una'Ware of the preVious work in the U.S. done by Straka
and Oesper). It may be that the method of synthesis used by straka and
Oesper resulted in a mixture of compounds which would account for a mix-
ture of colors at the endpoint.
* 17 *
 ' .. Discussion of R.esult§ in T&:bles 1 to 6. At ve.r1ou~ times
references to th€ tables have been' made end points exemplified ha'ifeb~e1~
discussed. To aVoid repetition, these points will no:t again be :fully
discussed in this section. Emphasis will bEf placed on the data not
already mentioned and the points illustrated by these.
Table 1 - Standardization of Ferrous Ammonium Sulfate
Solution with National Bureau of Standards' Potassium Dichromate.
As previously mentioned for reasons of convenience a solution of
ferrous ammonium sulfate was prepared for use as a primary standard.
Table 1 gives the results of standardizing that solution with a
sample of potassium dichromate (N.B.S. NO. 1,6) using diphenylamine
barium sulfonate, N-methyldiphenylamine barium sulfonate and a Beckman
MOdel G potentiometer for the endpoint determinations. The size of
the sample of dichromate taken appears to have no effect on the normality
factor of the solution. However, the values obtained with diphenylatlline
barium sulfonate as the indicator are erratic and much higher than
those obtained with the N-methyldiphenylamine barium sulfonate or the
potentiometer. This is a striking example of what Knop and hie co-
worker(5) refer to as the "greater resistance to oxidative cleavage
and correspondingly much smaller indicator correction of N-methyldiphenyl-
amine barium sulfonate". Each of the indica.tors referred to in the
table was added at the beginning of the titration, but the diphenylamine
barium sulfonate indicator being highly susceptible to OXidation began
to use up the oxidizing power of the dichromate. This resulted in less
ferrous· ammonium sulfate solution being reqUired for the reduction.
The longer the time the titration required, the more the indicator
exerted its influence on the dichromate. Consequently, values for
the normality of the ferrous ammonium sulfate solution with the indi-
cator wer~ high and erratic. This agrees with results reported by
Brennecke{ll) who also noted the strikingly greater indicator correc-
tion for diphenylamine barium sulfonate when dichromate was back
titrated with ferrous sulfate.
With N-methyldiphenylamine barium sulfonate, no difficulty
was experienced in obtaining reproducible results. The normality
factor determined with this indicator agreed closely with both the
theoretical and the potentiometric factQr,.
In Table 1, the dates of standardizing the solution are
given since subsequent standardization showed that the ferrous solution
had deteriorated. Inasmuch as the indicator titrations were performed
on the same day, the differences in results with each of the indicators
cannot be attributed to deterioration of the solution.

Table 2 - Standardization of Potassium Dichromate Solution.

Results are given showing the kind and quantity of primary standard
used, the method of reduction, and the method of determining the
endpoint. No results are reportt!d for i:he titration of the potassium .
dichromate solution with the standard solution of ferrous ammonium
sulfate as the primary standard using the indicator diphenylamine barium
sulfonate because a 'satisfactory normality factor for this standard could
not be obtained with this indicator (See Table 1).

* 18 *
 Reasonably good agreement between all methods of standardizing
the dichromate is shown in Table 2. Results with the iron ore were
slightly higher than those obtained with iron wire. Using the oondi-
tions previously outlined, the two methods of reduoing the iron wire
gave results in remarkably olose agreement. These points have been
disoussed previously and their signifioanoe will not be reiterated.

Table 3 - Standardization of Cerio Sulfate Solution. As in

Table 2 the kind and amount of primary standard used and the method of
determining the endpoint is indioated. In addition the date.of standardiza-
tion is shown sinoe it was observed that the solution deteriorated with
age. The differenoes in the values for the faotor of the solution may
be attributed to the passage oft1me rather than to the method of standard-
izing the solution. Satisfaotoryvalues were obtained with any primary
standard and by either means of determining the endpoint prOVided the
factors were determined within a d.q or two of eaoh other.

Table 4 - Dete~ination of. the FeriPs Content of Portland Cement.

This table shows a comparison of values obtained with potassium dichromate
and eerie sulfate as oxidiZing solutions. The indicators used with each
oxidant may be compared with" the corresponding potentiometric titrations.
Also the primary standards used to determine the Fe2 0 S factor for the
oxidant is indicated and comparisons between iron ore and iron wire as
the prima.ry standard may be made. The. metbod of reducing the sam.ples
is also shown in the table.
The volumes of the oxidizing solutions for the various
titrations were caloulated by two methods and are recorded in the table
in the columns designated f1A" and fiB". In column "A" the factor used
to calculate the Fe20 S percent was determined against a weight of
primary standard that yielded a titration apprOXimately equal to that
required by the cement sample. In column "B" an average factor
(reported in Table 2) was used to calculate the percent of Fe20 S • As
previously pointed out, the oxidant used, the method of standardizing
the oxidant, the method of determining the endpoint and the method of
reducing the samples had a greater effect on the calculated results than
did the use of an average Fe20s factor for the oxidizing solution.
The results in Table 4 are remarkably consistent resaraless
of the method of titrating the samples or the method of calculating
th~ results. These consistent results are believed to be due to the
fact that optimum conditions were maintained in conducting the analyses.
Such variables as acidity, temperature, dilution and rate of adding
the oxidant were kept as constant as possible using ordinary analytical
techniques. The solutions being titrated were stirred by a mechanical
(air motor) stirrer. While each of these details may seem to be of minor
importance, collectively they pan make the differenoe between consistent
and inconsistent results in routine analyses.

Sinoe it is evident by the results in Table 4 that satis-

faotory results can be obtained with any of the methods used in this
investigation, the selection of one method in preference to the others
for routine analysis should be based o~ convenience and econ~.

* 19 *
 Table 5 - Compari8on of the FeaO/i Content of Portland Cements
bl Different Methods. In this table the average results for each of the
indicators and the potentiometer with 'PotFl.ssium dichromate and for the
indicator and potentiometer with ceric mu1fate are shown. Also results
previously reported using the A.S.T.M. method for determining the FeaOs
content of the cements are given. The indicator diphenylamine gave the
highest results, and diphenylamine barium sulfonate the next highest.
N-methyldiphenylamine barium sulfonate and ferro in each gave about the
same values which were somewhat lower than results obtained with the other
indicators. Potentiometric results with eerie sulfate were about 0.02
percentage points higher than potentiometric results with potassium
dichromate. Also, results with samples reduced by the modified stannous
chloride procedure were about 0.02 percentage points higher than those
reduced in the silver reductor column. The significance of these points
has been discussed previously.

Table 6 - COmParison of the Highest and Lowest CalcUlated

Values for the FeaQs Content of Portland Cements. This table gives the
highest and lowest values from Table •4 for each of the cements. The
method of determining the factor used for calculating these percentages
of FeaOs is indicated. This table shows that the maximum difference
between the highest and lowest values for any cement, determined by any
method used in this investigation, was 0.08 percentage points.

SUMMARY
Determinations were made of the ferric oxide content of
portland cements with potassium dichromate, and the indicators
diphenylamine barium sulfonate and N-methyldiphenylamine barium sul-
fonate, as compared with similar determinat:J.ons made with eerie sulfate
and the indicator ferroin. Results obtained with the various indicators
were checked against potentiometric titrations. A silver reductor and
the stannous chloride method were used for reducing the iron. Also
various primary standards for determining the Fetps factor of the
standard soluti"ons were compared. Potassium dichromate solution
standardized against iron wire using the indicator diphenylamine barium
sulfonate is recommended as the most sui table method until N-methyldiphenyl-
amine barium sulfonate becomes commercially available.
All the points previously discussed will not be included in
this summary; but for purposes of comparison, the advantages and dis-
advantages of' the two OXidants, of the primary standards iron ore and
iron Wire, and of' the indicators diphenylamine barium sulfonate and
N-methyldiphenylamine barium sulfonate will be summarized.

* 20 *
1. Cerie Sulfate:
Advantages:
(1) It has a high oxidation potential~ being almost as
strong in oxidative power as potassium permanganate.
(2) It can be used for determining a number of· elements
besides iron, the most important of which, from the
standpoint of cement analysis, is calcium.
(3) In potentiometric titrations, the break in potential
is much greater than with potassium dichromate.
(4) It can be standardized against sodium oxalate which
gives a constant factor regardless of the weight of
the sample taken.
Disadvantages:
(1) The solution is much lesa stable than a potassium
dichromate solution.
(2) Because of the high oxidative power of eerie sulfate,
under certain conditions, it Will react with hydro-
chloric acid thus showing a $imilarity to potassium
permanganate. Therefore the concentration of the
hydrochloric acid must be more carefully controlled
than with potassium dichromate.
(3) Preparation of the eerie sulfate solution is a more
involved process in comparison with the preparation
of a potassium dichromate solution.
(4) Ammonium. nitrato-cerate from which the eerie sulfate
solution is prepared is more expensive than potassium
dichromate, and 10 times more of the cerate salt must
be used than to prepare the corresponding quantity of
dichromate solution. Also, the indicator, ferroin,
to be used with the eerie sulfate is costly.
2. Potassium Dichromate:
Advantages:
(1) It is the most stable of all oXidizing "solutions,
showing no appreciable change in titer for years.
(2) It is a simple aqueous solution without involVing the
expense and hazard of using sulfuric acid which
is necessary with eerie sulfate.
(3) It can be used with low cost indicators.

* 21 *
 (4) Potassium dichromate favors a hydrochloric acid medium
which is neceseary in cement analysis.
(5) There is evidence from a number of sources that the most
satisfactory results of iron determinations are obtained
with potassium dichromate. It was found to be true in
this study. Henry and Gelbach(19) recommend the use of
potassium dichromate rather than eerie sulfate. They
studied the use of potassium dichromate and eerie sulfate
in the determination of iron reduced by a silver reductor
and conclude that lithe stability, purity, low equivalent
weight and comparative low cost of potassium dichromate
render ita very desirable oxidizing agent". The results
reported by the 32 skilled analysts who analyzed the
National Bureau of Standards Sibley iron ore (No. 27b)
are also significant. The average of the dichromate
determinations was 68.22%; the average of the permanganate
determinations was 68.34%; the average of the eerie
sulfate determinations was 68.17%; the value recommended
by the National Bureau of Standards is 68.23%.
Disadvantages:
(1) Because of the lower oxidative power of potassium di-
chromate, the change in potential at the endpoint of a
potentiometric titration is not as easil;y recognized
with potassium dichromate as with eerie sulfate.
(2) The standardization of potassium dichromate is more
difficult than the standardization of eerie sulfate.

3. Iron Ore:
Advantages:
(1) The ore carries a certificate of analysis issued by
the National Bureau of Standards.
(2) As the ore is 68.23% Fe, errors in the weighing of samples
have less effect on the calculated factor of the
solution than would similar errors in weighing samples
of iron wire which are 99+10 Fe.
Disadvantases:
(1) The ore is not homogenous but a mixture of 'Various iron
OXides, some of which are very difficultly soluble in
hydrochloric acid.
(2) Weighing a sample of the ore and transferring it quanti-
tatively to a beaker is difficult because the fine
particles of ore cling tenaciously to whatever they touch.

* 22 *
 (3) Solution of the ore ie very time-consuming. Even if
solution is aided by the addition of stannous chloride,
the time required is much longer than that required
by the same weight of iron wire.
(4) Inasmuch as the certificate of anaJ.ysis for the ore shows
a wide variation in the results of the selected analysts
who tested the ore, it is very unlikely that the average
laboratory ohemist will obtain r.esults any more uniform
in his analysis with the ore.
4. Iron Wire:
Advantages:
(1) It is much more convenient to weigh samples of wire than
a powdered ore, or even orystals of a salt such as
sodium. oxalate.
(2) There is no difficulty in handling .the weighed sample
quantitatively; oonsequently, there is little likelihood
of errors due to loss of sample.
(3) Solution of the wire takes place very readily.
(4) The dissolved iron is largely in the ferrous state;
consequently, much less stannous chloride is needed
to reduce the solution. This IllaY be one of the
reasons why better results are obtainable with iron wire.
Disadvantages:
(1) The analysis of the wire is certified by a manufacturer
rather than the N.B.S. However, there are several
manufacturers who are noted for their integrity re-
garding the statement of purity on their products. ,
Consequently, any variation in the lab~:led EWld:actual
percent of Fe in the iron wire will ~J:'obably be less
than the variation in the ore shown by the certificate
of analysis. Moreover, in case of doubt, the iron wire
could be standardized.
(2) The wire when exposed to the atmosphere becomes oXidized
and must be cleaned with emery paper.

5. Diphenylamine Barium. Sulfonate:

Advantases:
(1) The color change is very easy to recognize. In fact, the
color change is so marked that the, titration of the ferrous
iron by the dichromate succeeds eVf}n \'Then the solutions
themselves have considerable color\20J.

* 23 *
 (2) The salt is soluble in water.
(3) The indicator may be used over comparatively wide ranges
of acidity.
(4) The hue of the indicator at the endpoint of the titration
is less affected by various substances such as tungsten,
stannic chloride and mercuric chloride, than is the
color change of diphenylamine.
Disadvantases:
(1) The indicator is very susceptible to oxidation, and,
consequently, has a high indicator correction.
(2) The solution of the indicator is affected by heat and
light and eventually becomes discolored. In this
respect it resembles diphenylamine.
(3) The indicator is not suitable for the titration of
dichromate with a ferrous solution. In this reverse
titration the indicator corrections are exceedingly
high end erratic.

6. N-methyldiphep.tlamine Barium Sulfonate.

Advantases:
(1) The indicator has the .lowest indicator correction of
all the diphenylamine series. Beaulta obtained with
this indicator closely agree with those obtained by
potentiometric titration.
(2) The sensitiVity of the lndicato~ is very great. A
concentration of 2 or 3 x 10- N of the potassium
dichromate is sufficient to cause the color change.
(3) The salt is readily soluble in water and its solutions
do not deteriorate with age, no doubt due to the
greater stability of the indicator.
(4) Moreover, because the indicator is more resistant to
oxidation than other members of the diphenylamine series,
it is suitable for the titration of dichromate with a
ferrous solution.
Disadvantages:
It is not yet available commercially.

* 24 *
 RECOMMENDATIONS
In concluB ion it is recommended that the following changes
be made in the current A.S.T.M. procedure (C 114-46) for determining
the ferric oxide content of portland cement: ,
(1) Use diphenylamine barium sulfonate as indicator in place
of diphenylamine. (This is already being contemplated
by the A.S.T .M. and has been adopted by the Federal
Specifications).
Reduce the strength of the stannous chloride solution
from lCl/o to 5'10. (This ha.s also been mentioned by
W. C. Hanna, Chairman, Working Oommittee on Methods of Chemical
Analysis, and has been adopted by the Federal Specifications.
Reduce the amount of saturated mercuric chloride solution
added to oxidize the excess stannous chloride from. 10 ml.
to 5 ml. It is important to add the mercuric chloride
before final dilution of the solution.
(4) Change the phosphoric-sulfuric acid mixture to a 1-1
phosphoric acid solution and reduce the amount to be
added to 10 ml. (This agrees with the new Federal
S-pecifications).
(5) Use a hydrochloric acid concentration of the solution to be
titrated of 1 to 2 N for best results, or approximately
10 to 12 ml. Hel per 100 ml. of solution.
(6) AlthoUgh it is logical from. a theoretical point of view to
consider iron ore as being a superior primary standard to
iron wire, in practical work the average analyst will
obtain better results with iron wire. Therefore, it is
recommended that iron wire at least be retained as an
alternate standard.
(7) A more reliable factor for the solution is obtained by
averaging the results of four or five determinations of
the factor (against 0.030 to 0.035 g. Fe) than to use
various factors determined by one or two titrations whose
volume approximates the volume required by the cement.
Individual determinations of the factor of the standard
solution may vary considerably even though titrated against
the same weisht of Fe, and therefore, single or even
duplicate determinations of the factor may not be accurate.

ACKNOWLEDGMENT
The author is indebted to Mr. C. L. Ford and Dr. L. S. Brown
of the Research Laboratories of the Portland Cement Association, and
to Mr. G. W. Schaeffer of the University of Chicago for information
and helpful suggestions concerning-this investigation.

* 25 *
 BEFEBENCES

(1) W. F. Hillebrand and G. E. F. Lundell, "Applied Inorganic Analysis",

John Wiley and Sons, New York, N. Y' J (1929).

(2) L. leMar, "Advantages of Newer Oxidation-Reduction Indicators

in Determining the Ferric Oxide Content of Portland Cement lt ,
Portland Cement Association Report (unpublished), (1945).

(3) I. M. Ko1thoff and L. A. Sarver, "Properties of Diphenylamine and

Diphenylbenzldine as Reduction-Oxidation Indicators",
Zeitschrift fllr Electrochemie, voL 36, p. 139 (1930);
Chemical Abstracts, vol. 24, p. 2078 (1930).

(4) F. J. Watson, "Diphenylamine as Indicator in the Titration of

Iron with Dichromate", Chemical Engineering and Mining
Review, vol. 22, p. 355 (1928).

(5) J. Knop and O. Kube1kova-J{nopova, "N-methyldipheny1pine-p-

Sulfonic Acid as a Redox Indicator", Zeitschrift fur Ane.lytische
Chemie, vol. 122, p. 183 (1941).

(6) G. F. Smith, "Cerate Oxidimetry", G. Frederick Smith Chemical

Co. (1942). . .

(7) G. F. Smith, V. R. Sullivan and G. Frank, "Hexanitrato

Ammonium Cerate as a Proposed Reference Standard in OXidimetry",
Industrial and Engineering Chemistry (Anal. Ed.), vol. 8, p. 449
(1936).

(8) H. H. Willard and P. Young, ftC eric Sulfate as a Volumetric

OXidizing Agent. 1. Preparation and Standardization of
Solutions. Determination of Calcium", Journal, Am. Chem.
Soc., vol. 50, p. 1322 (1928). '

(9) H. H. Willard and P. Young, "Ceric Sulfate asa Volumetric

Oxidizing Agent. II. Determination of Iron", Journal, Am.
Chem. Soc., vol. 50, p. 1334 (1928).

(10) 1. M. Ko1thoff and N. H. Furman:" "Potentiometric Titrations",

2nd. Ed. John Wiley and Sons, New York, N. Y. (1931).
" .
(11) W. Bottger (editor) et a1, "Newer Methods of Volumetric Analysis"..
translated by R. E. Oasper, 2nd printing, D. Van Nostrand Co.,
New York, N. Y. (1938).

(12) A.S.T.M. Committee C-l, "A.S.T.M. Standards on Cement", American

Society for Testing Materials, Philadelphia, Pat (1946).

(13) G. H. Walden, Jr., L. P. Hammett, and S. M. Edmonds, "Phenanthrollne-

Ferrous Ion. III A Silver Reductor. The Direct Determination of
Iron in the Presence of Vanadium", Journal, Am. Chem. Soc.,
vol. 56, p. 350(1934).

*26 *
 Research Laboratories of the
Portland Cement Association
33 W. Grand Ave., Chicago

Table 1 - STANDARDIZATION OF FERROUS AMMONIUM SULFATE SOLUTION

WITH NATIONAL BUREAU OF STANDARDS I POTASSIUM DICHROMATE

Values are for single determinations. Indicato!' and

potentiometric standardizations made on April 8, and
April 9, 1946, respectively.

INDICATORS POTENTIOMETER
Diphenylamine N-Methyldiphenylamine Weight
Barium Sulfonate Barium Sulfonate Dichromate Normality
Weight Weight
Dichromate Normality Dichromate Normality
0.0110 0.05133 0.0109 0.05029 0.0125 0.05048
0.0124 0.05109 0.0107 0.05040 0.0214 0.05045
0.0202 0.05186 0.0210 0.05021 0.0321 0.05047
0.0203 0.05092 0.0202 0.05024
0.0300 0.05065 0.0320 0.05020
0.0300 0.05056 0.0301 0.05040
Av. 0.05107 Av. 0.05029 Av. 0.05047

June 4, 1947
Series 339
 Research Laboratories of the
Portland Cement Association
33 W. Grand Ave., Chicago

Table 2 - STANDARDIZATION OF parASSIUM DICHROMATE SOLUTION

Values represent the average of two or more determinations.

Fe...O.,. FACTOR
Primary Indicators·
Standard Diphenylamine N-Methy1dipheny1amine Potentiometer
Barium Sulfonate Barium Sulfonate
Ferrous Ammonium.
Sulfate Solution
Vol.
5 ml. - 0.004014 0.004019
10 " - 0.004015 0.004019
15 " - 0.004035 0.004039
Av. 0.004021 Av. 0.004026
----- ---- - - - - --- --. - -- - - --- - - - - - -- - - -- -
Sibley Iron Ore
(SnC1a Reduction~
~
0.01 0.003995 0.004017 0.004023
0.02' 0.004032 0.004037 0.004028
0.03 0.004044 0.004047 0.004040
0.04 0.004038 0.004046 0.004052
0.05 0.004059 0.004057 0.004053
Av. 0.004034 Av. 0.004041 Av. . 0.004039
- - - .. _. - , - - - - - - - - - - - - - - - - - - - - - - - -------
- -liire
-Iron
.' SnC1a ReQ,ucti9n rj
~ .
0.01 0.004006 0.004003 0.004023
0.02 . 0.004023 0.004025 0.004020
0.03 0.004025 0.004032' 0.004038
0.04 0.004033 0.004046 0.004039
0.05 0.004033 0.004049 0.004049
Av. 0.004024 Av. 0.004031 Av. 0.004034
S:L1ver Reduator
g. Jfe
0.01 0.004008 0.004014 0.004018
0.02 0.004009 0.004030 0.004027
0.03 0.004030 0.004032 0.004031
0.04 0.004040 0.004048 0.004043
0.05 0.004036 0.004056 0.004052
IAv. 0.004025 Av. 0.004036 v. 0.004034

June 4, 1947
Series 339
 Research Laboratories of the
Portland Cement Association
33 W. Grand Ave., Chicago

Table 3 - STANDAIIDlZATION OF CERIC SULFATE SOLUTION

Values represent the average of two or more

determinations unless otherwise specified.

Primary Indicator (Ferroin) Potentiometer

Standard Date FeaOs Date Fe20s
(1946) Factor (1946) Factor
Ferrous Ammonium.
Sulfate Solution
Vol.
5 rol. 4-11 0.003913 4-10 0.003925
10 " 0.003940 0.003934
15 " 0.003946 0.003950
Av . 0 ..003933 Av. 0.003936
--------- - --
Sodium Oxalate
~ ~ - - -- -- - - - --
...

Wt. g.
0.03 7-30 0.003893* 4-11 0.003928*
0.03 0.00390]* 0.003937*
0.04 0.003943*
0.05 0.003934*
0.05 0.003926*
Av. 6.003897 Av. 0.003934
- - - -- -- ---
Sibley Iron Ore
I

(SnCl? Reduction)
g. ·Fe
0.01 6-18 0.003881 4-14 0.003893
0.02 0.003907 0.003912
0.03 0.003920 0.003929
0.04 0.003924 0.003932
0.05 0.003924 0.003933
Av. 0.003911 Av. 0.003920
Iron Wire
Silver Reductor
g. Fe
0.01 6-20 0.003870 6-16 0.003895
0.02 0.003912 0.003899
0.03 0.003912 0.003915
0.04 0.003932 0.003933
0.05 0.003940 0.003933
Av. 0.00391, Av. 0.003915
SnCla Reduction
g. Fe
0.05 7-30
* Single determinations only.
June 4, 1947
Series 339
 Ta-ole 4 - DErERMINATION OF THE Fe20s CONTENT OF PORTLAND CEMENTS, PER CENT BY WEIGHT /
Column A: Factor used to calculate the per cent of Fe2 0S was determined with a weight of primary standard that
. yielded a titration approximately equal to that required by the cement sample. .
Column B: Average factor used to calculate the Fe 2 0 S per cent.
Values represent the average of two or more determinations.
TITRATED WITH POTASSIUM DICHROMATE
INDICATORS POTENTIOMETER
Di he lamine Barium Sulfonate \ N-methyldiphenylamine Barium Sulfonate Iron Ore
Cement Iron Ore Used" Iron Wire Used Iron Ore Used Iron Wire Used Used as Iron Wire Used
Lot as Primary as Primary as Primary as Primary Primary as Primary
No. " Standard Standard Standard Standard Standard Standard
stannous Stannous Stannous Stannous Stannous Stannous
Chloride \ Chloride Silver Chloride Chloride Silver Chloride Chloride Silv i
Reduction ! Reduction Reductor*
I
Reduction Reduction Reductor* Reduction Reduction Reduc f J:
A1 B1 i A2 B2 As Bs A4 B4 A5 B5 A6· B6 A B A8- B8 A
1 954 0.36 0. 36 o. 3 o. 3 0 . 3 o. 3 II. 0 ~ 37 o. 37 o. 37 o. 37 0.37 0.37 0·35 0.35 0.35 0.35 0.36 O.
17123 2.11 2.11 ! 2.11 2.11 2.10 2.11 2.10 2.10 2.09 2.10 2.10 2.10 2.07 2.08 2.07 2.08 2.06 2.
17119
17126
3.0~ 3. 0 4 ! 3. , 3.0, 3.0~ 3.
0 ,1
0
3·7 3.7 5 I 3·7 3.7 3.7 3·7 I 3.7 3.73
3.0~ 3.03 3'701 3.0~ 3.02 3.02
3. 3 3·7 3.73 3.73
3.. 01 3~02 3.00 3.01 2.99 3.00
3·72.3.72 3.71 3·71 3.69 }.70
17120 4.70 4.68 I 4.67 4.67 4.68 4.67 4.69 4.68 4.67 4.67 4.67 4.67 4.66 4.66 4.66 4.66 4.63 4.64
17545 4
5.35 5.35 '\5.3 5.33 5.35 5.33 II', 5.35 5.35 5.35 5.33 5.35 5.34 5.34 5.33 5.33 5·32 5.30 5.29
l7348A 3·95 3.94 3.93 3.93 3.93 3.93
l7348B 3.58 3.57 ! 3.57 2.57 3.57 3.57
TITRATED WITH CERIC SULFATE

INDICATOR - (FERROIN) POTENTIOMETER

Cement Iron Ore Used as Iron Wire Used as Iron Ore Used as Primary Iron Wire Used as Primary
Lot Primary Standard Primary Standard Standard. Cement Solutiqns Standard. Cement Solutions end
No. (Sn012 Reduction) (Silver Reductor)* and Ore Reduced with SnC12 Wire Reduced in Silver Reductor
AIO B10 All Bll A12 B12 Al BI
16954 0.38 0.38 0.38 0.38 0.37 0.37 0.35 0.35
17123 2.09 2.09 2.09 2.09 2.09 2.09 2.07 2.07
17119 3~03 3.03 3·03 3.03 3.02 3.02 3.00 3.01
17126 3.74 3.73 3.73 3.73 3.73 3.72 3.72 3·72
17120 4.68 4.67 4.67 4.67 4.68 4.67 4.65 4.65
17545 5.36 5.34 5.37 5.34- 5.35 5.33 5.34 5·32
17348A 3.93 3.92 3.92 3.92
1 48B \ .59 3.59 I 3.59 .59 ere 339
* Silver reductor use 0 1'6 nce pr1mary s an
 Research Laboratories of the
Portland Cement Association
33 W. Grand Ave., Chicago

Table 5 - COMPARISON OF TID: Fe2 0 S CONTENT OF PORTLAND CEMENT

BY DIFFERENT METHODS, PER CENT BY WEIGHr

A.S.T.M. POTASSIUM DICHROMATE* CERIC(StTLFATE*

Method Indicators
..
Cement with Potentiometer Indicator Potentiometer
SnCl? Reduction (Ferroin) SnC12 Silver
tot Diphenyl- Diphe~rl- N-methyl- SnC12 Silver
No. lIDine SnC1 2 Reduc- Reduc-
amine dipheny1- Reduc- Reduc- Reduc- tion tor
.J.ndicator barium. amine tion tor
, (Previous tion
Results) Sulfonate Barium.
Sulfonate

16954 0.41 0.36 0·37 0·35 0.36 0.38 0.37 0.35

17123 2.15 2.11 2.10 2.08 2.06 2.09 2.09 2.07
17119 3.03 3.03 3.02 3.01 3.00 3.03 3.02 3·01
17126 3.78 3.75 3.73 3.72 3.70 3.73 3.73 3.72
17120, 4.67 4.68 4.68 4.66 4' 64 4.67 4.68 4.65
17545 I 5.44 5.34 5.34 5.33 5.30 5.35 5.34 5.33
17348A I
I
3.94 3.92
! .
I
17348B 3.57 3.59
I
* Average values from the respective method and indicator shown in Table 4.

June 4, 1947
Series 339
 Research Laboratories of the
Portland Cement Association
33 W. Grand Ave., Chicago

Table 6 - COMPARISON OF THE HIGHEST AND LOWEST CALCULATED VALUES

FOR THE Fs::Ps CONTENT IN PORrLAND CEMENTS, PER CENT BY WEIGHT

Cement HIGHEST VALUE LOWEST VALUE Maximum

Lot Fe:ps Solution Factor Fe 20 S Solution Factor Difference
No .. ojo Determined by: ojo . Determined by:

16954 0.38 Ce( S04)2 + 0.35 Potentiometer 0.03

Ferroin + with Ce(S04)2
all primary and K2Cr207
standards
I
I 2.06
I
17123 2.11 K2Cr207 + 0.05
Dipheny1amine
Barium Sulfonate
I K2cr20I with
Potent ometer +
Iron Wire +
+ Iron Ore I Silver Reductor
17119 3.04 Ditto 2.99 Ditto 0.05

17126 3.76 Ditto 3.69 Ditto 0.0'7

17120 4.70 Ditto 4.63 Ditto 0.07

17545 5·37 Ce( S04)2 + 5.29 Ditto 0.08

Ferroin +
Iron Wire

June 4, 1947
Series 339
 APPENDIX

DETAILED DESCRUTION OF METHODS OF DETERMINING THE

FERRIC OXIDE CONTENT OF PORTLAND CEMENT
 APPENDIX

Dm'AILED DESCRnTION OF METHODS OF DlL'l'ERMINING THE

FERRIC OXIDE COMENT OF PORTLAND CEMENT

*****
REAGENTS:

1. Standard Oxidimetric Solutions:

(a) A stock solution of potassium dichromate was prepared

by dissolving 24.53 g. of reagent grade K2Cr207 in
2 liters of distilled water. An approximately 0.05 N
solution was prepared by diluting 200 ml. of the
stock solution to one liter with distilled water.

(b) The ceric sulfate solution was prepared by adding

55 g. of reagent grade (NH4)age(NOs)6 to 60 m1.
of concentrated H2S0 4 contained in a3-liter beaker.
Water was gradually added in 100-ml. portions.
The mixture was stirred for a few minutes before
adding each succeeding lOO-ml. portion of water.
When the nitrato-cerate salt had been converted
to the sulfato-cerate salt as evidenced by the
absence of a precipitate, the solution w.as cooled
and the volume adjusted to 2 liters. The solution
was allowed to stand two weeks before standardizing.

2. Primary Standards: The following substances were used for

standardizing various volumetric solutions used in this study:

(a) Sodium oxalate oxidimetric standard furnished by the

National Bureau of Standards (standard sample No. 40)
was used for standardizing the ceric sulfate solution.

(b) Potassium. dichromate oxidimetric standard furnished by

the National Bureau of Standards (standard sample
No. 136) :was used for standardizing the ferrous
ammonium sulfate solution.

(c) Iron wire supplied by the Central Scientific Company

and carrying the manufacturer's assay of 99.89% Fe
was used for standardizing the ceric sulfate and
potassium dichromate solutions.

(d) Sibley iron ore (standard sample No. 27b) furnished

by the National Bureau of Standards was also used
for standardizing the ceric sulfate and potassium
dichromate solutions. Values on the N.B.S. certi-
ficate for the total iron content of the ore vary
from 68.l5~ to 68.43%. The recommended value of
68.23% was used in calculating the Fe2 0 s factors
of the potassium dichromate and the eerie sulfate
solutions.
 (e) An aJlproximately 0.05 N solution of ferrous ammonium.
sulfate was prepared by dissolving 20 g. of Baker's
special c.p. grade of FeS04.(NH4)2S04.6H20 in dis-
tilled water, addins 40 ml. of concentrated ~O" and
diluting to one liter with distilled water. The
solution was stored in a tightly-stoppered, brown
bottle.
3. Indicators: The indicators used were:

Name of Indicator Strength of Indicator Quantity Used for

Solution Solution Titrations
Recommended j Used Recommended . Used
Ortho-phenanthro1ine 0.025M 0.025M 1 to 2 1
ferrous complex or approx. drops drop
(Ferroin) 1.5~

Diphenylamine barium 0.005M 0.2'{o 3 to 5 3

sulfonate or drops drops
0.310
N-methyldipheny1amine 0.1% to O.l~ 0.1 to 0.2 cCI' 3
barium sulfonate 0.'4 (2 to 5 i drops
drops) ,
(a) Ferroin, purchased from the G. F. Smith Chemical Co.
is ready for use as it comes from the manufacturer.
(b) The diphenylamine barium sulfonate solution was pre-
pared by adding 0.12 g'. of pOWdered diphenylamine
barium sulfonate to 60 ml. of water. The mixture
was heated to boiling, digested for a short period
of time, allowed to cool and the undissolved portion
of the salt removed by filtration.
(c) N-methyldipheny1amine barium sulfonate is not available
commercially and was prepared in the laboratory of
the Portland Cement Association by the following
procedure:
To 10 ml. of concentrated H;aS04' 2 ml. of fuming
H2S0 4 (1510) was added and thoroughly agitated.
Tbis solution was then added dropwise while
stirring Vigorously with a mechanical stirrer
to 40 ml. of N-methyldiphenylamine contained
in a glass-stoppered flask. Stirring was
continued a few seconds after the final addi-
tion of the sulfuric acid solution, then the
flask was loosely stoppered and placed in an 011
bath for 6 hours at a temperature of 110°. to 120°0.
The reaction may also be carried out by placing
the flask in an oven at 110° to 115°C. for over-
night. This has the advantage of keeping the
react10n mixture protected from light which is the
source of decomposition products formed during the
reaction. *2*
 The reaction mixture was shaken thoroughly from. time
to time while digesting in the oil bath. At the
end of the digestion per10d, the reaction mixture
was cooled and po~ed into a liter distilling
flask containing 600 ml. of distilled water. The
oily layer of unsulfonated methyldiphenylamine
which rose to the surface was removed as far as
possible With the aid of a pipette and a suction
bulb. The mixture was then distilled until the
distillate in the water condenser began to appear
cloudy. Distillation was immediately interrupted
and the distillate diluted to 600 ml., heated to
boiling and cautiously neutraliZed with barium.
carbonate. The solution was cooled, then heated
again so that the precipitate of barium sulfate
became coarsely crystalline. The mixture was
filtered while hot With the aid of suction through
a Buchner funnel to remove the barium sulfate.
The filtrate was evaporated to about 100 ml.
whereupon crystals began to form. After cooling
the mixture was filtered with suction through a
Buchner funnel. The crystals of N-methyldiphenylein1ne
barium sulfonate were washed with cold water and
dried in an oven at 105°0. The filtrate from the
crystals was again evaporated to a small volume
(a.bout 90 ml.) and a second batch of crystals
obtained. Analysis of both batches indicated
that the barium salt of the mono-sulfonio acid
derivative of methyldiphenylamine had been
obtained. The yield was approximately 10 grams.

The indicator solution was prepared from the above

salt by dissolVing 0.06 g. of the salt in 60 ml.
of distilled water. Solution took place readily
without the aid of heat. No precipitate developed
on standing.

4. Miscellaneous Reagents:

(a) An approximately normal sulfuric acid solution was

prepared by adding 30 ml. of concentrated H2 S0 4 to
1020 ml. of distilled water.
(b) An approximately normal hydroc.hloric acid solution was
prepared by diluting 90 IDl. of concentrated Hel with
distilled water to make one liter.

(c) A phosphoric acid solution was .prepared by adding

500 ml. of concentrated BSP0 4 to 500 ml. of dis-
tilled water.
 (d) The current A.S.T.M. method(C 114-46) calls for a 10~
solution of stannous chloride. .One-half this strength
of stannous chloride solution was prepared by adding
2.5 grams of SnCle .2H2 0 to 12 ml. of concentrated HCl
and 38 ml. of distilled water. A globule of iron-free
tin was added and the mixture boiled until the solution
became clear. The solution was kept in a 60 ml. dropping
bottle containing metallic tin. It was prepared fresh
each week.
(e) An iodinemonochloride solution was used as a catalyst'
in the potentiometric titration of sodium oxalate
with ceric sulfate. It was prepared by dissolving
0.279 g. of XI and 0.178 g. of KIO s in 250 ml. of
distilled water. Immediately 250 ml. of concentrated
. HOl was added all at once. Dilute solutions of KI and
of KIO s were prepared by a.dding about one mg. of each
salt to two 25 ml. portions of distilled water. The·
iodine monochloride solution was adjusted electro-
metrically by adding a few drops of the dilute solution
of iodide or iodate until one drop of either solution
caused a large change in the oxidation-reduction
potential of the iodine monochloride solution. The
final endpoint was approached with potassium iodide.

APPARATUS:

1. Potentiometer: A Beckman Model G pH meter wa.s used for the

potentiometric t1 trations. The scale on this model may be
read either in pH or millivolts depending upon'how the pH
meter is adjusted. The millivolt scale was used for the
oxidation-reduction titrations.
2. Electrodes: A 8eaI.s6.; sleeve-type, saturated calomel electrode
was used as the .1~cf"3::"ence electrode. A sealed platinum elec-
trode (factory ma,;Le) '\-Tas used as the indicating electrode at
the beginning of the investigation but it was found that
this type was easily II poisoned" and hard to clean; conse-
quently, it was replaced by platinum wire electrodes. The
wires were made of c.p. platinum and varied in length from
li inches to .3 inches. Part of the wires had a B. & S.
gauge of 16 and the rest had a B. & S. gauge of 26. The
difference in the size of the wires had no effect on the
observed endpoint of the titration, but it seemed that the
smaller electrodes were more sensitive to ionic changes in
the solution. The platinum wire electrodes were connected
to the potentiometer by means of a small battery clip
attached to a rubber-coated, metallic-shielded, stranded
copper wire of the type used for microphone connections.
The metallic shield around the wire was grounded.
The platinum wire electrodes when not in use were
stored in chromic acid cleaning solution. Before using

* 4 *
 for a potentiometric titration, they were rinsed with dis-
tilled water, dried and flamed. The factory-made, sealed
platinum electrode was cleaned by dipping it in cleaning
solution, rinsing with water, and finally cleaned by electrolysis.
The electrolytic decontamination was accomplished by connecting
the electrode and a platinum crucible lid by means of battery
clips to a six-volt dry cell battery, placing the two pieces
of platinum one centimeter apart in a beaker containing
N!lO Hel and passing a current through the liquid for one
minute. By means of a two-waY switch, the current was
alternated every five seconds. The sealed platinum elec-
trode was seldom used in this study, the wire electrodes
being preferred.
3. Silver Reductor. A silver reductor was prepared by charging
a Jones reductor tube with 28 g. of metallic silver after
the prescribed method of Walden, Hammett and Edmonds(l).
The metallic silver had been prepared in the laboratory of
the Portlarid Cement Association by reclamation from silver
chloride residues by procedures outlined by Chase(2) and
Spies(3). The metallic silver thus obtained was heated in
'a muffle at 750°C~ to cause agglomeration of the fine
particles. The resultant spoIlBY' mass was ground in an
agate mortar to pass a No. 20 standard sieve. The granular
silver was suspended in water and the mixture poured into
the. reductor column. A plug of glass wool preViously
inserted in the lower end of the reductor column served
to retain the metallic silver. The silver was then washed
several times with dilute ammonia water (1-1) followed by
several washings with water, then washed with dilute
sulfuric acid (1-2). Care was taken never to allow the
liquid in the column to drain below the top of the silver.
The rate of flow through the column was 30 to 35 ml. per
minute depending upon the height of the liqUid in the
column. The silver reductor did not need to be regenerated
inasmuch as less than one-fourth inch of the metallic silver
was blackened with use during the entire investigation.

MEl'HODS:

1. Standardization of the Ferrous Ammonium Sulfate Solution: This

solution, which was used as a primary standard for the
oxidimetric solutions, was standardized against samples of
various weights· of the National Bureau of Standards'
potassium dichromate. TvlO indicators, diphenylamine barium
sulfonate and N-meth;Yldiphenylamine barium sulfonate, and a
potentiometer were used to. determine the endpoint of the

G. H. Walden, Jr., L. P. Hammett, and S. M. Edmonds, Journal

Am. Chem. Soc., v. 56, p. 350 (1934).
L. C. Chase, Chemist-Analyst, v. 18, No.4, p. 14 (1929).
J. R. Spies, Ind. & Eng. Chem.,(Anal. Ed.}, v. 7, p. 118 (1935).

*5*
 titrations. The weighed sample of K2Cr207 was dissolved
in 100 ml. of normal HOL If an indicator was to be used, .10 mL
of a 1-1 HsPO 4 solution were also added. After the \ ichromate
crystals were dissolved, the solution was titrated with the
ferrous ammonium sulfate solution using a 25 ml. calibrated
normax burette. The ferrous solution was added dro~se,
care being taken to allow the solution to tome to equilibrium
between drops, particularly as the endpoint was approached.
During titration, the dichromate solution was agitated With
a mechanical stirrer. The indicator endpoint was regarded
as being reached when one drop of the ferrous solution
caused a permanent fading of the indicator color. The
potentiometric endpoint was reached when one drop of the
ferrous solution caused a sudden break in potential.
The results of standardizing the ferrous ammonium
sulfate solution are shown in Table 1. .

2. Standardization of the Potassium Dichromate Solution: This

solution was standardized against the standard solution of
ferrous ammonium sulfate, iron ore and iron wire. The iron
ore was reduced with stannous chloride. The reduction of
the iron wire was carried out in a silver reductor, and by
a stannous chloride procedure. Two indicators, diphenylamine
barium sulfonate end N-methyldiphenylamine barium sulfonate,
and a potentiometer were used to determine the endpoint of
the titrations.
(a) Standardization of the Dichromate with Ferrous Ammonium
Sulfate. The method of standardizing the dichromate
using the standard ferrous ammonium sulfate solution
was as follows:
To 5, 10 and 15 ml. portions of the standard
ferrous ammonium sulfate solution measured, with
a 25 ml. calibrated burette, into 250 ml. beakers
were added 75 ml. of normal Hel. If an indicator
was used, 10 ml. of the phosphoric acid solution
were also added. The ferrous solution was then
titrated with the. potassium dichromate solution,
the former being agitated with a mechanical stirrer.
The dichromate was added dropwise. As the endpoint
was approached, as indicated by sudden flashes of
color in the solution or by sharp fluctuati9n of the
potentiometric needle, the time between adding each
successive drop was lengthened so that the color was
allowed to completely fade, or the needle allowed
to come to eqUilibrium. The indicator endpoint was
taken as that point at which one drop of dichromate
caused a permanent (5 minutes or longer) appearance
of color in the solution. The diphenylamine barium
sulfonate imparted a bright purple color to the
solution. N-methyldiphenylamine barium sulfonate
gave a rose-red color to the solution. The poten-

*6*
 tiometric endpoint was taken as that point at
which one drop of the dichromate solution caused
a sharp break in potential. The Fe20s factor
of the dichromate solution was calculated by the
following formula.:
. Fe20S factor :: normality of ferrouB solution
x 0.5585 x 1.4297
55.85 == atomic weight of iron (1941)
1.4297 = molecular ratio of Fe20 S to 2 Fe
The Fe20s factors of the dichromate solution
as reported in Table 2 for the N-methyldiphenylamine
barium. sulfonate and the potentiometer were cal-
culated from the normality factor of the ferrous
ammonium sulfate solution determined with the
corresponding indicator or the potentiometer.
(b) Standardization of Dichromate with Iron Ore. The
procedure followed for standardizing the potassium.
dichromate solution against iron ore is given below:
The iron ore was dried in an oven at 105°0.
overnight, cooled in a desiccator, and samples of
ore haVing weights of 0.0147, 0.0293, 0.0440,
0.0586 and 0.OT53 g. were weighed on a glass
weighing dish and transferred to a 250 ml. beaker.
These weights of samples were chosen because when
converted to grams of Fe they ranged from 0.01 to
0.05 g. of Fe in 10 mg. gradations. The samples
were weighed on e. glass dish counterpoised by a
similar glass dish, because it was found to be
impossible to quantitatively transfer the iron
ore from a metal pan to the 250 ml. beaker. A
25 ml. portion of concentrated HOl was added to
each of the s~ples of ore, the beakers covered
with watch glasses and placed on a hot plate.
When the mixture started to boil, the samples were
moved to a cooler section of the hot plate
until they simmered lightly. The samples were
then allowed to remain overnight on the hot plate
at a temperature of 50° to 60°0.
On the following morning the cover glasses
and sides of the beakers were rinsed with distilled
water, the lids were replaced and the iron ore
solution brOUght to boiling. The solution was
boiled lightly until the last traces of the ore
had gqne into solution. When the ore was in solution,
it· was reduced while hot with a '7'/0. stannous chloride
solution. Two drops in excess were added because
occasionally no precipitate was formed upon the
later addition of the mercuric chloride when only
one drop in excess was added.
*7*
 The ferrous solution was then cooled to room tem-
perature in a water bath. A 5 ml. portion of saturated
mercuric chloride solution "was added all at once and
the mixture allowed to stand for 5 minutes or if
titrated immediately, the solution was stirred vigorously
one minute with a mechanical stirrer.
After the excess stannous chloride was oxidized
by the addition of the mercuric chloride, the ferrous
solution was diluted to approximately 70 ml. with
distilled water. In the indicator method, 10 ml. of
the phosphoric acid solution were added and 3 drops
of indicator. In the potentiometric method this step
was omitted. The iron ore solution was then titrated
with potassium dichromate. The same precautions re-
garding the dropwise addition of the dichromate were
observed as in the titration of the ferrous ammonium
sulfate solution.
The Fe20s factor was calculated by the formula:
Fe ° factor = Vol.
2 s
Wt. of ore x 68.2~! x 1.4297
of dichromate
68.23 = ~ of Fe in ore
1.4297 = molecular ratio of Fe20s to 2 Fe
(c) Standardization of Dichromate with Iron Wtre: The method
which was used to standardize the potassium dichromate
against iron wire is as follows:
A length of iron wire (3 to 4 inches) was polished
bright wi th fine emery cloth. The wire was then wiped
clean with a soft cloth. A portion at each end of the
wire was cut off and discarded. Samples were taken
by cutting off pieces of the iron wire, the length
depending upon the weight of sample desired. Samples
of wire from 0.01 g. to 0.05 g. were taken in 10 mg.
gradations. The weighed samples were placed in 250 ml.
beakers and dissolved in 25 ml. 1-1 ReI. Solution
took place very readily and if desired, was hastened
by placing the beakers on a warm (60°0.) hot plate.
After the wire was dissolved, the lids and sides
of the beakers were rinsed with distilled water. The
samples were then "reduced by one of two methods. /l. 510
solution of stannous chloride was added dropwise, one
drop in excess being added to the solution of the iron
wire which had been heated to boiling. After reduction,
the samples were cooled to apprOXimately room temperature
in a water bath. A 5 ml. portion of the saturated mercuric
chloride solution was added all at once and the mixture
allowed to stand for 5 minutes or stirred Vigorously for
one minute. It was then diluted to approximately 70 ml.
with distilled water. If an indicator was used to

*8*
 determine the endpoint, 10 ml. of the phosphoric acid
solution and 3 drops of the indicator solution were
added. The mixture was then titrated with the solution
of potassium dichromate. The alternate method of re-
duction was carried out by passing a cool solution of
the iron wire through a silver reductor column. The
reductor column wa~ rinsed with 100 ml. of normal Bel
which was added in portions of 10 to 20 ml. at a time.
Each portion was drained to the top of the silver be-
fore the succeeding portion was added. The .rate of
flow through the column was approximately 30 ml. per
minute. The final volume was approximately 150 ml.
which was titrated with the dichromate. If an indicator
was used, 10 ml. of the HsP04 solution was added, 'ille
F:e20 3 'factor of the dichromate was calculated by the
formula:
Fe ° factor = Wt.Vol.ofofwiredichromate
2 S
x 99.8~ x 1 4297
•

99.89 ..; '10 Fe in the wire

1.4297 = molecular ratio of Fe20s to 2. Fe
The results of these methods of standardizing the
potassium dichromate solution are shown in Table 2.
3. Standardization of the Cerie Sulfate Solution~ The eerie
sulfate solution was standardized against the standard
ferrous ammonium sulfate solution, iron ore, iron wire,
and sodium oxalate. The indicator, orthophenanthro1ine
ferrous sulfate complex ("ferroin") and the Beckman Model
G pH meter, were used to determine the endpoint of the
titrations.
(a) Standardization of Ceric Sulfate with Ferrous Ammonium
Sulfate. The determination of the Fe20s factor of
the eerie sulfate against the standard ferrous
ammonium sulfate solution was conducted as follows:
To each of 5, 10, and 15 ml. portions of
the ferrous ammonium sulfate solution which had
been measured with a 25 ml. calibrated normax
burette into 250 ml. beakers was added 100 ml.
of no:rma1. Bel. One drop of ferro in was added
unless the titration was to be conducted poten-
tiometrically. The solution was agitated with
a mechanical stirrer and titrated with the
-ceric sulfate. In the potentiometric titra.tions,
the endpoint was exceedingly sharp. One drop
of the eerie sulfate solution (0.04 to 0.05 ml.)
was sufficient to cause a change in e.m.f. from
0.7 volts to 1.1 volts.
The indicator endpoint was fairly sharp; the
red fading, leaving a yellow-green colored solution.
At no time could the blue oxidized form of' the
ferroin be discerned.
*9*
(b) Standardization of Cerie Sulfate with Iron Ore: In standard-
iZing the' eerie sulfate against iron ore, the same pro-
cedure was followed that was used in standardizing the
potassium dichromate solution against iron ore. The
solution of iron ore was reduced with stannous chloride.
The indicator, ferro in, and the potentiometer were used
for determining the endpoint of the titrations.
(c) Standardization of Cerie Sulfate with Iron Wire: The
standardization of the eerie sulfate solution against
iron wire was conducted by the same procedure as was used
in standardiZing the potassium dichromate, except that in
the standardization of the eerie sulfate only the silver
reductor was used for reducing the solution of the iron wire •.
(d) Standardization of Ceric Sulfate with Sodium Oxalate plus Iodine
Monochloride: The potentiometric titration of sodium
oxalate with eerie sulfate solution was accomplished by
using an iodine monochloride solut.ion as catalyst. By
using iodine monochloride, it is possible to carry out the
titration at room temperature. The procedure used for the
standardization of the eerie sulfate solution by this
method was as follows:
Samples of sodium oxalate were placed in 250 ml.
beakers and dissolved iD. 100 ml. of normal HCI. A
10 ml. portion of the ICl solution was added and the
solution titrated potentiometrically with the eerie
sulfate solution. The break at the endpoint amounted
to 0.10 to 0.15 volts per 0.05 ml. of the eerie sulfate
solution. The Fe20S factor of the solution was eal-
culatedby the formula:
Fe20S = N(0.05585 x 1.4297) .
55.85 = atomic weight of Fe
1.4297 = molecular ratio of Fe20 S to 2 Fe
N= lit. ~ oxalate
0.067 x ml. of solution
0.067 = ~~o85 (g. eqUivalent wt. of N~C204)
(e) Standardization of Ceric Sulfate with Sodium Oxalate plus
Ferrous Ammonium SUlfate. The indicator method of standard-
izing the eerie sulfate solution against sodium oxalate was
performed with the aid of the ferrous ammonium sulfate solu...
tion. The procedure used 'Was as follows:
To each of the samples of sodium oxalate contained
in 250 ml. beakers, was added 70 inl. of normal sulfuric
acid. The oxalate solution was heated to approXimately
95°C. and then titrated with eerie sulfate solution
until it was evident by the yelloW color of the solution
that an excess of the eerie sulfate had been added. The
* 10 *
 solution was then cooled to room temperature. Fifteen IDl.
of the ferrous ammonium sulfate solution, whose equiva-
lent value in terms of the eerie sulfate solution had
previously been determined, and one drop of ferroin
were added. The titration was then continued to the
ferroin endpoint. The amount of eerie sulfate solution
required to titrate the 15 ml. of the ferrous ammonium
sulfate solution was then subtracted from the total
volume used for the titration. The remainder repre-
sented the amount required by the sodium oxalate.
The Fe20s factor was calculated by the same formula
as given for the potentiometric titration of sodium
oxalate with eerie sulfate.
The results of these various methods of standardizing
the eerie sulfate solution are given in Table 3.
4. Determination of the Ferric Oxide Content of Portland Cement: The
standard solutions of potassium dichromate and eerie sulfate were
used for determining the ferric oxide content of 8 cements ranging
from 0.4% to 5.4~ Fe20s' In the indicator method, the cement solu-
tions were reduced with stannous chloride. In the potentiometric
titration, the cements were reduced with stannous chloride, end
by passing the solution through a silverreductor.
(a) Titration of Cements with Potassium Dichromate: The
procedlu-e which was used in determining the FeaOs content
of tb. cements with potassium dichromate was as follows:
To 1 g. of the cement in a 250 ml. beaker was
added 4o ml. of water. The mixture was stirred with
a rubber policeman until the cement was suspended in
water, then 15 ml. of concentrated HOl was added. The
sample was placed ona hot plate end stirring continued
until the cement was dissolved. The beaker was covered
and the solution heated to boiling. In the stannous
chloride method of reduction, 8, 510 solution of SnCl2
was added drop by drop until the yellow color of the
solution disappeared, then one drop in excess was
added. The solution was placed in a water bath and
cooled to room temperature. Five ml. of saturated
HgC12 so1ut1on were added and the solution stirred
for one minute. A 10 ml. portion of RS P0 4 solution
was added and the cement solution diluted to approximately
75 ml. Three drops of indicatbr were added end titration
with the dichromate begun. As the endpoint was approached,
the drops of dichromate were added slowly to allow the
reaction to come to equilibrium before the succeeding
drop was added.
In the potentiometric titration of the samples,
the phosphoric acid and indicator were not added.

*11*
 · When the solution of the cement was to be .
reduced by passing it through a silver reductor,
filtration of the sample was necessary to remove
the undissolved silica. The hot solution of the
cement was filtered through a Whatman No. 31,
9 em. filter paper. The filtrate was collected
in a 400 ml. beaker. The 250 ml. beaker was
scrubbed with a policeman and washed several times
with small portions of water. The filter paper
and residue were carefully washed 10 times with
small portions ot water. The final volume in the
400 ml. beaker was 80 to 90 m!. This was poured
into the reductor tube and the beaker rinsed
with distilled water. The solution from the
reductorcolumn was collected into the same
beaker. The solution passed through the column.
at a rate of 30 to 35 ml. per minute. The
reductor column was rinsed with 150 ml. of
normal Hel which was added in 10 to 20 ml.
portions, each being drained to the top of
the silver before the succeeding portion was
added. The final volume was approximately 250 ml.
The samples reduced by this method were titrated
potentiometrically •.•

(b) Titration of Cements with Ceric Sulfate. Whenthe

Fe 20 S content of the cements was determined with
eerie sulfate, the same procedure of dissolVing
the cements and reducing the resultant solutions
was followed as with the potassium dichromate
solution. The indicator, ferro in, and the
potentiometer were used to determine the endpoint
of the t1trations. No phosphoric acid was used
with the ferro in indicator.

The results of the various methods of de-

termining the Fe 20 S content of the cements are
given in Table 4. The percent of Fe20s was cal-
culated by the following formula:

%Fe 20 S = (FeaOs factor)(ml. of solution).lOO

* 12 *
June., 1947